Capital Goods Global Primer: Digital Machinations

>> Employed by a non-US affiliate of MLPF&S and is not registered/qualified as a research analyst under the FINRA rules. Refer to "Other Important Disclosures" for information on certain BofA Merrill Lynch entities that take responsibility for this report in particular jurisdictions. BofA Merrill Lynch does and seeks to do business with issuers covered in its research reports. As a result, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision. Refer to important disclosures on page 136 to 137. 11665665

Capital Goods

Global Primer: Digital Machinations

Primer Equity | 09 September 2016

A Transforming World: The digital industrial world As part of our work on A Transforming World, we introduce a new Global Primer focused on the developments of industrial software, the industrial internet and the implications for the industrial automation landscape.

The software revolution digitising industrial automation Whether it is big data analytics, the industrial internet or the latest in connected wearables, companies are searching for the next level of productivity, and greater application of software offers the next revolution in industrial automation markets. Industrial software markets offer growth in a stagnant environment, so merit deeper analysis. We explore 8 critical trends that investors need to be aware of and identify companies we see as best positioned, most innovative or could face disruption. For conclusions on stocks, accompanying this Primer we publish Industry goes digital – Primer Picks.

It matters, and it matters now Investors need to care about the industrial internet because the management teams running the companies they own are already implementing industrial internet strategies. This means investment now, revenue and cost opportunities later. Industrial companies within our global stock coverage operate in end markets that account for c75% of the USD4-11tn pa in expected value accruing from industrial internet technology over the next decade. But, the landscape of industrial automation competitors is changing and fund managers need to be discerning in identifying the investment opportunities

8 trends for industrial automation We explore 8 key trends investors should be aware of driving industrial automation markets over the medium term. 1) Increasing software focus driven by need for capital efficiency and productivity, 2) The industrial internet; momentum is building in terms of adoption rates, 3) Big Data analytics; the starting point is connecting machines and analysing the data they produce, 4) Everything-as-a-Service; evolution of the industrial mind-set to sell availability rather than product, 5) Robotics, a continuing theme of penetration, 6) Distributed Control System (DCS) migration: disruption is a near term threat as major customers seek to change the status quo, 7) Cybersecurity; the biggest barrier to software adoption, 8) Disruptive technologies are closer than you think; blockchain, augmented reality, machine learning and connected wearables.

State of the union Discrete markets are typically early cycle and have been relatively flat since 2011, versus Process markets which are 10%+ ahead, but weaker of late because of low oil prices. PMIs suggest Programmable Logic Controller (PLC) market is likely to stay soft in the near term while DCS demand is affected by O&G capex contraction. Longer term, we believe industrial automation markets can at least maintain the 3-4% growth seen historically given the number of structural drivers. In China, near term growth is likely relatively mixed.

Pan European Capital Goods Mark Troman >> Research Analyst MLI (UK) +44 20 7996 4194 [email protected] Alexander Virgo >> Research Analyst MLI (UK) +44 20 7996 1221 [email protected] Andrew Obin Research Analyst MLPF&S +1 646 855 1817 [email protected] John P King, CFA >> Research Analyst MLI (UK) +44 20 7996 7265 [email protected] Michael Kaloghiros >> Research Analyst MLI (UK) +44 20 7996 1226 [email protected] Jeremy Elster Specialist Sales MLI (UK) +44 20 7996 0875 [email protected]

A Transforming World

Una

utho

rized

red

istr

ibut

ion

of th

is r

epor

t is

proh

ibite

d. T

his

repo

rt is

inte

nded

for

hsha

h7@

irevn

a2.c

om

http://research1.ml.com/C?q=2jZBfwClvDEY9BJoUi4-Xw

http://research1.ml.com/C?q=oEJvYJjn!ADGTWLbQayTQw


mailto:[email protected]






2 Capital Goods | 09 September 2016

Contents Industry goes digital 4

Viewpoint 4

The 8 key industrial automation trends 6

1. Software revolution 8

The Industrial Software market at a glance 9

Software grows faster as customers seek efficiency 11

Convergence of Operational Tech and Information Tech 16

GE’s differentiated approach 20

Software companies offer superior metrics to industrial automation vendors 24

Case Study: Daimler switching to Siemens from Dassault 26

2. Industrial internet changes the game 29

Assessing the potential economic benefit of the industrial internet 29

Industrial internet drives evolution in industrial automation 34

Changing competitive landscape in industrial automation 35

Industrial internet main applications and examples 38

3. Big data analytics 41

A natural starting point, but drives need for smarter field instruments 41

Investment focused initially on the easy wins and the basic building blocks 42

Next generation condition monitoring – APM software 42

“Edge” software capabilities push capability down the value chain 45

4. Everything-as-a-Service 47

Evolution of the service business 47

Can everything be provided as a service? 48

A change in mind-set is required 48

Getting paid is the critical question 49

5. The rise and rise of robots 51

Continuing to grow 51

US$24bn industrial robots market by 2025E 52

2015: Another record year 55

Prices of robots are dropping while labour costs are rising 57

China’s Robot Demand 58

Rise of the “cobots”: collaborative robots 61

Logistics are an important part of the equation 64

Warehousing operations: use of robots picking up 65

6. DCS disruption: real and present danger 68

Appointment of non-incumbent points to threat 68

7. Cybersecurity issues looming large 72

Capital Goods | 09 September 2016 3

Increasing occurrences of security breaches highlight the risks of a more connected world

72

8. Disruptive technologies nearer than you think 74

Block chain: potentially revolutionising the supply chain 74

Artificial intelligence: machines learning to think for themselves 75

Augmented reality: a tool for selling, training and enhancing service 76

Connected wearables: at the heart of personal and personnel safety 76

State of play in the end markets 78

PMIs suggest PLC market likely to stay soft, DCS demand affected by O&G capex contraction

81

China – mixed near term 83

Appendix A: The automation market 90

Defining the scope of the market 90

We see the size of the market at cUSD140bn pa 92

Process vs discrete manufacturing 92

Who is exposed to what? 94

Market shares by product grouping 97

How automation vendors built their software businesses 104

Siemens moved first, ABB and Schneider followed 105

Appendix B: Historical margin analysis 110

Historical trends: mid-teen margins, US stronger 110

Near term: op leverage, mix & execution 111

Appendix C: Product overview 113

ICS - Industrial control system 113

Industrial control equipment 115

Industrial interconnect equipment 118

Motion control – precise position & control 118

Software Definitions – Making sense of the acronyms 119

Appendix D: The industrial robots ecosystem 123

Where it fits in in our holistic view 123

The Industrial Robots market at a glance 123

Overview of the main industrial robot players 128

Autos: #1 sector for adoption of robots 129

Appendix E: Automation company overview 131


Industry goes digital Viewpoint This note is an industrial automation primer aiming to have considerable shelf life, but investors need to care about this market now principally because automation remains one of the few growth markets in a stagnant industrial environment. This growth is being driven by innovation and emerging technologies, which potentially enable a step change in manufacturers’ efficiency, flexibility, speed to market, and of course profitability. Automation investment is therefore shifting away from being capacity driven to being more about productivity and gaining a competitive advantage – for example being able to produce customised products at virtually no extra cost (mass customisation).

There are several potentially disruptive trends that offer both opportunity and threat to the current positioning of industrial automation companies in the medium term. Within the broader automation market the ‘industrial internet’ offers both revenue and cost saving opportunities for companies, potentially assuaging investor concerns about both growth and margin resilience. Customers’ increasing focus on capital efficiency requires industrial suppliers to continue to evolve the business model towards service, uptime performance, increased equipment longevity and reduced life-time costs.

Being an industry primer, this note does not make stock recommendations per se, incorporating valuation, consensus estimates, divisional growth drivers, internal initiatives etc. (for that, please refer to our note Industry goes digital – Primer Picks). However, to help investors navigate the industry-driven conclusions of this report, we highlight the following three categories in Table 1:

• Best positioned industrial software providers; intuitively, those companies with established and comprehensive industrial software suites should be best positioned to take advantage of the faster growth and higher margins offered by industrial software demand (e.g. in Product Lifecycle Management or PLM);

• Most innovative industrial companies; companies evolving proactively and embracing the potential of connectivity the Internet of Things (IoT). There is clearly overlap with the first category, but it is not just industrial software providers who offer ways for investors to play this theme;

• Potentially disruptive trends, particularly in process automation; companies for whom the IoT represents a threat or changing environment to their established market positions and competitive strengths. We are not saying these companies do not respond to the challenges, (in fact we would expect them to do so with vigour), but we wish to highlight the challenges nevertheless;

Table 1: Conclusions on industrial internet and automation trends Best positioned industrial software providers Most innovative industrials Potential disruption in DCS

Siemens: best in class industrial software SKF: SKF Insight changing the service game ABB: Leading DCS supplier, most exposed to OT Dassault Systemes: PLM market leader Atlas Copco: Full responsibility service model in CT Honeywell: Top 3 in DCS; incumbent DCS provider at Exxon GE: leading the charge on big data Dover: Remote monitoring of artificial lift systems Emerson: Top 3 DCS provider; largest in field devices Hexagon: highest % of revenues in software GEA: Partnered with SAP on food & beverage IoT Yokogawa: Major DCS supplier arguably lacking scale Rockwell: leading the debate around IT/OT ThyssenKrupp: Leading elevator connectivity GE: Looking to enter DCS aggressively PTC: CAD leader with supply chain management Bosch: Innovation leader in Industry 4.0 Schneider: Market leader in MES and safety Fanuc: World leading robotics supplier Note: IT/OT = Information Technology/Operations Technology; MES = Manufacturing Execution Systems, DCS = Distributed Control System; CT = Compressor Technique Source: BofA Merrill Lynch Global Research

We think the three key things that fund managers need to understand about the digitisation of industrial automation markets are as follows:

1. The industrial internet matters now: Investors need to care about the industrial internet because the management teams running the companies they own also care



and are already implementing strategies to try and exploit growth trends. This means investment now, revenue and cost opportunities later. Clearly, failure on the part of management to develop the right strategies now may well result in a weaker competitive position in the future;

2. Industrial companies well positioned for IoT, and the opportunity is significant: We think industrial companies within our global stock coverage operate in end markets that account for c75% of the expected value accruing from Internet of Things (IoT) technology over the next decade. This highlights the scale of the opportunity and the relatively favourable position industrial companies have as original equipment manufacturers (OEMs);

3. Disruptive changes are happening in automation markets: The landscape of industrial automation competitors is changing as customers seek better capital efficiency and new approaches to managing their asset base. Incumbent market participants need to evolve to face smaller and more nimble competitors and fund managers need to be discerning in identifying the investment opportunities.


The 8 key industrial automation trends Our key conclusions derive from our analysis of eight critical trends which we see affecting industrial automation markets over the medium term:

1. Increasing software focus: Industrial software markets are growing faster than the broader automation hardware market, driven by the advent of the industrial internet enabled connectivity, the need to increase productivity and reduce time to market and the evolution of the service offering (often involving real-time management of operating assets);

2. Industrial internet: A term that is undoubtedly experiencing a lot of hype, but in reality industrial companies we cover are exposed to c75% of the expected economic benefits, so cannot be ignored by investors. Momentum is building in terms of adoption rates and the potential for industrial companies to capture both revenue and cost opportunities is high. However, competition is intense and fragmented –we expect that smaller, more nimble industrial software suppliers will force larger incumbents to evolve or risk losing out. We would also expect further market consolidation, as we are seeing from the likes of Siemens and Hexagon;

3. Big data analytics and ‘edge’ computing: Collection and analysis of data forms the basis for much of the initial industrial internet developments, for example in condition-based monitoring of machines. The sheer scale of data generation is driving a trend towards pushing computing power and decision making to the ‘edge’, i.e. the point at which the data is generated. This means having to make field instruments and sensors smarter and arguably diminishes the need for a separate ‘control’ layer of industrial operational software;

4. Everything-as-a-Service: Service based business has been a focus for most industrial companies over the last decade or more as it typically delivers higher and more consistent returns. Furthermore, service growth remains faster than broader industrial automation demand. Digitally connecting the installed base of equipment increases the ability for suppliers to capture service revenue and deliver value via increased uptime. Real-time asset management requirements from customers, enhanced connectivity technology, and the desire to reduce through-life costs drive the evolution of the service offerings from the suppliers;

5. Robotics: These highly flexible machines have been extensively used in the automotive industry for over 30 years but innovative new lower-cost models are driving significant growth in general industry now as well. Emerging market penetration continues across a number of industries, while robots also have their part to play in smart flexible factories connected via industrial internet.

6. Old DCS migration: ExxonMobil’s appointment of Lockheed Martin as a lead systems integrator to develop a new Distributed Control Systems (DCS) architecture highlights the potential threat to incumbent providers. The capex constraints in the oil & gas/petrochemical industry and a customer desire to change the DCS business model for more flexibility represent a key challenge for the incumbent DCS suppliers (ABB, Honeywell, Emerson).

7. Cybersecurity: Given the development of the industrial internet, connected devices and the increasing prevalence of open architecture software, cybersecurity is a key concern for management teams. Recent high profile data breaches have lowered the risk appetite for new technology and continue to be an inhibitor to increased connectivity among industrial entities.


8. Disruptive technologies on the horizon: Block-chain might be the disruptive technology everyone is talking about in the electronics payment space, but the implications for the industrial supply chain could be huge in terms of inventory management and capital efficiency. Augmented reality offers scope for improved training and maintenance offerings as well as enhancing the design process.

We finish the report with an update on the state of play in the key industrial automation end markets and regions and with a deep dive on China.


1. Software revolution

The industrial software market is growing faster than the broader automation industry as it enables better productivity, improves time to market and increases operational efficiency.

The convergence of information technology and operational technology is at the heart of this growth, with the benefits of combining the two driving increased customer demand for holistic solutions and forcing incumbent industrial automation providers into competition with IT and software providers.

A higher portion of software is likely driving higher organic growth and operating margins for automation vendors, and is also warranting higher valuations.

Best placed – Siemens, Dassault, Hexagon; GE taking a differentiated approach Having a comprehensive range of different software building blocks in place, we think Siemens faces good prospects to offer an integrated solution across the whole value chain, combining factory hardware (e.g. Programmable Logic Controller or PLC) and engineering software (eg Product Lifecycle Management or PLM). The company is the only one within the capital goods space that is able to offer a full-scale PLM solution.

In contrast to Siemens, Dassault Systemes has bridged the gap from product development to manufacturing from the opposite direction. When used in combination, its strong position in design and simulation further supports its operations management solutions as it makes a larger amount of data available that can effectively be used in these processes.

Hexagon is approaching data integration through bridging the gap between design software (mainly for process industries such as oil & gas and marine) and precision measurement of the real world, be it in the factory or at the construction site.

General Electric’s digital’s pursuit targets building an industrials-first software platform that ties applications together with the Predix platform. We think Predix, will be the single most important driver of GE Digital’s growth strategy both as a standalone platform and a basis for development of applications by GE and others. Predix currently accounts for $50mn in sales (2016), but GE expects Predix sales to grow to $4bn by 2020. GE partners with leading technology companies to build out the rest of its capabilities and has listed Amazon, Cisco, Dell, EMC/VMware (through Pivotal), HP Enterprise, Oracle, and PTC as key technology partners to promote and support its Industrial Internet initiative.

How much software do industrial companies really have? Understanding exactly how much software exposure incumbent industrial automation vendors actually have is difficult as the tendency is to promote as large an exposure as possible given the appeal of the end markets and the interest of investors.

We have triangulated different sources of market and company information to try and establish who has the largest industrial software businesses and whether that capability resides more in the information technology (IT) space (which can be thought of as standalone software) or operational technology (OT) space (i.e. software that is typically embedded in products).

Please note that our software analysis is intended to be quite strict in scope, identifying software business alone whether embedded or standalone, excluding both the related hardware and service revenue. The analysis focuses on purely industrial applications of operational and information technology software. Applications of software “outside the


factory” such as aerospace (e.g. Honeywell avionics, GE avionics/jet engines) or healthcare (e.g. GE, Philips) are excluded.

Table 2: Pure software exposure for main incumbent industrial automation providers Software Software % of automation 2014, USDm OT IT Total revenue, BofAML est Siemens 35% 65% 3,348 31% Dassault 100% 2,820 89% ABB 66% 34% 1,175 25% Honeywell 90% 10% 560 19% Emerson 88% 12% 562 19% Schneider 50% 50% 560 18% Hexagon 100% 486 80% Rockwell 82% 18% 455 12% Source: BofA Merrill Lynch Global Research estimates

We conclude that Siemens has the largest and broadest industrial software business among the major global capital goods companies, but the answer also highlights the disparity between what the companies report and what the actual software exposure might be, suggesting the definition of what constitutes software revenue varies depending on application and scope of offering.

The Industrial Software market at a glance

We estimate the industrial software market has a size of c$52bn and is expected to grow at 5-10% p.a. for the next five years (2014-19 CAGR). It can be divided into four segments: Business Systems, Design and Simulation, Operations Management and Control and Supervision.

Our best estimate of the overall size of the industrial software market is around US$52bn and we expect it to grow by 5-10% p.a. on average in the coming years. Industrial software can be divided into four main segments:

• Business Systems includes Enterprise Resource Planning (ERP) and Supply Chain Management (SCM) software;

• Design and Simulation comprises Product Lifecycle Management (PLM) software;

• Operations Management includes Manufacturing Execution Service (MES) and Enterprise Asset Management (EAM) software;

• Control and Supervisory includes Supervisory Control and Data Acquisition (SCADA), Programmable Logic Controller (PLC) and Distributed Control System (DCS) software. This is the traditional focus of the major automation suppliers such as ABB and Siemens.

The quandary faced by incumbent industrial software providers is that their legacy expertise lies primarily in control & supervisory software, while the largest market and fastest growth is expected to come from more holistic systems at the business and operational level. While companies such as Siemens, Rockwell, Schneider and to a lesser extent ABB have made inroads into developing presence in markets such as PLM, MES and ERP, the competitive environment is changing driven by both customer requirements and the appeal of faster growth.

Companies such as SAP and Oracle, which traditionally dominated the business systems market, engineering software companies like Dassault and Hexagon and new competitors such as design software company PTC, are offering holistic solutions for customers to


benefit from simplicity and integration of business processes. This is the essence of the convergence of information technology (IT) and operational technology (OT).

For context, the industrial software market is the glue that holds the various operational and management layers together and enables companies to manage their production systems to better effect. A detailed description of the industrial automation hierarchy can be found in the appendix, but simplistically industrial software provides the brains of the manufacturing system, connecting the factory floor with system management, the finance department and the supply chain

Chart 1: Industrial software growth by control layer USDbn)

Source: Merrill Lynch Global Research Estimates

Historically these systems were independent and isolated, but advances in industrial software and IT are removing the barriers and customers are demanding more integrated and simplified software solutions to enable them to improve productivity, derive greater levels of efficiency and reduce costs. Block-chain for example provides a standardised solution for companies to connect to each other (essentially a common ledger) and could have far reaching implications for supply chain efficiencies, inventory management and visibility/predictability of factory loadings.

Exhibit 1: BofAML –software in the context of our holistic industrial automation view

Source: BofA Merrill Lynch Global Research

0%1%2%3%4%5%6%7%8%9%

05

1015202530354045

Business Systems Design & Simulation OperationsManagement

Control & Supervisory

2014 2019 2014-19 CAGR (rhs)

Motion control controllers, drives &

motors, actuators

Operation platformIncl MES (manufacturing execution system)

Enterprise applications

Engineering & design applications

PDM (product data management), models, drawings, P&ID’s, eq specs, PLM (product lifecycle management)

SCM (supply chain mgmt), content, ERP (enterprise resource planning), CRM (customer resource mgmt)

PLC (programmable logic controller - discrete)

DCS (distributed control system – process)

Robotics - including related products & software

Control valves & related actuation

Fluid power -pneumatics and hydraulic control equipment

Field instruments –pressure, flow, level, temperature sensors, machine vision, test & measurement, analytical equipment

Data centres

ICS - Industrial control system

HMI hardware -human machine interface equipment

Motors, drives + mechanical power transmission - mainly LV & MV AC drives and motors (can incl generators)

Industry specific products – stirrers, brakes, stabilisers, turbochargers etc

Machine tools

Industrial interconnect equip.Communication – fieldbus, ethernet, wireless

LV products - cable mgmt, energy dist.

Pumps Material handling incl RFIDMetering

Industrial control

equipment

Safety Systems / SCADA

Hybrid automation Industrial PC’s

Advanced software applications simulation, asset mgmt, mobile, analysis


Software grows faster as customers seek efficiency

Industrial software spending is expected to grow 5-10% pa. Customer demands for improved operational efficiency are driving innovation in the industrial software markets with the benefits being improved time to market, better asset management, reduced costs and a more flexible manufacturing system.

Historically, automation architecture has been mostly concerned about increased volumes of more simple or generic products as a route to lowering costs (the mass production concept with significant economies of scale), but customers want the best of both worlds in desiring more complex or bespoke products at ever greater levels of operational efficiency. This is the concept of mass customisation, where the ultimate aim is to be able to produce a ‘one-off’ customised product as cheaply as it would be to mass manufacture a common product. In theory, industrial software enables companies to approach this ideal state (we are already seeing it in automotive production) and we think this is why software growth rates are far higher than those for traditional industrial automation hardware.

As with many markets, customer demand is a critical factor in development and the industrial software market is no different. Whilst the aim of improved operational efficiency has always been at the heart of industrial automation, software advances are enabling ever greater levels of productivity while at the same time offering the prospect of less compromise on complexity.

As a result of decades of development, the basic levels of industrial software have developed to high levels of efficiency. PLC, DCS, SCADA and HMI systems have been around for decades and are well established in automating manufacturing processes. The software elements of PLC and DCS are growing with subsequent product generations and growth remains strong, especially as emerging market requirements in terms of automation remain high however the relative size of the associated software markets is small. ERP systems have also been around for some time and constitute the largest industrial software market overall. Solid growth in the market is expected over the medium term.

Exhibit 2: In order to remain competitive, manufacturing companies need to drive the next generation in operational efficiency

Source: Siemens

The higher level of control, establishing operational management from the perspective of the entire manufacturing system within a business is still growing fast as it offers the next level of productivity gains. PLM, MES and EAM systems have really only been in development for the last 10-15 years and offer a more holistic and integrated level of control. Ultimately this is where the most productivity gains are likely to be found.

Working capital management and overall capital efficiency has been a critical focus for industrial management teams over the last decade, particularly as top line growth has

Reducing time to market

Product and production integrated

- Shorter innovation cycles- More complex products- Larger data volumes

Enhancing flexibility

Flexible production

- Individualised mass production- Volatile markets- High productivity

Increase quality

Full process transparency

- Closed loop quality processes- Traceability and integrated

genealogy

Increasing efficiency

Optimised production resource

- Energy efficiency and resource efficiency as key competitive factors


been harder to come by. SCM software is likely to continue to offer strong growth driven by customer demands for reduction in working capital, more efficient supply chains and improved quality control.

Chart 2: Software vs non-software growth vs industrial software market segments

Source: Foster & Sullivan; ARC Advisory

Discrete manufacturers identify their top three business initiatives as lowering costs, improving productivity and improving business processes. Industrial software, and PLM software in particular, does just that.

Chart 3: Discrete manufacturers identify their top three business activities

Source: BofA Merrill Lynch Global Research, IDC

In an increasingly competitive but stagnant macro environment, time to market is critical for companies seeking to ensure they retain market position and customer loyalty. PLM software reduces the time, the cost and the efficiency of getting new products to market as it enables the design and simulation of the entire product lifecycle from drawing board to full scale production and subsequent use. It can also be used to optimise the aftermarket support for the product once in the field.

MES have evolved from independent factory control systems to cover the entirety of a company’s manufacturing system, with the benefit that production can be optimally managed. Linking the production system with the supply chain and the production systems of other companies is the next logical step and area for improved efficiency, though there is inherent caution in data and system integrity in the face of cybersecurity concerns.

5%

6%

7%

8%

9%

10%

0

5

10

15

20

25

30

DCS PLC SCADA HMI PLM MES ERP SCM EAM

2014 2019 Software 2014-19 CAGR (rhs) Non-software 2014-19 CAGR (rhs)

0%

15%

30%

45%

Lower costs Improving productivity Improving business processes


Chart 4: Benefits of implementing PLM in terms of range of reduction vs increased productivity (rhs)

Source: CIMData

The lifeblood of many industrial companies is the installed base of products once they have been delivered and the aftermarket revenue stream they offer. Improved asset management software offers companies scope to improve the management of their own installed base of equipment as well as the equipment they use within their own manufacturing systems.

Ultimately the application of industrial software brings considerable savings in time, money and energy – to the benefit of both the industrial supplier and the customer.

Exhibit 3: Illustrative benefits of digitalization of technologies and the application of software

Note: TIA = Total Integrated Automation Source: BofA Merrill Lynch Global Research, Siemens

Software spending is expected to increase across all industries… Unsurprisingly given the potential benefits, industrial software spending is expected to increase across a wide range of industries. According to Gartner, global vertical software spending is set to grow 7% pa from USD384bn in 2014 to USD529bn in 2019. The manufacturing & natural resources industries are expected to represent the highest total amount with USD93bn at the end of the forecasting period.

Manufacturing and natural resources in the Gartner forecast covers a wide range of end markets to which our industrial companies are exposed and is likely a catch-all term for the broader industrial supply chain. Our coverage companies would also have considerable exposure to other industries identified by Gartner, whether it is

0%

5%

10%

15%

20%

25%

0%

20%

40%

60%

80%

100%Tim

e to m

anufa

cturin

g

Engin

eerin

g cha

nge

proc

ess

Desig

n rev

iew pr

oces

s

Prod

uct d

evelo

pmen

tco

sts

New

part

numb

es

Time t

o find

infor

matio

n

Desig

n erro

rs

Time t

o des

ign

Trav

el co

st for

desig

n

Incre

ased

prod

uctiv

ity(rh

s)

FUTURE OF MANUFACTURING NETWORKED ENERGYSmart grid technologies enable

Siemens TIA portal reduces the integration of renewableengineering costs by up to 30% energy sources into the grid at

40% lower costs

Self learning software canPLM software and automation get predict the electricity outputproducts to market up to 50% faster from renewable sources over a 72hr

period with 90% accuracy

INTELLIGENT INFRASTRUCTURE HEALTHCARE IT

Intelligent building technologies Intelligent data managementreduce energy costs by up to systems can reduce lab

40% test errors by 73%

Intelligent traffic solutions can reduce Software accelerates cardiactraffic jams, accidents and emissions CT exam reporting by anby up to 20% average 77%


datacentres for the banking industry (Schneider), smart grid solutions for utilities (Siemens, ABB) or connected healthcare solutions (GE, Philips, Siemens, Smiths).

Given the anaemic growth environment generally, the prospect of solid mid-single digit growth for the medium term is attractive. With limited recovery in the capex cycle, software enables improvement in capital efficiency and asset longevity when applied to advanced maintenance techniques. Management teams have been reluctant to commit to large scale capex plans this industrial cycle, but software investment is generally smaller amounts with quicker payback, with subsequent upgrades also easy to implement.

Chart 5: Software spending is expected to rise across all sectors (USDbn)

Source: Gartner *Software includes application

Chart 6: With average 2014-2019 CAGR of 7%

Source: Gartner

Traditional capex dynamics require high levels of factory utilisation (>80% has historically been the tipping point in the US and Europe), while the prospect of increased demand and a benign macro environment are also generally required before management teams commit to significant capacity expansion or updating existing factory capabilities. While the cost of capital is at historically low levels, capacity expansion has been an emerging market characteristic, not one for developed markets.

Chart 7: Global automation market by region

Source: BofA Merrill Lynch Global Research, Frost & Sullivan

Chart 8: PLM and MES market growth by region


The combination of capex trends, increasing cost inflation and increases in product complexity have driven fast adoption of industrial automation in emerging markets. At

0

20

40

60

80

100

Manu

factur

ing &

Natu

ral

Reso

urce

s

Bank

ing an

d Sec

uritie

s

Gove

rnme

nt

Comm

unica

tion,

Media

&Ser

vices Ins

uran

ce

Retai

l

Tran

spor

tation

Utilit

ies

Healt

hcar

e

Educ

ation

Who

lesale

Tra

de

2014 2019

0%

2%

4%

6%

8%

Who

lesale

Tra

de

Educ

ation

Retai

l

Healt

hcar

e

Manu

factur

ing &

Natu

ral

Reso

urce

sCo

mmun

icatio

n, Me

dia&S

ervic

es Utilit

ies

Bank

ing an

d Sec

uritie

s

Insur

ance

Tran

spor

tation

Gove

rnme

nt

-

5

10

15

20

25

NA EMEA APAC LaTAM

USD

bn

2015 2016

6.1%

4.5%

3.1%

4.5%

4%5%6%7%8%9%

10%11%12%13%

2011 2012 2013 2014E 2015E 2016E 2017E

MES-NA MES-EMEAMES-APAC PLM-NAPLM-EMEA PLM-APAC


the same time, the opportunity to implement latest generation of industrial automation and software means that emerging markets are as large as developed markets in terms of demand.

Interestingly, the manufacturing sub-segments exhibit the highest growth rates among all industries. The detail underlying Gartner’s sector forecasts show that manufacturing industries could see the fastest growth in software markets. Given the potential efficiency gains, the short payback periods and generally anaemic demand growth, the appeal of software investment over capital investment is likely to prevail over the medium term.

Table 3: Manufacturing sub-segments like automotive and heavy industries (shaded) are expected to show the highest growth rates in terms of sector-specific software USDbn 2014 2015 2016 2017 ‘11-17

CAGR Banking & Securities Banking 13.9 14.6 15.4 16.2 4.9% Securities 6 6.3 6.5 6.8 4.2% Communications, Media & Services

Broadcasting and Cable 3.7 3.8 3.9 4.1 3.6%

Publishing and Advertising 1.7 1.8 1.8 1.9 2.8% Telecommunications 12.5 13.2 13.8 14.5 4.2% Entertainment 1.4 1.4 1.5 1.5 2.8% Information Technology Services & Software 1.6 1.7 1.7 1.8 3.2% Other Business, Technical and Consumer services 2.2 2.3 2.4 2.4 3.4% Education Higher Education 1.7 1.8 1.9 2 4.8% Primary and Secondary Education 1.3 1.4 1.5 1.6 5.2% Government National & International Government 12.1 12.6 13 13.7 2.8% Local & Regional Government 9.1 9.4 9.8 10.3 2.8% Healthcare Providers Physician 1.7 1.8 1.9 2 5.5% Hospital 6.1 6.5 7 7.5 5.9% Insurance Health Insurance (payer) 1.7 1.8 1.8 1.9 3.6% Insurance (other than health) 6.9 7.1 7.4 7.7 3.7% Manufacturing & Natural Resources

IT Hardware 1.1 1.1 1.1 1.2 2.4%

Natural Resources & Materials 3.4 3.6 3.8 4.1 4.8% Energy Resources & Processing 0.9 0.9 1 1 3.6% Consumer Non-durable Products 2.6 2.7 2.8 3 4.8% Automotive 1.6 1.7 1.8 1.9 5.9% Life Sciences & Healthcare Products 1.3 1.4 1.5 1.6 6.5% Heavy Industry 2.9 3.1 3.3 3.5 6.2% Retail General Retailers 1.0 1.0 1.1 1.1 3.6% Grocery 1.0 1.0 1.1 1.1 4.1% Restaurants and Hotel 1.0 1.0 1.0 1.0 3.3% Speciality Retailers 3.0 3.1 3.2 3.3 3.7% Transportation Air Transport 1.5 1.6 1.6 1.7 3.6% Motor Freight 2 2 2.1 2.2 3.7% Pipelines 0.2 0.2 0.2 0.2 3.5% Rail and Water 1.1 1.2 1.2 1.3 4.3% Warehousing, Couriers, Support Services 1.1 1.1 1.2 1.2 4.3% Utilities Water Utilities 0.8 0.8 0.9 0.9 4.6% Electric & Gas Utilities 3.9 4.1 4.4 4.7 4.7% Wholesale trade Wholesale Durable and Nondurable Goods 1.8 1.8 1.9 2 2.3% Source: Gartner

Software providers show stronger organic growth Unsurprisingly given the dynamics we have described, bottom up analysis of the market place shows average growth among software firms is generally higher than for general automation vendors and also overall global industrial production. When compared to automation vendors, software players also show a lower degree of cyclicality. This is due in part to the nature of their licence-oriented business model which creates a more recurrent stream of sales.


Chart 9: Software companies exhibit higher organic revenue growth than automation vendors and overall industrial production…

Source: BofA Merrill Lynch Global Research Software – Ansys, SAP, Dassault, Autodesk, Aveva Automation – Siemens, ABB, Schneider, Emerson, Honeywell, Rockwell

Chart 10: …outpacing peers by c30% and c45% since 2007 on a cumulated basis…


Convergence of Operational Tech and Information Tech

The convergence of information technology and operational technology is a continuing theme for industrial automation vendors as the widespread adoption of smart sensors and advances in industrial software enable greater operational efficiency and better business processes.

With advances in industrial software enabling greater integration through the manufacturing process and supply chain, the competitive environment for incumbent industrial automation providers is changing.

Schneider and Rockwell Automation may have been the first industrial companies to talk specifically about the convergence of IT and OT, but the reality is this trend has been a driver of the moves made by incumbent industrial automation providers into areas such as PLM, MES and EAM. Siemens was the first industrial company to see the potential of PLM with its acquisition of UGS back in 2007, however most of the industrial players have added capability through acquisition over the last 5-6years. The software providers traditionally focused on ERP and SCM software have also started to develop their portfolios into PLM and MES, resulting in more direct competition with industrial companies.

The concept of convergence is driven by the fact that software and control capabilities within industrial automation were based around the traditional automation hierarchy (see Appendix A), with providers being well-established industrial automation companies like Siemens, Schneider and Honeywell. The business management software, provided by the likes of SAP and Oracle, has traditionally been kept entirely separate (normally run by in-house IT departments). However, customer requirements for faster time to market, greater productivity and flexibility, improved capital efficiency and even reduction in numbers of suppliers is driving a need for more holistic and integrated automation. PLM, MES and EAM software offers this level of integration by managing product lifecycles, entire manufacturing systems or the installed base of equipment and the penetration of the internet into the industrial automation arena provides the standardisation necessary to develop these holistic and integrated solutions to the next level.

-15%

-10%

-5%

0%

5%

10%

15%

20%

2008 2009 2010 2011 2012 2013 2014 2015

Software Automation Global IP

80

100

120

140

160

2007 2008 2009 2010 2011 2012 2013 2014 2015Software Automation Global IP


Identifying the actual software capabilities within the portfolio is critical, in our mind, to establishing which companies are best positioned to compete, so we have tried to triangulate various sources (companies, automation market consultants, our own estimates) to set up a framework to enable us to see which industrial companies actually have genuine software capabilities and can take advantage of the evolving automation landscape. While we accept that much of the Operational Tech (OT) software cannot be sold without the underlying hardware (e.g. PLCs or DCS) we have done the analysis this way to highlight the difference between what companies identify as their automation exposure and the reality of where their software capabilities lie.

Please note that our software analysis is intended to be quite strict in scope, identifying software business alone whether embedded or standalone, excluding both the related hardware and service revenue. The analysis focuses on purely industrial applications of operational and information technology software. Applications of software “outside the factory” such as aerospace (e.g. Honeywell avionics, GE avionics/jet engines) or healthcare (e.g. GE, Philips) are excluded.

Table 4: Estimated revenue derived from industrial software for main automation suppliers

2014, USDm OPERATIONAL TECH (embedded

software only) % INFORMATION TECH (standalone software) %

Software % automation

Software revenue DCS PLC SCADA HMI CNC TOT PLM MES PAM EAM SCM ERP TOT TOTAL revenue, BofAML

e SAP (incl. Concur) 889 81 133 2,050 5,595 100% 8,749 81% Oracle 277 66 90 1,161 2,456 100% 4,049 80% Siemens 412 165 173 194 214 35% 2,052 128 11 65% 3,348 31% Dassault 2,806 14 100% 2,820 89% Infor 74 83 1,281 100% 1,438 80% Sage 1,336 100% 1,336 80% ABB 500 277 66% 178 23 65 132 34% 1,175 25% PTC 1,062 57 100% 1,119 78% Microsoft 16 998 100% 1,014 80% Autodesk 904 100% 904 80% Ansys 749 100% 749 80% IBM 217 118 408 100% 744 80% Kronos 568 100% 568 80% Emerson 339 154 88% 19 50 12% 562 19% Honeywell 350 154 90% 53 3 10% 560 19% Schneider 58 51 170 50% 253 13 16 50% 560 18% Workday 491 100% 491 80% Hexagon 486 100% 486 80% Rockwell 133 127 115 82% 62 18 18% 455 12% Fanuc 349 100% 0% 349 20% Aveva 331 100% 331 81% Aspentech 176 101 100% 277 66% GE 20 80 55% 81 45% 180 20% Mitsubishi 152 100% 0% 152 5% Yokogawa 60 63% 34 37% 94 5% Valmet 42 100% 0% 42 12% Alstom 40 100% 0% 40 12% Source: BofA Merrill Lynch Global Research estimates, ARC, Frost & Sullivan

The overwhelming conclusion is that Siemens has the broadest and largest portfolio of industrial software. It is clear that most of the traditional automation providers (ABB, Emerson, Honeywell, and Rockwell) have portfolios that are still dominated by OT software, while Schneider has followed Siemens’ lead with a more balanced portfolio. The second largest industrial provider of IT (that we cover directly) is Hexagon, though Schneider’s position at the top of the MES market positions it well for the next stage in industrial software development. The IT behemoths unsurprisingly dwarf their industrial competitors due to the size of the ERP/SCM market, but it is worth noting the overlaps, with SAP having a sizeable PLM business and both SAP and Oracle having MES businesses larger than Honeywell and Rockwell.


Moving up the manufacturing value chain From their legacy capabilities in PLCs and DCS, the likes of Siemens, Schneider, Rockwell and ABB have moved to add greater capability to the basic building blocks of industrial automation. In addressing customer requirements for greater integration in processes and more efficient management of the manufacturing process, all incumbent industrial automation vendors have developed some element of PLM, MES and EAM capabilities.

Siemens was the first to move, with the acquisition of UGS in 2007, which it has built into a top 3 global PLM player. The peers have followed in various shapes and sizes to a point where Schneider has the largest share of the MES market (following its acquisition of Invensys) and ABB has developed good positions in the EAM and PAM markets, and has bought in ERP capabilities with the acquisition of Ventyx. GE and Emerson lead the way in terms of plant management with their PAM offerings.

Moving down the manufacturing value chain One of the best examples of this evolution in capability is Dassault Systemes. Starting from its traditional strong position in PLM, Dassault’s acquisition of Apriso in 2013 added significant scope to its MES capabilities in an effort to connect front-end virtual engineering and operations management across consumer packaged goods, life sciences or high tech. Autodesk’s acquisition of Delcam also shows vertical expansion from being a provider of virtual design capabilities into the field of digital manufacturing.

Dassault’s takeover of Quintiq, a provider of operation planning and optimization software, represents a move into supply chain management software. Quintiq’s solutions are used to plan and schedule complex production supply chain at manufacturer such as Lafarge, AkzoNobel or ArcelorMittal.

Exhibit 4: Dassault moved into the operations management space by acquiring Apriso and Quintiq


Expanding both up and down the manufacturing value chain We have separated this into its own category as both Hexagon and GE are pursuing relatively unique strategies in areas of the manufacturing value chain new to both companies. Hexagon essentially bridges the gap between what exists in reality and what exists on computer - linking the digital twin by measuring a physical asset and providing software to manipulate or analyse the data and link it to the original product design. GE has come from a position of being hardly present at all in industrial automation to being a pioneer in the development of a common industrial internet software platform to harness the benefits of big data analytics.

Both Hexagon and GE enjoy historical brand strength in industry, whether it is Hexagon's metrology capabilities or CAD/CAM software or GE's installed base. The latter in particular is a great advantage when looking to roll out new capabilities as the installed base provides a natural market. But the success both companies have enjoyed, particularly the speed with which GE has built an industrial software business, should provide some food for thought about the future competitive landscape for the incumbent industrial automation providers. More than ever, speed of innovation will be critical to the development and maintenance of competitive position.


Extending into new industry verticals The development of new software capabilities has brought all providers into new industry verticals, from Rockwell Automation leading the way in developing hybrid automation (combining both process and discrete solutions, e.g. in the food & beverage industry) or Hexagon entering the mining automation market to GE developing the Predix software platform from scratch and little previous industrial automation presence.

With its acquisition of Gemcom in 2012, Dassault entered into the mining space, providing geological modelling and simulation software for customers such as BHP Billiton or DeBeers. In 2014, Dassault also increased its footprint in the chemistry and biology field through the acquisition of Accelrys, a provider of PLM solutions for the pharma and biotech space. The move of expanding PLM solutions into new verticals might ultimately form the basis for further penetration of the manufacturing value chain as the product design and simulation generate vast amounts of data which can in turn be used to streamline the manufacturing process (through the application of MES software).

Capturing new technologies and niche markets: In 2014 Autodesk acquired Within Technologies whose Enhance software is used to optimize design and reduce weights for 3D printing. The software will be integrated into Autodesk’s Fusion 360, a single CAD tool for mechanical and industrial design. Furthermore, Dassault’s takeover of RTT, a provider of 3D virtualisation software and 3D marketing levered the company’s offering in virtual design as digital assets to reuse it also for marketing and sales applications. Through its acquisition of ThingWorx in 2013, PTC acquired software that enables manufacturers to build IoT applications to service their products in the aftermarket.

Exhibit 5: PTC’s Thingworx provides Acme Mining with a platform set up a remote maintenance application via the cloud

Source: PTC

Various industrial firms have also teamed up with software vendors to extend its capabilities and better integrate their software products into the business environment.

Table 5: Manufacturers are partnering up (examples, not exhaustive), illustrating the fact that the best way of exploiting the industrial internet opportunity quickly is to do so in collaboration

Company Partnership Description

GE Microsoft, IBM Offer cloud solutions for its GE's industrial applications

GE PTC Combine PTC's PLM software with GE's MES solutions


Table 5: Manufacturers are partnering up (examples, not exhaustive), illustrating the fact that the best way of exploiting the industrial internet opportunity quickly is to do so in collaboration

Company Partnership Description

Siemens IBM Offer an integrated solution of PLM software and enterprise applications

ABB GlobalLogix Offer ABB's SCADA software as a service via the cloud

Siemens Bentley Systems Strategic partnership focussing on process industries combining Siemens engineering solutions and Bentley's virtual design and manufacturing capabilities

Rockwell Dassault Strategic partnership offering Dassault’s design and simulation capabilities to Rockwell customers

Source: Company data, BofA Merrill Lynch Global Research

Only Siemens has a full PLM offering, though other companies might follow For now only Siemens has a complete offering in digital manufacturing. However as we have shown, the competition is not far behind, though the scale of Siemens' business is more than 4x the size of its nearest industrial competitor (ABB). As a first-mover Siemens will likely enjoy advantages over its competitors whose software capabilities remain nascent or small in scale.

Software vendors have not targeted the equipment level yet Software companies have so far stayed out of the software which directly operates the automation equipment (such as PLC and DCS) and supervisory and control equipment (SCADA). In our view this is due to the lack of product know-how compared to the automation manufacturers and also due to the strong competitive position that manufacturers hold (they can sell their products and embedded software as an integrated solution, guaranteeing seamless interoperability). Hence, we do not believe in any competition going forward for automation vendors in this area.

GE’s differentiated approach – trying to leapfrog the competition GE Digital’s efforts stem from the recognition that in the coming years, software is going to play an increasingly important role in optimizing the use of GE industrial products. Also, over the years while software was an integral part of GE’s service offering the company has arguably fallen behind its industrial peers in software capabilities related to manufacturing operations (See Table 4).

Back in 2011, when GE started its Digital efforts in earnest, there was no off the shelf product optimized for (or even capable of) dealing with large volumes of unstructured industrial data. GE Digital efforts evolved from a defensive move by the company to build a “moat” around the company’s parts & service offering to a more comprehensive strategy trying to position GE as the key Platform-as-a-Service (PaaS) and Software-as-a-Service (SaaS) player in the midst of the ongoing “digitization” of the industrial vertical, harnessing the company’s unique domain expertise and scale. The strategy has also evolved to providing applications for optimization of the manufacturing process, trying to leapfrog existing industry offerings in the area of industrial automation.

We believe that GE’s efforts to date likely secure the company’s position as a provider of key software as part of the overall service package for its installed base. The company’s ability to establish itself as a major PaaS player with its Predix platform selling to a broader range of customers, alongside other traditional and emerging players remains to be seen. Overall, we believe GE’s significant investments to date, first mover advantage, and deep (likely unparalleled) domain expertise on rotating equipment position the company well to succeed.

GE’s PaaS offering, Predix, will be the single most important driver of GE Digital’s growth strategy both as a standalone platform and a basis for development of applications by GE and others. Predix currently accounts for $50mn in sales (2016), but


GE expects Predix sales to grow to $4bn by 2020. We view Predix status as a standalone independent platform as the area with the biggest upside, but also the one with the largest degree of uncertainty.

GE initially attempted to harness big data for industrial applications using existing platforms, but the existing offerings were insufficient, requiring GE to build out its software capability around asset-generated data. Creating an integrated capability for managing assets is an opportunity to build out new digital systems for industrial and healthcare verticals where GE has deep domain expertise that could be monetized inside and outside GE. As we have noted, longer term, the challenge for GE will be to apply its expertise outside of its core areas of rotating equipment and healthcare imaging.

Below we describe key efforts by GE

PaaS: Platform as a Service refers to providing a platform or operating system that allows customers to run applications without having to build or manage their own infrastructure, thereby reducing time to market for application development as more of the backend is being managed. PaaS can be on a public cloud (i.e. Amazon’s AWS, Microsoft’s Azure, IBM Bluemix, Salesforce’s Heroku , or Google’s App Engine), where the cloud provider sells the networks, storage, server hardware, OS and middleware and database services while the customer focuses on the specific software and applications. In other words, GE does not know exactly how a customer wants to operate, so the customer can tailor things to its own needs, using Predix as a platform or ecosystem. PaaS can also exist on a private cloud behind a customer’s own firewall to maintain security and privacy. Pivotal’s open source Cloud Foundry is a leading example of private PaaS and is the platform GE uses to build Predix and IBM uses to build Bluemix. Hybrid PaaS is becoming more prevalent as customers seek the flexibility and capacity of the public cloud and the cost efficiencies and security of a private cloud.

GE’s Predix platform is a PaaS targeting the industrial sector, built on Pivotal’s Cloud Foundry architecture. GE has built a platform hosted in the cloud that customers can access to deploy industry-specific applications. While other PaaS providers focus on software development for consumer or enterprise markets (i.e. developing apps for smartphones or sales & marketing software for businesses), GE’s Predix platform is optimized for the development of industrial specific applications. While robust development of consumer and enterprise cloud computing applications has occurred over the past decade, industrial uses remain in the early stages. GE’s Predix cloud offering includes the data infrastructure; Cloud Foundry open source architecture; analytics, security, operations software. Using Predix machine software, Predix can connect infrastructure assets (wind turbines, aerospace engines, etc.) to a customer’s enterprise system (ERP) and provides applications that customers can access to understand asset utilization and other trends to optimize performance. The Predix platform allows customers and third-party developers to build applications that analyze industrial machine data. We note that the introduction of the Predix-like system will disrupt the existing set-ups within industrial enterprises for data management and analysis (i.e. ERP, PLM, DCS) as it will potentially allow for a more seamless integration of data and analysis. We think that GE is one of the leaders in the industrial PaaS field, but this is a strategy with lower visibility v. SaaS and will depend on GE being able to attract independent developers onto its Predix platform in the next 2 years. We estimate GE PaaS operating margin opportunity to be at 50%+.

IaaS: Infrastructure as a Service refers to suppliers of computers (typically virtual machines – VM), servers, storage and networks. IaaS provides virtual hardware with adjustable scalability, with the customer paying for computation and storage capabilities. One of the key differences between IaaS and PaaS is that with IaaS, the customer still manages the server itself, while with PaaS, the server is done by the provider. Oracle, Microsoft, AWS, Google, IBM are key IaaS providers, with AWS holding the majority of the market.


In 4Q15, GE migrated its own software and analytics from AWS to the Predix Cloud. GE has noted that the Predix Cloud would be commercially available as an IaaS in 2016. The Predix Cloud is unique in its targeting of industrial customers and offering industrial oriented services (i.e. industrial specific security layer), which is the key differentiator for GE’s IaaS. Unlike AWS, which is a public cloud, GE’s Predix Cloud offers a private, managed cloud. This allows for greater security but also for shared services. However, customers can decide to deploy Predix on private clouds or on public clouds such as AWS. Compared to the cloud infrastructure strategy of Amazon, Google and Microsoft we view GE Predix Cloud as more of a niche effort, targeting specific needs of its industrial customers who might have different security and location requirements v. the public clouds.

Stack: A software stack is the full suite of SaaS, PaaS, and IaaS, from virtual machines, storage, networking, platform and software. In the past, prior to the emergence of cloud computing, companies would manage the full stack themselves, from networking through applications. With the emergence of IaaS and PaaS, companies can outsource the backend and focus on the application development and data. Some providers offer a full stack (i.e. Oracle), though most focus on certain parts of the stack and allow for integration where there are gaps (see below).

Predix is GE’s PaaS and IaaS layer. GE partners with leading technology companies to build out the rest of its stack. GE has listed Amazon, Cisco, Dell, EMC/VMware (through Pivotal), HP Enterprise, Oracle, and PTC as key technology partners to promote and support its Industrial Internet initiative.

Predix is GE’s PaaS and IaaS layer. GE partners with leading technology companies to build out the rest of its stack. GE has listed Amazon, Cisco, Dell, EMC/VMware (through Pivotal), HP Enterprise, Oracle, and PTC as key technology partners to promote and support its Industrial Internet initiative. GE can provide customers any range of cloud offerings, from simple SaaS, such as Asset Performance Management (APM), PaaS through Predix, and IaaS through Predix Cloud. Customers can decide how much they want from GE depending on scale and applications needed.

Table 6: GE Digital stack offering

SaaS Applications

Vertical Solutions

• Power generation & distribution (utilities, grid, renewables) • Oil & gas (APM, ops optimization) • Healthcare • Transportation (Aviation, rail) • Manufacturing (process, discrete, automation, MES) • Building management (HVAC, lighting, elevators) • Resources (water, mining)

Horizontal Solutions

• Automation (controls, SCADA, HMI) • APM (APM, availability maintenance, optimization) • Ops Optimization (logistics, fuel efficiency, service automation, scheduling, production optimization, field force intelligence) • Industrial Data Management (time-series data, analytics, historians) • Industrial Cyber Security (monitoring, threat detection, managed services/certification) • Managed services (data-driven equipment monitoring & remote diagnostics, predictive maintenance, service delivery, production optimization, security ops)

PaaS / IaaS Industrial Cloud

• "Community" cloud designed for industrial scale, security and data governance • Device connectivity and management from "edge-to-cloud" • Modern DevOps developer tools and processes • Big data and analytics capabilities optimized for industrial machine data • "Digital Twin" asset models linked to analytics and industrial workflows

Source: GE Company reports


GE has numerous technology and consulting partners to roll out the Predix platform. IT groups its partners into 3 main categories: 1) consulting services, 2) technology partners that work in conjunction with GE (i.e. Oracle, SAP, AWS, among others), 3) technology partners that are used as ingredients in GE Digital’s SaaS and PaaS offerings (i.e. Pivotal’s Cloud Foundry, PTC’s ThingWorx, VMware).

While GE Digital leverages GE’s existing installed base and key customer relationships, partners, including PWC, Infosys, Cisco, Tata Consulting, HP Enterprise, Wipro, Deloitte, among others, help develop new outcomes for customers by developing new apps for the Predix platform and driving app sales. GE has signed 31 partners YTD and targets 50+ by the end of ’16.

Table 7: GE Digital partners Partner Description Accenture Aiding in designing, developing, implementing Industrial Internet solutions for clients Altran Development of industrial apps for asset optimization, enterprise optimization and intelligent environments Amazon AWS is first cloud provider for GE's Industrial Internet platform AT&T Supplying IoT infrastructure to allow enterprises to connect assets with Predix and software BearingPoint Aiding in designing, developing, implementing Industrial Internet solutions for clients Bell Supplying IoT infrastructure to allow enterprises to connect assets with Predix and software Capgemini Leverage Predix and big data analytics for clients, 200 developers building Predix apps China Telecom Accelerate development of innovative Industrial Internet applications that will support China's industry transformation to Internet Plus Cisco Partnership to deploy Brilliant Factory architecture, building digital roadmaps, helping customers digitize Cognizant Digital Works - Foundry Methodology + Predix helps clients design digital transformation for clients Dell Develop use case blueprints to help customers accelerate IoT projects Deloitte Drive digital industrial transformation for mutual clients engro group A key digital industrial partner for GE in Pakistan, develop power gen apps on Predix Equinix GE Predix clouds will exist within Equinix data centers EY Aiding in designing, developing, implementing Industrial Internet solutions for clients Genpact End-to-end offering and joint go-to-market model to help clients transform through accelerated Industrial Internet adoption HP Enterprise Preferred storage and server infrastructure provider for Predix cloud technologies Infosys Developing new applications for Digital Twin, Brilliant Factory and AI for clients Intel GE software is optimized for Intel products and ecosystem helping spread intelligence across factories, machines Meridium Strengthen GE’s grid and asset management optimization offering Oracle Oracle Cloud Platform, Oracle Cloud Applications and ERP and Supply Chain Management applications can use Predix data Orange Supplying IoT infrastructure to allow enterprises to connect assets with Predix and software Pitney Bowes Develop customized APM applications for Pitney Bowes and its Enterprise Business Solutions customers Pivotal Predix built on open source Cloud Foundry, in-memory, Hadoop-based tech PTC ThingWorx-plant performance software available in GE's Brilliant Manufacturing Suite pwc Deliver end-to-end solutions that help clients achieve business outcomes Softbank 1st to license Predix, introduced IoT solution for LIXIL which optimized the installer-arrangement for home bathroom installation business Softtek Building apps on Predix, helping customers drive APM, Brilliant Manufacturing Tata Consulting Creating digital ecosystems using Tata's domain expertise and digital assets to enrich Predix Tech Mahindra Early adopter of the Predix platform, create innovative solutions that will address the need of the Power, Oil and Gas, and Transportation industries Verizon Securely connecting smart cities, railways and manufacturing plants to the Predix cloud, also Predix Connectivity-as-a-Service collaborator Vodafone Securely connecting smart cities, railways and manufacturing plants to the Predix cloud, also Predix Connectivity-as-a-Service collaborator Wipro Helping manufacturing clients with digital transformation, enabling apps on Predix Source: BofA Merrill Lynch Global Research, company report

Predix roll out timeline GE targets 20k+ developers on Predix by the end of 2016, with over 200k assets under management. GE aims to expand its partner list from >15 currently to 50+ and to generate $7bn in orders (up 30%). GE’s $15bn Digital sales target in 2020 is based on 100k+ Predix developers.

Table 8: GE Digital roll out 2016 2017 & beyond

Predix

• Release for general availability • Extend asset service to edge • Integration into edge management portal • Unified catalog • Digital twin enhancements • Ecosystem capabilities • Analytics library

• Edge for mobile • Extended analytics library • Ecosystem usage and social feedback tools and dashboards

APM • APM on Predix release • Maintenance optimization release

• APM 3.0 with knowledge management


Table 8: GE Digital roll out 2016 2017 & beyond

Manufacturing Suite • Efficiency analyzer on Predix release • WIP manager on Predix • Quality manager on Predix

• Routing Manager release

Mobile workforce/Ops Optimization • Operations Optimization beta • Operations Optimization general availability

• Mobile workforce release to general availability

Source: GE company report

Software companies offer superior metrics to industrial automation vendors In terms of profitability providers of software solutions perform better than their industrial peers. This is due in part to the higher proportion of recurring revenue and the higher value software has historically brought to the customer. Quick payback times help as well. Having said that, if we could split the industrial software businesses out from the broader automation revenue we believe industrial companies still "under-earn" on their software revenue. This is in part because software has historically been a bonus rather than a focus and is often part of a broader service offering. As industrial companies become better sellers of software we believe margins could start to improve.

Chart 11: …and also outperform the latter in terms of operating margin (FY 2014)


Software firms run their business more profitably than industrial challengers Siemens and Schneider run their software businesses less profitably than Hexagon and Dassault, we believe due to the high level of “service” attached to the business and differences in the way the companies go to market. This brings into question the effectiveness of manufacturers like Siemens to run software businesses alongside their manufacturing operations and highlights the domain expertise of software companies.

0%

10%

20%

30%

40%

50%

60%

Ansy

s

SAP

Dass

ault

Avev

a

PTC ….

Rock

well

Emer

son

Schn

eider

Hone

ywell

ABB

Siem

ens


Chart 12: Siemens runs its PLM software business on a lower EBIT margin than Hexagon and Dassault


We therefore think that industrials will put more focus on the way they operate their software capabilities to unlock the high margin potential and justify high the acquisition multiples paid. It is one of the reasons why industrial companies will acquire software companies in the first place - buying in the capability to run and sell software and was, we believe, one of the drivers for Schneider's attempted acquisition of Aveva last year.

Chart 13: Implied return on capital from software acquisitions is low.


Table 9: …as industrials pay high acquisition multiples Date Acquirer Target Deal Value EV/Sales EV/EBIT

Jul-13 Schneider Invensys £3.4bn 1.3x 16.0x

Nov-12 Siemens LMS International €680m 3.6x 23.9x

Jun-11 Schneider Telvent $2.0bn 2.6x 15.9x

May-11 ABB Mincom $450m 2.25x 25.0x

Jul-10 Hexagon Intergraph $2.125bn 2.8x 13.4x

May-10 ABB Ventyx $1.0bn 4.0x 33.3x

Jan-07 Siemens UGS $3.5bn 3.0x 15.9x

Source: Company Data

Software has traditionally traded at a premium to capital goods With respect to valuation, software companies have historically traded at a premium to their automation hardware peers. While we acknowledge that the 12 month forward P/E ratio is not the optimal metric to compare the two segments due to differences in taxation, accounting standards or financial leverage, it should still give a broad indication of valuation.

0%

5%

10%

15%

20%

25%

30%

35%

Siemens Software Schneider Software Hexagon Dassault

Operating Profit Margin

0%

1%

2%

3%

4%

5%

6%

ABB Siemens Schneider


Chart 14: Software provider have historically traded on a premium to automation vendors (12 month forward P/E)

Source: BofA Merrill Lynch Global research

Case Study: Daimler switching to Siemens from Dassault

A major example of vertical convergence and direct competition between software companies and automation vendors is illustrated by Siemens taking a major contract with Daimler from Dassault on the strength of its integrated capabilities.

In late November 2010 German automotive OEM Daimler decided to switch their CAD software which it uses to virtually design their new cars and trucks from Dassault’s CATIA to Siemens PLM’s NX. A main reason for Daimler deciding for a shift was the difficulty of integrating Dassault's latest CATIA V6 version with Daimler’s proprietary product database, a backbone for its engineering process (which relied on Siemens’ PDM solution Teamcenter).

0x

4x

8x

12x

16x

20x

24x

28x

06 07 09 10 11 12 13 14 16

Automation Software


Exhibit 6: Siemens NX 7

Source: Siemens

Exhibit 7: Siemens Integrated Solutions for Automotive OEMs

Source: Siemens

At the time when the decision took place Daimler had different choices on hand on how to deal with the integration problems of CAD and PDM software:

• Combine Teamcenter and CATIA V6 leading to significant costs and the constant culture clash between two competing technologies

• Make a switch from Teamcenter to Dassault’s Enovia V6 translating into twofold sorts of costs. Daimler would have needed to pay for the new software and incur further expenses relating to installations, translations, maintenance and training.

• Stick with Teamcenter and switch to Siemens PLM’s NX paying for the new CAD software

Daimler’s decision in favour of Siemens is less a judgement on the product quality of the two competitors, but rather highlights the importance of interoperability. Daimler prioritized its PDM platform as IT backbone to which the CAD solution has to adapt. Hence, we believe it is important for software vendors that they ensure an ease of integration of their solutions into the product space supplied by automation manufacturers and in general into heterogeneous IT environments.


Table 10: While Dassault still dominates the CAD space, Siemens is the top supplier in the CAM and PDM field for the automotive industry

Automaker CAD PDM CAM Toyota Dassault CATIA

PTC Creo Internal Dassault DELMIA Siemens Tecnomatix

GM Siemens NX Siemens Teamcenter Siemens Tecnomatix

Volkswagen Dassault CATIA PTC Creo Siemens Teamcenter Siemens Tecnomatix

Ford Dassault CATIA Siemens Teamcenter Siemens Tecnomatix

Hyundai Dassault CATIA PTC Creo

PTC Windchill Siemens Teamcenter

Dassault DELMIA Siemens Tecnomatix

PSA Dassault CATIA Dassault ENOVIA Dassault DELMIA

Honda Dassault CATIA Dassault ENOVIA Siemens Teamcenter


Nissan Siemens NX Siemens Teamcenter Dassault DELMIA Siemens Tecnomatix

FIAT Siemens NX Siemens Teamcenter Siemens Tecnomatix

Suzuki Siemens NX Siemens Teamcenter Siemens Tecnomatix

Renault Dassault CATIA Dassault ENOVIA Siemens Tecnomatix

Daimler Siemens NX Siemens Teamcenter Dassault DELMIA Siemens Tecnomatix

BMW Dassault CATIA PTC Creo

Internal Siemens Teamcenter


Volvo Dassault CATIA Siemens Teamcenter Tecnomatix

Jaguar Land Rover Dassault CATIA Dassault ENOVIA Siemens Teamcenter Dassault DELMIA


Siemens holds a powerful position in both CAM and PDM systems, whereas on average automotive OEMs still prefer Dassault’s CAD solutions when it comes to virtual designing. The example of Daimler shows how a strong position in PDM and CAM can be an opportunity to cross-sell CAD software.

Also Volvo’s IT Director Andreas Westholm stated at the beginning of the year that:

“Siemens Testcenter is the PLM/PDM backbone; we will not use Dassault's V6 if it requires double installations”

Volvo is currently still operating with CATIA V5 and manages its data in Teamcenter. The company has no plans to change their CAD software at this time. However, it is another proof that automakers emphasize the importance of its PDM applications and rather choose to adapt other areas to it.

Nonetheless, there are also examples of contrasting approaches. Jaguar Land Rover decided in 2010 to migrate to Dassault’s V6-architecture which should ultimately lead the way to the adoption of 3DEXPERIENCE, Dassault’s platform that provides software solutions for all business units within a company.


2. Industrial internet changes the game

Whether called the Internet of Things, the Industrial Internet of Things (IIoT), the Industrial Internet, Industry 4.0, the Internet of Everything or the Web of Systems they all boil down to one thing; the concept of the industrial internet changes the game for everyone, manufacturers, businesses and consumers. A bold statement, but in the same way that the PC on which we are writing this report changed the way we do business beyond all expectation, so too the advent of the industrial internet.

The potentially disruptive impact on the industrial sector, both from a revenue and cost opportunity, should not be underestimated. There are many estimates for the potential economic benefits of the industrial internet, all of which we suspect will prove inaccurate, but consensus appears to be for in excess of USD5-6tn pa by 2025, 75% of which our coverage companies in Europe and the US could access in some form or other - a large opportunity!

The fourth industrial revolution sounds like a grand concept, but the reality is that the next step in the evolution of the way industrial companies do business is a function of both competitive and customer pressures. Competitive as industrial automation vendors are coming into ever greater contact with new competitors in the form of information technology suppliers as well as more nimble, smaller software developers. Customer requirements for improved productivity, improved business processes and greater capital efficiency force manufacturers to reconsider their product capabilities, their approaches to market and their competitive positions.

Best placed - Siemens, Hexagon, GE As first movers in Industry 4.0 and the Industrial Internet respectively, Siemens and GE are likely best placed to harness the power of the industrial internet opportunity AND see that reflected in the investment thesis. Hexagon has the highest proportion of revenue derived from software of any of our coverage universe and should therefore almost be a default consideration for any investor considering how to invest in the industrial internet from an industrial perspective.

Quick note on nomenclature As a preface to our discussion, we will refer hereafter to the industrial internet as a broad concept, in lower case, to avoid confusion. The Industrial Internet Consortium and Platform Industry 4.0 recently agreed at the Hannover Messe industrial fair that they were both essentially talking about the same thing, just coming from different perspectives. Industry 4.0 comes from manufacturing roots, i.e. bottom up in our minds, whereas the Industrial Internet Reference Architecture (IIRA) has a more cross-domain oriented approach, or top down in our minds. Either way, what we are really talking about is the ability to connect equipment and systems of equipment to other equipment and systems to enable a real-time, multi-way communication of data, analysis and response.

Assessing the potential economic benefit of the industrial internet The most recent comprehensive study by Mckinsey's Global Institute (June 2015) estimates the economic benefit of the industrial internet to be between USD3.9tn and USD11.1tn per annum by 2025, but this is just the latest in a long line of studies and estimates that have been growing progressively larger. Helpfully, McKinsey have broken the larger number down in to various blocks of value according to industry sector, and we believe our coverage companies could touch c75% of that value. Given there has been a lot written on the subject of the industrial internet we are not seeking to simply repeat the basic explanations here. Rather than focus on one study or piece of research


in particular, the assessment we have done to put the industrial internet in context is to examine the progress of estimates.

Many studies were based on the easiest concept to start with - connecting machines. Again, estimates started high (Cisco) but have now levelled off somewhat to what appears to be consensus that c50bn machines and devices will be connected by 2025. As the value potential estimates have increased, though, the potential economic value per connected thing has increased rapidly reflecting the evolution in thinking about how to get the most from making the connections in the first place.

Chart 15: Independent estimates for the potential size for the industrial internet in terms of value have accelerated again in recent reports but the number of things connected by 2025 is relatively consistent

Source: BofA Merrill Lynch Global Research, Cisco, PwC, McKinsey, GE, Gartner, IDC, Machina

Chart 16: The economic value per connected thing has therefore improved rapidly through recent research reports by independent consultants

Source: BofA Merrill Lynch Global Research, Cisco, PwC, McKinsey, GE, Gartner, IDC, Machina

There have not been many follow up studies which have updated surveys or estimates, but PwC has published two surveys, one in 2014 and one in 2016. It is instructive to compare the change in awareness of the industrial internet between the two surveys. Manufacturing companies appear to have seen the greatest advances in awareness in the last two years, though the biggest step up in expectations in spend have come the in electronics industry.

Chart 17: Comparing awareness of the industrial internet by industry between PwC surveys in 2014 and 2016

Source: BofA Merrill Lynch Global Research, PwC

Chart 18: Comparing % of annual revenue expected to be spent on the industrial internet between PwC surveys in 2014 and 2016

Source: BofA Merrill Lynch Global Research, PwC

We do not find the change in expectations around spend particularly surprising - the electronics includes much greater exposure to consumer markets, which tend to adopt new technologies much quicker than businesses do. However, we are encouraged by the

0

50

100

150

200

250

02468

101214161820

Intel

2009

Cisc

o 201

1

GE 20

12

Mach

ina 20

12

McK

2013

Gartn

er 20

13

IDC

2014

McK

2015

Cisc

o 201

5

Market Size (lhs, USDtn) Things Connected (unit bn)

$0$50

$100$150$200$250$300$350$400

Intel

2009

Cisc

o 201

1

GE 20

12

Mach

ina 20

12

McK

2013

Gartn

er 20

13

IDC

2014

McK

2015

Cisc

o 201

5

Value per Connected Thing

0%

10%

20%

30%

40%

50%

60%

70%

Manufacturing Automotive ProcessIndustries

Electronics

2014 2016

0%

1%

2%

3%

4%

5%

6%

7%

8%

Manufacturing Automotive ProcessIndustries

Electronics

2014 2016


increase in awareness among manufacturers that the industrial internet is something they need to think about.

More than just smart manufacturing McKinsey Global Institute estimates the economic value of the industrial internet to be between cUSD4tn and USD11tn per annum by 2025, with factories and manufacturing generally accounting for the largest part of that - between USD1.2tn and USD3.7tn pa. The reality of the end market exposure of our coverage universe though is that the potential value they could access is much broader in scope.

Chart 19: McKinsey estimates the economic value of the industrial internet to be between USD3.9tn and USD11.1tn pa by 2025, with factories accounting for the bulk

Source: BofA Merrill Lynch Global Research, McKinsey Global Institute

If we map the sub-sectors identified by McKinsey to the end market exposure by revenue of BofA Merrill Lynch industrial coverage universe, we believe our companies could potentially access some 75% of the USD4tn-USD11tn in value McKinsey identifies.

200 210

930 560 170 160 410

1,210 350

740

1,660

850

1,590

930 1,160

3,700

-

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

Home Vehicles Cities Outside Human Worksites Retail Factories


Exhibit 8: Mapping industrial company exposure to the end market value opportunities for the industrial internet

Source: BofA Merrill Lynch Global Research, McKinsey Global Institute

Greatest demand from emerging markets Regionally, the greatest demand according to IDC in terms of both connected devices and related spend appears likely to come from emerging markets, and Asia in particular. However the spend per device is expected to decline the most in Asia. We assume this is due principally to the cost of the actual devices, but there may also be regional differences in terms of services or usage of the connections.

Near term growth is expected to be very strong, with device connections growing north of 25% pa in the next couple of years in EMEA as Industry 4.0 initiatives pick up.


Chart 20: The number of connected things worldwide is excepted to be dominated by Asia, in USDmn

Source: IDC

Chart 21: … while growth in connected devices is expected to exceed 15% pa for the medium term

Source: IDC

Growth rates in terms of USD value are expected to be slightly slower, reflecting the price declines in the cost of devices themselves particularly on the electronics side, but the regional mix is expected to become more balanced over the medium term with Asia currently estimated to account for c55% of total annual IoT related expenditure falling to 47% in 2020.

This could favour European companies over their Asian and US counterparts of the near term in terms of growth prospects, but ultimately we believe there is enough opportunity for everyone at present. Longer term, perhaps the weight of spending in Asia may mean domestic competition develops faster. At present, the dominance of Asia in the mix of spending is most likely to due to the willingness to adopt new technologies, particularly from the standpoint of Greenfield investment, whereas developed market customers have to change the way they think or operate and then make adjustments for their existing installed base, capabilities or portfolio of products.

Chart 22: IDC estimates the annual IoT related expenditure will grow to over USD1.4tn by 2020…

Source: IDC

Chart 23: … with annual growth rates remaining well north of 10%

Source: IDC

Near term spending dominated by sensors/devices, but growth is in software The current investment in the industrial internet remains focused on the "blocking and tackling" of simply connecting devices, machines and systems to each other and developing the data capture, analysis and response capabilities to deal with the new systems. In the early stages, sensor/module investment accounts for some 26-27% of total spend, dropping to sub-23% in 2020 and likely falling further in the years beyond.

-

5,000

10,000

15,000

20,000

25,000

30,000

35,000

2014 2015 2016 2017 2018 2019 2020

Asia/Pacific EMEA Americas

0%5%

10%15%20%25%30%35%40%

2015 2016 2017 2018 2019 2020

Asia/Pacific EMEA Americas Worldwide

- 200 400 600 800

1,000 1,200 1,400 1,600

2014 2015 2016 2017 2018 2019 2020

Asia/Pacific EMEA Americas

0%

5%

10%

15%

20%

25%

30%

35%

2015 2016 2017 2018 2019 2020

Asia/Pacific EMEA Americas Worldwide


Chart 24: Analytics software spending is expected to grow at 24%, the fastest among all types of spends, in USDbn

Source: IDC

The development and application of software is critical to the ability to maximise the value derived from connectivity. IDC forecasts growth in software investment to be c400bps pa faster than the growth in sensor/device investment over the 5yrs 2015-2020 and by the software market will be c15% bigger than the market for sensors/devices by 2020 (compared to similar size today). While growth in the investment in sensors/devices is expected to be good (c14% CAGR 2015-2020), there is clearly a normalisation effect as the number of devices connected reaches saturation levels. Software can continue to evolve and offers longer term growth prospects to those companies innovative enough and fast enough to get involved.

Industrial internet drives evolution in industrial automation

Industrial automation companies are being forced to be more nimble as the industrial internet offers flexibility, greater integration and improved productivity and opens up the competitive landscape

The industrial internet affects traditional automation systems through the integration of industrial devices onto the enterprise network. The increased application of Ethernet technology allows wider integration of connected devices and control systems and opens the factory or process plant to the opportunities of the internet. In itself, automation networks already provide the proper setup and infrastructure to extract the benefits from the industrial internet, but the several methods of machine/system communication that have evolved over the years has led to different standards and specifications that make it difficult to switch between or communicate between different suppliers' product offerings. Customer requirements are changing, with efficiency and productivity at their core and industrial automation providers need to evolve accordingly. Ethernet and the advent of the industrial internet offer the opportunity to work to a global set of standards.

-

200

400

600

800

1,000

1,200

1,400

1,600

2014 2015 2016 2017 2018 2019 2020

Module/sensor Connectivity Content service Application software

IT services Platforms Analytics software Security software

Servers Storage Security hardware Other hardware


Exhibit 9: The traditional automation system is at the heart of a broader IoT setup


Increasing move to industrial Ethernet Traditional communication across automation networks has developed through different communication protocols within the plant. Even within the same product layer, due to different industry specific demands and the proprietary solutions of large manufacturers, several bus solutions have emerged. Standardisation was started in the 1990’s to relieve the problem, however established fieldbus technologies already have such a strong position that the result is a number of standards instead of one.

Exhibit 10: Conventional industrial network vs industrial Ethernet network

Source: Renesas,

Increasingly, however, the market is moving towards industrial Ethernet across automation networks, which uses standard computer networking communication technologies. Typically industrial Ethernet allows integration into the office environment, connecting and addressing more devices over a wider area, using real time data, with more standard equipment and stronger inter-operability. There are a number of industrial Ethernet protocols, with Profinet, EtherNet/IP and Modbus TCP/IP the most popular (differences largely at the protocol layer).

Changing competitive landscape in industrial automation

Software and technology companies will compete with automation vendors going forward as IT and OT converge. Some of the incumbent suppliers of industrial automation are well positioned, while knowledge


acquired in building the industrial assets in the first place ought to provide early competitive advantages if used appropriately.

As the industrial internet adds a software layer on top of the typical automation setup, traditional automation vendors are moving from providing more than just the physical equipment to offering software/service based solutions for their customers. However, as the industrial internet value opportunity lies in an area of overlap between manufacturing and software, incumbent industrial automation players are facing competition from the software and technology firms that have traditionally dominated the IT layer in the industrial automation pyramid.

Having said that, industrial automation providers like Siemens and GE score well in recent surveys about competitive positioning in the provision of industrial internet offerings. Many companies are resorting to partnerships to enable quicker development of product offerings, largely in recognition that different skillsets are required to maximise the value opportunity, however we are wary that this may empower the IT providers with sufficient underlying knowledge of the physical assets that the one initial competitive advantage the incumbent industrial automation providers have is essentially lost.

Table 11: Top twenty most important IoT firms according to IoT Institute survey of its members, April 2016 Name % of votes IoT Platform Survey comment M&A Google 42% While the company has invested aggressively in its cloud platform it trails Amazon, Microsoft and IBM in adoption Nest, Dropcam

Microsoft 42% Azure IoT Suite Its Azure IoT Suite is second only to Amazon in terms of the size of its cloud market Cisco 35% Cisco IoT System Emerging as a leader in Smart Cities Jasper

Amazon 27% AWS IoT Growing number of partners eg Uber, Fitbit, Spotify, Nest. Cloud division revs in 2015 were USD8bn IBM 26% Watson IoT Launched global IoT headquarters in Munich in Dec 2015. Connected car issue w/ Continental

Intel 26% Launched IoT business unit in 2013 Arynga, Yogitech

Siemens 18% Mindsphere Recently launched Cyber Security Operation Centres in Lisbon, Munich and Milford, OH CD Adapco AT&T 15% Industrial clients include Tesla, Audi, ROK, EMR. Connected car business has cUSD1bn revs

GE 15% Predix Targeting to be a top 10 software company by 2020; partnered w/ Microsoft, IBM and Intel Honeywell 15% Niagara 4 Owns Tridium; partnered w/ McAfee on cyber security solutions platform

HP 12% Edgeline IoT System HP Enterprise focused on edge computing technology

Oracle 11% More suited to remote equipment monitoring, maintenance for industrial applications and asset tracking Rockwell Automation 10% Partnered with Microsoft, Cisco and Symantec

Schneider 10% Open software designed for the IoT—primarily for smart cities and asset management applications. OpenFog member

Texas Instruments 9% M2.COM Recently, TI collaborated with Advantech, ARM, Bosch Sensortec, and Sensirion to develop the M2.COM IoT sensor platform

Johnson Controls 9% Panoptix Leader in IoT applications of HVAC and energy storage; Tyco has developed a prominent IoT platform for fire, security, and retail uses

SAP 7% SAP HANA Cloud Partnered w/ Siemens and Intel

Verizon 7% ThingSpace Focused on building IoT tools. Developed inexpensive LTE modems to enable developers to link to Verizon network

Dell 7% Edge Gateway computer pioneering edge computing and "distributing" the capability. Strategic partnership w/ 25+ other IoT participants

Bosch 6% Bosch IoT Cloud c5m devices already linked to software suite (source: Reuters) Source: IoT Institute Survey 2016

Industrial companies such as Siemens and GE are at the forefront of the development of the industrial internet. GE has made considerable effort to develop its Predix platform and Siemens was the first mover of the traditional providers of industrial automation to develop PLM and the concept of the smart factory (with the acquisition of UGS in 2007). Companies such as Honeywell, Schneider and Rockwell Automation have also developed industrial internet offerings, often partnering with their IT competitors to mitigate gaps in their skill-sets.

Software companies such as Dassault Systems and PTC have entered the industrial software space. PLM technology is already an integrated solution for companies to be


able to manage the design, production and lifecycle of their products and is a natural platform for the expansion of industrial internet capabilities as at its heart is the concept of a virtuous circle of communication and data analysis.

Technology firms like Google, Amazon or Cisco are levering their traditional capabilities to develop solutions for the industrial internet. Often these are based on their existing cloud infrastructure offerings, extended by services for advanced data analytics to increase asset efficiency for clients.

An example for these offerings is Microsoft’s Azure which leverages cloud technology for asset monitoring and advanced asset analytics. ThyssenKrupp teamed up with Microsoft to increase uptime of its elevators. Together with independent consultancy firm CGI, Microsoft created a connected asset monitoring system on the cloud to analyse data coming from sensor-equipped elevators. The benefits for ThyssenKrupp were increased reliability through predictive maintenance and rapid remote control diagnostics.

Domain expertise will be key for success When we compare the different players, we believe that automation vendors have edge over their software/tech peers due to their knowledge of the installed equipment and hence also the data to be analysed. Despite software players’ superior know-how of analytical capabilities, we think that the crucial step of extracting efficiency gains and value from the industrial internet setup is to properly contextualize the data coming from machines. This is further underpinned by a survey conducted by Cisco which showed that companies value deep industry knowledge over analytics cross-industry experience.

Exhibit 11: Automation vendors dominate the competitive landscape as industrial domain expertise is key for success


The critical step for industrial companies will likely be a change in mind-set. Industrial companies are still traditional very siloed in terms of expertise, routes to market and customer relationships with a concept that IT is dealt with by the IT department. Overcoming this mind-set and harnessing the breadth of latent expertise will be crucial to industrial companies' success and ability to extract value from the industrial internet.


Chart 25: Industry knowledge dominates analytics capabilities

Source: Cisco

Industrial internet main applications and examples The main industrial internet applications can be summarised into two main categories: asset optimization and operations optimization. Sensor technologies and analytical software are the two most crucial parts that enable those functionalities and, as a result, enable the reduction of costs, improved uptime, increasing efficiency of allocation of production resources and increased visibility and versatility of the production process.

Table 12: IoT main applications summary

Main Applications Description Benefits Industries Examples Asset Optimization

Critical Assets Monitoring Use sensors to monitor the health and viability of the

asset in real time. Detect and diagnose problems remotely

• More accurate maintenance, reduce costs • Increase revenue by increasing uptime

O&G, Mining, Manufacturing, Power, Marine, Railway

Aero engines, Oil drills

Predictive Maintenance Use big data analytics to analyse the real-time data collected by sensors. Planning maintenance

• Save costs: prevent operations interruptions • Increase revenue by increasing uptime

O&G, Mining, Manufacturing, Power, Marine, Railway

Aero engines, Oil drills

Asset Tracking Use barcode, RFID or other connectivity techs to track the

location and status of everything in real-time, providing new dimensions of information for control system

• Enable operation optimizations that use location and status data General

Inventory tracking, Mining vehicle

tracking Operations Optimization

Production Optimization Obtain real-time data on products at every production stage to increase process visibility and transparency,

better allocate resources.

• Increase efficiency • Improve quality and reduce cost related to defects: detect defects at early stages • Enable more customized products, reduce batch size • Reducing problem solving time

General Siemens EWA

Supply Chain Optimization Use the location and profile data in RFID tags, optimize the supply chain from suppliers to customers.

• Just in Time, improve efficiency • Leaner inventory, reduce cost related to inventory • On-time delivery, accurate delivery schedule • Accelerate decision process: more transparent operation

General Raw material.

Tracking, Inventory tracking

Remote Control Operate and inspect facilities from a remote location. • Reduce labour cost • Increase safety O&G, Mining, Power Offshore oil field

Fleet Management Obtain real-time data from sensors to improve asset identification, tracking, utilization and logistics operations

• Increase fleet efficiency, utilization rate • Save costs such as energy cost

O&G, Mining, Logistics, Construction Mining fleet mgmt.

Enhance Collaboration Use the latest communication methods to enable real-time collaboration within firms or externally on subjects

from product design to production.

• Increase efficiency: real-time problem solving • Reduce travel costs and costs related to time

General PLM, Remote assistance


Deep industry knowledge

30%

Both industry knowledge and

analytics experience

43%

Analytics cross-industry experience

24%

N/A 3%


Stanley Black & Decker: Stanley Black & Decker Inc. is a leading global hand tools, power tools and related accessories manufacturer. Reynosa plant, Mexico, is one of the largest tool manufacturing plants with 40 multiproduct manufacturing lines and thousands of employees; the plant produces millions of power tools each year. For plants at this scale, managing the manufacturing complexity and increasing transparency of the process are key challenges that need to be overcome. By attaching RFID tags to all material and integrating them with PLCs, the floor managers are constantly aware of each line’s output, whether production needs to speed or slow to meet daily targets, and how quickly employees are completing their respective stages of production. As a result, the overall equipment and labour efficiency has been greatly improved and the company now can provide accurate delivery schedules to customers. Rexroth Homburg Plant: At the Rexroth plant in Homburg, a new pilot assembly line which is integrated into the “Bosch Production System” (BPS) enables semi-automatic production of disc valves for tractors – in six basic types and over 200 variants. RFID tags are used to detect the required product variant and communicate the necessary materials and processes to the line. This reduces the batch size to ideally one, thus enabling more customized products to clients. Dundee Precious Metals: Dundee Precious Metals (DPM) is a Canadian-based, international mining company involved with precious metals. DPM’s flagship mine, in Chelopech, Bulgaria, produces a gold, copper, and silver concentrate. The mining industry always face challenges on communications: tracking miners’ locations, equipment location and status, limited real-time communication capability etc. To tackle those problems, wireless access points were deployed in the tunnel and RFID tags were attached to both employees and equipment, providing a real-time data stream to analytical software so that it could optimize resource allocation immediately. Now the utilization rate is close to 100%, and safety has greatly improved. Shell Smart Field: Shell has recently initiated a smart field pilot project to test how the industrial internet could add value to increasing critical machinery’s operational efficiency and tackle safety issues. Sensors were added to critical rotating equipment such as drillers, compressors, generators and pumps. These sensors then capture information on temperature, pressure and other measurements, and send it out to control centres. Advanced algorithms then analyse the data and decide whether the equipment needs to be repaired, preventing unexpected shutdown and increasing uptime.

40 Capital G

oods | 09 September 2016

Table 13: IoT case studies: summary table

Clients Solution Providers Challenges Solutions Results

Manufacturing

Black & Decker Cisco

• Managing large scale of production and manufacturing complexity • Transparency of real-time production to schedule

• Wi-Fi RFID tags are attach to virtually any material and provide real-time location and status to assembly workers, shift supervisors, and plant managers. • Wi-Fi tag is integrated with the PLC, good and bad production results are immediately sent and analysed. • Floor managers are constantly aware of each line’s output, whether production needs to speed or slow to meet daily targets, and how quickly employees are completing their respective stages of production.

• Overall equipment effectiveness increased by 24%. • Keep inventory as low as possible and consequently lower the relevant costs • Providing accurate delivery schedules to customers. • Achieved 10% greater labour efficiency and improved utilization rates from 80% to 90%. • Quality improvements: defects per million opportunities were reduced 16%

Siemens Amberg Plant Siemens • Improve efficiency

• Improve quality

• High-performance PLM program helps developers design and simulate new products. • Through its unique barcode, each product continuously provides numerous sensors with information regarding its status, its requirements, and the next production steps. An individual PLC can take corrective action to avoid damage in the production process, and automatically replenishes individual parts in order to meet delivery deadlines

• Production volumes eightfold in 25 years. • Production quality at EWA is at 99.9988 percent. • Material is brought from warehouse to machine in 15 minutes. • 50 million pieces of process data are generated daily to inform management decision-making.

Bosch Rexroth Homburg Plant Bosch • Improve efficiency and

versatility of the plant

• Use RFID to detect the required product variant and communicate the necessary materials and processes to the line. • The individual workplaces automatically adapt to the needs of their operators. This is done using Bluetooth tags which are carried by each employee and which are read out by the assembly station. In this way, the illumination of the station or the font size and language on the monitor are adapted accordingly.

• Reduce throughput time and enable optimum use of resources. • Weak points are quickly identified. This reduces down times and allows greater productivity. • Reduce batch size and increase the capability of customization

Mining

Dundee Precious Metal Cisco

• Difficulties to track important info of miners and equipment underground • Real-time communication and collaboration. • Safety issues.

• Wireless Access Points provide coverage along 50 kilometres of tunnels. The network connects everyone and many things. • RFID tags attached to miners’ caps and to vehicles. The wireless network picks up signals from the tags and transmits them to the control center.

• 400% increase in production thanks to greatly enhanced production transparency. • Improved safety. Blast system integrates with the people-tracking system. • Improved asset efficiency. Vehicles utilization rate close to 100%. Predictive analysis reduces downtime. • Lower communication and energy costs.

O&G

Shell GE

• Harsh environment, difficult to monitor the conditions of mission critical equipment • Safety issues

• Installed thousands of sensors on its equipment, such as valves and pumps. • The sensors capture on temperature, pressure and other measurements, and sends it out to control centres back on land. • Engineers then read the measurements and monitor production in real time so they can optimize each individual process.

• Increase production • Reduce downtime • Improve the overall recovery of O&G while reducing costs and minimizing safety risks.



3. Big data analytics

Given the scale of the installed base of industrial equipment and the maturity of the condition monitoring market, big data analytics is the natural starting point for industrial companies in seeking to benefit from increased connectivity and the industrial internet. Data generation is becoming significant, leading to problems with storage, speed of analysis and efficacy of data. As a result, more data analysis and response is likely to happen at the "edge", or the device level with a bar bell effect on the middle layers of industrial automation.

Big data analytics is essentially the first generation of the industrial internet and offers the potential for any industrial company to access the market by making their products “smart”, i.e. capable of generating data.

Best placed – GE, Siemens, device manufacturers GE appears to be the pioneer with big data analytics at the heart of GE's Digital Industrial model, but perhaps more interestingly, non-automation companies such as SKF may also be well positioned as they develop smart capabilities at the point where the data is generated. Both GE’s Predix platform and Siemens Sinalytics platform are designed to capture the benefits of big data analysis. Companies that may lose out could be the providers of the current middle layer of automation capabilities such as DCS and SCADA as the functions of those control systems are replaced by more capable control devices (PLCs, actuators, valves etc.) and the overarching PLM, MES or PAM software.

A natural starting point, but drives need for smarter field instruments The sheer volume of installed equipment globally and established condition monitoring techniques mean data detailing the operational life of machines is already available for analysis. The industrial internet allows this to be done on a grand scale, in real time and in the context of a broader network of machines to allow response to data analysis (as opposed to legacy condition monitoring, which was essentially just one-way).

Exhibit 12: Schematic illustrating the process of data analytics

Source: Gartner

The installed base of industrial equipment runs to billions of individual pieces of equipment, all of which generate data (e.g. pressure, temperature, revolutions per minute, vibration etc.). More recently installed equipment was likely delivered "smart", i.e. capable of being connected to a computer network in some manner to be able to deliver that data for analysis but retrofit of data measurement and transmission is also cheap and easy. Hitherto, condition monitoring extended to pre-emptive maintenance, with data produced by on-location sensors to allow the prediction of maintenance requirements. Now, real time analysis and the ability to change the machine's operating

Data

DescriptiveWhat happened?

DiagnosticWhy did it happen?

PredictiveWhat will happen?

Analysis Human input

PrescriptiveWhat should I do?

Decision Action

Decision supportDecision automation


parameters accordingly introduces improved abilities to optimise the operating cycle of a given machine or system to the benefit of energy efficiency, equipment longevity or productivity.

With the various estimates of connected machines at >50bn by 2025, the amount of data generated will be enormous, but only some of it will be useful – at present, only 8% of businesses studied by Verizon/Oxford Economics are using more than 25% of their IoT data. Ultimately, this means that a considerable amount of data is generated (and has to be transmitted, analysed and stored) that may not be used - a concept known as digital exhaust. In response, the development of edge computing will drive equipment to be much smarter, capable of making initial analysis, data processing and response at the point of generation rather than relying on the current model of transmission to some central processing point (i.e. a cloud, or physical data storage and processing system). In our mind this means the device level in the traditional industrial automation hierarchy become smarter - PLCs become Programmable Automation Controllers, or PACs, or actuators become Distributed Control Nodes, or DCNs. Automation decisions are made earlier and closer to the industrial process and the overall process is managed by the PLM or MES system.

Investment focused initially on the easy wins and the basic building blocks Some 50% of the current investment related to the industrial internet is directed to modules/sensors and connectivity. Enabling equipment to monitor operating characteristics through sensors and then connecting the resultant smart machines to the internet is not only an easy win in terms of building on existing remote monitoring capabilities it also forms the basic requirements for enabling the industrial internet and data analysis.

Once machines and equipment has been connected, and data generated, the analysis and response to the analysis can be implemented and the software to do this represents the fastest growing market for investment spending over the next five years. In terms of relative size, though, analytics software is likely to remain a small part of overall industrial internet investment in the near term, at least until the pace of penetration of smart sensors has slowed.

Chart 26: IoT spending by product (USDbn)


Chart 27: … o/w analytics software at >20% pa


Next generation condition monitoring – APM software Industrial companies have been using condition monitoring techniques for better maintenance practices for some time, though the means of acquiring and processing the data has evolved (we fondly remember an internship listening to recordings of ship propellers for signs they were running off balance sent to us through the post on radio

-

200

400

600

800

1,000

1,200

1,400

1,600

2014 2015 2016 2017 2018 2019 2020

Module/sensor Connectivity Analytics software Other

0%

5%

10%

15%

20%

25%

30%

2015 2016 2017 2018 2019 2020

Module/sensor Connectivity

Analytics software Other


cassettes!). The traditional use for condition monitoring has been to improve maintenance practices, pre-empt failures and develop a better understanding of the operational performance of installed equipment but the process has been limited to one way communication and a convoluted feedback look. Siemens talks about Digital Services, GE talks about the Digital Industrial but the basic premise is to get more out of the data machines generate to improve the operational performance on a real time and proactive basis. The growth expectations differ somewhat between the two companies, but the reality is that growth is likely to be far stronger than underlying industrial demand.

Exhibit 13: Siemens big data approach is being offered through its Digital Services initiatives and its Sinalytics/Mindsphere platform

Source: BofA Merrill Lynch Global Research, company data

Exhibit 14: GE has developed the Predix operating system internally to exploit the installed base of GE equipment and the data it produces


The last decade has seen industrial companies take a more proactive approach to maintenance and the development of their service businesses, especially following the financial crisis and the onset of the current anaemic growth environment. The maintenance business model is evolving, as we will discuss further in a later section, but at its heart is greater knowledge of the underlying asset base and the analysis of the data it generates. As things stand, the concept of big data analytics remains firmly rooted in the collection of data from a given machine or set of machines, analysis of the data and then implementation of adjustments to improve the operating performance of the machine – typified by GE improving LNG complex reliability for RasGas (Qatar) or Siemens delivering 99.9% train availability on the Barcelona-Madrid high speed rail service where only 1 in 2,300 trips is >15mins late (requiring full passenger refund).

The Asset Performance Management market (APM, sometimes also referred to as Plant Asset Management (PAM), represents the next generation of condition monitoring software. GE places APM software at the heart of its industrial internet approach, which revolves around big data and has presented several examples where the APM capability has been critical to the contract delivery.

SINALYTICS/MINDSPHERE

300,000+ connected machines

-

0.5

1.0

1.5

2.0

2016e 2020f

15% + CAGR

PREDIX

225,000+ connected machines

-

1.0

2.0

3.0

4.0

2016e 2020f

199% CAGR


Exhibit 15: The core of GE’s industrial internet strategy revolves around big data, its collection, analysis and use, for which APM software is key

Source: GE

Exhibit 16: Specific examples of Predix winning business based on the strength of benefits brought about by applying APM

Source: GE

The advent of big data analytics really builds on existing condition monitoring capabilities by making them smarter, real time and with two way communications at its core. By connecting the installed asset base and making the communications two-way, the OEM and the customer have the opportunity to better understand the operating parameters of the equipment. This concept is likely to still face considerable hurdles in terms of adoption despite the quick and significant potential benefits due to growing concerns around cybersecurity and system integrity, which we will discuss in a later section.

Chart 28: Global condition monitoring market by region (USDmn)


Chart 29: Growth accelerating across all regions


In the first instance this can enable better use of the equipment to improve longevity but in a broader context it can improve the way systems run, saving energy, money or time (GE have already applied big data analytics to save customers money, energy or time; link). In the second instance it can involve real time feedback to enable adjustment of the machines or systems to have an immediate impact on operational performance. Lastly, condition monitoring can provide the longer feedback loop into product design and provide a link to PLM software.

Outcomes

Technology

Talent & culture

"As a serv ice" model

Asset PerformanceManagement (APM

EnterprisePartner

ecosystem

Cyber-security

Operating system and

cloud (Predix)

Edge devices

Digital twin

Digital thread

Domainknowledge Tech skills FastWorks

- Customer: BP - Customer: SRP - Customer: RGSolution - APM software and - Enterprise-wide - APM, Predix; 1k+

Predix APM software GE and non-GE kit

- Provide analytical - Shift from unplanned - Increasing reliabilityinsights at enterprise outages to planned by 1%scale through knowledge outages - Reduce volumetric

Outcome sharing on demand Significantly reduced downtime- Eliminate disruption maintenance costs - Decrease excess flashby being proactive & - Optimised asset life gasanalyse problems - Manage maintenancemore efficiently costs

-

500

1,000

1,500

2,000

2,500

2012 2013 2014 2015 2016 2017 2018 2019

NA EU APAC ROW

0%1%2%3%4%5%6%7%8%9%

2013 2014 2015 2016 2017 2018 2019

NA EU APAC ROW

http://research1.ml.com/C?q=t3I0vm9r9Zpb1bqBExMBTg


Chart 30: Global condition monitoring market by sector (USDmn)


Chart 31: Growth accelerating across all sectors


What we find most interesting when considering the condition monitoring and APM markets is the fact that they are a) fragmented and b) the largest players are not necessarily the incumbent automation providers. Condition monitoring capabilities have been developed by the likes of Meggitt and SKF due to their position at the heart of machines (Meggitt in aero-engine sensors, SKF in bearings) and which means these companies’ products are uniquely well positioned to provide data on the operational performance of the larger machine they are part of. This means that the first iteration of connectivity and the industrial internet is accessible by any industrial OEM and many of the sub-suppliers, simply by accessing and connecting with their own installed base. By making their products “smart” the suppliers may be able to capture as much of the market as the larger assemblers.

Chart 32: Condition monitoring market share (2014)


Chart 33: APM market share (2015)


“Edge” software capabilities push capability down the value chain We are firm believers in the fact that the initial stages of the industrial internet and big data analytics will provide more than enough market opportunity for the broader industrial supply chain. However, the developing trends of big data and the links with PLM and MES software mean that data analytical capability and the ability for equipment and machines to respond to what the data might be saying is going to be pushed further down the value chain. The need to sift through what data might actually be relevant, data storage and

-

500

1,000

1,500

2,000

2,500

2012 2013 2014 2015 2016 2017 2018 2019

O&G Power Gen.A&D Process&ManufacturingOthers

0%

1%

2%

3%

4%

5%

6%

7%

2013 2014 2015 2016 2017 2018 2019

O&GPower Gen.A&DProcess&ManufacturingOthers

GE Bently Nevada Meggitt Sensing

SKF Condition Monitoring Rockwell Automation

Emerson CSI Others

GE Emerson SKF

ABB Rockwell Meridium

Bentley Systems Schneider Siemens

Others


management capacity requirements and the time taken to transmit/process/respond will all drive capability down to the device/field instrument level.

A simple illustration would be in the control of autonomous cars – the car may be operating within an envelope of data transmitted and received regarding where it is going, its operational performance, fuel efficiency etc, but may need to react very quickly to a road accident. In an industrial context, simple controls within certain parameters could be undertaken by smart actuators without need for constant recourse to the central “cloud” supervisory/control software.

The need to limit data transmission (a combination of storage capacity and the need or desire for quick response times in data analysis) means computing power and software capability will need to be pushed down to the field devices and primary control devices in any given industrial system, factory or plant. In our mind this puts the incumbent suppliers in a stronger relative position due to their inherent understanding of their own products compared to the software providers who have limited or no field device product offering. In practical terms, this means Emerson’s acquisition of Pentair’s valves and controls business strengthens Emerson’s position as one of the few process automation companies with a strong presence in field devices. ABB, Honeywell and Rockwell Automation have all acknowledged they have stronger presence in the mid-level control software than the primary device level.

Exhibit 17: Defining the “edge” in the context of the existing process automation architecture

Source: Honeywell

Longer term, we think the capabilities that currently sit with the mid-level control software (like DCS or SCADA) end up being pushed down into the device level and up to integrate into the PLM/MES/PAM industrial control software. This could be to the detriment of the incumbent providers such as ABB or Honeywell, though both have indicated their M&A strategies could encompass field instruments in acknowledgement of the potential gaps in their current portfolios. It is worth noting that at their recent Honeywell Users Group Americas in June Honeywell launched the ControlEdge PLC, designed to integrate with its existing DCS. The intent is to provide customers with the potential to buy from a single vendor of DCS and PLC as well as the PAM and data historian software.

Data consumptionLocal user access & control

Data access for complex OEM

analy tics

Ancillary data for improv ed analy tics

Remote user access & control

Secure data

Advanced analytics

Data aggregation

Basic analytics

Data collection

THE

EDGE

Data access for complex end-user analy tics

Data access for compliance, commerce of environmental

reporting

Secure gateway

DCS server Historian SCADA

DCS controller PLC RTU

Common smart protocol input/output Traditional input/output Asset data Non-critical process data

Edge gateway

Stranded asset data acquisition

edge device

Ancillary process data acquisition

edge device

Edgecontroller,

PLC or RTU

Traditional control system Full capability edge devices


4. Everything-as-a-Service

Power by the Hour, Software as a Service, Product as a Service… Everything as a Service.

At first glance this might seem an odd concept; after all we cannot virtualise power generation, however the industrial business model is evolving. Instead of providing maintenance services on a machine the customer pays for up front and wholly owns industrial companies are starting to sell products as part of service packages based on uptime or output availability. In simple terms, the industrial equivalent of the concept of Power by the Hour in aerospace terms or Software as a Service in technology terms.

Best placed – service “mentality” is critical It is difficult to make a call on the “best placed” industrial to benefit from the development of everything as a service. Both GE and Siemens have made significant moves in developing their installed base, but companies such as SKF and Atlas Copco have very well developed service concepts and are also changing the way they bring these concepts to market. Investors focus on highest proportion of sales from service as a means of identifying companies with the highest level of recurring revenue. Penetration of the installed base is unlikely to be enough on its own in the future, but provides the perfect channel to “up-sell” greater breadth of service.

Evolution of the service business Service is evolving as a concept in the industrial world to encompass significantly more than the traditional model of providing an engineer or technician to turn up and get the equipment working again. Industrial companies are applying more broadly the tried and tested power by the hour concept in aerospace. While industrial companies offer a broad range of service options, the leading edge business model is one where the industrial company commits to providing an output for the customer rather than selling the equipment that produces something and then the spare parts/expertise to maintain it. In practical terms, Atlas Copco commits to providing a given flow rate of compressed air to its customer rather than selling a compressor and returning when the bearings need replacement or the energy efficiency drops. The logical conclusion to this development would be that Atlas ultimately never sells a compressor, but sells the supply of compressed air and gets paid a monthly retainer in much the same way we, as consumers, pay for our utilities. This is some way off!

Exhibit 18: GE’s view of the evolution of service business

Source: BofA Merrill Lynch Global Research, GE

Increasing the service business and aftermarket capability has been a consistent mantra of industrial management teams for much of the last decade in recognition of the

1980s Big Iron - Partners- Field engineers

1990s Repair Technology - Advanced repairs- Part upgrades

2000s Contractual Services - Risk sharing- Monitoring & diagnostics

Today Digital Industrial - Analytics-enabled uptime- Software-enhanced performance- Remote inspections

Technology application


greater resilience, sustainability and profitability of the business model. Investors also value companies which derive a large part of their revenue from aftermarket business due to the lack of cyclicality, good visibility and the recurring nature of the revenue. As a result, service has increased from 28% of revenue in 2005 to 33% in 2015 across our coverage universe (for the companies where we can track it). However, to date the development has been largely driven by penetration (i.e. capturing more of the installed base) and some increase in sophistication (i.e. scope of the service contract or offering predictive maintenance capabilities).

Can everything be provided as a service? In short, no. However, we think the trend is for greater and greater service content as customers seek to reduce their own exposure to non-core capabilities and industrial companies look to strengthen the customer relationship by offering additional capability as part of the broader service package. In essence this means that the old business model of selling the equipment up front and then providing a service contract as part of the aftersales service is history and companies look to provide the output of their respective equipment as a service.

In practice, this could be Atlas Copco committing provide a given flow rate of compressed air or SKF a given availability of rolling equipment. Siemens could commit to providing a given power output, uptime or thermal efficiency in selling gas turbines and Sandvik could sell cutting tools on the basis of amount of metal cut per hour rather than the up-front cost of the tool. Weir can commit to ensuring a given flow rate through its installed mining slurry pumps or Schneider to a targeted energy saving per month on a commercial building.

Exhibit 19: Examples of GE using the Predix platform to win business beyond the machine based on improving enterprise efficiency


Exhibit 20: Examples of GE using the Predix platform to win business in new markets based on improving enterprise efficiency


The changes in business model are significant, but there is likely to be a limit. Non-critical applications are unlikely to represent enough of a return on the investment in smart capability to justify making the change. For example, it will be critical to ensure back-up capacity can function despite not being in constant use. Although critical in its function, such a machine does not need the same sort of real time monitoring and service commitment to enable a monthly or quarterly payment structure as the equivalent machine in the actual process.

A change in mind-set is required While service has been a critical element of strategy for most industrial management teams over the last decade, it remains largely focused on maintaining an asset once it has been delivered. Management teams are starting to think about the process of selling the end product as a service but manufacturing organisations are still largely set up to produce a product most efficiently and then sell it rather than think about that product’s output as a source of revenue. Service organisations are generally separate, with different managers and management structures, which tend to deal with the product once it has been installed, rather than interacting with the design office or the product

- Customer: P&G - Customer: engro - Customer: ExelonSolution - Smart automation and - Operations - APM, Predix; OO,

digitised manufacturing optimisation (OO) business optimisation

- Increasing manufacturing - Increasing operational - Increased analyticsproductivity, facilitating efficiency and reducing velocity as part of digitalspeed of innovation fuel consumption transformation

Outcome - Improved relianility withAPM- Reduced SCRAMS andimproved performance- Better power forecasting

- Customer: Lixil - Customer: Toshiba - Customer: Pitney BowesSolution - Predix & scheduler - Predix & application for - Predix & asset health

application predictive maintenance application

- Reducing cost and - Predix is the architecture - Up to 10% machineproject delays platform supporting efficiency- Improving controller Toshiba IoT solutions & - Up to 20% increase in

Outcome productivity services machine yield- Enhancing customer - Up to 15% savings on satisfaction parts replacement

- Up to 10% tech supporttime savings


sales people. Key account managers within industrial companies still often think about the product sale with a service agreement attached, as opposed to providing a service to their customers of which the provision of the original equipment is part and ensuring it remains available to the customer and operating optimally is the bulk of the value.

Recognition that service offers more stable growth, greater revenue visibility and improved proximity to the customer has driven most industrial management teams to make developing the service business a key element of their strategies. To that end, service as a proportion of sales has grown steadily over the last 10 years and now represents over 35% of revenue on average for our coverage universe. Even this might be the wrong way to think about it, in light of what we have said, as identifying service revenue separately reinforces the concept of initial equipment sale and subsequent service contract.

Chart 34: Service as proportion of sales for available data across our coverage universe

Source: Company data, BofA Merrill Lynch Global Research estimates

Chart 35: 10yr CAGR for service business growth for selected European capital goods companies (2005-2015)

Source: Company data, BofA Merrill Lynch Global Research estimates

The companies with the highest proportion of revenue derived from service are well known, often being identified by investors as premium or high quality companies – companies like Kone, Atlas Copco and Wartsila. However, in terms of changing the mind set and adjusting the way product and service are sold, the list of companies changes slightly. Atlas Copco has talked about selling compressed air, rather than a service contract attached to a compressor sale, GE talks about the “power of 1%” – the concept of making everything 1% more efficient in order to generate cost savings.

Getting paid is the critical question The biggest issue for industrial companies in changing the way they sell is ensuring they get paid for the value they are bringing to the table. This is not just about making sure the customer pays for the service. Structuring any contract related to guaranteeing equipment availability or the provision of a service requires an adjustment in the traditional ways industrial equipment has been sold. It also requires customers to understand what the equipment manufacturers are selling.

We believe companies are still early in the process of establishing the best contract structure and in the first instance many agreements have been based on energy efficiency with both customer and supplier sharing the savings. Clearly accounting for these types of agreement is complex; companies and investors will be wary of the Rolls Royce example.

25%

27%

29%

31%

33%

35%

2005 2010 2015

Capital Goods Avg.

0%

5%

10%

15%

20%

Hexa

gon

Rotor

kAt

las C

opco

Siem

ens

Spec

trisMe

tsoAs

sa A

bloy

Weir

GEA

War

tsila

Kone

ABB

SKF

Sand

vik


Chart 36: The customer still thinks about the product price of an Atlas Copco compressor as an up front cost


Chart 37: Atlas Copco looks to break down the through life cost of the various initial options for the customer


For more than ten years SKF has been running an audited Documented Savings Programme that delivered an audited cost savings number derived from changes SKF had made for a given customer. The annual savings of >SEK4bn can then be a concrete number to form the basis of contract discussions between SKF and its customers. On top of this, the company has been changing the way it sells to customers with the establishment of SKF Solution Factories, where customers are invited to see the practical implications of the solutions SKF can bring to the table. The interactive and tangible nature of the sale allows SKF to drive the conversation towards the value it brings to the table and beyond the simple purchase of a bearing.

Chart 38: SKF’s Documented Savings Programme annual savings since launch (SEKm)


Chart 39: SKF’s Solution Factory network has been built up from nothing in 2008


050

100150200250300350400

MBC FP PM TR

Customer labour Parts supply Product price

Maint

enan

ce by

cu

stome

r

Fixed

price

Prev

entat

ive

maint

enan

ce pl

an

Total

resp

onsib

ility

plan

050

100150200250300350400

MBC FP PM TRProduction loss Repair risk OverhaulsExtra energy Management ExecutionParts supply Product price

0500

100015002000250030003500400045005000

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

Savings (SEKm)

0

10

20

30

40

50

60

2008 2012 2013 2014 2016f

Factories


5. The rise and rise of robots Continuing to grow Today, there are more than 1.4mn industrial robots all around the world (source: IFR). We believe they are a key enabler of the Industrial Internet and Industry 4.0. Integrating robots, big data, and software analytics can optimize the full manufacturing chain, connecting virtual design, simulation, manufacturing and supply chain management. Today, they perform 10% of all manufacturing activity, but we see adoption accelerating on the back of several growth drivers:

• Energy efficiency and new materials

• Global competiveness requires increased productivity and quality

• Growing consumer markets require expansion of production capacities

• Decreasing life-cycles of products and increasing variety of products

• Robots improve the quality of work by taking over dangerous, tedious and dirty jobs that are not possible or safe for humans to perform

Global industrial robot penetration stands at 10%; however, 72% of all US executives say their company will invest in additional automation or advanced manufacturing over the next five years (Source: BCG)

Robots will be a key for the “industrial internet” and “industry 4.0” Industrial robots have been a key enabler of the global shift towards greater electrification of the manufacturing value chain. According to our European Cap Goods team, the combination of cheaper sensors, big data, the cloud, analytics, and the resulting explosion of Internet of Things (IoT), have made robotics uptake within industrials an area of fast growth.

• The industrial internet is a term coined by GE to describe a network of connected machines. The network generates data via sensors connected to machines which are part of an open protocol IT network. The data can then be stored and processed in real-time on the cloud using predictive analytics. This last step is crucial in order to contextualize the information coming from the machines and translate them into efficiency gains. GE attempts to monetize these gains by providing analytics capabilities such as its Predix software and its corresponding Productivity solutions, providing asset management and process optimization services.

• Industry 4.0 is the German-led equivalent to GE’s Industrial Internet and invented by acatech (National Academy of Science and Engineering) and manufacturer Bosch. The concept of Industry 4.0 is combining previously non-compatible units such as humans and machines to improve the manufacturing value chain, create efficiency gains optimize costs. In contrast to the Industrial Internet, Industry 4.0 has a bigger focus on engineering driven by the central role of the Smart Factory which represents the full virtualization and integration of the manufacturing value chain in order to obtain efficiency gains at all steps from product initiation to product launch. PwC estimates that Industry 4.0 will account for more than 50% of all capex investments over the next five years, rising to €140bn per annum in Europe alone. While only one in five industrial companies has digitized its key processes, 85% will have implemented Industry 4.0 by 2020.


Manufacturing will be the key beneficiary of the Internet of Things (IoT) Manufacturing will be the key beneficiary of the move to the ICT-enabled IoT. According to Cisco, two-thirds of the IoT value stake will be created from the manufacturing segment (including power, and oil and gas), the retail sector and the public space. Manufacturing will account for 34% of the total value over the next decade. IoT creates value for a stylised manufacturer, highlighting the particular areas where efficiency gains occur and how these gains manifest.

Exhibit 21: The IoT yields value across the whole manufacturing value chain

Source: Cisco,

The Smart Factory will digitalise & optimise the manufacturing value chain The Smart Factory is the new term for the full integration of all steps in the manufacturing value chain, connecting virtual design, simulation, manufacturing and supply chain management. It generates efficiency gains at all phases of the production process by gathering real-time data and using feedback loops between them. As an example, feedback between virtual design and manufacturing allows for simulating the production process for a digitally designed product in order to anticipate potential shortcomings instead of incurring the full cost by launching a prototype.

Exhibit 22: The Smart Factory is digitalizing the entire manufacturing value chain


US$24bn industrial robots market by 2025E In 2014 the value of the global industrial robot market stood at US$10.7bn, which represented a new peak, according to the IFR. BCG projects total robot spending to increase to US$67bn in 2025, representing a CAGR of 9.5% from 2015. Expenditures for industrial robots are projected to rise by US$13.4bn accounting for one third of the


overall increase, reflecting a CAGR of 8.5%. The International Robot Federation also states that the worldwide stock of industrial robots stood at c1.48m in 2014 and is expected to rise to over 2.3mn by 2018, reflecting a CAGR of c12%).

Chart 40: The industrial robot market stood at $9.5bn in 2013 and is estimated to reach $13.4bn by 2018E, CAGR of 8.5% (thousands)

Source: BCG

Chart 41: The global installed base of industrial robots is expected to rise above 1.6m and 2.3m in 2016 and 2018,respectively (thousands)

Source: IFR

15% CAGR through 2018E, 2.3mn robots in operation Global robot installations are expected to increase at a CAGR of 15% from 2014 to 2018. In the US and Europe, growth is estimated at 9-11% during the period. In Asia and Australia, growth is expected to be 18-21% per annum. This means the worldwide stock will increase from 1.48mn units in 2014 to 2.3mn by 2018 (source: IFR 2015).

Chart 42: Worldwide annual supply of industrial robots 2008-18* (units)

Source: IFR

Chart 43: Worldwide estimate operational stock of robots (thousands)

Source: IFR

Growth to continue to accelerate Given the convergence of falling prices, performance improvements and wage inflation, robotic adoption could be reaching an inflection point. Globally, only 10% of manufacturing is being performed by robots, although a much higher proportion could be achieved. For instance, only 0.5% of tasks have been automated in the chemical and primary metals industry in Italy, versus 14-16% that could be automated. In the US, only 8% of tasks in the transportation-equipment industry are automated, versus 53% potential. However, as economic and technical barriers fall, growth is likely to continue to accelerate (source: BCG).

0

20

40

60

80

2000 2005 2010 2015E 2020E 2025E

Personal Commercial Industrial Military

-

500

1,000

1,500

2,000

2,500

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

e20

16e

2018

e

0

50,000

100,000

150,000

200,000

250,000

300,000

350,000

400,000

2008

2009

2010

2011

2012

2013

2014

2015

2016

e

2017

e

2018

e 0

500

1,000

1,500

2,000

2,500

1973

1983

1990

1995

2000

2005

2010

2014

2015

*

2018

*

2016-2018: +18% per year on average


Chart 44: Global robotics installations to increase at a CAGR of c10% over the coming decade Estimating potential market for advanced industrial robots at c14 million units in 2014 and growing c2%-3% per year with manufacturing output. Anticipate c20% of all tasks will be automated by 2025.

Source: IFR, BCG * 2016 IFR projection as reported in its “World Robotics – Industrial Robotics 2013” report. 2017 IFR projection as reported in its “World Robotics – Industrial Robotics 2014” report. Note: Market size is estimated from an evaluation of jobs within US industries that may be automated, extended to global manufacturing output by industry.

Up to 45% manufacturing penetration in 10 years While adoption will vary by industry, the transportation equipment, computers & electronics, electrical equipment, and machinery industry are expected to reach 40-45% penetration. They will account for 75% of all advanced robotics installations over the next 10 years. Areas such as food products, plastics, fabricated metal will be slower given the technical difficulty to automate and low wages (source: BCG).

Up to US$600bn-US$1.2tn in economic impact by 2025E Adoption of industrial robots is expected to yield US$600bn-1.2tn in positive economic impact by 2025E. Developed countries will see higher rates of robotic penetration as 15-25% of worker tasks are automated within the manufacturing, packing, construction, maintenance, and agriculture spaces. That figure will be 5-15% for Ems. In aggregate, 30-60mn full-time employees could be replaced by automation, equivalent to a 75% improvement in productivity per unit of work automated (source: McKinsey).

Manufacturing labour up to a third cheaper by 2025E For the world’s 25 largest goods-exporting countries, manufacturing labour costs will be 16% lower on average by 2025E than they would have been without large-scale take-up. This means output per worker could rise by 10-30% above productivity gains that would come from other measures. This will be the most pronounced for countries with the highest robotics adoption such as South Korea, China, the US, Japan, and Germany, which together make up 80% of all robotic sales. There, the productivity-adjusted labour costs will be 18-33% cheaper (source: BCG).

Countries slow to adapt will be losers Conversely, countries that have been slow at adapting are expected to see their cost competitiveness continue to deteriorate. These include France, Italy, Belgium, as well as emerging markets like Brazil, India and Mexico (source: BCG)

0

2,000

4,000

6,000

8,000

2007 2010 2013 2016 2019 2022 2025

glob

al st

ock o

f ope

ratio

nal r

obot

s by i

ndus

try (u

nits

)

Baseline Projections Aggressive Projections

Low Projections IFR Projections

Actual

Projected Actual


2015: Another record year The worldwide sales of industrial robot just achieved a new record number of 248,000 units in 2015 (12% yoy growth). The global market size amounted to US$10.7bn in 2014, while the overall robot systems market, including software, peripherals and systems engineering is now valued at US$32bn.

15% CAGR to 2018 Industrial robot market has been growing rapidly since 2009. According to IFR, this growth is expected to continue at 15% CAGR reaching 379,000 units annual sales volume by 2018.

Chart 45: Worldwide annual supply of industrial robots 2008-18E*(units)

Source: IFR Statistical Department *forecast

Chart 46: The industrial robot market valued at US$9.5bn in 2013 and is estimated to reach US$13.4bn by 2018, CAGR of 8.5% (thousands)

Source: BCG

Concentrated market: Japan has the highest penetration The global market is highly concentrated, with China, the US, Japan, South Korea and Germany dominating 70% of the global market. Japan currently has the biggest installed base of industrial robots accounting for 20% of the global total (source: IFR).

Chart 47: Installed base of multipurpose industrial robots at year-end in selected countries (2014)

Source: IFR * America includes Brazil, North America and Other America

Chart 48: China is expected to account for 38% of global shipments of multipurpose industrial robots in 2018

Source: IFR * America includes Brazil, North America and Other America

China: #1 buyer for two straight years: 38% of units by 2018E China has been the largest buyer of robots for two years in a row, and now makes up 25% of the global market. China’s global market share stands at 13%. However, when

0

50,000

100,000

150,000

200,000

250,000

300,000

350,000

400,000

2008

2009

2010

2011

2012

2013

2014

2015

2016

E

2017

E

2018

E

Japan 20%

America* 17%

Europe excl. Germany

16%

China 13%

South Korea 12%

Germany 12%

Asia/Australia excl. China, Japan and

South Korea 8%

Other & Africa 2%

China 38%

America* 12%

Asia/Australia excl. China, Japan and

South Korea, 11%

Europe excl.

Germany, 10%

Japan 10%

South Korea 10%

Germany 6%

Other & Africa 3%

0

20

40

60

80

2000 2005 2010 2015E 2020E 2025E



compared in terms of annual shipments, the country clearly stands out against other markets with 57,096 units sold in 2014. The IFR estimates that by 2018E, China will account for 38% of annual units sold. India will see the highest global growth albeit from a much lower starting point.

Chart 49: China and India to account for the highest growth in annual demand for new units

Source: IFR

Huge room for growth: only 66 robots per 10,000 employees globally Globally, it is estimated that robots perform around only 10% of all manufacturing tasks, despite having been used in factories for decades (source: BCG). Robot density is still relatively low for general industry at only 66 robots per 10,000 workers, versus the highest penetration rates such as the Japanese automotive sector at 1,520 robots per 10,000 workers (source: IFR). This signals huge potential for more upside, especially in emerging markets, along with modernization in developed markets.

Chart 50: Number of multipurpose industrial robots (all types) per 10,000 employees in the manufacturing industry – 2014

Source: IFR 2014

Chart 51: Operational stock of industrial robots 2014 and 2018E

Source: IFR 2015 ** Germany, Italy, France, Spain, UK

0%

5%

10%

15%

20%

25%

30%

35%

0

10,000

20,000

30,000

40,000

50,000

60,000

China

North

Ame

rica

Japa

n

Kore

a

Germ

any

Taiw

an Italy

Thail

and

Spain UK

Fran

ce

India

Units sold in 2014 '14-'18 CAGR rhs

050

100150200250300350400450500

Repu

blic o

f Kor

eaJa

pan

Germ

any

Swed

enBe

lgium

Denm

ark

Italy

USA

Taiw

anSp

ainFr

ance

Finlan

dAu

stria

Cana

daNe

therla

nds

Slov

akia

Slov

enia

Switz

erlan

dAu

strali

aCz

ech R

epub

licUn

ited K

ingdo

mCh

ina

Average robot density - 66

189

313

237

614

371 323

0

100

200

300

400

500

600

700

China EU_5** North America

'000 o

f uni

ts

2014 2018*


Prices of robots are dropping while labour costs are rising One of the key drivers for adoption is the declining cost of robots. For instance, the cost of an advanced robotic welder fell from US$182,000 in 2005 to US$133,000 in 2014, or a 27% decline. Prices are expected to continue to drop a further 22% by 2025E (source: BCG). The latest models of the Baxter ”cobot” retails for as low as US$22,000 (source: Baxter). At the same time, performance of robotic systems is expected to improve by 5% each year, which makes the economics increasingly compelling.

Reaching the inflection point on costs The inflection point where manufacturers replace human workers with robotics is estimated to be when an automated operating system becomes 15% cheaper than employing a human worker. This explains the high penetration of robots in the automotive sector, where it costs US$8/hour to use a robot for spot welding, versus US$25 for a human labourer (source: BCG). The cost advantage gap will continue to widen for such industries, while other sectors are also quickly reaching that point.

Chart 52: Manufacturing labour costs in select provinces and countries, 2014 (US$/hour)

Source: Economist Intelligence Unit

Wage inflation: labour costs are rising Wage inflation has been a reality globally, but is particularly pronounced in emerging markets (like China), which are global manufacturing hubs. Per-hour earnings in Chinese regions like Henan, Anhui have been expanding at a 16-17% CAGR since 2001 (source: Economist Intelligence Unit).

124.3 54.9 502.2 54.1 27 41.4 64 29.9 66.6 41.1 41.4 37.4 41.9 39.5 24.9 48.3 14 9.1 38.2 64.2 48.1 12 75.3 27.6 20.2 52.9 18.3 11.5 110.9 5.3 26.4 65.9 155.7 44.8

0 5 10 15 20 25 30 35 40 45 50 55 60

IndonesiaNigeria

IndiaVietnam

JiangxiHebei

HenanYunnan

ShandongHunan

PhilippinesHubeiAnhui

ThailandLiaoningSichuan

MalaysiaKazakhstan

ZhejiangGuangdong

JiangsuShanghai

RussiaTurkey

South AfricaMexicoPolandTaiwan

BrazilPortugal

South KoreaJapan

USGermany

Labor cost (US$/hour)

Manu

fact

urin

g lab

or co

sts i

n se

lect p

rovin

ces a

nd co

untri

es, 2

014 (

US$ /

hou

r)

Other countries ChinaLabor force

(mn)


We believe industrial robots will continue to see strong demand going forward given their associated benefits such as lower operating costs, improved process quality, higher manufacturing flexibility and increased workplace safety.

Non-automotive segments gain ground Industrial robots were traditionally mainly used within the automotive sector. Already in 1978, Kuka started supplying leading automotive manufacturers like Ford and Daimler-Benz with its first industrial robots IR 601/60. Going back 11 years from now, the automotive industry displayed the most dominant end-market for industrial robots, accounting for almost half of all shipped units.

Chart 33 shows a split of the different end-markets for industrial robots in 2013, underlining the still high demand in the automotive space (c.39% of total annual supplied units), but also from the electronics sector that accounted for c.20% of the overall space and that benefits from the emergence of new types of robots (e.g. collaborative robots which are particularly well-suited for small parts assembly and packaging and testing of small goods such as mobile phones).

In recent years, this dominance has somewhat decreased as non-automotive segments have seen high growth rates, driven by plenty of new products (typically smaller, lighter and increasingly customizable robots) (Chart 34). Overall, non-automotive sectors have outgrown their auto peers by c.3.2% on average each year since 2004.

Chart 53: The automotive sector’s share of shipped robot units has decreased over the past from c. 46% in 2004 to 43% in 2014

Source: IFR

Chart 54: Non-automotive segments (CAGR: 9.4%) outgrew their auto peers (CAGR: 8.4%) by around 1% per year on average

Source: IFR

The global industrial robot market is mainly dominated by four main players: Kuka, ABB, Fanuc and Yaskawa which together account for c.53% of the total market for units sold in 2013 (Chart 10). Concerning the different geographies, IRF projects that by 2017 the majority of shipments of industrial robots will go to Asia and Australia (186,000 units), which represents a CAGR of c17.1%, underlining the region’s significance as a driver of demand. In particular China stands out in this region, expecting to account for c.1/3 of total demand by 2017, reflecting average annual growth of c.30%.

China’s Robot Demand China’s share of global manufacturing output almost stands at 25%. In 2013 the country became the biggest robot market in the world with 36,560 industrial robots sold. Between 2008 and 2013 supply of industrial robots increased on average by 36% per year.

Automotive 43%

Electronics/Electrical

21%

Metal 9%

Chemical, Rubbers and

Plastic 8%

Food 3%

Other 5%

Unspecified 11%

0%10%20%30%40%50%60%70%80%

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Automotive Non-Automotive


Chart 55: China has grown to become the biggest robot market by annual shipments

Source: IFR

Nevertheless structural factors hint at a continuation of strong demand: China’s population is ageing with the share of people older than 65 years now at 9.7% (compared to 5.6% in 1990). The population control in form of the one-child policy has been a drag on growth which is likely to continue going forward.

• This shortage of labour has also resulted in upward pressure on wages leading to an increased outperformance of average industrial wage growth when compared to the increase in GDP per capita

• Although China represents the biggest robot market in absolute terms, its density (the number of robots per 10,000 workers) is still significantly below the level of competitors in the manufacturing space like Germany or Japan. With only c.23 robots per 10,000 employees, compared Japan’s 323 or Germany’s 282 figure, this suggests great potential for further installations.

• In 2014 China’s government put “the industrial-robot revolution” on their agenda and started supporting the sector through VAT refunds as well as direct subsidies, and tax advantages for robot purchasers. Recently it also outlined its strategy “Made in China 2025” which represents a plan to foster automation and IT technology through R&D funding, tax benefits and intellectual property protection

Chart 56: The decline in childbirth due to China’s policies and aging society will lead to a lack in labour force

Source: National Bureau of Statistics of China

Chart 57: Wage growth continues to be higher than GDP per capita growth

Source: National Bureau of Statistics of China

020,00040,00060,00080,000

100,000120,000140,000160,000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2018

e

(units)

020000400006000080000

100000120000140000160000

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

2012

2014

0-14 15-64 >65

0%

4%

8%

12%

16%

20%

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

GDP per capita growth Average industrial wage growth


China only a fraction of global installed base, but is the market leader With regard to geographies, Japan currently has the biggest installed base of industrial robots accounting for 23% of the total. China’s share stands at 9%, however when compared in terms of annual shipments, the country clearly stands out of the other geographies with 66,700 units sold in 2015. IFR estimates that by 2017, China will account for 1/3 of annual units sold.

Chart 58: China only accounts for 9% of the global installed robot base…

Source: IFR

Chart 59: …but is representing the highest annual demand for new units.

Source: IFR

Chart 60: China robot penetration (per 10,000 workers) is still very low

Source: IFR

Chart 61: Robot density in EMs is still far behind developed economies

Source: IFR

Japan 23%

Germany 12%

America 17%

South Korea 11%

Europe excl. Germany

17%

China 9%

Asia excl. China, Korea

and Japan 11%

0%

10%

20%

30%

40%

0

10,000

20,000

30,000

40,000

China

North

Ame

rica

Japa

nKo

rea

Germ

any

Taiw

an Italy

Thail

and

Spain UK

Fran

ceInd

ia

Units sold in 2013 '13-'17 CAGR rhs

0100200300400500

South

Kor

ea

Germ

any

Denm

ark

USA

Italy

Finlan

d

Cana

da

Nethe

rland

s

Slov

akia

Czec

h UK 0200400600800

1000120014001600

Japa

n

Germ

any

US UK

China

Braz

il

India

Automotive General Industry


Rise of the “cobots”: collaborative robots Robots have been used for a long time in many industries to handle complex operations. Traditional models are wide-spread among manufacturers, operate strictly separated from humans and are most effective for repetitive work and large scale production that can be found for example in the automotive industry. But with the emergence of Big Data – robots are evolving for greater utility. Recent versions are characterised by increased autonomy, flexibility and the ability to cooperate. Especially this last attribute has helped phrase the term “cobot” (collaborative robot), a new generation of robot which is typically smaller, cheaper and easier to program than its traditional industrial peers.

Exhibit 23: Kuka industrial welding robots…

Source: Kuka

Exhibit 24: …and robots that can work together with humans (Cobots)

Source: Kuka

Traditionally robots have been seen as a substitute for simple and repetitive operations at low wage costs. In an effort to standardize quality, increase precision and drive down costs further, manufacturing companies gave resorted to the alternative to automate tasks via robots. With the emergence of collaborative robots (which we discuss later) opportunities in new industries will open up, as robots are able to perform tasks in an unstructured environment (through sensing technologies) and can apply logic to their operations. This development can be seen by the expected outperformance of human workers by robots (Chart 10). Over time technological advance and lower costs will make low-wage industries as well as sectors with technological limits accessible for robotic penetration, as can already be seen in the example of China.

Exhibit 25: Robots are expected to outperform human workers in terms of price/performance in the electrical equipment and furniture industry

Source: BCG

Exhibit 26: …showing how technological advances make even low wage areas accessible for increased robot penetration going forward

Source: BCG


Cobots differ from their traditional peers in terms of their smaller size, their lower price and their ease of configuration and reprogramming. The emergence of cobots presents a remedy for manufacturers in emerging markets which are confronted with wage inflation, and also for firms in developed economies that seek to lower their production cost in an effort to increase productivity.

Below we present key advantages of collaborative robots that, we believe, will lead to increased demand going forward:

• Less expensive when compared to traditional industrial robot. Although the cost for traditional industrial robots have come down over the past and are also expected to decline further (see Chart 8), the advantage of cobots is that they do not require expenses for safety systems/barriers and also have significantly lower cost in terms of installations and programming (which represent about 2/3 of overall costs). Also payback times for cobots are significantly lower than those of traditional ones. Universal Robots states their collaborative robots pay off after 195 days.

Chart 62: Example of cost for a traditional industrial robot system ($’000)

Source: BCG

• Increased flexibility for programming enables customers to reprogram and configure the robot in order to carry out a multitude of different task as opposed to one specific activity. Through this, manufacturers are able to quickly react to changes in demand and save on costs as different activities can be carried out by the same robot and do not require several mono-functional ones.

• Cobots are smaller than their industrial peers and also perform more controlled motion as they are equipped with technology that can sense its environment and therefore prevent potential conflicts with human co-worker. Hence, the robot provides a higher level of safety which ultimately also leads to increased productivity.

Demand from the top 10 countries accounts for a 80% of the total According to Kuka, demand for cobots is strongly driven by the food, electronics and also arc-welding segment. Besides their smaller size and lower weight, cobots are preferred in these industries due to low energy consumption high flexibility. In terms of geographies, 80% of demand is currently concentrated on the top 10 countries, which we believe might change going forward as other countries might also enjoy the benefits coming from the usage of industrial robots (such as decreased production cost) as wages are rising across all regions.

182 156

133 117

104

0

40

80

120

160

200

2005 2010 2014 2020E 2025E

Robot Peripherals System Engineering Project Management


Exhibit 27: 80% of the market is concentrated on the top 10 customers, while cobots are in high demand from the food and electronics space, but also in the arc-welding space

Source: Kuka

Case Study - YuMi In the first quarter of this year ABB commercially launched their collaborative robot called YuMi which costs around 40,000$. It is designed to work safely, in a seamless way with humans ensured by technology that stops the robot when it is getting too close to its human co-worker.

Exhibit 28: YuMi – IRB 14000

Source: ABB

Exhibit 29: A view on YuMi’s integrated controller

Source: ABB

• Safety: The robot’s magnesia skeleton arms are covered with a floating casing

wrapped in soft padding. This cushion is combined with embedded force-sensing technology that enables YuMi to sense collisions with co-workers and to pause any motion within milliseconds.

• Productivity: The same soft-padded arms also ensure a high degree of resistance and absorb unexpected forces. This reduces downtime, allows the robot to run faster and hence increase productivity.


• Cost-efficiency: Due to the integrated controller, the amount of space consumed is minimized and the robot is easier to relocate and transport. Furthermore costs on floor wiring and potential cable maintenance can be saved.

• Precision: Moving at a speed of 1,500 mm/sec to the same point over and over again with an accuracy of 0.02mm compares with Universal’s UR3 cobot and Roberta, a cooperation between ABB and Gomtec, which both operate at a 0.1mm accuracy at lower speed.

• Ease of programming: Besides traditional programming via coding for more complicated tasks, YuMi is able to adopt less complex motions simply via co-workers guiding its arms through a series of movements. The waypoints and sequences are logged on a paired tablet using the YuMi app. This ease of programming reduces time to ready the robot for the productions process and saves costs on potential software installations, hence increasing flexibility in the manufacturing process. Moreover, through the use of ABB’s RobotStudio, YuMi can be programmed and optimized without the need to interrupt operations.

• Smartness: YuMi displays features like build-in cameras (integrated vision) to locate products and deal with a less structured environment, or sensors that regulate the appropriate amount of force the robot should apply when putting small parts together (integrated force control).

Best-placed: ABB, Kuka, Fanuc, Yaskawa

The big-four provider of robots might continue to profit from the abovementioned forward. A potential catch-up of China in terms of robot penetration might pose an opportunity that ABB might be able to lever to its service-oriented business. The Swiss firm is also well-placed to profit from the adoption of robots in other industries and among smaller manufacturers, enabled by the emergence of cobots (ABB’s current split between automotive and general industry stands at c. 50/50 compared to 80/20 in the past). Moreover ABB is able to include its robots into its broader automation offerings, thereby exploiting cross-selling opportunities.

Logistics are an important part of the equation Automation of logistics, including packaging, material handling, and storage are vital components of a smart factory.

Exhibit 30: Mobile robot applications in logistics and distribution, Global 2014

Source: Seegrid, Frost & Sullivan

Logistical systems have three major types of flow: information, control and material. Automation of logistics processes will affect each of these. But the most important by far is the material flows which are at the core of logistical operations. Degrees of autonomy have been developing since the 1950s and currently close to 47% of logistics companies say that they are already using robots (source: Roboscan 2012). It appears


though that while many companies have product-processing and packaging operations that are at a high degree of automation, transportation and storage still largely requires the use of manually operated forklifts (source: EU 2015; DHL).

Table 14: Technology timeline

Year Technology Description 1950 Check Weighing A weighing apparatus that checks if the package has been filed with the correct amount of the

product. 1952 Bar Code The Woodland and Silver “bull’s eye” was first patent for a bar code 1960s RFID Radio Frequency Identification (RFID) was leveraged for commercial use for electronic article

surveillance (EAS) 1970s Commercial Use

Of Bar Code First commercial implementation of bar coding was for grocery distribution in 1970 and was soon made an industry standard by UGPIC, eventually evolving in the present day UPC code

1980s-90s

Warehouse Mgmt System (Wms)

Control the flow of inventory into, within, and out of a company’s DC in order to track inventory at all times.

1995 E-Auctions FreeMarkets launches the first e-Auction online negotiation platform, giving the supplier market feedback to improve their proposal

1995 E-Commerce Future king of eCommerce launched in 1995. Also in this year AuctionWeb launched their site that is later renamed eBay.

1998> Mega Dc’S Mega DC's emerge to serve mega markets to gain a competitive advantage at locations are which optimize multiple regional markets.

2000s Robotics/Automation Palletizing Software, Automated Case Pick, KIVA Rovers, Autonomous Fleet-Based Storage & Fulfilment developed, UAV delivery emerging

Source: Symbotic; BofAML Global Research

Logistics, packaging & materials worth US$31bn by 2020E The market for sales of robots used in logistics, packaging and materials handling will be worth US$18bn in 2015, and this is set to expand at a 10.1% CAGR and reach US$31.3bn by 2020E (source: Deloitte, WinterGreen Research).

Tremendous opportunities to enhance the efficiency of goods flow There is tremendous potential to enhance the efficiency of goods flow. 70% of logistics companies feel that there is a “High” need for investment in robotic solutions with 71% feeling that they would like to upgrade their current robotic solutions with intuitive robot programming. Important criteria for the implementation of robotic solutions are seen to be (1) adaptability to changing parameters, (2) ease of integration with existing technology and (3) ease of use.

Chart 63: How do you assess the need for investments in robotic solutions within the next five years?

Source: Roboscan 2012

Chart 64: Could you imagine upgrading your existing robotic solution with intuitive robot programming?

Source: Roboscan 2012

Warehousing operations: use of robots picking up In recent years, the use of robots – such as autonomous or self-driving robot vehicles – in warehousing has picked up.

70%

24%

1% 5%

High Low No need NA

71%

14%

15%

Yes No NA


Table 15: Current robots in use in warehousing operations

Operation Robot/Project Developer Description

Autonomous loading and transport

Open Shuttle KNAPP Free-moving vehicle that uses laser navigation technology and can be used for transport and picking activities involving cartons and containers.

RoboCourier Swisslog Ideal for tight and spaces it uses laser navigation and has a 360-degree turn range MultiShuttle Move Fraunhofer IML / Dematic Self-driving vehicles handling small load carriers and pallets that can operate in a swarm Auto Pallet Mover Jungheinrich Selfdriving pallet mover that can also be used in conjunction with manual vehicles. MOVEBOX kits BALYO Can convert regular electric forklifts into self-driving vehicles requiring no additional

infrastructure FTF out-of-the-box Jungheinrich/Götting

KG/University of Lübeck/IPH Hannover

Autonomous forklift equipped with a 3D camera and intelligent image recognition software

Assisted order picking

FiFi BÄR Automation/KIT Self-driving vehicle that follows the order picking process using vision guidance, and responds to human gestures

Kiva warehouse automation system

Acquired by Amazon Mobilizes warehouse shelves by attaching an autonomous vehicle to the bottom so that the entire shelf can be transported

Outdoor logistics operations SaLsa research project - Testing autonomous transport vehicles using sensors installed in the yard infrastructure Altenwerder Container Terminal

- 84 driverless vehicles transport containers between the wharf and the storage areas via 19,000 transponders that are installed in the ground

Source: DHL

Strong navigation and situational analysis capabilities are essential for autonomous vehicles. Today the most common method is to rely on a mixture of depth cameras and lasers which can constantly scan the environment to identify position and obstacles. It is suggested that the logistics industry will adopt self-driving vehicles faster than most other industries as it is possible for logistics vehicles to move around in a secure private zone. There are applications across autonomous loading and transport, assisted order picking and outdoor logistics operations (Source: DHL)

Table 16: Expanding AGV capabilities, Global 1990-2022

Year Name Description 1990 Nissan Forklift QX2 Series Warehouse control systems: Forklift fleet with data tracking sensors 2012 RapidStore Miniload ASRS Miniload technologies: Automated storage and retrieval systems 2016 Mercury xPRESS RFID: Voice recording pickup 2018 Seegrid Fully automated warehouse: Order picking systems 2022 Jafza Aerial Delivery Automated unmanned aerial logistics and distribution Source: Frost & Sullivan


Amazon Kiva robots Amazon in 2012 paid US$775mn to acquire Kiva systems, which makes small, fast, flexible warehouse robots. The 30,000 robots spend 65-75% of their day walking within Amazon’s 103mn ft2 of fulfilment centres (Source: Deloitte). The robots move in grid patterns on the warehouse floor, and move large merchandise from shelves to the packing & shipping areas. This helps Amazon with a low-cost way to offer one- and two-day guaranteed delivery and can reduce fulfilment costs of an average order by 20-40% (Source: BCG). Amazon is currently also experimenting with delivery drones.

Exhibit 31: Amazon Kiva robots

Source: South China Morning Post

The EU PAN-ROBOTS project

Six partners across five EU countries have come together to develop a generic automation system for factory logistics. Bolstered by €3.33mn in funding from the EU this project aims to enable the ‘Factory-of-the-Future’ (FoF) by incorporating the use of advanced Automatic Guided Vehicles (AGV). It involves four main work areas: exploration systems for 3D-mapping of the plant; advanced perception systems on-board the AGVs; a modern control centre for monitoring them, and cooperative infrastructure laser scanners distributed about the facility. AGVs in combination with these advanced infrastructure systems will be able to transport material and products using autonomous on-board path planning and navigation. The installation time and costs is expected to be reduced via the use of a localization approach utilizing already existing landmarks and an advanced pallet handling system. Half of all European factories will be operating with AGV fleets by 2030. (Source: European Union, 2015)

Exhibit 32: Warehousing operations of the future

Source: DHL


6. DCS disruption: real and present danger

Most process plant is over 20yrs old and the average age of plant assets is closer to 35. To date, the traditional model has been for systems to be replaced in entirety every 10-15yrs, but now ExxonMobil (XOM) is looking to change that.

The appointment by XOM of Lockheed Martin (LMT) to a lead systems integrator role in the development of a new DCS architecture highlights to us that DCS technological disruption is happening now and that it could have considerable consequences for the incumbent providers. Saudi Aramco appears to have developed similar ideas independently and concurrently, suggesting the threat to the current competitive landscape is real and current.

Best placed – Lockheed Martin, Schneider and Emerson The issue with calling who is best placed to benefit from this disruption is that the logical outcome is that none of the incumbent providers (Honeywell, ABB, Emerson, Schneider, Yokogawa) will fare well. Newer players in the market such as Rockwell Automation may prove more open to new methods and could be able to build competitive position. At face value, LMT as the system integrator in XOM’s new architecture could be a prime beneficiary however the fact that neither XOM nor Aramco appear to have included safety systems within the scope of their vision for future process automation architecture might help Schneider. Similarly, the envisioned architecture appears to be based on increased capabilities at the field instrument/device level, which ought to favour vendors with a higher proportion of that equipment (Emerson) or the direct providers thereof, e.g. IMI (valves), Flowserve (pumps, valves, actuators), Rotork (actuators).

Appointment of non-incumbent points to threat Although only at the early stages, the appointment of LMT by XOM to develop a new DCS system represents a tangible and visible threat to the incumbent vendors of DCS software such as ABB, Honeywell, Emerson and Yokogawa. XOM’s intention is to have a commercially available prototype system by 2019, suggesting that the threat to the current industry structure is both tangible and visible from the perspective of an investment horizon. While the decisions made by such a substantial industry participant as XOM make us sit up and take notice on their own, the fact that independently Saudi Aramco has been investigating strikingly similar changes suggests to us that the threat to the current industry structure and vendors is significant.

A typical process plant includes four key components: inputs (typically from field devices like temperature and pressure sensors); outputs (typically control valves); DCS controller that collates, translates and analyses the inputs and the outputs, and; the HMI through which the operators interact with the DCS. Incumbent DCS vendors such as ABB and Honeywell believe the value resides in the DCS and its related software and the complexity of such systems has historically led to long system life, reluctance on the part of customers to upgrade or switch and the market to be relatively concentrated.

To date, most of the change in the existing provision of process automation has been on development of more advanced automation systems which has resulted in the development of the role of Main Automation Contractor, which could extend to Main Automation and Information Contractor (MAIC) with the advent of the industrial internet. Most of the automation suppliers hold annual customer days where the event is geared towards developing better solutions to the operational problems (e.g. Honeywell Users Group or Emerson Exchange). However, what is most interesting about the XOM developments is the appointment of LMT to the lead role – an industry outsider has been preferred to an incumbent vendor. XOM clearly believes in order to evolve to the next generation in process automation they need to turn to somebody without a vested interest.


Exhibit 33: Exxon Mobile is looking to simplify the structure of automation


Exxon’s basic arguments have been set out by Sandy Vasser (Senior Adviser at the ExxonMobil Production Company) and Don Bartusiak (Chief Engineer Process Control at ExxonMobil) at the last two ARC Industry Forums (February 2015 and February 2016), with the basic issue being current DCS are too difficult to replace and do not provide sufficient value in their current form. These issues are echoed by Saudi Aramco, with the added frustration of the limits of the DCS controller, which have natural limits on the available compute, storage and sensor Input/Output (I/O) resources. In both instances the customer appears to be frustrated by the rigid hierarchies of existing DCS architecture (field device, controller, server) and the inflexibility of available vendor solutions - interoperability and compatibility is an oft-pedalled notion but the reality is very different.

It is striking to us that both of these critical customers in the process industry are expressing frustration at the lack of flexibility and yet the "solutions" we have seen from incumbent providers even at recent user group presentations (e.g. Honeywell in June 2016) continue to refer to "one-stop shop" capabilities. It is obvious that incumbent vendors wish to protect the installed base, but the reality appears to be that customers want a fresh approach, and in Exxon and Saudi Aramco, these are not just any old customers. Neither Exxon nor Aramco make reference to safety systems in their respective "ideal" solutions, suggesting companies like Schneider which, through Invensys, have a more safety system bias may escape the initial disruption to the existing industry structure.


At its heart, the new process automation architecture suggested by both Exxon and Aramco appears to be open. Commercial Off the Shelf (COTS) systems and open software to enable interoperability seems to be at odds with the professed openness of incumbent suppliers' solutions. While we understand the natural reaction of incumbent providers to protect their installed base, it seems to us that there is a fundamental gulf between industry titans such as Exxon and Aramco and their automation vendors, manifested in the choice of LMT as the systems integrator for Exxon 's new generation DCS platform. Historically the 12-15yr life of automation system lifecycle led to long customer relationships, significant replacement cost in terms of time and money and a sales structure that was opaque in terms of pricing. Ultimately, both Exxon and Aramco are looking to simplify their options - standardisation, openness, interoperability are the key tenets of their visions for future process automation with much greater capability at the device or field instrument level.

Exxon’s vision sees the replacement of the existing reliable but inflexible DCS hierarchy with nodal control - the concept of a Distributed Control Node (DCN) while the Aramco architecture appears to be based on Smart Junction Boxes and increased capability around the remote I/O sensors. Ultimately this amounts to the same thing, with control residing in a more open environment at the server level. Greater reliability and security is available with COTS systems than has been the case historically.

Exhibit 34: …With a single, open architecture straddling the whole operation

Source: Company data

To our mind, these developments represent a significant and near term threat to incumbent process automation suppliers. Granted, it will take some time for widespread


adoption to happen but the fact that Exxon and Aramco are the driving forces lends considerable weight to the issue. Exxon desire a commercially available system by 2019 and Aramco has, according to ARC Advisory, already run small scale pilot projects to test the feasibility of its system architecture. As an additional consideration, it would seem that as neither future vision considers much change to existing safety systems and focuses on improved capabilities further down the automation hierarchy, incumbent suppliers with higher safety system exposure (Schneider) and higher field device/instrument revenues (Emerson) might be slightly better insulated. The changes proposed by Exxon and Aramco represent fundamental changes to the way process automation has been implemented historically and we think investors should not underestimate the scale of the proposed changes.

Chart 65: DCS market share, 2014


Chart 66: Process automation sales, 2015 (USD bn)

Source: BofA Merrill Lynch Global Research estimates

The company that would appear most at risk from fundamental changes to the existing process automation market structure would be ABB, with Honeywell the next in line by virtue of existing market shares. Having said that, we do not imagine any of the existing suppliers are standing still in terms of any response to the Exxon or Aramco initiatives. The main issue we see is the continued adherence to the notion of "one-stop shop". Honeywell's recent User Group conference (June 2016) launched the EdgePLC to address some of edge computing capabilities required by the advent of the industrial internet, but there was undoubtedly an emphasis on the ability of Honeywell to meet all the requirements of the potential process automation customer.

The proliferation of standards that characterises the process industry actually highlights the lack of openness of existing architectures that is at the core of the frustration of Exxon and Aramco. Although professing to be open, interoperable and flexible, incumbent vendors' solutions are only flexible if a customer buys everything from one supplier. This appears to be contrary to the prevailing customer requirements if Exxon and Aramco are representative of the industry (which they should be!). The recent agreement of the Industrial Internet Consortium (IIC) and Industrie 4.0 that they share common visions and plan to develop global standards is very much aligned with the core of the Exxon /Aramco vision for more open DCS architecture. While incumbent suppliers are involved in these discussions, the resultant presentations of IoT-ready solutions suggest little has changed in terms of vendor mindset. One of the key motivations in the Exxon /Aramco approaches is the desire for flatter system architecture and a move away from the field device>server>controller hierarchy, which is a cumbersome constraint of legacy systems in light of current technology but which remains a key feature of even the most modern DCS offerings.

ABB 21%

Honeywell 15%

Siemens 14%

Emerson 14%

Yokogawa 13%

Schneider 6%

Rockwell 5%

Alstom 2%

Valmet 2%

Others 8%

0%10%20%30%40%50%60%70%80%

01234567

SIE ABB SU ROK HON YOKO EMR

DCS Sales Other Process Automation Sales

Process as % of total sales


7. Cybersecurity issues looming large

Cybersecurity is cited as one of the main inhibitors to adoption of new connected technologies and the industrial internet. As systems become more open the potential for data breaches, system hacks and proprietary technology leaks becomes greater. Developments in system integrity and security will need to evolve at the same pace as the technological developments of the industrial internet to ensure companies are sufficiently confident to adopt new solutions.

Best placed – integrated providers We doubt there is a single solution for cyber security given the nature of most industrial networks, but we think vendors offering robust protection processes (rather than just a product) as part of its overall control/safety offering will be well placed. Ultimately we think this issue comes down to customer trust and is probably indistinguishable from the outside from an investment perspective. It is likely to come down to the partnership decisions the various providers make more than anything.

Increasing occurrences of security breaches highlight the risks of a more connected world With greater connectivity, open software architecture and the usage of cloud-based storage the issue of cybersecurity is becoming ever more important. Increasing numbers of high profile security breaches have raised the profile of cybersecurity as an issue for industrial companies, particularly with respect to endangering life (hacking connected cars resonates with consumers!) but also with respect to corporate intellectual property (IP) and infrastructure security (e.g. power stations). Our thematic investing colleagues have written a substantial report on cybersecurity (link), so we will not revisit the details and trends again here, but we note that data security is one of the biggest inhibitors to adoption of new technologies in the industrial sector generally and this is even more so with the advent of the industrial internet.

Chart 67: ICS-Cert cyber incident reports by sector (2011)

Source: BofA Merrill Lynch Global Research, Frost and Sullivan, Schneider

Table 17: Examples of cyber-attacks on control systems

March 2000, Australia SCADA compromised at Maroochy Shire Sewage

Jan 2003, US Cyber attack on David-Besse Power station of First Energy

Jan 2008, Poland Public tram system hacked remotely

Dec 2010, Iran Stuxnet attack in Iranian nuclear plant

Nov 2011, Iran Duqu attacks

Source: BofA Merrill Lynch Global Research, Frost and Sullivan, Schneider

From 2006 to 2012, the number of cyber security incidents in total (not just automation related) reported to the US Computer Emergency readiness team increased by 782%. Detailed information on industrial automation related issues is limited (as end users are wary of reporting them), but there have been some well-publicised examples in adjacent automation markets, which underlines the risk. Perhaps the best known issue was the Stuxnet worm, which is reported to have included a PLC instruction to destroy a centrifuge in an Iranian nuclear power plant.

water 41%

energy 16%

govt 6%

nuclear 5%

chemical 4%

cross sector 25%

other 3%


Returning to our comparison of recent PWC surveys on the industrial internet we note that unresolved questions around data security appear to have risen in the last couple of years, where benefits of implementing the industrial internet have become clearer and problems around talent availability and common standards appear to be lessening. Concerns around some form of data security account for 50% of respondents' key concerns around the implementation and growth of the industrial internet. Chart 68: Change in PwC survey respondents answers to top challenges or inhibitors to building digital operations capabilities, 2016 vs 2014

Source: PwC, BofA Merrill Lynch Global Research

Chart 69: PwC survey respondents’ key concerns in terms of data security

Source: PwC, BofA Merrill Lynch Global Research

Both Siemens and Schneider have co-authored whitepapers addressing the cyber-security issues inherent in any more open or connected automation architecture. In our view, the security issue is an inhibitor to adoption, but not a differentiator in and of itself. All automation providers need to develop robust systems and the ability to defend themselves and, more importantly, their customers, from outside access. Customer trust is therefore critical to the adoption of new technologies, but this is not a new issue in terms concerns automation vendors have to deal with. Granted, security is becoming more complex and more high profile, which makes the solution more complex in turn, however we do not believe it becomes a differentiating factor between industrial automation providers save perhaps for the speed with which the vendor might be able to convince its customers to adopt new solutions.

Exhibit 35: Security involves a number of layers

Source: Verizon

0%5%

10%15%20%25%30%35%40%45%50%

Unclearecononmic

benefits

Insufficienttalent

Lack ofstandards

unresolvedquestions ondata security

2014 2016

0%10%20%30%40%50%60%

Oper

ation

al dis

rupti

on

Liabil

ity th

roug

h data

loss

Unau

thoris

ed da

taex

tracti

on

Dama

ge to

repu

tation

Misu

se of

data

Loss

of IP

Viola

tion o

f reg

s/law

Enda

nger

ment

of life

Governance, Risk Threat Authentication Professionaland compliance management and privacy security servicesPrepare to Protect the Trust the Respond tomanage risk perimeter ecosystem the threats

Access governance Security configuration Data discovery Rapid responsThreat vector analysis management M2M security servicesPenetration testing Vulnerability scanning Managed certificate Digital forensicsPartner security programme Application scanning Application securityPCI Compliance programme Content scanning Smart credentials

Cloud assessment SSL certificates


8. Disruptive technologies nearer than you think

Four additional nascent technologies that could have a meaningful impact on the industrial consciousness are block chain, artificial intelligence or machine learning, augmented reality and connected wearables. Their potential impact is both a revenue and cost opportunity for industrial companies and in some instances companies are already taking their first steps, e.g. Philips using connected wearables for remote patient monitoring or Honeywell offering enhanced service capabilities through augmented reality.

Perhaps what is most interesting in this context is the scope for new market entrants to develop new applications. To illustrate this point simply we recall attending an industrial internet conference in late 2015 where we heard:

• Orange talking about connected healthcare;

• Shell talking about connected cars;

• AT&G talking about connected logistics; and

• Samsung talking about the connected home (more adjacent than the others).

The most striking observation for us is that not many of these companies would be considered "natural" inhabitants of the industry in which they are offering a connected solution. It highlights nicely how disruptive a pervasive internet could be and therefore the potential implications of technologies such as block chain or machine learning.

Block chain: potentially revolutionising the supply chain Block chain is a form of mutual distributed ledger (MDL). It is the underlying technology that Bitcoin works on but it is not Bitcoin. While we will not launch into a full scale discussion here, we think the technology could have potentially far reaching effects in bringing the industrial supply chain to the next level of efficiency and just-in-time production. Our colleagues have written extensively on block chain (link), though mostly in reference to the finance industry, but we believe there is considerable potential for working capital efficiency to be gained in applying block chain to the industrial supply chain. SAP recently articulated its digital economy strategies for industries where they identified block chain as one of the five critical tech trends.

In simple terms and in reverse order:

A ledger is simply a set of records. The records can relate to a multitude of items such as bank balances, votes, health records, payment transactions and so on.

These ledgers are typically held centrally; banks, for example, hold a ledger with bank balances for their customers. The National Health Service holds the health records for patients in the UK. The government holds the electoral register with a list of citizens registered to vote. Generally, a central party decides what data is recognised in the ledger and provides trust to users of the database that the data held is correct and has not been tampered with.

With a distributed ledger, there is no single master database held by a central party. Instead, various parties hold copies of the whole database. These parties/nodes1 either have a vested interest in maintaining that database or are provided with some form of economic incentive for their effort. Instead of a central party making decisions on what goes into the ledger, cryptography techniques and consensus among a certain user

1 Nodes are independent computers in the blockchain network that perform calculations and keep the ledger. The number of participants in the network can be significantly higher than the number of nodes.

http://research1.ml.com/C?q=zL5YC53G-Bm0V1nJA8Gd-A


group, or potentially anybody from within the parties holding the database, are used to decide what data is recognised in the ledger. Trust in the accuracy of the ledger entries is obtained by mutual consent rather than through a central ‘trusted’ third party.

In a World Economic Forum survey, 73.1% of respondents noted that they expect taxes to be collected using a block chain by 2025, while 57.9% said they expect 10% of global GDP to be stored on block chains by this time. The potential is considerable but the time to adopt is likely at least 3-5yrs if not more, though we understand Toyota is one of a small number of industrial companies who have investigated the application of block chain to their supply chain.

Block chain in supply chain management

Block chains can be used to share provenance, demand and inventory data more efficiently between the different parties in a supply chain.

As an example, consider the semiconductor sector. Within the supply chain, multiple parties such as distributors, OEMs etc separate semiconductor companies from end customers. Hence semiconductor companies generally tend to have very little visibility on true end market demand and a significant amount of demand is driven by inventory related fluctuations in the supply chain. This is one of the key reasons behind the amplified demand cyclicality. Using block chains to record/share inventory and demand data can enable just in time manufacturing to be better aligned with true end market demand and potentially help reduce inventory-driven demand cyclicality. Similarly, tracking provenance data using block chains can increase transparency over the true origin of products and help with ethical sourcing.

Artificial intelligence: machines learning to think for themselves Our US technology research colleagues recently published an in-depth look at the progress being made in terms artificial intelligence and the ability to harness its potential (link). In simple numbers, the 2015 USD2bn AI market is expected to grow to USD36bn by 2020 as the advances made in machine learning are applied across a number of industries. The ad service, investments, retail and media markets are expected to be early adopters, but the implications for smarter manufacturing and increased productivity are considerable.

Combining machine learning with advanced robotics such as the new ABB Yumi robot that can work alongside human workers has considerable implications for productivity improvements and the speed of programming new tasks. Connecting factories and machines is generating the data sets required to develop predictive AI to better manage the production response to changing demand. Adoption rates, particularly in the more cautious industries such as oil & gas, may be slow, but according to our colleagues' research four big industrial segments account for a quarter of AI revenue in 2015.

http://research1.ml.com/C?q=nBcpfE4lkXL9-sf7PauJKA


Chart 70: Artificial Intelligence revenue end-market 2015

Source: Tractica

Augmented reality: a tool for selling, training and enhancing service Digi-Capital expects the AR market to be worth USD90bn by 2020 with enterprise applications a significant share of that and our thematic research colleagues have published a very timely primer on AR as part of their continuing work around The Transforming World (link). The potential for AR to enable selling, training and service enhancement is considerable, with industrial applications already starting to enter mainstream corporate thinking. Several companies have developed AR-based products which will help improve the design process, the training process and also enhance service provision. This has implications for productivity as well as personnel safety, making training much more realistic and ensuring trainees can go out into the field with a much better idea of what to expect. AR could also have an influence on the sales process, allowing a customer to experience the product before buying.

Both Hexagon and Honeywell have developed AR products. Hexagon's unique position in measuring and matching the physical world with the digital world means the data clouds it generates are well suited to the use of AR. Its SmartPlant technology allows users to walk through the digital plant as if they were in the real version, which enables training, disaster simulation and design optimisation. Honeywell has been developing AR capabilities in both its aerospace and defence businesses.

PTC's development of the Vuforia platform places it at the heart of the industrial application of augmented reality. PTC bought Vuforia last year (October) and now believes it is the most widely used platform for AR development for enterprise applications. Vuforia's AR technology is integrated with PTC's PLM capabilities to enable 3D depiction of models via AR headsets. A commercial application would be an AR overlay of the original 3D design for a repair technician.

Connected wearables: at the heart of personal and personnel safety The proliferation of products such as FitBit and the ability to use the internet to monitor our every move should we so wish has practical implications for industry. Such technologies have obvious application in remote healthcare (which Philips is developing for example) but also have broader industrial applications in personnel tracking and incidence response. Wearable devices could be used to improve the productivity and capability of any employee used to working “deskless” (e.g. service technician) but who could benefit from access to information at the same time as keeping hands free. Arguably an extension of AR/VR technology, connected wearables could be a much broader route to productivity and worker safety.

Ad Service Tech Investment Retail

Media Other Oil & Gas

Manufacturing Automotive Agriculture

http://research1.ml.com/C?q=Q!67FxVleDNiGh4B4yRURQ


Philips' strategy is now very much focused on HealthTech and one of the areas they are seeking to develop is connected healthcare, with an emphasis on remote patient monitoring. This can range from remote alarm activation in the event of an elderly person falling at home or having a stroke to more complex monitoring of drug delivery or vital signs (e.g. pacemaker or blood sugar levels). Philips and Accenture have also demonstrated technical feasibility of using Google Glass in the operating theatre integrated with Philips healthcare software.

Honeywell have developed an industry leading personal safety business which could be seen as the forerunner of connected wearable and which is clearly aligned with the development of the connected wearables market. The Honeywell Connected Worker solution, in conjunction with Intel, is based on a broad range of sensors that combine to provide an accurate representation of the worker environment, vital signs and location. Tracking of personnel through a plant has been around for some time, but combining the basic tracking capability with AR or AI can be used to better protect lives and also respond more appropriately to an incident.

Oracle has built interfaces for its JD Edwards ERP system with smart glasses, smart watches and smart phones to enable better worker inter-activity. Tata has developed a smartwatch which monitors health and environmental risks for factory workers and which is being piloted by crane operators at Tata Steel in India.

GE has developed Smart Helmets in its Oil & Gas business that directly connects field engineers to more experienced technicians back in the office to deal with the joint problems of an aging, but experienced workforce and the urgency of fixing a problem.


State of play in the end markets In this section we review historical growth rates by region and product, demand over the next 6-12 months and longer term growth drivers for automation markets. In the near term, we think Discrete automation markets, particularly in the US and Europe, are best placed to benefit from the recent stronger global IPs and industry orders. Discrete markets are typically early cycle and have been relatively flat since 2011, versus Process markets which are 10%+ ahead, but weaker of late because of low oil prices.

Near term - Discrete automation markets would benefit the most from near term pick up in PMI’s

Longer term, we believe industrial automation markets can at least maintain the 3-4% growth seen historically given the number of structural drivers. Higher productivity requirements, energy efficiency, increased outsourcing, quality/ regulation and higher EM wage inflation are all meaningful tailwinds to growth in our view, which should continue to support growth above global IP.

Historical trends: 3-4% growth, above global IP

In the next section we look at the growth rates achieved historically for each region relative to sector growth rates, regional and global IP. While the major players in each region are global, most do have a bias to domestic markets. Where we can, we have tried to strip out the most relevant automation related specific division, although for some companies the data used will be distorted by none automation related sales and is not available on a long term consistent basis due to changing divisional definitions.

Longer term – Industrial automation markets should be able to continue to outperform Global IP given the number of structural drivers – productivity, wage inflation, quality, energy efficiency and the greater connectivity.

In general, the main players in each region have grown faster than local IP and global IP growth with 3-4% growth vs global IP growth of slightly under 3% (2002-12). Niche players show a wide range of growth rates – although have typically been slightly earlier cycle, more cyclical but seen faster through-cycle growth.

Europe – major players in line with sector, range at niche In Europe the major players have grown at broadly similar rates to the rest of the Capital goods sector – averaging 3.7% growth 2002-2012. In general, the major stocks have been slightly later cycle than the sector, although not by more than 3-6 months, largely due to process exposure.

The niche automation businesses have shown a large range of growth rates. The businesses/divisions are rarely pure plays, but should be driven by automation related trends. Since 2007, IMI has grown at a slower rate than the sector (in part impacted by the truck exposure), Spectris’ relevant divisions (Inline Instrumentation and Industrial Controls) have grown broadly in line with peers – although the shape of the group has changed following the acquisition of Omega, and Hexagon’s Metrology business has clearly outperformed the automation markets with 10%+ growth. KUKA has also performed strongly, seeing just under 10% headline growth since 2007.

The major global players have typically outperformed local and global IP. Considerable range of growth profiles at niche companies, depending on the nature of products


Chart 71: Europe – automation growth rates vs European & Global IP and Capital goods avg growth

Source: BofA Merrill Lynch Global Research Estimates

Chart 72: Europe – range of growth rates at European niche players


US – large players saw 3-4% growth historically, similar cyclicality Similar to the large European peers, the large multi-product US names have grown at 3.4% c. 2002-2012 (5-6% ex 2009), with a relatively limited difference between companies (although Rockwell has seen slightly higher growth). Like Europe, this is well ahead of regional IP, demonstrating that the large US names are typically global franchises. Both the large European and US automation players have also on average seen similar levels of cyclicality with +/-15/20% growth around 2009.

Chart 73: US – automation growth rates vs US & Global IP – similar growth and cyclicality to large European peers


Chart 74: US – range of niche markets – Parker and National Instruments early cycle, strong growth at Aspentech


Within the niche names, Aspentech (software) has grown significantly faster than the sector (although this is a headline growth). Parker Hannifin and National Instruments have been relatively early cycle and have higher levels of cyclicality, while Flowserve has been relatively later cycle with a more limited cycle given its process related exposures.

-30.0%

-20.0%

-10.0%

0.0%

10.0%

20.0%Q2

2000

Q2 20

01Q2

2002

Q2 20

03Q2

2004

Q2 20

05Q2

2006

Q2 20

07Q2

2008

Q2 20

09Q2

2010

Q2 20

11Q2

2012

Q2 20

13Q2

2014

Q2 20

15Q2

2016

Euro IPEuro - Cap gds sector growthGlobal IPEuro - Automation growth -60%

-40%

-20%

0%

20%

40%

60%

Q2 20

05

Q2 20

06

Q2 20

07

Q2 20

08

Q2 20

09

Q2 20

10

Q2 20

11

Q2 20

12

Q2 20

13

Q2 20

14

Q2 20

15

Q2 20

16

Hexagon MetrologyIMI - Fluid PowerKukaMetso - AutomationSpectris - InlineSpectris - Ind C

-0.3

-0.2

-0.1

0

0.1

0.2

Q2 20

00Q2

2001

Q2 20

02Q2

2003

Q2 20

04Q2

2005

Q2 20

06Q2

2007

Q2 20

08Q2

2009

Q2 20

10Q2

2011

Q2 20

12Q2

2013

Q2 20

14Q2

2015

Q2 20

16

US - Automation growthUS IPGlobal IP

-60%

-40%

-20%

0%

20%

40%

60%

Q2 20

07Q4

2007

Q2 20

08Q4

2008

Q2 20

09Q4

2009

Q2 20

10Q4

2010

Q2 20

11Q4

2011

Q2 20

12Q4

2012

Q2 20

13Q4

2013

Q2 20

14Q4

2014

Q2 20

15Q4

2015

Q2 20

16

AspentechFlowserveParker HannifinNational InstrumentsDanaher


Japan – higher cyclicality, recent pick up Similar to the other regions, we believe the major Japanese automation players have typically out grown both global and domestic IP although the numbers used for the Japanese names are headline, rather than underlying organic growth. The larger companies have also seen considerable cyclicality, with significant declines seen in 2009.

Aside from Fanuc, growth rates across the Japanese names have been reasonably similar historically, with slightly higher cyclicality at Yaskawa and SMC, and lower cyclicality at Yokogawa. More recently all players have seen solid 10%+ growth.

Chart 75: Japan – headline automation growth rates vs Japanese & Global IP – similar growth profile but higher cyclicality


Chart 76: Range of headline growth rates in 2010 – 2011 – with stronger growth at Fanuc, Yaskawa and SMC


Products – Discrete has outperformed Process Many of the larger companies supply a broad range of products across a number of different areas in automation markets. However, below we try to look at relative product growth rates based on companies with the major exposures to each product grouping. While the analysis is not perfect due to the relatively small sample size for each product area and product overlaps within each division, it should give a rough idea for relative growth rates.

Chart 77: PLC markets have outgrown DCS markets since 2005. DCS recent weakness derives from O&G capex contraction

Source: BofA Merrill Lynch Global Research Estimates Discrete = Siemens Ind Auto, Rockwell, Mitsubishi, Schneider. Process – ABB, Honeywell, Emerson, Yokogawa, ISYS

Chart 78: Fluid Power typically one of the earliest cycle product areas

Source: BofA Merrill Lynch Global Research Estimates. Field instrum – Danaher, National instruments, Spectris, Omron. Fluid Power – Parker, IMI, SMC. Robots – Kuka, Fanuc, Yaskawa. Motors & drives – ABB & Siemens

-50%

-30%

-10%

10%

30%

50%

70%

Q2 20

04

Q2 20

05

Q2 20

06

Q2 20

07

Q2 20

08

Q2 20

09

Q2 20

10

Q2 20

11

Q2 20

12

Q2 20

13

Q2 20

14

Q2 20

15

Q2 20

16

Japan - Automation growth

Japan IP

-70%-50%-30%-10%10%30%50%70%90%

Q2 20

04

Q2 20

05

Q2 20

06

Q2 20

07

Q2 20

08

Q2 20

09

Q2 20

10

Q2 20

11

Q2 20

12

Q2 20

13

Q2 20

14

Q2 20

15

Q2 20

16

Fanuc (calendarised)OmronYaskawaYokogawaSMC

-25%

-15%

-5%

5%

15%

25%

Q2 20

02Q2

2003

Q2 20

04Q2

2005

Q2 20

06Q2

2007

Q2 20

08Q2

2009

Q2 20

10Q2

2011

Q2 20

12Q2

2013

Q2 20

14Q2

2015

Q2 20

16

PLC DCS

-60%

-40%

-20%

0%

20%

40%

60%

Q2 20

05

Q2 20

06

Q2 20

07

Q2 20

08

Q2 20

09

Q2 20

10

Q2 20

11

Q2 20

12

Q2 20

13

Q2 20

14

Q2 20

15

Q2 20

16Field instrumentsMotors & drivesFluid powerRobots


Since 2002, discrete/PLC markets have grown 100bps faster than process markets, although much of this is due to the significant rebound seen in 2010/11. Since 2005, in PLC markets Schneider and Rockwell have grown the fastest, while the Japanese players have grown at the slowest rates. In DCS markets, ABB and Emerson have grown the fastest, while Honeywell and Yokogawa have lagged. Robots have been one of the fastest growing markets, and fluid power has been one of the earliest cycle parts of the automation market.

PMIs suggest PLC market likely to stay soft, DCS demand affected by O&G capex contraction In the near term, most indicators are pointing downwards as PMI’s continue to look weak. We don’t expect a significant bounce in IP, but a pick-up in these indicators should benefit the shorter cycle Discrete/PLC markets (particularly In Europe) which have historically correlated well with PMI’s.

Chart 79: US, European and Chinese PMI’s – Global PMI around 50, but weak China depresses overall reading


Chart 80: Euro PMI vs Siemens automation organic growth


Global industrial & automation specific data points Below we look at the relevant indicators for near term demand in global automation markets;

Global PMIs have come off their highs with US and the Global PMI around 50 and China well below and Europe still above 50. The main drag to the global PMI readings have been significantly weaker China and US readings.

Industry data – the NEMA electro-industry business confidence index suggests that conditions in 6 months time should be better than current conditions across all regions. While the data series can be quite volatile, there is a notable spread between conditions now and those expected in the next 6 months in Europe, suggesting an improvement here – see Appendix 9.2.

Emerson 3 month order data – Emerson’s 3 month rolling order data (Chart 37) shows a significant deterioration in process automation (impacted by low oil price) and industrial automation orders (weak industrial spending, upstream Oil&Gas and Europe).

Japan export machine tool orders – have turned very negative following a positive stretch (double digit growth from Dec13-Feb15). The last readings have been declining to -31% in Aug from -18% in July and -11% in June (Chart 38).

Company commentary – although there have been some exceptions (Siemens, Parker) and management teams in Discrete markets have limited visibility, overall company commentary has been more positive on near term demand.

30

35

40

45

50

55

60

06 07 08 09 10 11 12 13 14 15 16

US ISMEuro PMIChina PMI HSBCGlobal PMI

-35%

-25%

-15%

-5%

5%

15%

25%

35%

45%

25

30

35

40

45

50

55

60

65

08 09 10 11 12 13 14 15 16

Euro PMI Siemens automation growth


Chart 81: Emerson 3 month rolling order data vs US automation growth


Chart 82: Japan machine tool orders have turned very negative since June 2015


Discrete most likely to benefit in the near term When indicators rebound, we think discrete automation markets are the most likely to benefit in the near term given the nature of products/customers, route to markets and historical early-cycle characteristics. Chart 39 shows that through the recent slow-down in demand, discrete demand is down 1-2% since the end of 2011, while process markets are 10%+ higher as companies have executed on backlogs.

Chart 83: Recent relative performance between PLC and DCS markets


-35%

-25%

-15%

-5%

5%

15%

25%

35%

04 05 06 07 08 09 10 11 12 13 14 15 16

Emerson - process autoEmerson - ind autoEurope - automation gr

-90

-40

10

60

110

160

210

260

03 04 05 06 07 08 09 10 11 12 13 14 15 16

90%

95%

100%

105%

110%

115%

120%

125%

Q2 20

12

Q3 20

12

Q4 20

12

Q1 20

13

Q2 20

13

Q3 20

13

Q4 20

13

Q1 20

14

Q2 20

14

Q3 20

14

Q4 20

14

Q1 20

15

Q2 20

15

Q3 20

15

Q4 20

15

Q1 20

16

Q2 20

16

PLC DCS


China – mixed near term Below we include a separate section on China, given competition related issues in other Electrical end markets, strong growth (alongside other local industrial markets) and China is now large in the context of global automation markets.

In the near term, we expect growth to be relatively mixed. Recent PMI’s indicate an improvement for discrete/PLC markets, but high investment levels over the last 3-4 years are likely to be a headwind for process related/DCS markets.

In terms of competition, we think the large global automation players have done a good job at establishing strong market positions, with no large domestic players. Competitive dynamics do differ from other end markets which have been impacted by Chinese competition (power, trains, and construction equipment) and we think the automation markets are likely to remain relatively stable, at least in the medium term.

Chart 84: DCS market growth in China – $1.4bn market

Source: BofA Merrill Lynch Global Research estimates, Gongkong

Chart 85: : PLC market growth in China – US$1.2bn

Source: BofA Merrill Lynch Global Research estimates, China Automation Research, Gongkong

Historical trends: 10% growth, 10-15% of automation market Chart 76 and 77 above show the size and growth rates of the Chinese DCS and PLC markets according to Gongkong. Despite a pull-back in 2012, DCS market continued to grow into 2015 at 7% while PLC market hasn’t fully recovered from it.

Historical trends: 10% growth, 10-15% of automation market =>original Chart 76 and 77 above show the size and growth rates of the Chinese DCS and PLC markets according to Gongkong. Despite a pull-back in 2012, they both show historical and forecast growth of 10%+, in excess of global automation markets.

Hollysys indicate that the DCS market in China is in total c.$23bn – although we suspect this is a total process automation market size including other process equipment (like controls valves, safety systems etc) and also power related DCS equipment. We think the pure DCS market is more like $2bn of the total $15bn global DCS market (10-15% of total market). Hollysys believe the total PLC market is around $1.5bn – around 10-15% of the global PLC market.

-2%

0%

2%

4%

6%

8%

10%

12%

14%

-

2,000

4,000

6,000

8,000

10,000

12,000

2009

2010

2011

2012

2013

2014

2015

E

2016

E

2017

E

2018

E

in R

MB m

n

-20%-15%-10%-5%0%5%10%15%20%25%30%

- 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000

2009 2010 2011 2012 2013 2014 2015E

in R

MB m

n


Chart 86: DCS – main end markets in 2014


Chart 87: PLC – project market (c.35% of total PLC market) by industry in 2014


Chart 88: PLC – OEM market (c.65% of total PLC market) by industry in 2014


Chart 89: DCS market share in 2014


Chart 90: PLC market share in 2014


The main end markets for both process and discrete markets are shown in Chart 78-80. End markets typically mirror those seen elsewhere – with DCS suppliers selling into Chemical, power, and petrochem markets, and PLC suppliers selling into a range of machinery markets. Overall, we believe the market is roughly split 25% control equipment (DCS, PLC’s, and HMI), 25% motors and drives, 10% general motion control, 15% instrument and sensors and 25% control valves/ actuators.

Critically requirements are different from US and European markets – with Chinese markets typically more focused on value, high product robustness, use-ability and cost driven technology vs Developed markets focused on plant availability, minimised lifecycle costs, features and flexibility. However, many of the structural growth drivers mentioned previously are likely to be powerful tailwinds – including moving up the quality curve, skills shortage, wage inflation, and energy consumption.

Chemistry

34%

Power 25%

Petrochem

14%

Municipal

7%

Cement 4%

Paper 4%

Metallurgy 3%

Others 9%

Metallurgy

18%

Automobile

14%

Municipal 10%

Transportation 9%

Chemistry

8%

Power 7%

Food & Beverage

6%

Infrastructure 6%

Petrochem

6%

O&G 4%

Building Mats 4%

Mining 3%

Paper 3%

Others 2%

Texile 22%

Package 11% Machine

tool 7% Elec. Eq

5%

Elevator 5% AC

4%

Rubber 3%

Pharma 3%

Printing 3%

Const Eq 3%

Plastic Eq 3%

Wind 3%

Cement 2%

Glass 1%

Others 23%

Supcon 17%

Hollysys 15.7%

Honeywell 13% Emerson

14%

ABB 8%

Yokogawa 6%

Siemens 7%

Invensys 7%

Zhishen 3%

GE 2%

Metso 1%

Xinhua 1% Winmation

1% Others

4%

Siemens 38%

MHI 13%

Rockwell 11%

Omron 10%

Schneider 9%

Delta 4%

GE 3%

Panasonic 1%

Koyo 1%

LS 1%

Others 9%


Chart 91: Chinese PMI has been underperforming recently


Chart 92: AirTAC monthly sales (RMB) – rolling sales rates have improved through the year


Near term: PLC over DCS markets In the near term, a recent weakening in Chinese PMI’s and weaker end market data points and suggest that short cycle/PLC markets should only see an improvement in demand once worries abate. We are more cautious on near term capex spending, given the level of investment over the last 3-4 years and current concerns over the economy. For example, the chart below shows machine tool orders in China relatively to Japan.

Chart 93: CNC machine tool orders China vs Japan => source file is a picture…

Source: METI, CEIC, Shenyang MT, BofA Merrill Lynch Global Research Estimates

Who is exposed? The largest players in China are the integrated global automation providers. In DCS markets, we believe ABB has the largest share, closely followed by Emerson, Siemens and Honeywell, with GE having a position in power markets through its acquisition of Xin Hua. In PLC markets, Siemens is dominant with a c.50% share, followed Mitsubishi (c.20% share) and Rockwell. In other areas we believe that Mitsubishi is particularly strong in laser beam machines (75% share).

In both these markets, and other automation areas, there are a number of Chinese and Asian based competitors, who have a large exposure to China;

Hollysys – the largest domestic automation player (60% of sales), with a co estimated mkt share of 12-13% (incl nuclear markets).

AirTAC – main operations in China, Fluid power equipment, with 85-90% of sales in China. No 3 market share in China behind SMC (38% share) and Festo (15%).

40424446485052545658

06 07 08 09 10 11 12 13 14 15 16

China PMI

-50%

-25%

0%

25%

50%

75%

100%

125%

09 10 11 12 13 14 15 16Rolling 3m average YoY% YoY %

0

100

200

300

1980 1985 1990 1995 2000 2005 2010

Japan China

(1,000 units)


Hiwin Technologies – motion control equipment, 50% sales to Asia.

Nanjing Technical equipment – motion control equipment.

Wuhan Huazhong Numerical Control – motion control equipment.

Siasun - robot supplier, with c 70% of sales in China and around a 10% share of the domestic market (ABB c.20%, Fanuc and Yaskawa both 15%).

China Automation Group – supplier into Safety Systems markets, where the company believe they have a 70% market share, versus Honeywell and HIMA.

Supcon – leading local DCS provider alongside Hollysys.

Threat of low cost competition? More limited Low cost competition has impacted a number of end markets – particularly power generation and transmission, rail and consumer goods. While competition will inevitably increase in automation markets, we think existing global suppliers are relatively well positioned versus other capital goods related end markets.

There are likely to be some risks around niche product areas and regionally focused manufacturing (for example AirTAC have taken their share from low single digit into the teens based on 40-50% lower costs) but in the large products segments, we think there are a number of differentiating barriers to entry which offer protection in China, but also to defend current market positions outside of China in developed markets .

Table 18: Key competition issues Nature of products Operating structure Nature of competitor/ customers - Technology differentiation - Cost structure - manufacturing, sourcing - Ownership - customers/competitors, govt vs private - Regulation/accreditation - end mkt requirements - Distribution - route to mkt, service, aftermarket - Fragmentation of customers - Local mkt product - multi brand, mid mkt - Regional tariffs - input & export - Positioning in domestic & foreign markets - Political sensitivity (e.g. defence) - incl financing req. - Ownership - JV's, financial & IP - Stability of peers - size, strategy, share - Niche - low demand home mkt/small % of cost - Management - track record/ flexibility - Growth in domestic market Source: BofA Merrill Lynch Global Research estimates

In automation specifically, we would note that;

• On the discrete side of automation markets, customers tend to be fragmented and not state owned. In process markets, customers may be state owned, but are typically governed by regulation.

• Chinese players aren’t huge and typically don’t have higher government ownership.

• Customer funding is typically not a feature of automation markets.

• There is Software content across automation systems.

• The large automation players have manufacturing capacity in China, so typically have a similar cost structure to local market peers.

There are threats at the low technology end of the product market, and as other Emerging markets expand, unless the Developed suppliers have cost competitive mid-market offering. IP transfer is also a risk, but we think the impact of low cost competition from China or elsewhere is likely to be in-line with the rest of the sector.


Chinese automation peers overview Table below shows the Chinese automation players (detailed information on each could be found in Chinese automation overview table). Our main takes are: 1) Local players generally don’t have strong presence, reflecting the strong positions of foreign players. 2) In certain markets such as DCS, drives and servos, however, Chinese competition is not negligible. 3) After a series of supported policies being issued, more and more Chinese companies entered in robotics market. However, the total market share of local players is expected to be less than 10% and they mainly compete with each other in low-end market.

Table 19: Chinese automation peers positioning map Control System Control Equipment Interconnect Motion Control

PLC DCS SCADA CNC HMI Adv. Software Motors Drives Field

Instruments Control Valves Actuators Fluid

Power Communication LV Products

Controller, Drives & Motors Robots

Hollysys Supcon Sciyon

Zhishen Chint/ Xinhua

Etrol Estun GSK

HZNCC Siasun Robot

STS STEP Efort

Inovance INVT

V&T Tech Megmeet Hiconics

China Automation Chuanyi Automation

Welltech Googol

Eura Drive Wecon HMI

Kinco Haiwell Yangli

Source: BofA Merrill Lynch Global Research Note: Light grey means some market presence. Dark grey means strong market presence.

Table 20: Chinese automation peers financial overview

Revenue in USD mn Gross Margin (%) Name 2011 2012 2013 2014 2015 5-y CAGR 2011 2012 2013 2014 2015 5-y avg.

Hollysys* 20 30 34 36 34 11% 35% 31% 37% 38% 36% 35% Sciyon 36 36 39 50 67 13% 41% 39% 40% 44% 43% 41%

Chint/ Xinhua* 16 38 41 38 24% 41% 49% 37% 36% 41% Etrol 31 40 54 69 86 23% 50% 51% 46% 43% 40% 46% Estun 74 63 73 82 76 1% 29% 31% 32% 32% 34% 32%

HZNCC 99 67 81 95 87 -2% 24% 23% 25% 28% 31% 26% Siasun Robot 120 164 214 246 266 17% 30% 27% 33% 36% 34% 32%

STEP* 0.1 13 56 NA 26% 43% 34% 34% Inovance 161 187 278 361 437 22% 55% 52% 52% 50% 48% 51%

INVT 106 116 150 170 171 10% 40% 41% 42% 42% 42% 41% V&T Tech* 16 17 20 21 15 -1% 42% 61% 51% Hiconics 11 13 13 13 15 7% 38% 37% 33% 36% 39% 37%

China Automation 306 350 376 311 261 -4% 38% 39% 36% 32% 30% 35% Welltech 20 20 21 20 15 -5% 43% 39% 39% 40% 35% 39%

Chuanyi Automation 474 512 515 550 499 1% 24% 24% 25% 25% 27% 25% Source: BofA Merrill Lynch Global Research, Company reports Notes: * Hollysys figures only take its industrial automation division. Chint figures only take its Control system division. STEP figures only take its Robot and motion control division. V&T Tech figures only take its Inverter and servo division. (no margin breakdown before 2014)


Table above is a financial overview of Chinese peers. We found that 1) Chinese DCS players have outperformed the market over the past 5 years, growing at avg. 18% vs. 5% of the overall market. This implies foreign players might lose market share to locals. 2) The weakest automation product among all is field instrument. Companies’ sales drop 3% on average over the past 5 years and margins trend downward. Table 21: Automation related policies

Date Issuance Entities Policy names Key Takes May-12 MIIT 12th Five-Year Plan for High-end

Equipment Manufacturing Automation and Intellectual Manufacturing Equipment industry are key areas of development. Automation rate reaches 30% by 2015 in key industries

Dec-13 MIIT Instruction on Robotics Industry Development

By 2020, cultivate 3-5 international players. Domestic products' market shares in high-end market reach 45%. Increase R&D in key components: servo, reduction gear etc.

May-15 State Council Made in China 2025 Numerical control and robotics are key areas of development. Integrating information tech and industry

Dec-15 State Council Internet Plus Promotion Instruction (2015-2018)

Strengthen the integration of internet and manufacturing. Obviously increase the localization of high end intellectual industrial equipment. Improve R&D capability in automation software and hardware (MES, PLM, DCS)

Dec-15 MIIT Instruction on Intellectual Manufacturing Standard Set up complete standards system for intellectual manufacturing

Apr-16 MIIT Robotics Industry 13th Five-Year Plan

Industrial Robot annual production 100k units, ow 6 axis > 50k. (not incl. foreign brands production in China) Industrial Robot density > 150 units per 10k workers. (currently c. 36 units) Service robot annual sales > 30bn RMB Key components: reduction gear, servo motor, drive, controller, achieve the same level as foreign peers, >50% mark. Share in >6 axis industrial Robot market. Cultivate 3+ leading players in industrial Robot market. industrial Robot installed base 800k units.(communicated in press)

Apr-16 MIIT 2016 Intellectual Manufacturing Demo. Projects

60+ Intellectual Manufacturing projects to be started in 2016.Targets: in 2-3 years, Opex down 20%, products development time down 20%, Productivity up 20%, deficiency rate down 20%. Energy efficiency up 20%

Source: BofA Merrill Lynch Global Research, MIIT, State Council of China


Table 22: Chinese automation peers business overview Name Ticker Products Business Brief

Hollysys HOLI US DCS, PLC, Software, Actuator Leading domestic player in DCS, c. 15% market share. Sales split: industrial Automation (DCS, PLC) 40%

Supcon Private DCS, Software, Actuator Leading domestic player in DCS, c. 17% market share. Strong R&D capability. (Owned by Zhejiang University)

Sciyon 002380 CH DCS, PLC, Software, Actuator Leading automation producer focusing on thermal power industry. Sales split: DCS (68%), Robotics (11%), Process software (11%)

Zhishen Private DCS, PLC Leading DCS producer, c.3% market share. State owned, belongs to Guodian Group

Chint/ Xinhua 601877 CH DCS, PLC, LV Products Acquired Xinhua control's automation business in 2012. Sales split: DCS, PLC, CNC (2%) Etrol 300370 CH DCS Leading RTU products and solution provider. Etrol also has presence in DCS market.

Estun 002747 CH CNC, Servo, Robotics CNC system integrator and industrial robot producer. Sales split: CNC (84%), Industrial robot (16%)

GSK Private CNC, Robotics, Servo #1 local CNC system manufacturer. Largest research and production base of CNC system in China. (Strong R&D capability, R&D spending c.8% of sales)

HZNCC 300161 CH CNC, Servo, Robotics Leading manufacturer of high/ medium -end CNC markets. Sales split: CNC System (39%), CNC Machine (28%), Motor (17%), Robot (6%)

Siasun Robot 300024 CH Robotics Leading industrial robot manufacturer, integrator. Owned by China Academy of Science and Technology. (Strong R&D) Sales Split: System (33%), Robot (31%), Logistics (25%)

STS Private CNC, Robotics Leading industrial robot OEM focusing on General, Pharma and Electronics sectors. State owned.

STEP 002527 CH Servo, Robotics Elevator control products producer. Entered robotics market in 2013. Sales split: Elevator control products (37%), Robotics (23%)

Efort Private Robotics Industrial robot OEM and System integrator. Specialized in Metal processing and Automotive (Acquired CMA robotics in 2014).

Inovance 300124 CH Servo, PLC, LV products, HMI Largest local low-voltage inverter producer in China. Strong R&D Background. Sales split: LV inverter (51%), Servo (13%), PLC (3%)

INVT 002334 CH LV products, Servo, PLC, HMI Leading LV inverter producer. (Market share c. 4.5% in China LV inverter market) Sales Split: Inverter (55%), Servo (8%)

V&T Tech 300484 CH LV products, Servo Leading EV motor controller producer. Strong R&D background. Sales split: L/MV Inverter (27.4%), Servo (3%)

Megmeet Private LV products, Servo, PLC, HMI Strong R&D, founded by senior employees from Huawei and Emerson. Specialized in Medical equipment, Telecom, Electronics and Transportation sectors.

Hiconics 300048 CH Servo Specialized in HV transformer. Acquired a servo drive player in 2014. Servo business now accounts for c. 12% of total sales

China Automation 569 HK Control Valves Leading player in petrochemical and O&G process control system integration and products. Sales split: system (35%), Valves (38%)

Googol Private CNC, Motion Controller, HMI Specialized in Motion controllers and systems.

Eura Drives Private PLC, HMI, Drives, Servo Specialist in Mid-LV AC drives. Customers mainly in Pharma, Paper, Textile, Petrochemical and O&G sectors.

Wecon HMI Private HMI, PLC Leading HMI brand. Also has presence in PLC market.

Kinco Private HMI, PLC, Servo Leading HMI brand. Products also incl. PLC and Servo. Key end markets: Textile Machinery, Packaging, Pharma, Electronic Manufacturing and High-end Healthcare

Haiwell Private PLC, HMI, Servo Leading HMI brand. Also manufacture PLC, CNC control system and servo motors. Key end markets: packaging, textile, food, medical and pharma sectors.

Welltech 002058 CH Sensors, Actuators, Valves Leading Field Instruments and Actuators manufacturer and system integrator. Key products: Electromagnetic Flowmeters.

Chuanyi Automation 603100 CH Actuators, Control Valves, Field

instruments Leading actuators, transmitters and valves manufacturer. (82% of Sales)

Yangli Private Actuators, Valves State owned electrical actuators manufacturer. Source: BofA Merrill Lynch Global Research, Company reports


Appendix A: The automation market Under our relatively broad market definition, we estimate the industrial automation market is worth around $140bn pa. The market often focuses on the controller portion of the market (PLC vs DCS), but we estimate these products only cover 20% of the market, and we look at the 13 product segments including estimates for market size, key competitors and shares in each area.

Typically there are 3 types of automation competitor – multi industry companies with typically a large automation presence across a number of products areas (for example ABB, Siemens, Emerson, Honeywell, GE, Mitsubishi), automation pure plays across a number of product areas (Rockwell, Yokogawa) and niche product specialists (Kuka, Belden, AirTAC).

Defining the scope of the market Automation is one of the major end markets for the global electrical/multi industry and niche players in the sector. The companies involved typically give a wide range of market sizes across process and discrete markets, depending on the particular areas they compete in with little consistency, making comparisons difficult. A lot of end market studies also typically tend to focus on the pure control side of the market (PLC and DCS), but for many of the stocks, automation products include a very wide range of equipment, often in areas that do not compete directly.

To gain a better understanding of the market and to allow better comparison between companies, we have broken the automation market down into different product groupings. We think these groupings should give a better understanding of competitive position, market share and difference in business model between the major vendors.

We have arrived at these product groupings based on components with similar technology/characteristics, within similar areas in the automation system, based on competition and how the range of automation companies typically divide the market. These product groupings are not supposed to be definitive, and there will be some overlaps (e.g. in motors – large motors in the Motors, drives and small generators grouping, and high precision servo motors in Motion Control for machine tools), but we think this splits the automation market into sensible units.


Exhibit 36: BofAML – definition of industrial automation market


Product areas based on similar product characteristics, competitors and company definitions

Not included within our definition of the automation market are machine tools (typically different companies), pumps (broad market, not necessarily connected with factory automation, typically driven by electric motors or fluid actuation which is included), metering (increasingly used, but wider market, different range of competitors), material handling (large part of the market not involved with factory automation, different set of competitors) or UPS providers (different technology focus and competitors). The industrial software market is difficult to define due to the given the number of players and overlap with pure software companies (IBM, SAP) and we have not included software above Advanced applications (i.e. not included PLM, ERP, MES software).

A more detailed description of what is included in each product grouping, with examples is included in Section 7 at the back of the note.


motors, actuators












Data centres





Machine tools




Industrial control

equipment





We see the size of the market at cUSD140bn pa The total industrial automation market is large – and under our definition is estimated to be worth $140bn pa. We have tried to define the size of each product grouping, based purely on exposure to industrial automation (some of the end markets compete in other areas – for example, Fluid Power) based on our research in to the end markets and company commentary. Where possible, we have also tried to give estimated end market split, geographical focus and process/discrete focus for each product grouping. While PLC’s and DCS’s are typically a main focus for investors, they represent only 20% of the market, with other major product areas being motors/drives and field instruments.

Table 22: Main automation product segments - estimated market size, product, geographical and manufacturing exposure

Industrial Control System Industrial Control Equipment Ind. Interconnect Motion Control

Product PLC DCS Safety

Systems Ind PC's/ HMI Adv software

Motors & drives

Field Instru-ments

Control valves

Fluid power Comm-

unication LV products

(ind)

Controller, drives & motors

Robots

Est. global market size

$12bn ($8bn hybrid)

$15bn ($8bn hybrid)

$5bn $2-3bn $10bn $35bn $15-20bn $5bn (excl power gen)

$5-10bn $4bn $10-15bn (total $70bn)

$8bn $8.5bn pure robots

Detail

85% hardware (incl bundled

software)

Est. 50% services, 30%

hardware, 20% software

Split - integrated & stand-alone providers, process,

hybrid and discrete

specialists

HMI software can incl with

PLC's

No of product areas -

simulation, asset mgmt,

mobile, analysis/

optimisation

$15bn motors, $10bn drives, $5bn mech

transmission, $5bn

generators

Sensors $10-15bn,

machine vision $1bn,

(Global sensor $65bn)

70% valves, 30% other (actuators

etc). Process actuator mkt

incl power $1-1.5bn

30% valves, 30%

actuators, 15% fittings

Industrial Ethernet $800mn,

cellular M2M $500mn,

wireless $1bn

LV breakers & switches,

connectors, enclosures, DIN rails, LV

systems, accessories, dist boxes

70% servo systems (low

and high end). Stepper used in mid range.

20% weld, 35% mats

handling, 8% painting, 10% other pro, 15%

arc welding

End markets

5% Chems, 15% Food &

bev, 5% water, 10% auto, 5%

Semis', 15% machinery

manuf

Ex power gen - 30% chums, 15% refining,

20%O+G, 15% pulp and

paper, 5% pharma

10% auto, 10% semis,

15% machinery, 10% food &

bev, 5% pharma, 5%

water

Based on ISYS - 40%

gen ind, 25% oil/ gas, 10%

discrete.

$25bn industry, $5bn utilities, <$5bn infrastructure,

<$5bn transport

5% pharma, 40% chems, OGP, 15%

food and bev, 5%. 30% base

sensors

Ex power gen - 25% chems,

30% O+G, 15% refining (

50% OEM, 50% MRO. Total fluid power mkt >$120bn

35-45% end users, 25-35%

OEM'a, 25-35% MRO

40% end users, 30% OEM's, 30%

MRO

15% food/ bev, 15%

machine tools, 15% semi-con

Geographical split

20% N Am, 40% EMEA,

40% Asia

15% N AM, 40% Euro,

40% Asia, 5% RoW

10% N Am, 50% EMEA,

35% Asia, 5% RoW

30% N Am, 24% Euro,

40% Asia, 5% RoW

25% N Am, 35% Euro,

35% Asia, 5% RoW

31% Americas, 38% Euro, 31% Asia

10% Americas, 45% Euro, 45% Asia,

NAM 14%, Euro 26%,

Asia 53%, 7% RoW

Process vs discrete

Discrete Process Both Higher discrete

Higher process

Both Higher process

Process focus Both Both Both Discrete focus Discrete


Process vs discrete manufacturing

A customer’s manufacturing model, process (continuous flow of product) vs discrete (stand-alone product) drives controller selection

One the most important distinctions in automation markets is driven by the nature of a customer’s manufacturing model – whether a customer is producing discrete product (like a car, or stand-alone machined product), or whether they are producing a product which is a continuous flow (e.g. oil in a refinery plant, beer in a brewery). The major differences between discrete and process manufacturing models/products are shown in the table below

…drives controller selection – DCS vs PLC’s… The customer’s manufacturing model dictates the type of automation control system and related product requirements – with discrete manufacturing typically using PLC’s (Programmable Logic Controllers - rugged industrial programmable control units) and process automation typically using DCSs (Distributed Control Systems – a plant wide custom control system). On top of the type of controller typically used, the type of manufacturing model used by the customer can also impact other product requirements – for example, control valves are more likely to be used to control flow in process


automation applications, and precise servo motors are more likely to be used in motion control application in discrete automation. A more detailed description of DCS’s and PLC’s is included at the back of the note.

Table 23: Discrete vs process automation

Discrete Automation Process Automation

Description Manufacture & assembled parts, components to finished product

Continued process converting raw materials to finished products

Typical industries Auto, general mfg, consumer electronics Oil & gas, chems, pulp & paper, mining,

metals

Supplier offering Programmable logic controllers (PLC), drives, motors, sensors, HMI, software/eng

Process controllers (DCS), measurement, actuators, (motors, drives, valves), sensors, electrification, software/eng

Ind penetration Low to moderate - more diverse markets & applications, customers with less expertise and investment capacity

High, long history, maximise rtn on large capex

Channels A mix of direct, distributors, OEM's and systems integrators

Mainly direct sales or through EPC's and system integrators

Competitive landscape

Global and local/regional Dominated by global players

Source: ABB, BofA Merrill Lynch Global Research Estimates

Exhibit 37: Process automation is typically on long lead time projects (1)

Source: BofA Merrill Lynch Global Research Estimates, Flowserve

Table 24: Process automation is typically on long lead time projects (2) Project type Example Project time - inception to start up Large Refinery 5 - 7 years Medium CCGT 3 - 5 years Small Mining brownfield 1 - 2 years Aftermarket Spares/parts 24 hrs to 3mths Aftermarket Alliances multi year Source: BofA Merrill Lynch Global Research Estimates, Flowserve

… and defines end markets & business model for automation players Beyond different end market growth rates, the difference in manufacturing model in discrete and process markets also impacts the business model for the automation suppliers;

Discrete vs process customers also defines end markets, nature of growth profile and cyclicality

1 – DCS markets (and related products) tend to be longer cycle – with larger project sizes and lead times. In general for the automation suppliers into process markets, the business tends to be more driven by order books, with margins more impacted by project execution, level of bought in equipment and competitive tendering rather than pure volumes, leverage and efficiency within their own plants. 2 - PLC markets tend to be shorter cycle, more economical sensitive and less dependent on large projects, with margins driven more by operating leverage rather than contract execution.

Hybrid – relevant for a number of customers end markets

Hybrid manufacturers use both PLC and DCS

With customers moving towards more flexible manufacturing, automation customers typically use a combination of PLC and DCS control. For example a large process plant


may use a DCS to control its main pharma manufacturing process, but PLC based control for around the rest of the plant. Below we show markets which typically use hybrid automation systems, and products typically used in each end market for control (and safety monitoring) of the manufacturing process.

Exhibit 38: Process vs Discrete manufacturing – by product type, end markets and complexity

Source: Invensys , BofA Merrill Lynch Global Research Estimates

Who is exposed to what? Total exposure to automation markets as a % of group sales is shown below, although these numbers are based on a number of different product areas.

Chart 94: Automation related sales as a % of group sales for the major global players

Source: BofA Merrill Lynch Global Research Estimates Note: Numbers as of FY2015/FY2016

To allow for a better comparison between the different automation players in the sector, we have broken down each company’s sales exposure to each product grouping;

1 - The tables split between the large global players who typically compete in a number of product segments, and the niche players in each region who have lower market shares or exposed to a lower number of product areas. We have tried to include both listed and the larger private companies in each region.

0%

20%

40%

60%

80%

100%

Rock

well

SMC

Yoko

gawa

Fanu

c

Emer

son

ABB

Mitsu

bishi

Elec

tric

Parke

r H

Siem

ens

Schn

eider

Hone

ywell

Eaton GE


2 - We have differentiated between those that have a large market share in a product area or where it’s a large part of the group (dark grey), those that have some presence (light grey) and those with no/limited exposure (clear box).

3 - Where possible we have also included for the major players an estimated sales exposure (as a % of total automation sales) in each product category. Typically most of the big players have a bias towards process or discrete markets, which can be seen through the exposure to DCS or PLC controllers.

At the back of the note we have include a more detailed breakdown for each company – including divisions exposed to automation markets, geographical focus, size and profitability.

Larger global automation players Below we show the main global automation players;

• The industry is dominated by developed markets, largely multi industry electrical companies, with ABB, Siemens, Emerson, Rockwell, Mitsubishi and Yokogawa the main players.

• In terms product bias - ABB, Invensys, Emerson, Honeywell and Yokogawa are all biased towards process/DCS markets – while Schneider, Mitsubishi and GE are more biased to PLC/discrete markets. Siemens and Rockwell are more balanced between PLC and Discrete markets.

Exhibit 39: Main automation players – product exposures

Source: BofA Merrill Lynch Global Research Estimates Note: Invensys and Schneider are now a combined entity, however we show them separately for more granularity

• There tends to be grouping based on product background – for example Invensys is focused on control (DCS/Safety/software), Eaton electrical style products (PLC, HMI, drives, field, communication), Emerson largely a product business (sensors, field equipment, valves), Rockwell, Hitachi, Toshiba and Mitsubishi more focused on electronics (PLC and drives etc).

• We have included some product specialists within the main automation players – Fanuc (robots and motion control) and SMC (fluid power), both based in Japan, due to their market shares in their respective markets.

• Many of the companies include related products in their automation divisions which we have tried to strip out in our sales exposures. For example, ABB include generators and turbo-charging, Invensys include nuclear safety systems, Siemens include Metals Technology and large motors for other markets (wind, rail).

Company PLC DCS Safety Systems/SCADA

Industrial PC's/HMI

Adv software

Motors Drives Field Instru-ments

Control valves

Ind specific/

other

Fluid power

Comm-unication

LV products


Robots

Main playersEuropeABB 1% 15% 3% 5% 22% 16% 8% 1% 3% 1% 16% 10%Invensys 33% 24% 22% 12% 9%Schneider 22% 3% 7% 7% 4% 13% 12% 6% 22% 4%Siemens 12% 9% 2% 1% 6% 19% 10% 5% 25% 3% 4% 4%USEaton 3% 3% 25% 4% 35%Emerson 15% 5% 35% 25% 10% 10%GE 10% 5% 60% 5% 15% 5%Honeywell 15% 60% 25% 8% 7% 15%Rockwell 20% 10% 5% 15% 20% 5% 15% 10%JapanFanucHitachiMitsubishiSMCToshibaYokogawa

65%

Control System Control equipment Interconnect Motion Control


Niche automation players - Europe Below, we show the European niche automation players – covering the listed companies in the sector, and some of the privately held companies. The European market can be split into 3 areas; 1) the listed companies who are typically focused on a specific niche within each product area (IMI – Norgren fluid power, Metso – largely pulp and paper DCS’s), 2) the private companies (indicated by * next to name) with a control electronics bias (Beckhoff, B&R, Lenze) or sensors and communications focus (Endress & Hauser, SICK and Pepperl & Fuchs), and 3) companies where automation markets are one of a number of markets they sell into (Legrand), or automation is one of a number of drivers (Spirax and Rotork).

Exhibit 40: Niche automation players – Europe - (dark grey - significant market share/exposure, light grey – some exposure)


North America The smaller players in the North American market are typically either; 1) focused on niches (Aspentech – software, Parker – fluid power), 2) biased towards communication (Belden, Moxa, Sierra), or 3) test and measurement focused (Ametek, Danaher, National instruments, Teledyne).

Exhibit 41: Niche automation players – US - (dark grey - significant market share/exposure, light grey – some exposure)


Japan The niche Japanese players have more of electronics focus – with IDEC and Omron supplying PLC’s, HMI, Field instruments and communication equipment, Keyence focusing on sensors and Yaskawa focused on drives, motion control and robots.


Industrial PC's/HMI

Adv software


Control valves

Fluid power

Comm-unication

LV products


Robots

Niche EuropeBeckhoff *Bosch Rexroth *B&R Industrial *Endress & Hauser *Festo *HexagonIMIKukaLegrandLenze *MetsoPepperl & Fuchs *Phoenix contact *RotorkSICK AG *SpectrisSpirax



Industrial PC's/HMI

Adv software


Control valves

Fluid power

Comm-unication

LV products


Robots

USAmetekAspentechBeldenDanaherFlowserveNational InstrumentsMoxa *Parker HannifinPentairRoper industriesSierra Wireless



Exhibit 42: Niche automation players – Japan - (dark grey - significant market share/exposure, light grey – some exposure)


RoW We have also included some of the more significant automation ROW players in the exhibit below– which are spread across a number of different segments. Largely these companies are currently exposed to domestic or adjacent markets.

Exhibit 43: Niche automation players – RoW - (dark grey - significant market share/exposure, light grey – some exposure)


Market shares by product grouping The previous tables showed product overlap by company, below we look at market share in each product grouping in more detail. A number of the more generic product segments (Fluid power, motors, sensors, control valves) are sold into many end markets – but where possible we have tried to focus on shares in the pure industrial automation area.

PLC – Programmable Logic Controllers

Major PLC players – Siemens, Rockwell and Mitsubishi driven by very strong “home” market shares

Table 17 below shows the major global PLC suppliers, which is dominated by Siemens (from its strong European base), Rockwell (from its US base) and Mitsubishi (50%+ share in Japan). Other regional players not included in the table below include; ABB (in Europe <5% share), Eaton (<5% share in Europe), Sharp (<5% share in Asia), Hitachi (<5% share in Japan), Yokogawa (<5% share in Japan). Second tier suppliers include Beckhoff, Bosch, Delta Group, Hollysys, IDEC, JTEKT and Toshiba.

Table 25: PLC - estimated market shares Mkt share Regional position Siemens Mid 20%'s US (#2), Euro (#1, 50%+ share), Asia (#2) Rockwell Around 20% US (#1, 50%+ share), Euro (#2/3), Asia (#3), Latam (#1, 40%+ share) Mitsubishi Low teens Asia (#3, Japan 50%+ share) Schneider Mid/high single digit Euro (#2/3) Omron Mid single digit Japan (#2) B&R industrial Mid single digit GE Low/mid single digits US (#3) Source: BofA Merrill Lynch Global Research estimates


Industrial PC's/HMI

Adv software


Control valves

Fluid power

Comm-unication

LV products


Robots

JapanAzbil Group IDECKeyenceOmronYaskawa



Industrial PC's/HMI

Adv software


Control valves

Fluid power

Comm-unication

LV products


Robots

RoWAirtac (Taiwan)China Autom. Grp (China)Hiwin (Taiwan)Hollysys (China)Siasun (China)WEG (Brazil)



We believe there are also slightly different shares depending on the size of the controller – with Rockwell, Mitsubishi strong in larger controllers, and Siemens, Omron and Schneider stronger at the smaller end.

DCS - Distributed Control Systems

Major DCS players - ABB, Honeywell, Siemens – although depends on end market focus

The tables below quantifies the broad market shares of the major DCS vendors, based on company commentary over time. Despite the long cycle nature of the DCS business, shares can move around between years, but the table should give a rough idea of relative size. The table is based on process automation markets – market shares can look different if power related DCS shares are included (e.g. we estimate 40% of Siemens DCS business is related to power markets, Alstom has no presence in process markets but does sell into power DCS markets etc). Due to the importance of end market focus, we also show the key competitors in each end market segment.

Table 26: DCS - estimated market shares and regional positioning Market share Regional position

ABB High teens US (#3), Eur (#1/2), China (#1/2), Asia (#2), LATAM (#1)

Honeywell High teens US (#2), Eur (#3), China (good), Asia (#3), LATAm (#3) Yokogawa Mid teens Jap (#1), Asia (#1) Emerson Low teens US (#1), Eur (#4), China (good) Siemens Low teens Eur (#1/2), China (#1/2), Asia (#4), LATAM (#2) Invensys High single digit US (#4) Rockwell Mid single digit Azbil Low single digit Jap (#2) Source: BofA Merrill Lynch Global Research estimates

Table 27: Main DCS competitors per end market vertical Refining Pulp and Paper Oil and Gas Chemical Petrochem Pharma

Honeywell ABB ABB Siemens Yokogawa Emerson Invensys Honeywell Honeywell Honeywell Honeywell Honeywell

Yokogawa Valmet Invensys ABB Emerson Siemens Emerson Yokogawa Yokogawa Emerson Invensys ABB

ABB Siemens Emerson Yokogawa ABB Invensys Source: BofA Merrill Lynch Global Research estimates, ABB

Outside the major players, competitors with smaller market shares include GE, Valmet (mainly pulp and paper and power) and SMAR (Brazil)

Safety Systems – process, hybrid and discrete markets Within the Safety Systems market, we believe there are differences in shares between the discrete market vendors, stand-alone suppliers for process markets, and integrated process safety providers, where safety systems are incorporated as part of a wider process offering (typically the larger DCS providers).

Overall Invensys is the market leader in this segment with a mid-teens market share, followed by Rockwell, Honeywell with low/mid-teens shares, and Yokogawa, ABB, HIMA and Emerson with <10% shares (ranking also shown in table below). Table 28: Safety systems - main players and exposure Process Integrated providers Siemens, Yokogawa, Emerson, ABB,

Honeywell Stand alone Pilz, GE, Rockwell, HIMA, Invensys Hybrid GE, Schneider, Siemens, Rockwell Discrete Schneider, HIMA, Rockwell Siemens, Source: BofA Merrill Lynch Global Research estimates

Table 29: Industrial PC's and HIM hardware - main market players

Large Mid - sized Regional/other Siemens, Advantech (Taiwanese), B&R (German), Kontron

(German), EVOC (China)

NEC (Japan), Schneider, Meidensha, Contac

GE, Rockwell, National instruments, Eaton,

Spectris

Source: BofA Merrill Lynch Global Research estimates

Industrial PC’s & HMI hardware (human machine interface) We believe that the Industrial PC’s/HMI market is relatively broad but fragmented market. There are a large number of competitors with different backgrounds – from large multi-product automation players to small niche product specialists and large electronic hardware providers with a PC background. This segment includes products from rugged-ised PC’s running applications on standard windows operating systems to specific software running on embedded processors.


The tables below show the general market competitors by size, with most companies also competing outside of factory automation markets (power, water, building automation etc). Market shares can differ between end markets and geographies depending on specifications/installed base.

Advanced software – range of applications Given the range of providers, markets, overlapping market definitions and relative immaturity of the market, there is a range of estimates for the size and scope of advanced applications/software automation market. Below, we show Honeywell’s view of the main players and relative rankings within the advanced applications area – with the large automation process vendors and Aspentech the main players. We also show Invensys’ definition of the market - competitors include a number of industrial software specialists (Aspentech, Aveva) and pure software related businesses (IBM, SAP). Table 30: Honeywell (HPS) - estimated market positions across a number of its markets - including Advanced applications DCS Safety ctrl Adv apps Services ABB 1 5 2 1 AspenTech 4 E&H Emerson 4 7 7 4 HPS 2 2 1 3 Invensys 6 1 3 6 Rockwell 7 3 5 5 Siemens 3 6 6 2 Yokogawa 5 4 8 7 Source: BofA Merrill Lynch Global Research estimates, Honeywell

Table 31: Invensys - market definition of its exposure to Advanced Software markets Line of business Market size Competitors Design, simulation & optimisation

$1.5bn AspenTech, Aveva, Honeywell, ISYS

Trading, planning & scheduling

$300mn AspenTech, Honeywell

Manufacturing Ops Mgmt $2.1bn Apriso, Rockwell, SAP, Siemens, Werum

Asset mgmt $1.8bn ABB, IBM, Infor, SAP, Schneider

HMI & supervisory $1.2bn Copadata, GE, Inductive, Indusoft, OSI, Rockwell,,

Siemens Source: BofA Merrill Lynch Global Research estimates, Invensys

Chart 95: Market shares within Aspentech’s defined Engineering software market

Source: BofA Merrill Lynch Global Research estimates, Aspentech

Chart 96: Market share within Aspentech’s defined Manufacturing software market


Chart 97: Market share within Aspentech’s defined supply chain mgmt software market


The charts above show the companies that Aspentech considers competitors in its definition of the industrial software markets. For reference, we also show includes Aspentech’s definition of supplier chain management software – this maybe an area the automation players look to expand, although we don’t include it within the scope of this note.

Given the data above, we believe the major players in advanced software are Honeywell, ABB, Aspentech, Rockwell, Siemens, Schneider and Emerson.

Industrial control equipment

Fluid power Similar to other end markets, Fluid Power providers supply into a number of end markets outside of pure factory automation (incl aerospace, auto/truck, med-tech, rail, construction etc). In the factory automation space we believe that the 2 largest players are SMC with a market share of 25-30% and Festo (German, private) with a 15-20%

2% 3% 3% 3%

8% 39%

0% 20% 40%

HoneywellEnergy SolsSPT Group

GL NobelInvensys

AspenTech

5% 5% 7% 9% 9%

17%

0% 20% 40%

OSI SoftEmersonInvensys

ABBAsptenTech

Honeywell

3% 3%

6% 9%

26% 28%

0% 20% 40%

LogilityOM Partners

JAOracle

AspenTechSAP


market share – both sell a broad range of products across a number of regions. Elsewhere in the Factory automation related fluid power space, we believe that Parker Hannifin (no 3), Bosch Rexroth, AirTAC and IMI Norgren have significant shares.

Control valves, related actuators and positioners The major suppliers across the global control valve (and actuator market) including Emerson, GE, Flowserve, Nihon Koso (Japan – private), Samson (German – private), Metso, CCI (IMI), Tyco and Azbil. We believe power is a substantial market for Control valves, so the table below lists the main players in each of the process automation markets relevant to this note (for example we believe CCI’s main focus is on power markets). In most end markets and geographies Emerson is the market leader with a 20-30% market share (we believe the other main players typically have a 5-10% market share).

Table 32: Main players in the relevant control valve markets

End market Main players Chemicals Samson, Emerson, Flowserve Oil and gas Emerson, Flowserve, Metso Refining Emerson, GE, Nihon Source: BofA Merrill Lynch Global Research Estimates

In the standard process spring diaphragm actuator market (used to typically drive process control valves) Emerson, GE and Flowserve are the larger players. Other players with smaller shares in this product area include ABB, Auma (actuators), Invensys (positioners), Rotork (actuators), SPX and Yokogawa.

Motors, drives, generators and mechanical power transmission This category includes a number of large markets and given the nature of the products, the products are also sold into end markets outside of industrial automation. The table above shows the major players in the motor, drive and generator markets (according to ABB), but below we break down the major players in each area in more detail;

1. Drives – we believe there are distinct differences between low and high voltage drives. In Low voltage areas - ABB is the market leader (est. 15% share), followed by Siemens (est. 10%) and Danfoss in markets that are typically quite fragmented. In medium/high voltage markets the market leaders are Siemens and ABB (both with 20%+ share) followed by Converteam and Rockwell.

2. Motors – similar to the other product segments electric motors go into a number of end markets – not just factory automation – notable HVAC, power generation and auto end markets.

In the industrial segment in the US, we believe that ABB are the largest player following the acquisition of Baldor (mid 20% share), followed by Regal Beloit (low teens), Nidec (c.10%, following the acquisition of Emerson’s motors business), GE (c.10%), Siemens (high single digit) followed by a tail of regional competitors.

In Europe, Siemens are the dominant supplier with a c.20% share, followed by ABB (low teens), Nidec (low teens), then a range of companies with mid-single digit shares (ATB, VEM, TT electric, Danfoss).

Table 33: ABB - major players in motors, drives and generator markets 1 2 3 LV motors ABB Siemens WEG Large AC induction motors Siemens ABB WEG LV drives ABB Siemens MV drives Siemens ABB Converteam Generators Emerson ABB Cummins Source: ABB, BofA Merrill Lynch Global Research Estimates


Outside the US and Europe, the major players are TMEIC (50% Toshiba, 50% Mitsubishi), WEG, TECO and Yokogawa.

Like the Drives market, motors can be broken down into a number of segments – in terms of size (small or fractional, medium and large motors), type (AC and DC) and efficiency rating. We believe is market leader in low voltage, while Siemens is dominant in mid/high voltage markets.

3. Mechanical transmission – a very fragmented market, with competition varying by product type (gearboxes, couplings, belts etc), region and often including companies outside of pure automation markets.


Field instruments – sensors, machine vision, test & measurement Within the field instruments product grouping, the main competitors within each product segment are;

• Sensors – in general this is a large global market ($60bn+), selling into a number of different end markets. Within industrial markets, ABB, Emerson, Honeywell, Siemens, Endress & Hauser, Invensys and Yokogawa are the major global players although we believe the market is relatively fragmented, with market shares and smaller competitors varying between the main flow, temperature, level and pressure product area (30% of the market) and also regionally. While the analysis perhaps over simplifies the markets given the wide range of technologies and other sensor types, Table 14 shows the main competitors in the main sensors categories.

Table 34: Sensors - main global player, and product specialists

Flow (mkt $6bn) Temp (mkt $5bn) Level (mkt $5bn) Pressure (mkt $6bn) Main global competitors - Emerson(#1 - 27% share), Endress & Hauser (#2), Yokogawa (#3) Honeywell (#4), ABB (#5),

Siemens (#6), ISYS (#7) Azbil, GE Epicos (TDK), FLIR, GE,

Iroon (Danaher) Hyde Park (Schneider), STI

(Omron) Druck (GE)


• Machine vision – there is overlap between machine vision and RFID markets, which is not included in our definition of industrial automation markets (RFID – tracking systems that use electromagnetic fields to transfer data). We believe the market leader in machine visions is Cognex (c.20% market share), followed by Omron (10%), with other players including Keyence, National Instruments, Spectris and Matrox.

• Test & measurement, and analytical equipment – this is a very large segment with a range of applications and technologies. We believe the market leaders are Danaher and Agilent, with competition from some of the larger automation players in some markets (ABB, Yokogawa, Siemens in gas analysis) and more specialist companies in other areas (Teradyne, Advantest, LTX Credence - focused on the semis market).

Communication – network equipment, industrial Ethernet, M2M We have broken this product grouping into 3 products areas – general network equipment, industrial Ethernet switches and M2M products. We think there is inevitably some overlap with low voltage markets, but below lists the major players and estimated share in each section where available;

• General network equipment – includes a wide range of companies from the broad global automation players (Siemens, ABB, Emerson, Honeywell), to the small players with a variety of backgrounds (Phoenix Contact, Pepperl & Fuchs, Controltech, Endress & Hauser, Relcom, National instruments).

• Industrial Ethernet – key players in this market are Belden, Spectris (Ntron and Sixnet), Ruggedcom (Siemens), Moxa and Hirsch. The chart below shows the main providers in industrial communication markets under Belden’s definition of the market (we think this includes industrial Ethernet, but may also include some cabling products) – in which the company believe they have 18% market share. Westermo and Korenix (Beijer) and Garretcom are also included as competitors.


Exhibit 44: Industrial data communication market – Belden definition

Source: Belden

M2M modules - In M2M markets, we believe the largest provider of cellular embedded modules is Sierra Wireless with a market share of c.30%, followed by Gemalto, Spectris and Telit – although we believe most products are relatively end market/application specific.

Low voltage products Low voltage products are typically used in a number of end markets (resi, general construction, utility markets) although we believe industrial markets are around 20-30% of the market. Market shares are very regional given the nature of the products (lower ticket price, subject to regional preferences). The large low voltage competitors by product are shown in Table 15, although in industrial markets we believe Schneider, Siemens and ABB are the market leaders.

Table 35: Low voltage markets - major competitors in each product area defined by ABB Wiring accessories LV systems Enclosures & DIN rail prod. Control products LV breakers and switches

Legrand Eaton Eaton Phoenix contact Eaton Leviton GE Hager Rockwell Mitsubishi

Schneider Schneider Schneider Schneider Schneider Siemens Siemens Siemens Siemens

Source: BofA Merrill Lynch Global Research Estimates, ABB

Motion control

Motors, drives & controllers The main players for the precise motion control market are shown below. There is some overlap with smaller motors but based on industry commentary;

- Top 3 - we believe the top 3 players are Yaskawa (est. 15% share), Siemens (est. 13% share) and Mitsubishi (est. 11% share).

- Mid tier – this is followed by a number of companies with mid to high single digit shares – including Bosch, Danaher, Rockwell and Schneider.

- Regional tail – the market then consists of players with regional positioning Emerson in the US, Lenze and B&R in Europe, and Fanuc and Omron in Asia.


Robotics We believe the overall market leader in industrial robot markets is Fanuc, with close to a 20% market share. The market is relatively consolidated, with ABB, Kuka and Yaskawa all having mid-teens shares, resulting in the top 4 players making up 2/3 of the market.

Other players tend to be more regional or end market in focus, and include Siasun (China), Kawasaki, Duerr (paint and sealing robots for auto markets), Reiss, Comau and Staubli.

How automation vendors built their software businesses We think that there are five different reasons why capital goods firms built on their business model from just providing the physical devices (or physical devices with embedded software given away for free) to selling an integrated package consisting of hardware and software:

• Reaction to software companies’ push into manufacturers’ area of domain expertise. We highlight this point later in our analysis, discussing the software companies’ angle in the convergence of automation hardware and software.

• Generate a more recurring stream of revenue from software licenses compared to more lump sum payments through hardware sales. For example, Aveva has in the past moved its business model from charging customers a lump sum upfront for its software package towards a licence-based solution, in which customers pay for the “right to use” the software. This latter model provides a strong recurring base of revenue for Aveva and highlights how also the business model of automation vendors is going to change in the future.

Chart 98: The bulk of Aveva’s revenue comes on a recurring basis

Source: Aveva filing

• Remedy pricing pressure that comes along with commoditization of equipment through addition of software which is harder to replicate by competitors and even allows for price premiums

• Increase switching costs to raise the barriers for new entrants. Vendors of automation hardware might increase their customers’ entrenchment with their installed base by providing additional software capabilities. We believe customers are less likely to switch since the added software is particularly tailored to the respective hardware and since the adoption of a holistic hardware/software package might result in significant switching cost relating to training and installation of external solutions.

• Find additional sources of revenue by cross-selling software to hardware customers (Hexagon was able to sell virtual design capabilities from the acquisition

0

50

100

150

200

250

FY2009 FY2010 FY2011 FY2012 FY2013 FY2014 FY2015 FY2016

Recurring Revenue Initial Licence Fees Training and Service


of Intergraph to its existing client base of measurement instruments) or vice versa (following its acquisition of Telvent, Schneider was able to sell its low-voltage equipment in order to complement Telvent’s energy management capabilities.

Exhibit 45: Schneider managed to cross-sell its power equipment to Telvent’s installed base

Source: Schneider

Siemens moved first, ABB and Schneider followed

Siemens acquired UGS and its virtual design and manufacturing software in 2007. ABB (2010) and Schneider (2011) followed, buying asset management software capabilities. The latter stepped up its efforts recently to complete its digital manufacturing offering through the takeover of Invensys.

Automation vendors have sought different ways to enter the software space:

• Siemens is the front-runner among automation vendors to enter the software space through its acquisition of UGS in 2007. The firm continued to build up a full digital manufacturing offering through the takeover of various small software players, which it integrated into its industrial business.

• Schneider is Siemens’ closest follower and has been discussing the convergence of IT and OT for some time. With the acquisition of Invensys Schneider added considerable software capability that it tried to build on with the attempted acquisition of Aveva.

• ABB has still to define a clear software strategy, though several high profile software deals (e.g. Ventyx) have added capability to the portfolio.

• Hexagon has transformed the nature of its business model and moved to become a software company itself through the acquisition of Intergraph in 2010.

• GE has forged a partnership with PTC to combine its PLM and its own MES software, and also heavily invested in predictive analytics capabilities with the launch of its Predix platform


Exhibit 46: Siemens moved first towards more digital continuity, Schneider recently sought to catch up through its acquisition of Invensys and Aveva


Siemens was first to acquire significant software capabilities in 2007 Through its acquisition of UGS in 2007, Siemens managed to become the first supplier of software and hardware across the complete product and production life cycle. The deal provided Siemens significant capabilities in the field of computer-aided design (CAD), computer-aided manufacturing (CAM), computer-aided simulation (CAE) and product data management (PDM). Going forward Siemens actively integrated UGS capabilities into the overall group to form operate the software business alongside its manufacturing capabilities. UGS was mainly focused in the discrete manufacturing industries such as automotive, aerospace and defence, consumer goods, electronics and machinery.

Exhibit 47: Through the UGS acquisition, Siemens acquired PLM capabilities in the aerospace, consumer goods, electronics and machinery space

Source: Siemens

Siemens completes digital manufacturing through various small acquisitions Siemens gradually completed its digital manufacturing offering between 2008 and 2014 by making a series of small acquisitions. The most recent example was its takeover of CD-Adapco, adding Computational Fluid Dynamics software capabilities to its PLM portfolio. Camstar, whose MES capabilities further complemented Siemens’ strong PLM footprint and broadened its capabilities from product initiation to manufacturing was bought in 2014.


Exhibit 48: Camstar reduces time to value and the need for coding and customization due to high out of the box functionality

Source: Siemens

Table 36: Siemens acquired various small software players to build up its digital manufacturing capabilities Year Target Area of Business 2014 CD Adapco Computational Fluid Dynamics (CFD)

2014 Camstar MES software for electronics, semiconductor and medical devices

2013 TESIS Integration of PLM and enterprise applications

2012 Kineo Computer aided motion software

2012 IBS Industrial quality and production management

2011 Active MES software for pharma and biotech industries

2011 Vistagy CAD design for production of composite materials

2009 Elan Software Systems MES software for life science

2008 Innotec Digital engineering for process industries

Source: Company data

ABB and Schneider followed Siemens from 2010 by adding asset management

solutions In 2010, ABB acquired Ventyx, a software provider for energy solutions in the utilities vertical, followed by the takeover of Mincom, offering asset management solutions for the mining space, in 2011. In comparison to Siemens’ move towards digital continuation, ABB’s actions were more geared towards “add-on” vertical services to its overall solutions in the process industry rather than developing a complete offering across the entire digital value chain which can be seen complementary to its offering of supervisory and control (SCADA) software embedded within its DCS control systems (such as its flagship product 800xA).

Exhibit 49: The Ventyx deal makes ABB the leading provider of energy solutions in areas such as plant operations management, load & revenue forecasting and mobile work management

Source: ABB

Schneider’s acquisition of Telvent in 2011 provided the company with real-time management capabilities for critical infrastructure (such as energy and operational efficiency management) in the process automation space, giving it exposure to the segments of utilities, oil and gas and water. Schneider’s first action can rather be understood as a move into the energy-intensive process automation space, extending its focus from a traditionally strong position in discrete manufacturing. The transaction complemented Schneider’s product offering in low-voltage (which is needed to connect the devices on the factory floor with the monitoring tools) and also drives.


Exhibit 50: Telvent’s software capabilities complement Schneider’s position in energy management

Source: Schneider

Schneider broadened its end-market exposure by acquiring Invensys in 2013 Schneider represents one of the most attempts to complete the digital value chain integrating the design process, manufacturing operations until the launch of the end-product. Its acquisition of Invensys in 2013 displays a crucial step, though can be also be seen again as an extension of its DCS capabilities due to Invensys’ large installed base in DCS control systems (Foxboro). The move into process industries complements Schneider’s energy management software (such as the Telvent suite) and its products like drives and low-voltage equipment which are particularly important for energy intensive industries. Traditionally, Schneider was rather focused on “lighter” industries in the discrete manufacturing space such as the automotive segment.

Exhibit 51: Via its acquisition of Invensys, Schneider is able to extend its end-market reach to the process automation space and…

Source: Schneider

Exhibit 52: …to provide a more complete digital offering in all key areas,

Source: Schneider

In 2010 Hexagon transitioned to a hybrid hardware/software business Hexagon’s acquisition of Intergraph in 2010 displays a special case when compared to the aforementioned transactions as it can rather be seen as a merger of software and hardware companies than the mere acquisition of software capabilities of an automation vendor. This can partly be seen by the relative size of the transaction, where Hexagon carried out a rights issue and took on large amount of debt (Hexagon’s net debt/EBITDA ratio stood at 5.7x vs a 2014 target of 3.8x in 2009 and a target of 2.5x for 2014) in order to finance the transaction. In turn Intergraph added more than 50% to Hexagon’s existing revenue, making a material addition to the overall business. Intergraph added PLM capabilities to Hexagon’s existing portfolio of measurement instruments. The combined portfolio offers a significant value-add, as data from measurement instruments can be levered for digital design and manufacturing operations.


Exhibit 53: Hexagon levers its measurement equipment to provide the input data for the virtual design software acquired from Intergraph

Source: Hexagon

GE and Schneider chose alternative strategies to enter the software space GE announced a partnership with software maker PTC in 2013, combining its own MES software capabilities (Proficy for Discrete Manufacturing) with those of PTC (Manufacturing Process Management) which will also provide its PLM solutions (PTC Windchill) within the joint offering. GE intends to provide a closed loop of information including design, engineering, manufacturing and service for the end-product. The step displays a first step for GE towards digital continuity, however lacks the commitment seen for example in the case of Siemens and Schneider. This is mainly due to the fact that GE so far only engages in a commercial agreement with PTC and since its main focus is probably on data analytics.

Exhibit 54: GE’s partnership with PTC addresses key customer demands such as shorter product release cycles, connection of design and manufacturing and real-time insight into operations

Source: GE


Appendix B: Historical margin analysis

Margins across the sector are likely to be driven by growth, associated leverage, mix and execution. Given higher product focus, Discrete markets tend to higher more consistent operating leverage, around 30%

Below we look at regional margin differences, typical operating leverage and consider barriers to entry for the automation companies. In general, automation markets are a mid-teens margin businesses for the large integrated players, with niche players able to achieve margins in the mid-twenties, if execution and positioning is strong. We think there are very few restructuring cases in the sector, with margins from here being driven by growth and associated leverage, which can vary depending on product/software split, execution and discrete/ process focus. Discrete biased companies should be well positioned in the near term, given typically good leverage to growth.

Longer term, we expect margins to be defendable given the importance of the installed base of automation equipment, high direct and indirect switching costs for customers, typically conservative customers, technical requirements, DCS domain knowledge and associated engineering requirements.

Historical trends: mid-teen margins, US stronger Since 2006/2007 average regional margins across the sector (ex Fanuc) have been relatively similar. In the period between 2004 and Q2 2008, on average operating profit margins for the large automation players steadily moved higher from 10-12% to 13-15% in 2008 (Chart 49). Since 2011, margins in North America have continued to improve, while margins in Europe are somewhat flattish and Japan has just about regained its peak. Within each region;

- North America – in general recent improvements in margins have been across all the major automation suppliers, with margins now in a relatively tight range (15-20%).

In general, industrial automation markets are mid-teen margin businesses, with higher margins in the US

- Europe - since the end of 2010, margins have tracked lower but have recently recovered. This prior decline was mainly driven by flat/declining margins at most of the larger players. There has been more volatility in QoQ margins, with a larger margin range than the US names.

- Japan - most companies made a loss in the 2009, saw a substantial rebound in 2010, but have declined since and recently recovered once more. There is a substantial margin range within the Japanese names – with Fanuc on 35-40% margins, SMC in the mid-twenties/low thirties and Yokogawa in the mid-single digits.


Chart 99: Regional operating profit margins - US leads Europe and Japan


Across the niche names in the sector, margins range from mid-single digits to the mid 20%’s – we think based on market position and operational management. The largest range is within Europe, with Spectris, IMI and Hexagon all with margins that have grown into the high teens/low 20%’s, while Kuka has mid-single digit margins. In the US, niche margins are typically in low to mid-teens (Flowserve, Parker Hannifin, National Instruments) with some volatility at Aspentech.

Near term: op leverage, mix & execution

From here, we expect sales growth and operating leverage to be the main margin drivers – Discrete product margin leverage is typically higher than Process software/integration sales, which carry additional risks

With most companies at or around peer margin levels, respectable returns for most companies and management teams not signalling additional restructuring measures in the near term, margins for most names are likely to be driven by operating leverage, mix and execution.

For Discrete automation related products, which are typically more product focused (PLC’s, sensors, fluid power etc), we believe operating leverage is c.30% and execution risk is relatively low, given the standard operating model. Below, we show organic growth vs operating leverage for the more Discrete exposed names – while there are a number of other issues that impact leverage and cause volatility (FX, mix, cost initiatives), operating leverage here has averaged c.30%, and ranged from 10-50%.

-10%

-5%

0%

5%

10%

15%

20%

25%

30%Q2

2004

Q2 20

05

Q2 20

06

Q2 20

07

Q2 20

08

Q2 20

09

Q2 20

10

Q2 20

11

Q2 20

12

Q2 20

13

Q2 20

14

Q2 20

15

Q2 20

16

North America Europe Japan Japan (ex Fanuc)


Chart 100: : PLC – organic growth (x axis) vs operating leverage (y-axis)

Source: BofA Merrill Lynch Global Research Estimates PLC companies/divisions used Siemens Ind auto, Rockwell, Mitsubishi, Schneider

Chart 101: DCS - organic growth (x axis) vs operating leverage (y-axis)

Source: BofA Merrill Lynch Global Research Estimates DCS companies/divisions used – ABB, Honeywell, Emerson process, Yokogawa, ISYS

In contrast, operating leverage for the Process automation players varies based on mix (and associated operating leverage by product line) and execution. In general, those companies more exposed to Systems/solutions style work (particularly ABB and Honeywell), have lower leverage and higher operating risk on contract execution, and those that are exposed more to Products (e.g. Emerson) are likely to see higher leverage. We show organic growth versus operating leverage for some of the more Process facing automation suppliers – with lower average leverage in general (17%) and a wider spread of operating leverage.

0%

10%

20%

30%

40%

50%

60%

70%

-30% -20% -10% 0% 10% 20% 30%-20%-10%

0%10%20%30%40%50%60%

-20% -15% -10% -5% 0% 5% 10% 15%


Appendix C: Product overview In this section, we describe the major product groupings within our definition of industrial automation markets in more detail

ICS - Industrial control system ICS is a general term used to describe control systems used to production processes. The main control systems are PLC’s (used in discrete manufacturing), DCS (used in process applications) and SCADA (supervisory control).

PLC – Programmable Logic Controller A PLC is a digital unit used for the control of electro-mechanical processes/ products, typically in discrete manufacturing processes (typical end market exposures are shown below). Modern PLC’s are typically programmed by software on PC’s and are used to operate a set programme using a number of inputs and outputs. PLC’s are used to control motion, relay control, process control, within distributed control systems and networking. The processing power is broadly similar to a modern PC, but PLC’s have considerably more stable operating systems and are designed to operate in a wide range of operating conditions on the factory floor (temperature, moisture, dust etc).

PLC’s can normally connect to a wide range of products – sensors (incl analog), actuators, machine visions products, motors, pneumatic or hydraulic control products, relays, solenoids, SCADA systems and be controlled by HMI’s on the factory floor. PLC applications are highly customised, so the cost of the PLC is relatively low versus the cost of specific custom built controller application/design. PLC’s can be bundled with associated software (incl HMI software).

Chart 102: End market exposure to type of process control

Source: BofA Merrill Lynch Global Research Estimates, Siemens

DCS – Distributed Control System A DCS is a system where the control elements are not located centrally, but are distributed throughout the whole manufacturing system. A DCS typically uses a number of specific digital controllers (rather than a standard product adapted via programming, like a PLC) each operating a number of regulatory feedback loops. The controllers can be connected to each other and a number of different devices (like HMI, switches, motors or actuators) either directly, through a field network (which can potentially be custom depending supplier) and/or through intermediate systems (e.g. SCADA system).

A DCS is used to control a continuous manufacturing process (e.g. oil refining, petrochemicals, power stations pharma, food and beverage plants). Processes aren’t limited to fluid flow, and may include paper machines (quality controls), mining operations and ore processing facilities. DCS systems typically include a degree of redundancy, which makes them more reliable.

0%

20%

40%

60%

80%

100%

Elec

tronic

s

Autom

otive

Food

&Bev

erag

e

Phar

ma

Metal

formi

ng

Ceme

nt&Gl

ass

Pulp&

Pape

r

Oil&

Gas

Chem

icals

Wate

r

Refin

ery

Ener

gy

Manufacturing Hybrid Process

Manu

factur

ing a

utoma

tion

PLC

base

d

Process automation DCS based


Table 37: PLC's vs DCS - typical differences PLC "controlling a machine" DCS "controlling a plant" Market introduction 1960's 1975 Replacement of… Electromechanical relays Pneumatic & single loop controllers Products manufactured… "things" - discrete items "stuff" - transformation/combination of raw materials, high batch value Classic application Automotive Refining Speed of control High speed (scan rates less than 10millis) More modest response speed (100-500 millis) Type of control Discrete Regulatory Up-front cost $$ $$$$ Size Compact Large Redundancy May not be justified Typically requirement given value of plant Engineering requirements Custom programming normally required Typically configuration of standard logic function blocks Operator Handles exceptions Req. to make decisions to run efficiently Driver of control decisions PLC controller contains all logic to move products HMI - central control room System Open Closed (proprietary) Source: BofA Merrill Lynch Global Research Estimates, Siemens

The main applications for DCS systems are batch control, continuous process applications (the majority of systems), quality systems and SCADA (Supervisory control and data analysis, both standard and industry specific). SCADA is an industrial control system (ICS, similar to the DCS’s and PLC control system’s used above) which is focused on collecting and supervising data typically over multiple sites and large distances. SCADA systems typically have DCS components and can use PLC’s to measure, and override control of the process.

Hybrid systems – the main differences between PLC and DCS’s are shown in the table above. In reality most manufacturing plants use a combination of both DCS’s and PLC’s for control manufacturing processes. For example in large process plants DCS’s generally control and manage the core processes, while PLC’s are used to control non-core processes. With convergence over time of PLC and DCS functionality and increasing manufacturing complexity, we expect the market for hybrid automation systems to continue to grow.

Chart 103: Process system vs HMI directed intervention vs Safety System. Why do Process Safety Incidents Occur?

Source: Honeywell, BofA Merrill Lynch Global Research Estimates

Chart 104: HMI equipment – screens and control panels

Source: Mitsubishi, BofA Merrill Lynch Global Research Estimates

Safety Systems Safety systems are similar to control systems, but are designed purely to intervene when there is an unsafe process - when a worker, the plant or the environment is threatened. Safety systems can be related to fire and gas, emergency shutdown systems and burner management systems. While the equipment can be similar to that used in DCS systems, Safety systems typically have their own set of parameters and standards that exist in a separate system.

Industrial PC’s, HMI Industrial PC’s – are PC’s for industrial environment, designed to deal with harsh environments, long operating times and difficult conditions. They can be used for

Time

$

$$

$$$$

$$$$$$$$Equipment Failure

OfferingHPS

Position

#2

#1

#1

#2

N/A

Control System

Advanced Control

Alarm System

Operator Training

Safety System

Response,EvacuationRescue

NormalOperation

UpsetCondition

CriticalSituation

Safety System Controlled

Human Controlled

Process Control

Control Action

Operator Action

Emergency Shutdown Action

#1

$$$$

$$$

$


controlling a process, data collection, and be the front end to other control hardware (for example in a DCS). Software can be customised or off the shelf depending on the application. HMI hardware (HMI) - is the equipment that provides a control and visualisation interface between the operator and the manufacturing process. They are typically control panels or screens which allow the operator to control, monitor, diagnose and manage the process.

Advanced software applications This category encompasses software to help improve efficiency and manage the plant more effectively, including;

• Simulation and operator training

• Asset management and preventative maintenance software

• Mobile applications

• Power metering, carbon and energy management

We have not included broader manufacturing and software used to integrate into the wider enterprise (e.g. MES – manufacturing enterprise software, PLM – product lifecycle management), due to the range of different competitors involved in this market.

Industrial control equipment Motors, drives, generators, mechanical power transmission This category of products is controlled by the industrial control system and provides motion in the manufacturing system;

Chart 105: Typical drive train layout – drive, motor & power transmission

Source: ABB

1Drives (variable speed drive, inverter drive) - electronic device used to regulate the performance of an electronic motor. A Drive typically controls the power, frequency and current that the motor draws from the grid. Drives are used in small equipment to large mine mills and compressors. In terms of route to market, in North America, customers typically look to the Drive manufacturers to provide complete packages, where as in Europe, Drives are typically sold through 3rd parties.

2 - Motors – convert electrical energy into mechanical energy. There are a range of sizes and specifications (IEA definition below);

- Small motors, <0.75KW – are used in a variety of small applications, mostly outside industrial markets in residential and commercial sectors. In general they are used in packaged applications like extractor fans and computer hard drives.

- Mid size motors, 0.75KW to 375KW – there are a range of different technologies in this category, but AC induction motors are the most popular.

Powersupply

Conversionand/or regulationof electricalenergy supplied

Conversion ofelectrical intomechanical(rotary) energy

Transmission,conversion and/orregulation ofmechanical energy

Drive Motor Mech.PT

Workingmachine


Products are typically sold to OEM’s and sold as part of a pump, fan or compressor or solid stand alone. Mid-range motors are most often found in industrial applications, but can also be found in residential and infrastructure markets.

- Large motors, >375kW – normally high voltage AC motors and given the size, are normally custom designed and built to specification. While not a big market, they are estimated to consume 23% of the electricity used across electric motor markets

Table 38: Installed base of motors industrial motors - mainly medium sized and move to middle

mn of units % of energy consump. Small <750W 2000 9% Medium 0.75KW to 375 KW 230 68% Large >375KW 0.6 23% Total 2,236 Source: IEA

3 - Generators – the reverse of a motor – a device that’s converts rotating mechanical movement into electric power.

4 - Mechanical power transmission – equipment to convert and connect mechanical motions – products include gearboxes, couplings.

Field instruments – sensors, machine vision, test & measurement Field instruments include a number of different products which help to monitor the industrial process. Main products include;

• Sensors – products detecting various factors in a manufacturing process. The table below shows the characteristics that are typically measured. Sensors can range from very simple, low value products, to large high value complex solutions.

Table 39: Sensors - parameters that are typically measured

Physical parameters Chemical parameters Acoustic, pressure, temperature, flow, humidity Gas, liquid, solid Mechanical parameters Electromagnetic Absolute position, level, force/deformation. Inertial, position/ displacement, proximity/ distance Current, voltage, magnetic field


The main sensor types are pressure, temperature, flow and level, which account for around 30% of the total market. More application specific sensors can include vibration, torque, strain, speed, acceleration, image, gas, rain, moisture and humidity sensors.

• Machine vision – technology used to provide image based automatic inspection for quality, process control (e.g. sorting, material handling) and robot guidance systems.

• Test & measurement, and analytical equipment – products that are used to test and measure within the manufacturing process (online testing, rather than offline). This includes a wide range of end markets from industrial PC’s to infrared measurement of food and bulk materials, gas analysis instrumentation, process data loggers and recorders.

Control valves and associated equipment Valves used to control the flow, pressure, temperature or level of a liquid/gas by fully or partially opening. Typically controlled by an actuator in process industries, according to a pre-defined reference point or process variable. There are various types of control valve (globe, rotary, sliding cylinder) which offer different characteristics for different applications. Within this category we include the associated equipment; 1) valve


automation (multi turn, part turn, linear actuators, positioners) which can be electrical (using motor, sensors, gearing, controls), hydraulic or pneumatically driven, 2) transducers, supply pressure regulators, manual operators, limit switches etc.

Exhibit 55: Typical control valve – globe valve and diaphragm actuators

Source: Emerson

The main actuator technology for process control valves is pneumatic spring diaphragm (fluid controlled), although Rotork, through there CVA actuator range are introducing electric actuators into the market which require lower maintenance, lower infrastructure (air lines, clean air) and service requirements.

Fluid power – hydraulics and pneumatics Fluid power products use fluid under pressure to generate, control and transmit power. Fluid power is split into; 1) hydraulics, which use liquids such as mineral oil or water, and 2) pneumatics, which uses gas. A typical fluid power control circuit is shown below

Exhibit 56: A typical industrial pneumatic device system

Source: Misumi

Typical components in a fluid control circuit are hydraulic pump/ air compressors (mechanical to fluid power), valves (which control the direction, pressure and rate of flow), actuators/ positioning systems (convert fluid to mechanical power), filters, regulations and lubricators (condition the flow), manifolds, hoses, fittings (transmit the flow), instruments (pressure gauges, flow meters, sensors etc), vacuum components and related electrical connections. Typical applications include material handling (palletizers, conveyors, automated storage), food processing (packaging, handling and manufacturing products), electronics (pick and placed assembly) and in machine tools (standard machine components).


Industrial interconnect equipment In this category we include electrical interconnect equipment – both the networks and low voltage equipment that make the physical link for moving data around, but also communication related equipment which defines what the data is and what it means.

Communication related equipment There a number of communication network protocols used for control in factory automation. The traditional system is organised as a pyramid, with the system management/SCADA systems at the top, connected through the plant/ process/ controller Network to the process level equipment (PLC’s, robots, HMI, CNC’s), which in turn is connected through the Field network/bus to field level equipment and potentially Actuator/sensor specific interface/ buses to connect to sensors, switches, motors, actuators etc.

• Fieldbus is the family of communication protocols typically used for real time control in automation systems (it is still c.75% of new industrial network connections). The fieldbus typically connects the control system (PLC, DCS) with the control equipment (sensors, actuators, motors etc) and allows one cable to connect the field devices to the controller (rather one cable for each device). Equipment here includes power supplies, wiring components (incl spurs), controller interfaces (ensure communication), device coupler (junction between cable and field device), terminators diagnostic and test equipment.

• Industrial Ethernet – with the increasing move to Industrial Ethernet (same communication protocol uses across the automation network – described previously), we also include Ethernet related products in the Communication part of the market – including specialised cabling, Ethernet switches, converters, security, communication processor/ interfaces, network transitions.

• There is also a move toward wireless and longer range cellular connectivity (machine to machine), helping to remove physical cabling requirements. Both can improve flexibility, reduce costs and easy planning, and allow remote monitoring. Wireless and cellular industrial automation products include device managers, routers, bluetooth devices, radios, I/O gateways & nodes, modems and servers.

Low voltage products Products used for the protection, control and measurement of low voltage electrical installations. Products include cable management (trunking, ducting, routing systems, cable trays), energy distribution (circuit breakers, surge protection, distribution panels) and control (switches).

Motion control – precise position & control Motion control is where the position or velocity of a machine is controlled by a hydraulic or electric motor. There are clear product overlaps with other product areas within our definition of the automation market, but motion control involves high precision, can be stand-alone from broader plant automation and is an important part of robotics and machine tool markets. Motion controls components are typically used in material handling, packaging and robotic applications.


Exhibit 57: Techno CNC machine tool – including major motion control products

Source: TechnoCNC

The core components of a motion control system are;

1. a controller (multi axis, servo, single axis, stepper)

2. drive and motor (servo, stepper, linear, micro)

3. amplifiers

4. actuators (linear, high speed, miniature, pneumatic)

5. feedback devices (position and speed sensors)

Other items include gantry systems, guides/rails, positioning tables, nano/micro positioning and other machine specific components. While we include the major control components, we do not include CNC (computer numerical control) machine tools in our definition of the automation market given the different nature of the business model and the different companies typically involved.

Robotics Factory Robots typically use many of the other automation control components – including motion control equipment, sensors, machine vision, although typically include tooling for a specific set of tasks (welding, painting etc). We deal with this subject in more detail in its own section.

Software Definitions – Making sense of the acronyms

In order to cut through the complexity of the software various software segments, the following passage will give an introduction to the most relevant types of industrial software and describe how they are deployed.

Product Lifecycle Management (PLM): Virtual Design, Manufacturing and Simulation From an initial phase of virtual design via Computer-Aided Design (CAD) software, the manufacturer might model his product with the help of computer tools and even run simulations through Computer-Aided Engineering (CAE) software in order to test the robustness and viability of his construction. In another step also the manufacturing process can be modelled by using Computer-Aided Manufacturing (CAM) software in an effort to test feasibility of the actual execution step. Through Product Data Management (PDM) software manufacturers can manage product and process-related information in a single, central system.


Dassault, the leader in the space, currently values the core PLM market (automotive, aerospace, industrial equipment) at US$10.6bn, but expects to extend its scope of penetration, applying its software to other industries such as consumer goods, life science or high tech. The total new addressable market is projected to amount $11.2bn in 2018, representing annual growth of 9%.

Chart 106: Dassault is the clear market leader in the PLM segment

Source: Dassault

Chart 107: New industries such as life science and high tech will amount to US$11.2bn in 2018 (’13-’18 CAGR: 9%)

Source: Dassault

Example: In order to produce its award-winning Qashqai model, Nissan used Siemens’ NX software for design, simulation and manufacturing. Through the virtualization of the design process and digital analysis and stress-testing of the data, Nissan was able to reduce its time-to-market by c.50% without the creation of any hand-built prototype. Moreover, through the usage of Siemens PDM software Teamcenter, Nissan was able to share real-time data from the design and manufacturing process among its geographically dispersed workforce and enabled them to collaborate faster and more efficiently. Dassault leads the PLM segment with an overall share of 30%.

Manufacturing Execution Services (MES): Operation Management MES software helps the manufacturer to analyse factory floor activities including the production plan changes, capacity and changing material requirements. Via its integration with PLM, data from the actual manufacturing execution can be fed back into the models which run product simulations in order to improve these.

Chart 108: The global MES market is very fragmented, with many providers of niche solutions

Source: Frost & Sullivan

Exhibit 58: Rockwell’s Track&Trace Application for automotive manufacturers allows for faster recalls and identifies quality concerns along the supply chain by keeping track of equipment

Source: Rockwell

Dassault 30%

Siemens 19% PTC

10%

Autodesk 9%

Other 32%

0

2

4

6

8

10

12

2013 2015 2018

Schneider 8% ABB

6% Siemens

4% SAP 3%

Oracle 2%

Rockwell 2%

Other 75%


Example: Biopharmaceutical maker Ferring Pharmaceutical uses Rockwell’s Factory Talk Pharma Suite to reduce its lead time and to better handle its quality compliance testing. Before the introduction, the company conducted its paper-based quality test that consumed up to 44 days between the manufacturing of a batch and its release. Via electronic batch recording (eRB) software, Ferring is able to detect deviations from the batch recipe in real-time and can also conduct quality reviews during the manufacturing process thereby significantly reducing its time to market.

Enterprise Asset Management (EAM): Maintenance management and asset

efficiency Maintenance Management Traditional maintenance management functions include planning, organizing and implementing maintenance activities. This come in the form of tracking and scheduling maintenance work, optimizing the material procurement and inventory management, and also the financial management of the maintenance process.

Chart 109: IBM holds the top spot in the global market for maintenance management

Source: IDC

Predictive Maintenance The market for predictive maintenance (PdM) is still relatively young and lacks a clear definition. The term describes the usage of data in order to improve the performance and utilization of assets in order to increase productivity and reduce maintenance spending. According to SAP companies that actively engaged in data analytics showed a 10% higher capacity utilization through active monitoring of real-time conditions within a plant which results in increased uptime.

IBM 16%

SAP 10%

Oracle 7%

Infor 5%

ABB 5%

Other 57%


Chart 110: The traditional market for predictive analytics software was dominated by specialists (e.g. SAS) or big technology companies (e.g. IBM), but…

Source: IDC

Chart 111: … traditional industrial companies like GE (with its Predix software platform) are driving up their own software capabilities significantly

Source: GE

Example: German air compressor supplier Kaeser Kompressoren uses SAP’s HANA software in order to get real-time data from the compressor at client sites such as temperature, pressure level or the amount of delivered and compressed air amounting to more than a one million data points per day. Through predictive maintenance i.e. the identification of real-time usage patterns and its comparison with allowed levels to predict machine health, Kaeser Kompressoren can deliver improved customer service. Unnecessary maintenance work can be avoided and unplanned machine outages can be pre-empted. The result is higher customer satisfaction and lower service expenses for Kaeser. Through its web-based access the software is also readily available at customer sites and service engineers can operate using real-time analytics.

SAS 33%

IBM 16% Microsoft

3% FICO 2%

Other 46%

0

1

2

3

4

5

6

2014 2015E … 2017E

GE Predix software revenue ($bn)


Appendix D: The industrial robots ecosystem Where it fits in in our holistic view

Exhibit 59: BofAML – definition of industrial automation market


Demand for industrial robots is increasing steadily in many industries. In order to give an overview of the functions and applications that robots are mainly performing today, we provide a short 101 below.

The Industrial Robots market at a glance

The Industrial Robot market had a size of $9.5bn in 2013 and is expected to grow by 8.5% until 2025.

Although China only accounts for c9% of the global installed robot base, it is expected to account for 1/3 of annual shipments by 2017.

In 2013 the value of the global industrial robot market stood at $9.5bn, which represented a new peak, according to IFR. BCG projects total robot spending to increase to $67bn in 2025, representing a CAGR of 9.5% from 2015 (Chart 29). Expenditures for industrial robot are projected to rise by 13.4bn accounting for one third of the overall increase, reflecting a CAGR of 8.5%.


motors, actuators












Data centres





Machine tools




Industrial control

equipment





The International Robot Federation also states that the worldwide stock of industrial robots stood at c1.4m in 2013 and is expected to rise to over 1.6m by 2016, reflecting a CAGR of c14%).

Chart 112: The industrial robot market stood at $9.5bn in 2013 and is estimated to reach $13.4bn by 2018, CAGR of 8.5% (thousands)

Source: BCG

Chart 113: The global installed base of industrial robots is expected to rise above 1.7m in 2016 (thousands)

Source: IFR

Four main types of robot We divide the industrial robot space into four types or segments, namely articulated robots, SCARA robots, 124artesian robots and cylindrical robots. Table 40: Overview of the four main types of industrial robots

Robot Type Description Cost Size Payload Cycle Time Applications

Articulated Robot with several rotary joints ranging from to 2 to 10. Suitable for various applications within industrial production given its flexibility. Can move in all three dimensions.

High Big High Fast Welding, assembly, machine tending, painting

SCARA “Selective Compliant Assembly Robot Arm” is a simpler from of the articulated robot which only works on the horizontal axis as it lacks rotary joints.

Medium Small Medium Fast Assembly, drilling, pick and place

Cartesian

Can move in all three dimensions, however only in a straight line without rotating. Cartesian robots operate behind security barriers and are well-suited for heavy payloads

Highest Biggest Highest Medium Heavy duty applications (e.g. automotive assembly)

Parallel Consists of a joint base which holds three connected arms. The arms are arranged in a parallelogram shape, helping to maintain the direction of the end-effector.

Low Small Low Fastest Light weight, pick and place


0

20

40

60

80

2000 2005 2010 2015E 2020E 2025E


- 200 400 600 800

1,000 1,200 1,400 1,600 1,800

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

e20

16e


Exhibit 60: Articulated robot

Source: Kuka

Exhibit 61: SCARA Robot

Source: Omron

Exhibit 62: Cartesian robot

Source: Toshiba

Exhibit 63: Parallel robot

Source: Fanuc

Articulated and Cartesian robots account for 84% of the market According to technology research firm TechNavio, the articulated and 125artesian robot markets together account for c.84% of the overall global industrial robot space, which reflects the still strong dominance of traditional industries such as the automotive field where high payloads are required of robots and where these robots operate separated from humans due to safety reasons.


Chart 114: Articulated and Cartesian robots still the dominating types given their demand from traditional heavy industries

Source: TechNavio

Three main functions in industry Robots are used for three main functions in industry: material handling, welding, and other uses such as painting and assembly.

Material Handling • Pick and Place: Robots are used to pick up elements and placing them in a new

area. The robot is able to increase efficiency and accuracy in these repetitive processes.

• Machine-Tending: Robots oversee machines while it is running and feed parts in and out. While this is a dull and dangerous job for humans, it still requires a high level of consistency and is therefore well-suited for the application of industrial robots.

Welding • Laser Welding: Laser Welding robots use servo-controlled, multi-axis mechanical

arm which have a laser cutting head mounted to the faceplate of the arm. Deploying industrial robots for laser welding can improve repeatability and drive up quality of the welds.

• Arc Welding: Arc Welding robots use a process which applies intense heat to metal at a joint, causing the metal to melt and intermix. Advantages of robot welding include higher weld consistency, shorter cycle times and enhanced efficiency and safety.

Other • Painting: Painting robots are used in order to protect human workers in unsafe

environments where they would be exposed to isocyanides, carcinogens etc.

• Assembly: Assembly robots replace workers at repetitive assembly line jobs, increasing consistency and speed and lowering production costs.

Articulated Robot 62%

Cartesian Robot 22%

SCARA Robot 11%

Other 5%


Chart 115: Material handling and welding accounts for the majority of functions performed by industrial robots

Source: TechNavio

Exhibit 64: Robotic material handling can replace dangerous and repetitive manufacturing jobs previously done by human workers

Source: T-Tek

What are industrial robots made of? In order to understand what the key components for industrial robots are, we compare them to the individual parts of the human body and their corresponding functions. Furthermore, we also highlight the key supplier of the individual parts.

Exhibit 65: Mobile robot – industry value chain, global 2014

Source: Frost & Sullivan

Analogies with the human body Equivalent to the human intelligence, robots derive the information on which they operate through hardware controllers and software. Human arms and hands are represented by end-effectors through which the robot ultimately performs its functions such as pick & place, welding and material handling. Actuators reflect human muscles which set other parts like arms and end effectors in motion and transfer mechanisms act as robotic legs. Finally, sensors, which are comparable to human senses, receive feedback from the arm and end effector about its position and velocity, and transmit this information back to the controller.

Handling of Materials

39%

Welding & Soldering

26%

Assembling &

Diassembling 11%

Dispensing & Painting

4%

Cutting, Milling &

Other Processes

3%

Others 17%


Table 41: Especially the big Japanese players manufacture the components of their robots in-house

Human Robot Who makes it?

Intelligence Information, PLC, HMI, Software Kuka, ABB, Fanuc, Yaskawa, Schneider, Siemens, Delta

Arms Arms Kuka, Fanuc, ABB, Yaskawa

Hands End Effectors ABB, Kuka, Fanuc, Yaskawa, Applied Robotics, Anver, Krones

Muscle Actuators Mitsubishi, Yaskawa, Fanuc, Delta, Hiwin

Legs Transfer Mechanism

Sense Sensor (External Sensor, Internal Sensor)

Jenoptik, Panasonic, Nitta, Interlink, Murata


Overview of the main industrial robot players The global robot market is dominated by four main players which account for more than half of the total revenues. These companies are Fanuc, ABB, Yaskawa and Kuka.

Chart 116: Fanuc, ABB, Yaskawa and Kuka are the four biggest industrial robot vendors that account for over half global revenue

Source: IFR

Table 42: Smaller robot vendors are specialized on individual end-markets or technologies Company Country Description Adept US Intelligent, vision-guided robotic systems

Nachi Japan Mainly exposed to the automotive sector

Denso Japan Specialized manufacturer of small assembly products

IHI Japan Focus on material handling

Intuitive Surgical US Specialized on surgical robots

Lincoln Electric US Focus on laser cutting and arc welding

Hiwin Taiwan One-axis, Parallel and SCARA robots producer, mainly focussed on the electronics, food-processing and general manufacturing segment

Delta Taiwan Focus on electronics and general manufacturing


Industrial robot producers outgrow industrial production, but are also more cyclical As can be seen in the Charts below, the companies in the robot segment grow on average c.3x as fast as global IP per year. Between 2006 and 2014 the firms in the robotics space displayed average growth of 7.2% vs 2.4% for global industrial activity annually. And although being more cyclical than the overall sector activity, robot manufacturers also outperform global IP in terms of compounded annual growth which is almost twice as high for the former at 4.8% vs. 2.6% during the for the previous 8 years. The high growth rates which are supported by the expected overall spending on robotics increased spending as well as the emerging demand from non-automotive segments reflect the value potential that robot producers inherit and which should be supportive of a continuation of the strong upward trend on revenues in the longer term.

Fanuc 16%

ABB 14%

Yaskawa 12%

Kuka 11%

Others 47%


Chart 117: Companies in the robotics segment (CAGR: 4.8%) grew almost twice as fast as global IP (CAGR: 2.6%) between 2006 and 2014

Source: BofA Merrill Lynch Global Research Robot companies include Kuka, Fanuc (robot division), ABB (robot business, only disclosed until 2009, onwards BofA estimates), Yaskawa (robotics division)

Chart 118: Fanuc outperforms its peers through superior technologies and full vertical integration


Table 43: European robot maker are less vertical integrated than their Japanese peers , but have a higher exposure to general industries

Company Business Degree of Integration Auto/General Industry Sales Split Installed Base

Fanuc Mainly Arc Welding, Learning, and Genkotsu Robots

Full Vertical Integration, Manufacturer of Core Components (CNC, Servo Motors etc.) 60%/40% 250,000

ABB Full Product Line in Industrial Robots Industry Assembly and System Integration 50%/50% 250,000

Yaskawa Specializes in Arc Welding, Spot Welding, Painting and Material Handling

Full Vertical Integration, Manufacturer of Core Components (Controllers, Servo Motors etc.), System Integration in China

70%/30% 300,000

Kuka Offers Automation Solutions in all Payload Classes Assembly and System Integration 50%/50% 200,000


Autos: #1 sector for adoption of robots The automotive industry has been the leading sector for the robotics industry since 2010. In 2014, nearly 100,000 robots, or nearly one out of two industrial robot sold, was for the autos sector. Between 2010 and 2014, robots for autos rose at a 27% CAGR, with demand coming from both EMs and developed markets in terms of modernising their factories (source: IFR 2015). Adoption has been swift given the cost benefits. In markets like the US, it costs US$8/hour to use a robot for spot welding, versus US$25 for a labourer (source: BCG). Other sectors are also quickly reaching that turning point, which we believe will drive growth.

-60%-50%-40%-30%-20%-10%

0%10%20%30%40%

2006 2007 2008 2009 2010 2011 2012 2013 2014

Global IP Median Robotics Organic Growth

-20%

-10%

0%

10%

20%

30%

40%

50%

2006 2007 2008 2009 2010 2011 2012 2013 2014

Kuka ABB Yaskawa Fanuc


Chart 119: Estimated worldwide annual supply of industrial robots at year-end by main industries 2011-14

Source: IFR World Robotics 2015

Chart 120: Robot density in 2014 per 10,000 employees

Source: IFR 2015

In auto plants, robots are often given nicknames. Tesla names theirs after Marvel comic book characters such as Wolverine, Professor X, and Iceman. Nissan names theirs after anime characters such as Doraemon and Vegeta

Exhibit 66: “Wolverine” a Fanuc robot used in Tesla’s Fremont factory

Source: Business Insider

Exhibit 67: Kuka robots in Tesla’s Fremont factory

Source: KALW

0 40,000 80,000 120,000

not specified by industries

other

Pharmaceuticals & cosmetics

Food

Rubber and plastics

Metal

Electrical/electronics

Automotive

2014 2013 2012 2011

0

200

400

600

800

1,000

1,200

1,400

1,600

Japan Germany United States Republic ofKorea

robo

ts p

er 10

,000 e

mpl

oyee

s

Capital G


131

Appendix E: Automation company overview Chart 121: Europe – main automation players – overview


Local FX, mn Main players FY15 numbers Schneider

Group sales 26,640Group Op profit 3,641Margin 13.7%

Relevant divisions Discrete Auto. & Process Automation Low Voltage Software Industrial Automation Industry Process Industries and Digital FactorySales 9127 6374 6547 260 948 5696 9894 9956as a % of group sales 26% 18% 18% 13.8% 50.5% 21% 13% 13%Profit 991 593 909 62 72 975 536 1738Margin 10.9% 9.3% 13.9% 23.8% 7.6% 17.1% 5.4% 17.5%

Automation products 25% LV drives, 15% power electronics and MV drives, 45% motors and generators, 15% robotics

10% Marine, 20% Metals and minerals, 25% Oil & gas, 15% Measurement & control tech, 5% Pulp & paper, 15% Turbocharging, 10% Perfomrnace services

30% LV breakers, 15% control products, 25% DIN & rail products, 20% LV systems, 10% wiring accessories

Design, sim, optimisation software (28%), trading, planning, scheduling(2%), op mgmt (32%), HMI and supervisory (38%)

Systems (73% - safety and control sys for process mkts), Nuclear (12% - nuclear systems), field devices (measurement and instrumentation)

40% motion, drives and contactors, 33% HMI, PLC's and sensors, 11% CST, 16% solutions

Manufactruing and process automation, product lifecycle mgmt

Motors, drives, numercial control, converters

Process vs Discrete Process 7% Discrete, 93% process

Process (3% discrete) Discrete Process 30%

Geo exposure 38% Europe, 27% Asia, 32% Americas, 3% Middle East & Africa. 35% Emerging markets

39% Europe, 27% Asia, 22% Americas, 12% Middle East and Africa (45% emerging markets)

56% Europe, 28% Asia, 9% Americas, 7% Mid East & Africa, 50% EM

27% Euro, 43% North America, 6% South America, 18% Asia

20% Euro, 29% North America, 9% LATAm, 27% Asia, 15% Africa/Mid east

37% EM, 63% mature 67% EMEA, 18% Americas, 15% Asia

64% Europe, 13% Americas, 23% Asia

End mkts 20% Discrete mfg, 10% Power gen, 5% Buildings, 10% OGP, 25% Minerals/metals, 10% Transport, 5% Water, 15% Other

35% Chemical, oil and gas, 25% Metals and mining, 15% Marine, 10% Pulp and paper, 5% Power generation, 10% Other

25% Residential, 40% ind, 30% construction, 5% utility

39% Gen ind, 24% oila nd gas, 7% discrete, 6% utilities, 6% food and bev, 4% petrochemc, 14% other

38% Oil and gas, 23% gen ind, 19% utilities & power, 6% petrochem, 3% discrete, 11% other

32% Gen industrial, 22% auto, 27% light processind, 9% electronics, 10% commerical and ind buildings

30% Process, 70% manufacturig (incl PLM). No end mkt more than 10% of sales.

40% manufacturing industries, 30% process industries, 30% energy/ infra-structure (wind + rail)

Co defined market size and position

No 1 or 2 in all mkts, est mkt size $125bn

$120bn market no1 in Oil and gas, no 2 in power and P&P, no 4 in chem and pharma

$70bn market £4.6bn market £16.1bn process control and safety systems, £7.5bn field devices and valves

#1 PLC, #1 ind switching technology, #2 industry software

$63bn mkt, no1 in inverters and controls, no1 in motors and generators, no2 in gears and couplings

Route to market 20% distributors, 45% OEM's/system integrators, 5% EPC, 20% direct, 10% ABB internal

55% direct, 30% OEM's/sys integrators, 15% EPC's

50% distributers, 10% OE's, 15% panel builders, 10% contractors, 5% end users, 10% ABB internal

64% direct, 36% partner (distributor, integrator)

85% direct 15% indirect (field devices)

Brands System 800xA (DCS) Simsci, Wonderware, Avantis

Safety - Triconex, DCS - Foxboro

PLC - Modicon PLC + DCS - Simatic

Type of sales 85% Products, 15% service

35% Systems, 40% Service, 25% Products

80% products, 5% service, 15% systems

51% license, 27% services, 22% support

81% products, 19% solutions

80% product flow, 20% product, service and solutions

10mn installed base of large motors and inverters

Installed base 200k installed base Triconex (£3bn installed base, 13k systems), Foxboro (top 3, £9bn DSC installed base)

SIMATIC PCS 7 (DCS)

ABB

35,4813,0498.6%

Invensys

1,878160

8.5%

Siemens

75,6367,2189.5%

132 Capital G


Chart 122: Europe – niche automation players – overview


Local FX, mn Niche players FY15 numbers Hexagon IMI Legrand Metso Rotork Spirax

Group sales 3,044 1,557 4,810 2,977 546 667Group Op profit 656 239 930 555 125 152Margin 21.6% 15.4% 19.3% 19% 22.9% 22.8%

Relevant divisions Indust Enter Solution Fluid Power Robotics Systems Industrial related Flow control Controls In-line Industrial Controls SteamSales 1537 662 910 1472 337 778 287 255 219 515as a % of group sales 50% 43% 31% 50% 7% 26% 52% 21.4% 18.4% 77.1%Profit 389 118 100 115 110 85 37 35 115Margin 25.3% 17.8% 11.0% 7.8% 14.2% 30% 14.4% 16.1% 22.3%

Automation products c.50% Geosystems, 50% Metrology - measurement of components - split stationary and portable products

Pneumatics solutions to control air, water, oil and other fluids. 45% sector related (largely outside factory automation), 30% aftermarket,

Manufacturer of robots for any sector and application

Robot based automation solutions for welding, gluing, sealing, converting, assembling, testing, material handling

Low voltage products and systems - control and command, cable mgmt, energy distribution, VDI - 7% industrial sales

Automation systems, control valves, measurement and control - energy, pulp and paper industries. 33% Flow control, 23% process automation,

Electric valve actuators - focus on oil+gas and power markets inpart process automation, but also specific process automation mkts

Process analytical measurement, asset monitoring and inline controls for primary processing and converting industries

Automation and control products for discrete manufacturing industries

Products and services to control the efficent use of steam. Many products are mechanical based - but moving towards electrical and electro-

Process vs Discrete Process Process Process Discrete

Geo exposure Metrology -35% Europe, 20% Americas, 35% Asia, 10% LATAM

Division - 58% Europe, 23% North Am, 15% Asia, 4% RoW

Group - 24% France, 13% Italy, 18% Rest of Europe, 17% USA, 28% RoW

Division - 25% Europe, 19% Americas, 47% Asia, 9% other

35% Europe, 26% North America, 33% Asia. 6% RoW

67% North America, 20% Europe, 11% Asia, 2% ROW

Group - 44% Europe, 29% Americas, 27% Asia

End mkts Metrology - 9% autom, 7% aero, 9% manufacturing, 1% heavy vehicles, 12% other

Sector sales - 16% trucks, 7% life sciences, 6% food & bev, 4% energy, 2% rail

48% residential construction, 45% commerical. 7% industrial - factory automation related.

41% pulp and paper, 49% energy, oil and gas, 10% power.

41% oil and gsa, 18% water, 30% power, 8% industrial, 3% other

38% pulp and paper, 25% converting, 15% energy, 17% machine OEM's, 5% machine builders

12% foods, 3% brewing & distilling, 11% HVAC, 11% OEM machinery, 9% refining & petrochem, 8%


Metrology SEK37bn, 15% share, no 1

2/3 of sales generated from areas where no 1 or 2 positions

Process automation - no3 in P&P control, no 9 in power. Control valves - P&P No 1, oil & gas no 4, refining no 4, chemical no 6

No 1, 18% share No 1 in Steam Service

Route to market Sector and aftermarket direct. Factory automation through distributers

Distributors Direct or through partners (EPC or valve makers)

Mostly direct Mostly direct Mostly direct

Brands Different brands in segments

Norgren Legrand, Bticino, local and specialist product brands

Metso Rotork Beta Laser Mike, Bruel & kjaer, BTG, NDC infrared, Servomex

Red Lion, NTRon, Sixnet, Omega< Microscan

Spirax

Type of sales Mostly products Products Products and software

Products Products Products

In 80% of its markets, Spectris is no 1 or no 2

Spectris

34% Germany, 23% Rest of Europe, 25% North America, 18% Asia/other

46% automotive 37% general industrial, 17% service

28% share in auto, 6% gen ind

1,190181

15.2%

Kuka

2,966136

4.6%

Capital G


133

Chart 123: US – automation players – overview


local FX, mnFY15 numbers Honeywell GE Eaton Parker Hannifin

Group sales 38,581 108,796 20,855 12,712Group Op profit 7,256 17,966 3,133 1,838Group Margin 18.8% 16.5% 15.0% 14.5%

Relevant divisions PMT - Process Solutions - (24%

div)

Architecture & software

Control Products & Solutions

Energy Management

Process management Industrial Automation Electrical Systems and Services

Industrial Technologies Test & measurement Industrial North America and RoW

Automation sales 2,749 2,750 3,558 7,600 8,516 4,121 1,460 10,457 (auto rel 3660)as a % of group sales 7% 44% 56% 7% 38% 18% 7% 29%Automation profit NA 809 552 270 1,493 600Margin NA 29.4% 15.5% 3.6% 17.5% 14.6% 23.0% 17.7% div 14.7%

6688.9Automation products Control systems

(est 45%), field equipment (10%), advanced solutions (10%) and lifecycle solutions (35%)

Controllers, Visualisation and software, industrial networrks, motion, safety and sensing

Ind control products, motor control & drives, motor control centers, enginereed sys and sol, services

Sensors & measurement devices, ind. control and interconnect products, motors, services

45% Measurement devices, 25% Systems and solutions, 30% Final controls

21% Power gen, 34% Motors and drives, 12% Mech power transmission, 16% Fluid automation, 10% Power distribution, 7% Ind equip

Of Electrical Products - 19% Ind, 12% machine builders of division. motor control centres, variable drives, HMI, control, connectivity, instrumentation. Rest of electrical IT, resi, non resi, utility.

Motion control (25%), Product identification (25%)

Ind automation relevant - Fluke $1bn of sales

Ests splits of Industrial - 35% industreial auto, 13% seals, 3% instrumentation, 36% hydraulics, 13% filtration

Process vs Discrete Process focus (>30k installations, >8k sites)

Process focus

Geo exposure 20% US, 10% LATAM & Canada, 37% Europe, 9% Mid East & Africa, 24% Asia

39% US, 24% Asia, 20% Europe, 9% Mid east, 8% LATAM

41% North Am, 16% Asia. 36% Europe, 4% Mid east, 3% LATAM

Electrical products - 59% US, 21% intl developed, 20% EM

Motion - North America 50%, EU 40%, RoW 10%

Fluke - 45% NA, 25% EU, 30% RoW

Group - 60% north America, 28% EMEA, 8% Asia, 4% LATAM

End mkts 29% oil and gas, 17% chemcials, pulp and paper 14%, metals and mining 6%, power 6%, pharma 2%

Energy sector focus

40% Oil and gas, 21% Other, 16% Chemicals, 14% Power and utilities, 9% Refining

24% Com & Ind, 13% HVAC, 12% power, energy & utilities, 8% building and construction, 6% oil and gas, 5% auto, 32% other

Motion - machine, equipment producers, medical, A&D, mobile offhighway, clean energy

Industrial, process and biomed technicians. Metrology/calibration, electronics engineers

Group - 27% indusrtrial, 35% mobile, 12% energy, 12% aero, 6% defence, 4% air con, 4% life sciences


# 2 in oil & gas, #1 in refining & pulp and paper, #4/5 in power gen, #3/4 chemicals

$80bn ($35bn process)

Motion $14bn, Product ID $5bn

Div - market size $18bn (Ind auto relevant $4bn)

Motion control $120bn market, 20% market share

Route to market Distributor main channel (sys integrator, OEM's and end users other channels)

Motion - 85% direct, 15% distribution

50% OE, 50% aftermarket

Brands DCS - Experion, TDC2000/3000. Safety - Safety Manager

PLC - Allen Bradley, DCS - PlantPAX

DCS - Proficy, PAC8000, PLC - PACSystems

DeltaV (DCS) PLC - PS4 Kollmorgen, Thomson, Portescap

Type of sales 15% products, 85% systems

75% instruments, 25% consumables/service

70% instruments, 30% service + software

25% Systems75% products, 25% systems

49% US, 7% Canada, 21% EMEA, 15% Asia, 8% LATAM (22% EM)

65% products, 25% solutions,10% services

Rockwell

6,3081,36121.6%

85% discrete. 15% process Process focus

Emerson

22,3043,42515.4%

62% heavy (oil and gas, metals, mining, P&P, semi, water), 28% consumer (food & bev, home & personal care, life sciences), 15% transportation (auto, tire, off-road), 5% other

1,4077%

Danaher

20,5633,46916.9%

134 Capital G


Chart 124: Japan – automation players overview


Local FX, mnFY15 numbers (FY16 for Japan) Mitsubishi Electric Omron SMC Azbil Group

Group sales 4,394 834 476 257Group Op profit 301 62 134 17Margin 6.9% 7.5% 28.2% 6.7%

Relevant divisions Factory automation Robots Industrial Automation Systems

Industrial automation business

(IAB)

IA and Control Test & Measurement Robots Motion control Advanced automation (AA)

Automation sales 170 188 1,322 336 476 367 23 154 188 94as a % of group sales 27% 30% 30% 40% 100% 89% 6% 37% 46% 36%Automation profit 159 48 134 37 2 15 22 5Margin est 50% est 20% 12.0% 14.3% 28.2% 10.0% 10.2% 10% 12% 5.3%

Automation products CNC systems, lasers, servo motors

Robots 44% FA/56% Auto parts. Factory automation - 80% Equipment (PLC, Servo Motor, Breaker), 20% Mechatronincs (CNC, Laser cutting machines, Electric discharge machines)

Sensors (vision, photoelectric, proxmity), Control equipment (temp controllers, PLC's, power supply units), Display (HMI), Motion Control (servo motors + drives, inverters, position control)

Pneumatic equipment - pressure and flow sensors, valves, pumps, filters, remp control, vacuum equip, electric actuators, instrumentation

Software, DCS, PLC, hybrid control, SCADA, controllers, field instruments

Meters, basic measuring instruments, optical and portable measureing and test equipment

Robots - Arc and spot welding, painting, handling, semicon robots

AC servo drives and controllers, general AC drives, linear motors and controllers, high speed motors, power conditioning, motor drive systems

Industrial control systems - DCS, fieldbus. Ind automation products - transmitters (pressure, temp, level), flowmeters, controllers, control valves. Control products - sensors, switches, controllers

Process vs Discrete Discrete Discrete Discrete bias Process Discrete ProcessGeo exposure Japan 38%, US 9%,

Europe 8%, Asia 44%

Japan 13%, US 46%, Europe 23%, Asia 17%

Ests - Japan 40%, China 25%, Taiwan 10%, Others 25%

Group - Japan 46%, US 11%, Europe 20%, Asia 23%

19% North Am, 22% Europe, 55% Asia

Japan 28%, US 22%, Europe 18%, Asia30%

Japan 42%, US 18%, Europe 8%, Asia 29%

Group - 49% Asia, 38% China, 7% NA, 4% Europe

End mkts Mainly Machine tools Auto 50% Semi/LCD 35%, Auto 15%, Others 50%

Semi/LCD 20%, auto 20%, machinery 15%, others 45%

DCS - petrochem, refining, oil and gas, chems

Auto 70%, Semi/LCD 10%, others 20%

Semi's, LCD, electronics, machine tools, injection moulding, metal forming

Mkt position 60% share for CNC system

15-20% share for robots

16% for AC servo-motors, 21% for PLC, 60% for Laser

40% of control and sensing devices in Japan

Japan - 63% mkt shr, N Am 19%, 17% Europe, 44%

15-20% share for robots

18% share for AC servo-motors, 16% share for invertors

Route to market Direct Direct Mainly distributors Mainly distributors DirectBrands Fanuc Fanuc PLC - Q, L, F ,

Alpha seriesPLC - Zen, CPM DCS - Centum,

Safety - Prosafe SLSYaskawa Yaskawa (Previously called

Yamatake). HarmonasDEO (DCS)

Fanuc Yasakawa

Group - 41% Japan, 26% Asia, 8% Europe, 6% N Am, 8% Mid East

41137

8.9%

Yokogawa

41440

9.6%

623216

34.6%

Capital G


135

Chart 125: RoW – automation players overview


Local FX, mn RoWAirtac (Taiwan) Hiwin (Taiwan) WEG (Brazil)

Group sales 8,797 15,087 9,760Group Op profit 2,073 2,848 1,158Margin 24% 18.9% 11.9%

Relevant divisions Motor Systems automation

Automation sales 24,600 5,400as a % of group sales 50.6% 11.1%Automation profitMargin

Automation products Pneumatics equipment - compressors, dryers, seperators, filters, regulators, valves and actuators

30% Ball screws, 59% linear guideways + equip, 12% Ind robot

Motors - low and high voltage motors AC/DC motor controls, PLC's, servo controllers, switches, circuit breakers, relays, inverters

Industrial motors, drives, controls and panels

Process vs Discrete Geo exposure 86% China, 8% Taiwan, 6% overseas 47% Asia, 30% Europe, 16% Taiwan,

7% US49% Brazil,

End mkts 15% machinery + machine tools, 19% elec eq, 8% packaging, 8% construction, 5% autos, 3% environ, 3% glass, 3% plastic

30% machine tools, 30% electronics, 10% automation, 10% medical, 10% energy, 10% aero

Mkt position Chinese Fluid power mkt RMB12bn, 13% share

10% in linear guidaways (behind THK and NSK)

Route to marketBrands AirTAC

Type of sales 40% replacement Installed base

Fluid machinery, haulage, petrochem, iron and steel, plastic and rubber, machine and tooling, construction and mining, power gen

50% Taiwan, 20% Americas, 10% China

Teco (Taiwan)

48,5993,7887.8%


Disclosures Important Disclosures

FUNDAMENTAL EQUITY OPINION KEY: Opinions include a Volatility Risk Rating, an Investment Rating and an Income Rating. VOLATILITY RISK RATINGS, indicators of potential price fluctuation, are: A - Low, B - Medium and C - High. INVESTMENT RATINGS reflect the analyst’s assessment of a stock’s: (i) absolute total return potential and (ii) attractiveness for investment relative to other stocks within its Coverage Cluster (defined below). There are three investment ratings: 1 - Buy stocks are expected to have a total return of at least 10% and are the most attractive stocks in the coverage cluster; 2 - Neutral stocks are expected to remain flat or increase in value and are less attractive than Buy rated stocks and 3 - Underperform stocks are the least attractive stocks in a coverage cluster. Analysts assign investment ratings considering, among other things, the 0-12 month total return expectation for a stock and the firm’s guidelines for ratings dispersions (shown in the table below). The current price objective for a stock should be referenced to better understand the total return expectation at any given time. The price objective reflects the analyst’s view of the potential price appreciation (depreciation). Investment rating Total return expectation (within 12-month period of date of initial rating) Ratings dispersion guidelines for coverage cluster*

Buy ≥ 10% ≤ 70% Neutral ≥ 0% ≤ 30%

Underperform N/A ≥ 20% * Ratings dispersions may vary from time to time where BofA Merrill Lynch Research believes it better reflects the investment prospects of stocks in a Coverage Cluster.

INCOME RATINGS, indicators of potential cash dividends, are: 7 - same/higher (dividend considered to be secure), 8 - same/lower (dividend not considered to be secure) and 9 - pays no cash dividend. Coverage Cluster is comprised of stocks covered by a single analyst or two or more analysts sharing a common industry, sector, region or other classification(s). A stock’s coverage cluster is included in the most recent BofA Merrill Lynch report referencing the stock. BofA Merrill Lynch Research Personnel (including the analyst(s) responsible for this report) receive compensation based upon, among other factors, the overall profitability of Bank of America Corporation, including profits derived from investment banking. The analyst(s) responsible for this report may also receive compensation based upon, among other factors, the overall profitability of the Bank’s sales and trading businesses relating to the class of securities or financial instruments for which such analyst is responsible. Other Important Disclosures From time to time research analysts conduct site visits of covered issuers. BofA Merrill Lynch policies prohibit research analysts from accepting payment or reimbursement for travel expenses from the issuer for such visits. Prices are indicative and for information purposes only. Except as otherwise stated in the report, for the purpose of any recommendation in relation to: (i) an equity security, the price referenced is the publicly traded price of the security as of close of business on the day prior to the date of the report or, if the report is published during intraday trading, the price referenced is indicative of the traded price as of the date and time of the report; or (ii) a debt security (including equity preferred and CDS), prices are indicative as of the date and time of the report and are from various sources including Bank of America Merrill Lynch trading desks. The date and time of completion of the production of any recommendation in this report shall be the date and time of dissemination of this report as recorded electronically on the platform in which the report was accessed. Officers of MLPF&S or one or more of its affiliates (other than research analysts) may have a financial interest in securities of the issuer(s) or in related investments. BofA Merrill Lynch Global Research policies relating to conflicts of interest are described at http://go.bofa.com/coi. "BofA Merrill Lynch" includes Merrill Lynch, Pierce, Fenner & Smith Incorporated ("MLPF&S") and its affiliates. Investors should contact their BofA Merrill Lynch representative or Merrill Lynch Global Wealth Management financial advisor if they have questions concerning this report. "BofA Merrill Lynch" and "Merrill Lynch" are each global brands for BofA Merrill Lynch Global Research. Information relating to Non-US affiliates of BofA Merrill Lynch and Distribution of Affiliate Research Reports: MLPF&S distributes, or may in the future distribute, research reports of the following non-US affiliates in the US (short name: legal name, regulator): Merrill Lynch (South Africa): Merrill Lynch South Africa (Pty) Ltd., regulated by The Financial Service Board; MLI (UK): Merrill Lynch International, regulated by the Financial Conduct Authority (FCA) and the Prudential Regulation Authority (PRA); Merrill Lynch (Australia): Merrill Lynch Equities (Australia) Limited, regulated by the Australian Securities and Investments Commission; Merrill Lynch (Hong Kong): Merrill Lynch (Asia Pacific) Limited, regulated by the Hong Kong Securities and Futures Commission (HKSFC); Merrill Lynch (Singapore): Merrill Lynch (Singapore) Pte Ltd, regulated by the Monetary Authority of Singapore (MAS); Merrill Lynch (Canada): Merrill Lynch Canada Inc, regulated by the Investment Industry Regulatory Organization of Canada; Merrill Lynch (Mexico): Merrill Lynch Mexico, SA de CV, Casa de Bolsa, regulated by the Comisión Nacional Bancaria y de Valores; Merrill Lynch (Argentina): Merrill Lynch Argentina SA, regulated by Comisión Nacional de Valores; Merrill Lynch (Japan): Merrill Lynch Japan Securities Co., Ltd., regulated by the Financial Services Agency; Merrill Lynch (Seoul): Merrill Lynch International Incorporated (Seoul Branch) regulated by the Financial Supervisory Service; Merrill Lynch (Taiwan): Merrill Lynch Securities (Taiwan) Ltd., regulated by the Securities and Futures Bureau; DSP Merrill Lynch (India): DSP Merrill Lynch Limited, regulated by the Securities and Exchange Board of India; PT Merrill Lynch (Indonesia): PT Merrill Lynch Indonesia, regulated by Otoritas Jasa Keuangan (OJK); Merrill Lynch (Israel): Merrill Lynch Israel Limited, regulated by Israel Securities Authority; Merrill Lynch (Russia): OOO Merrill Lynch Securities, Moscow, regulated by the Central Bank of the Russian Federation; Merrill Lynch (DIFC): Merrill Lynch International (DIFC Branch), regulated by the Dubai Financial Services Authority (DFSA); Merrill Lynch (Spain): Merrill Lynch Capital Markets Espana, S.A.S.V., regulated by Comisión Nacional del Mercado De Valores; Merrill Lynch (Brazil): Bank of America Merrill Lynch Banco Multiplo S.A., regulated by Comissão de Valores Mobiliários; Merrill Lynch KSA Company, Merrill Lynch Kingdom of Saudi Arabia Company, regulated by the Capital Market Authority. This research report: has been approved for publication and is distributed in the United Kingdom (UK) to professional clients and eligible counterparties (as each is defined in the rules of the FCA and the PRA) by MLI (UK) and Bank of America Merrill Lynch International Limited, which are authorized by the PRA and regulated by the FCA and the PRA, and is distributed in the UK to retail clients (as defined in the rules of the FCA and the PRA) by Merrill Lynch International Bank Limited, London Branch, which is authorized by the Central Bank of Ireland and subject to limited regulation by the FCA and PRA - details about the extent of our regulation by the FCA and PRA are available from us on request; has been considered and distributed in Japan by Merrill Lynch (Japan), a registered securities dealer under the Financial Instruments and Exchange Act in Japan; is issued and distributed in Hong Kong by Merrill Lynch (Hong Kong) which is regulated by HKSFC (research reports containing any information in relation to, or advice on, futures contracts are not intended for issuance or distribution in Hong Kong and are not directed to, or intended for issuance or distribution to, or use by, any person in Hong Kong); is issued and distributed in Taiwan by Merrill Lynch (Taiwan); is issued and distributed in India by DSP Merrill Lynch (India); and is issued and distributed in Singapore to institutional investors and/or accredited investors (each as defined under the Financial Advisers Regulations) by Merrill Lynch International Bank Limited (Merchant Bank) (MLIBLMB) and Merrill Lynch (Singapore) (Company Registration Nos F 06872E and 198602883D respectively). MLIBLMB and Merrill Lynch (Singapore) are regulated by MAS. Bank of America N.A., Australian Branch (ARBN 064 874 531), AFS License 412901 (BANA Australia) and Merrill Lynch Equities (Australia) Limited (ABN 65 006 276 795), AFS License 235132 (MLEA) distribute this report in Australia only to 'Wholesale' clients as defined by s.761G of the Corporations Act 2001. With the exception of BANA Australia, neither MLEA nor any of its affiliates involved in preparing this research report is an Authorised Deposit-Taking Institution under the Banking Act 1959 nor regulated by the Australian Prudential Regulation Authority. No approval is required for publication or distribution of this report in Brazil and its local distribution is by Merrill Lynch (Brazil) in accordance with applicable regulations. Merrill Lynch (DIFC) is authorized and regulated by the DFSA. Research reports prepared and issued by Merrill Lynch (DIFC) are done so in accordance with the requirements of the DFSA conduct of business rules. Bank of America Merrill Lynch International Limited, Frankfurt Branch (BAMLI Frankfurt) distributes this report in Germany and is regulated by BaFin. This research report has been prepared and issued by MLPF&S and/or one or more of its non-US affiliates. MLPF&S is the distributor of this research report in the US and accepts full responsibility for research reports of its non-US affiliates distributed to MLPF&S clients in the US. Any US person receiving this research report and wishing to effect any transaction in any security discussed in the


report should do so through MLPF&S and not such foreign affiliates. Hong Kong recipients of this research report should contact Merrill Lynch (Asia Pacific) Limited in respect of any matters relating to dealing in securities (and not futures contracts) or provision of specific advice on securities (and not futures contracts). Singapore recipients of this research report should contact Merrill Lynch International Bank Limited (Merchant Bank) and/or Merrill Lynch (Singapore) Pte Ltd in respect of any matters arising from, or in connection with, this research report. General Investment Related Disclosures: Taiwan Readers: Neither the information nor any opinion expressed herein constitutes an offer or a solicitation of an offer to transact in any securities or other financial instrument. No part of this report may be used or reproduced or quoted in any manner whatsoever in Taiwan by the press or any other person without the express written consent of BofA Merrill Lynch. This research report provides general information only. Neither the information nor any opinion expressed constitutes an offer or an invitation to make an offer, to buy or sell any securities or other financial instrument or any derivative related to such securities or instruments (e.g., options, futures, warrants, and contracts for differences). This report is not intended to provide personal investment advice and it does not take into account the specific investment objectives, financial situation and the particular needs of any specific person. Investors should seek financial advice regarding the appropriateness of investing in financial instruments and implementing investment strategies discussed or recommended in this report and should understand that statements regarding future prospects may not be realized. Any decision to purchase or subscribe for securities in any offering must be based solely on existing public information on such security or the information in the prospectus or other offering document issued in connection with such offering, and not on this report. Securities and other financial instruments discussed in this report, or recommended, offered or sold by Merrill Lynch, are not insured by the Federal Deposit Insurance Corporation and are not deposits or other obligations of any insured depository institution (including, Bank of America, N.A.). Investments in general and, derivatives, in particular, involve numerous risks, including, among others, market risk, counterparty default risk and liquidity risk. No security, financial instrument or derivative is suitable for all investors. In some cases, securities and other financial instruments may be difficult to value or sell and reliable information about the value or risks related to the security or financial instrument may be difficult to obtain. Investors should note that income from such securities and other financial instruments, if any, may fluctuate and that price or value of such securities and instruments may rise or fall and, in some cases, investors may lose their entire principal investment. Past performance is not necessarily a guide to future performance. Levels and basis for taxation may change. This report may contain a short-term trading idea or recommendation, which highlights a specific near-term catalyst or event impacting the issuer or the market that is anticipated to have a short-term price impact on the equity securities of the issuer. Short-term trading ideas and recommendations are different from and do not affect a stock's fundamental equity rating, which reflects both a longer term total return expectation and attractiveness for investment relative to other stocks within its Coverage Cluster. Short-term trading ideas and recommendations may be more or less positive than a stock's fundamental equity rating. BofA Merrill Lynch is aware that the implementation of the ideas expressed in this report may depend upon an investor's ability to "short" securities or other financial instruments and that such action may be limited by regulations prohibiting or restricting "shortselling" in many jurisdictions. Investors are urged to seek advice regarding the applicability of such regulations prior to executing any short idea contained in this report. Foreign currency rates of exchange may adversely affect the value, price or income of any security or financial instrument mentioned in this report. Investors in such securities and instruments, including ADRs, effectively assume currency risk. UK Readers: The protections provided by the U.K. regulatory regime, including the Financial Services Scheme, do not apply in general to business coordinated by BofA Merrill Lynch entities located outside of the United Kingdom. BofA Merrill Lynch Global Research policies relating to conflicts of interest are described at http://go.bofa.com/coi. MLPF&S or one of its affiliates is a regular issuer of traded financial instruments linked to securities that may have been recommended in this report. MLPF&S or one of its affiliates may, at any time, hold a trading position (long or short) in the securities and financial instruments discussed in this report. BofA Merrill Lynch, through business units other than BofA Merrill Lynch Global Research, may have issued and may in the future issue trading ideas or recommendations that are inconsistent with, and reach different conclusions from, the information presented in this report. Such ideas or recommendations reflect the different time frames, assumptions, views and analytical methods of the persons who prepared them, and BofA Merrill Lynch is under no obligation to ensure that such other trading ideas or recommendations are brought to the attention of any recipient of this report. In the event that the recipient received this report pursuant to a contract between the recipient and MLPF&S for the provision of research services for a separate fee, and in connection therewith MLPF&S may be deemed to be acting as an investment adviser, such status relates, if at all, solely to the person with whom MLPF&S has contracted directly and does not extend beyond the delivery of this report (unless otherwise agreed specifically in writing by MLPF&S). MLPF&S is and continues to act solely as a broker-dealer in connection with the execution of any transactions, including transactions in any securities mentioned in this report. Copyright and General Information regarding Research Reports: Copyright 2016 Bank of America Corporation. All rights reserved. iQmethod, iQmethod 2.0, iQprofile, iQtoolkit, iQworks are service marks of Bank of America Corporation. iQanalytics®, iQcustom®, iQdatabase® are registered service marks of Bank of America Corporation. This research report is prepared for the use of BofA Merrill Lynch clients and may not be redistributed, retransmitted or disclosed, in whole or in part, or in any form or manner, without the express written consent of BofA Merrill Lynch. BofA Merrill Lynch Global Research reports are distributed simultaneously to internal and client websites and other portals by BofA Merrill Lynch and are not publicly-available materials. Any unauthorized use or disclosure is prohibited. Receipt and review of this research report constitutes your agreement not to redistribute, retransmit, or disclose to others the contents, opinions, conclusion, or information contained in this report (including any investment recommendations, estimates or price targets) without first obtaining expressed permission from an authorized officer of BofA Merrill Lynch. Materials prepared by BofA Merrill Lynch Global Research personnel are based on public information. Facts and views presented in this material have not been reviewed by, and may not reflect information known to, professionals in other business areas of BofA Merrill Lynch, including investment banking personnel. BofA Merrill Lynch has established information barriers between BofA Merrill Lynch Global Research and certain business groups. As a result, BofA Merrill Lynch does not disclose certain client relationships with, or compensation received from, such issuers in research reports. To the extent this report discusses any legal proceeding or issues, it has not been prepared as nor is it intended to express any legal conclusion, opinion or advice. Investors should consult their own legal advisers as to issues of law relating to the subject matter of this report. BofA Merrill Lynch Global Research personnel’s knowledge of legal proceedings in which any BofA Merrill Lynch entity and/or its directors, officers and employees may be plaintiffs, defendants, co-defendants or co-plaintiffs with or involving issuers mentioned in this report is based on public information. Facts and views presented in this material that relate to any such proceedings have not been reviewed by, discussed with, and may not reflect information known to, professionals in other business areas of BofA Merrill Lynch in connection with the legal proceedings or matters relevant to such proceedings. This report has been prepared independently of any issuer of securities mentioned herein and not in connection with any proposed offering of securities or as agent of any issuer of any securities. None of MLPF&S, any of its affiliates or their research analysts has any authority whatsoever to make any representation or warranty on behalf of the issuer(s). BofA Merrill Lynch Global Research policy prohibits research personnel from disclosing a recommendation, investment rating, or investment thesis for review by an issuer prior to the publication of a research report containing such rating, recommendation or investment thesis. Any information relating to the tax status of financial instruments discussed herein is not intended to provide tax advice or to be used by anyone to provide tax advice. Investors are urged to seek tax advice based on their particular circumstances from an independent tax professional. The information herein (other than disclosure information relating to BofA Merrill Lynch and its affiliates) was obtained from various sources and we do not guarantee its accuracy. This report may contain links to third-party websites. BofA Merrill Lynch is not responsible for the content of any third-party website or any linked content contained in a third-party website. Content contained on such third-party websites is not part of this report and is not incorporated by reference into this report. The inclusion of a link in this report does not imply any endorsement by or any affiliation with BofA Merrill Lynch. Access to any third-party website is at your own risk, and you should always review the terms and privacy policies at third-party websites before submitting any personal information to them. BofA Merrill Lynch is not responsible for such terms and privacy policies and expressly disclaims any liability for them. Subject to the quiet period applicable under laws of the various jurisdictions in which we distribute research reports and other legal and BofA Merrill Lynch policy-related restrictions on the publication of research reports, fundamental equity reports are produced on a regular basis as necessary to keep the investment recommendation current. Certain outstanding reports may contain discussions and/or investment opinions relating to securities, financial instruments and/or issuers that are no longer current. Always refer to the most recent research report relating to an issuer prior to making an investment decision. In some cases, an issuer may be classified as Restricted or may be Under Review or Extended Review. In each case, investors should consider any investment opinion relating to such issuer (or its security and/or financial instruments) to be suspended or withdrawn and should not rely on the analyses and investment opinion(s) pertaining to such issuer (or its securities and/or financial instruments) nor should the analyses or opinion(s) be considered a solicitation of any kind. Sales persons and financial advisors affiliated with MLPF&S or any of its affiliates may not solicit purchases of securities or financial instruments that are Restricted or Under Review and may only solicit securities under Extended Review in accordance with firm policies. Neither BofA Merrill Lynch nor any officer or employee of BofA Merrill Lynch accepts any liability whatsoever for any direct, indirect or consequential damages or losses arising from any use of this report or its contents.

Documents

Capital Goods Global Primer: Digital Machinations