Disruptive technologies review citi july 2016

Citi GPS: Global Perspectives & Solutions

July 2016

DISRUPTIVE INNOVATIONS IVTen More Things to Stop and Think About

Citi is one of the world’s largest financial institutions, operating in all major established and emerging markets. Across these world markets, our employees conduct an ongoing multi-disciplinary global conversation – accessing information, analyzing data, developing insights, and formulating advice for our clients. As our premier thought-leadership product, Citi GPS is designed to help our clients navigate the global economy’s most demanding challenges, identify future themes and trends, and help our clients profit in a fast-changing and interconnected world. Citi GPS accesses the best elements of our global conversation and harvests the thought leadership of a wide range of senior professionals across our firm. This is not a research report and does not constitute advice on investments or a solicitation to buy or sell any financial instrument. For more information on Citi GPS, please visit our website at www.citi.com/citigps.

Citi GPS: Global Perspectives & Solutions July 2016

Andrew Baum, MD Global Head of Healthcare Research +44-20-7986-4498 | [email protected]

Graeme McDonald Japaenese Machinery & Shipbuilding Analyst +81-3-6270-4732 | [email protected]

Jason B Bazinet U.S. Entertainment, Cable, & Satellite Analyst +1-212-816-6395 | [email protected]

Yigal Nochomovitz, Ph.D. U.S. Mid/Small Cap Biotechnology Analyst +1-212-816-1441 | [email protected]

Dennis Chan Asian Handset Sector Analyst +886-2-8726-9086 | [email protected]

Michael Rollins, CFA U.S. Telecom and Telecom Infrastructure Analyst +1-212-816-1116 | [email protected]

Elizabeth Curmi, Ph.D. Global Thematic Analyst +44-20-7986-6818 | [email protected]

Ashwin Shirvaikar, CFA U.S. Payments, Processors & IT Services Analyst +1-212-816-0822 | [email protected]

Mukhtar Garadaghi European Oilfield Services Analyst +44-20-7986-4417 | [email protected]

Thomas Singlehurst, CFA Head of European Media Research Team +44-20-7986-4051 | [email protected]

Amit B Harchandani Head of European Technology Research Team +44-20-7986-4246 | [email protected]

Jim Suva, CPA U.S. IT Hardware & Tech Supply Chain Analyst +415-951-1703 | [email protected]

Dan Fox Homan European General Retail Analyst +44-20-7986-4112 | [email protected]

Martin Wilkie European Capital Goods Analyst +44-207-986-4077 | [email protected]

Arthur Lai Greater China Hardware Sector Analyst +852-2501-2758 | [email protected]

Anthony Yuen, Ph.D. Global Commodities Strategist +1-212-723-1477 | [email protected]

July 2016 Citi GPS: Global Perspectives & Solutions

© 2016 Citigroup

3

DISRUPTIVE INNOVATIONS IV Ten More Things to Stop and Think About Every morning my inbox is filled with a mixture of emails – request emails that

require me to take action and informational emails to keep me up to date about

what’s new in the world. My strategy with request emails is the same as my strategy

playing tennis – when one comes into my court the best thing to do is to reply

quickly and ‘get it back over the net’ into the senders court. But the information

emails are different. Sometimes they’re read right away, sometimes they’re

skimmed and deleted and sometimes they’re read and filed. Whatever happens to

those emails, the information in them gets tucked somewhere in my brain.

One thing that’s different about my daily inbox over the past four years is that the

informational emails are more and more filled with stories about disruptive

innovation and even more excitingly, about the disruptive innovations that we’ve

highlighted in past Citi GPS reports. Whether it be 3D printing, virtual reality gaming,

immunotherapy or digital banking, disruption and accelerating technological change

is increasingly part of our everyday conversation.

So where is all of this innovation happening? With innovation clusters scattered

throughout the world, in places like Silicon Valley, Boston, London, Stockholm, and

China we wanted to see if there was a recipe for success that other regions could

follow. In the first chapter of the report, we look at the factors affecting different

regions and find that the majority of the successful clusters thrive in an open, global

environment where knowledge sharing is at the core of their activities.

In our new report, we once again look at some of the newest concepts across

sectors and identify new products which could potentially disrupt the marketplace.

Some of the concepts are familiar, i.e. big data, robotics and on-line retailing, but

the application of these concepts is very different than in the past. Big data is

playing a significant role in the energy sector and being used across the entire

energy value-chain, from finding energy faster and producing it more inexpensively

in the upstream, to more customized energy at the downstream user level. Robots

are not just being built and utilized in the automotive industry, but are now being

produced with open source software which will boost collaborative development and

potentially shifts the use of robots in a much wider range of end markets. And in

retail, beyond the shift from physical retail to online retail, we are now exploring

direct-to-consumer marketplace for fashion where consumers can order direct from

the manufacturer and bypass the retailer. Contextual Commerce is also changing

the nature of e-commerce by introducing content and intent to the online experience

and could lower the rate of online shopping cart abandonment.

Immunotherapy was the newest disruptive treatment for cancer treatment but now

epigenetics are being developed to increase the effectiveness of immunotherapy as

well as become another tool in a doctor’s cancer toolbox. Doctor’s will also be

helped by new delivery systems for the eye – next gen ocular drug delivery – which

increase effectiveness of eye treatments and the ease of delivery. New innovations

in smart box technology to create Home Hubs and innovations in screen materials

that will allow mobile phones to move away from their current shape and potentially

have no shape are both looked at in the pages that follow. Finally, we look at new

advances in subsea pipes that will reduce the cost of deepwater drilling and new

developments in wideband semiconductors which will increase performance with

every smaller sized devices.

Can’t wait for these to hit my inbox!

Kathleen Boyle, CFA

Managing Editor, Citi GPS

© 2016 Citigroup

There’s a big opportunity out there

Wide Bandgap SemiconductorsWide bandgap

semiconductors

let devices operate

at much higher

temperatures,

voltages, and

frequencies while

being smaller and

more reliable.

Home Networking Consumer media

devices could be

the focal point of

the “connected

home”, integrating

a variety of services

and connectivity

into one location.

The Future Look of Devices Consumer devices by

2021 could look like a

thin and flexible piece

of paper through the

use of flexible OLED

technology.

EpigeneticsEpigenetic approaches

in cancer treatment

could become a $10bn

market by 2025.

Energy: The Big Data RevolutionBig data analytics

would make producing

oil/gas faster and

cheaper, renewables

forecasting more

accurate, and the

transport-generation-

storage model more

integrated.

Open-Source RoboticsThe use of open-

source software

in robots can

accelerate robot

penetration by

lowering customer

adoption cost.

Contextual CommerceIncreasingly, online

purchases will be

suggested and

transacted through

non-traditional

e-commerce sites

such as social media.

Direct-to-Consumer MarketplaceMoving from proximity-

sourced product to a

direct-to-consumer

marketplace would

create a $200bn annual

revenue opportunity for

apparel manufacturers.

Thermoplastic Subsea PipesSwitching from

traditional steel pipes

to new thermoplastic

pipes decreases subsea

costs by 30-40% and

total deepwater costs

by 10%, enough to

lower the breakeven

oil price by $4/bbl.

Next Gen Ocular Drug DeliveryNew delivery

methods will

increase the ease

and effectiveness

of drug delivery

for the growing

number of people

with ocular disease.


© 2016 Citigroup

6

Contents Clusters of Innovation 7 Finding the Successful Recipe 7 1. Big Data Disruption 15 The Big Data Revolution in Energy 15 2. Contextual Commerce 22 What is Contextual Commerce? 22 3. Direct-to-Consumer Marketplace 29 Another Threat to Retail 29 4. Epigenetics 33 Immunotherapy’s New Best Friend 33 5. The Future Look of Devices 37 Reshape, No Shape 37 6. The Smart Box 41 Battle to Be the Heart of the Connected Home 41 7. Next Gen Ocular Delivery 49 Sustained Drug Delivery – The Future of Ocular Medicine 49 8. Open Source Robotics 55 Open Source to Drive Robotics Beyond Its Automotive Roots 55 9. Thermoplastic Subsea Pipes 61 Enabling Stranded Deepwater Assets 61 10. Wide Bandgap Semiconductors 65 Powering the Future 65


© 2016 Citigroup

7

Clusters of Innovation Finding the Successful Recipe In recent years there has been a growing interest on the importance of regional clusters in fostering the ‘next’ economy. This discussion is not new – Michael Porter in 1998 wrote an article for the Harvard Business Review in which he argued that the economic map of the world was dominated by what he called clusters: ‘critical masses- in one place- of unusual competitive success in particular fields’. Porter defined clusters as ‘geographic concentrations of interconnected companies and institutions in a particular field’.1 According to the author they include an ‘array of linked industries and other entities important to competition.’ Breekal et al. (2015) extend the definition of clusters to areas that attract the brightest talent and are seen as areas of high innovation due to their ability to ‘endogenously generate and diffuse knowledge’.2

Silicon Valley is the most famous innovation cluster; home to the semiconductor, computer software, and related electronics industries it has become one of the most innovative areas in the world, accounting for 52% of California’s patent registrations in 2013, attracting highly skilled people (47% of residents have a professional degree compared with 29% in the U.S. itself) and creating over 57,000 new jobs between 2013 and 2014.3 Boston’s Route 128 corridor is also another well-known regional cluster which has nearly 30,000 scientists directly involved not only in the high-tech industry but in biotechnology, pharmaceutical. and clinical research.4

There are many other regional clusters that are less globally well-known but are extremely important for the local and regional economy. We use two proxies to identify important regional clusters (1) the amount of venture capital/investment in the region and (2) the number of patents filed in different regions. These include the Raleigh-Durham area in North Carolina for life sciences (pharmaceuticals and bio-technology), Stockholm for information and communication technology (ICT), and London for financial technology (FinTech) amongst others. What we found is that the success of these regional clusters depend on a number of issues including: (1) the availability of adequate skills; (2) good local/regional policy that targets fields with a long comparative advantage and encourages a good open immigration policy that enables access to global talent; (3) low cost structure especially in the early stages of development such as low cost office space and rent; (4) the availability of funding (venture capital, angel investment, federal and state funding); (5) good infrastructure and good lifestyle offering; and (6) proximity to markets and geographical location.5

1 Porter M.E. (1998), Clusters and the New Economics of Competition, Harvard Business Review, November/December 1998 issue 2 Breekel T, Fornalh D, Morrison A (2015) Another cluster premium: Innovation subsidies and R&D collaboration networks, Research Policy, Vol 44, issue 8 pp 1431-1444 3 Joint Venture Silicon Valley, 2015 Silicon Valley Index, People, Economy, Society, Place, Governance 4 Growth Commission, London Stansted Cambridge Corridor, Corridors and Tech regions: International Case-studies, 5 Innovation Clusters: Understanding life cycles, A briefing paper from the Economist Intelligence Unit

Elizabeth Curmi, Ph.D. Global Thematic Analyst


© 2016 Citigroup

8

Regional clusters- important for different sectors Venture capital (VC) funding and patents can provide some information as to where innovation is occurring in the world. Venture capital investment increased by $30.4 billion in 2014 when compared to 2011 figures. The San Francisco Bay area attracted the largest VC investment reaching a total of $24.8 billion in 2014, an increase of over $10 billion when compared to 2013. The city of Beijing increased its VC investment from $ 0.9 billion in 2009 to a staggering $7.7 billion in 2014. In Europe, the UK, France, Germany, and Switzerland attracted the most VC funding with a total of $7.4 billion. Some of the regions that received the most VC funding such as the SF Bay Area, Beijing, New England (Boston-Worchester-Manchester area), New York, and Germany (Munich and Stuttgart) also filed the most patents6 in 2013 (Figure 2). However, this relationship didn’t hold true across the board. Tokyo and the Shenzhen-Guangdong region in China filed the most patents in 2013 however they did not receive much VC funding. Two potential reasons for this are that it was well-established companies filing patents in these areas which do not necessarily attract VC funding or that patents are being filed in main cities rather than the region in which the innovation occurs.

Figure 1. VC Investment in Different Regions (US$ Billion) Figure 2. Top 20 Regions for Patent Filing in 2013 Area 2009 2011 2013 2014 Bay Area (US) 8.8 14.2 13.9 24.8 Beijing (CHN) 0.9 2.9 2.3 7.7 NY Metro (US) 1.8 3.3 3.6 5.3 New England (US) 3.3 3.9 3.8 4.8 S. California (US) 2.3 4.0 3.3 4.1 Germany 1.0 0.8 2.0 2.9 UK 1.5 2.0 2.4 2.7 Bengaluru (IND) 0.1 0.3 0.6 2.2 Shanghai (CHN) 0.4 1.2 0.8 2.1 Israel 0.9 2.0 1.8 1.9 Canada 0.6 1.2 1.1 1.4 France 1.0 1.3 1.2 1.3 Illinois (US) 0.4 0.8 0.5 0.9 Potomac (US) 0.7 1.6 1.2 0.8 Switzerland 0.4 0.4 0.4 0.5

Source: Ernst & Young7, Citi Research Source: Ernst & Young, Citi Research

Some of these regions are becoming specifically important for innovation in certain sectors, similar to the way Silicon Valley has become extremely important for the tech industry. Successful regional clusters that include a network of similar companies are forming in many different areas, sometimes crossing over different state and national boundaries. We use the above proxies to highlight a number of important regional clusters for three particular sectors – FinTech, Life Sciences and ICT.

Clusters Important for FinTech – A Case Study of London Since the financial crisis, there has been a rise in start-ups concentrating on financial technology. Like other disrupters from Silicon Valley (AirBnB, Uber etc.), FinTech companies are growing extremely fast. California, the U.K., and New York are the three largest markets for start-up FinTech companies (Figure 3) and the

6 These patents are filed under the Patent Cooperation Treaty (PCT) which assists applicants in seeking patent protection internationally for their inventions. 7 Ernst & Young, Venture Capital Insights – 4Q14, Global Investment Landscapes, January 2015.

0 4000 8000 12000 16000

Tokyo (JPN)Shenzhen - Guangdong (CHN)

San Jose-San Francisco-Oakland (US)Beijing (CHN)

Kanagawa (JPN)Boston-Worcester-Manchester (US)

Seoul (KOR)San Diego-Carlsbad-San Marcos (US)

Gyeonggi-do (KOR) New York-Newark-Bridgeport (US)

Aichi (JPN)Los Angeles-Long Beach-Riverside (US)

Houston-Baytown-Huntsville (US)Osaka (JPN)

Seattle-Tacoma-Olympia (US)Chicago-Naperville-Michigan City (US)

Kyoto (JPN)Minneapolis-St. Paul-St. Cloud (US)

Noord-Brabant (Einhoven)- NLShizuoka (JPN)

Total No. of Patents filed in 2013


© 2016 Citigroup

9

sector is also growing extremely fast in China. We note that VC funding in the area is second to only California, higher than New York and the UK (Figure 4).

The UK has raised over $680 million in venture capital, which is low when compared to California, where over $4.7 billion was raised over the same period8. However the UK (in particular London) has produced some of the largest and most successful FinTech companies. These include companies such as Transferwise – an online international money transfer platform that has raised over $90 million in venture capital – and Funding Circle – a peer to peer lending platform that has raised over $273 million in investment. Both companies are now valued at over $1 billion.9

Figure 3. Market Size and FinTech Staff in Different Regions

Place Market size (Billion £) FinTech Staff California 4.7 74,000 UK 6.6 61,000 New York 5.6 57,000 Germany 1.8 13,000 Hong Kong 0.6 8,000 Singapore 0.6 7,000 Australia 0.7 10,000

Source: Ernst & Young10, Citi Research

Figure 4. FinTech Investment By Region (£ Millions) Figure 5. Number of FinTech Hubs By Type

Source: Ernst & Young, Citi Research Source: Ernst & Young, Citi Research

There are a number of reasons why London has become a hub for FinTech including: (1) a highly skilled work force and entrepreneurial talent pool, (2) a strong technology cluster (such as Silicon roundabout in Old Street); and (3) proximity to one most competitive financial centers in the world. Over 61,000 people are employed in the FinTech industry in the UK, with most of these positions based in 8 Ernst & Young, Commissioned by HM Treasury, UK FinTech, On the cutting edge, An evaluation of the international FinTech sector. 9 David J., Grimmelmann L., McCarthy B., Pietrella S., and Whiting G. (2016), The UK’s FinTech Cluster, Leveraging the UK’s strengths in financial services, technology and entrepreneurship. Macroeconomics of Competitiveness, Spring 2016, Harvard Business School. 10 Ernst & Young, Commissioned by HM Treasury, UK FinTech, On the cutting edge, An evaluation of the international FinTech sector.

0

500

1000

1500

2000

2500

3000

3500

4000

0

2

4

6

8

10

12Accelerators Incubators


© 2016 Citigroup

10

London (Figure 3). The UK has also the highest number of incubators and accelerators for the FinTech sector when compared to other regions (Figure 5). An example of this is Level39 at Canary Wharf which is Europe’s largest FinTech accelerator, housing more than 190 start-ups. This accelerator gives young companies access to technology investors, industry experts, and experienced entrepreneurs. It has recently expanded to two further levels, extending the desk-space for early-stage businesses.11 Although Competition with other countries in the FinTech space is growing (i.e. China); London has the potential to continue expanding in this sector and to cement its place as a world renowned center for FinTech.

There are some issues of concern for London in retaining its FinTech strength, in particular access to talent and the availability of funding. The UK is only ranked 43rd in the world for tech talent. Visa policies are also becoming more stringent for immigration of labor, and this could increase with Brexit, given that EU citizens might also be restricted easy entry.12 There is also a gap in funding when compared to other countries such as the U.S. at the VC and initial public offering levels. Other issues include high labor and living costs when compared to China or Eastern Europe. On the plus side, London has the world’s highest concentration of global financial institutions which are both competitors and customers of FinTech companies, good infrastructure, and is surrounded by world-renowned research centers that can ultimately bring the needed talent to the sector.

Life Sciences Pharma and biotech regional innovation clusters are usually based in the same geographical locations – in fact the majority of top regions that filed the highest number of patents in pharma also filed the highest number of patents in biotechnology (Figure 6 and Figure 7). As the life sciences industry shifts from ‘Big Pharma’ companies to small & medium-sized enterprises, innovation, partnership, M&A activity, and good environmental for start-up companies have become extremely important.13 Therefore it comes as no surprise that the Boston-Worchester-Manchester region in Massachusetts has filed the most patents in both biotech and pharmaceutical sector. The area is known for its leading universities (there are 114 colleges in Massachusetts, including Harvard and MIT); it also had the third largest venture capital investment in the U.S. in 2014 and the highest concentration of life-science research workers in the U.S.14 However U.S. regions such as San Diego, Raleigh-Durham, Washington DC and others are also becoming huge hubs for life sciences.

11 CBRE, London: a leading FinTech cluster. 12 David J., Grimmelmann L., McCarthy B., Pietrella S., and Whiting G. (2016), The UK’s FinTech Cluster, Leveraging the UK’s strengths in financial services, technology and entrepreneurship. Macroeconomics of Competitiveness, Spring 2016, Harvard Business School. 13 JLL (2014) Life Sciences Cluster Report. 14 Growth Commission, London Stansted Cambridge Corridor, Corridors and Tech regions: International Case-studies.


© 2016 Citigroup

11

Figure 6. Top 20 Regions Filing Biotech Patents in 2013 Figure 7. Top 20 Regions Filing Pharma Patents

Source: OECD, Citi Research Source: OECD, Citi Research

A Case Study of the Raleigh-Durham Area The Raleigh-Durham-Cary region in North Carolina has filed approximately 100 patents in both biotechnology and pharmaceuticals. An important life sciences cluster has developed in the region known as the Research Triangle due to the geographic locations of the area’s leading universities – Duke University, North Carolina State University and the University of North Carolina at Chapel Hill. In the center of this region is the Research Triangle Park, an industrial/research park created by the state to help bring research and development into the area. Some of the big names in pharma have set up base in this area attracted by the availability of high skilled people due to the proximity of well-known universities, together with the high quality of life in this region. The region is home to four of the five top agri-tech companies including Syngenta, Bayer, CropScience, BASF and Monsanto. In 2013, total employment in life sciences reached over 27,000 and over $1 billion was invested in the area in the same period (over 77% came from NIH funding, the rest from VC funding).15 There is plenty investment being done in the area from big biotech firms, for example Syngenta announced plans to extend its research campus by 2018 which will attract 150 new jobs and will bring new technology and office facilities to the campus. However there are a number of factors that could become a problem in the area with challenges including: (1) the lack of investment for start-ups; in fact VC funding is rather low when compared to other areas and (2) the lack of available infrastructure for example the existing supply of quality lab spaces is quickly diminishing in the area. Plans are however in place to invest in new research facilities over time.

15 JLL (2014) Life Sciences Cluster Report.

0 100 200 300 400 500 600 700 800 900

Boston-Worcester-Manchester (US)San Jose-San Francisco-Oakland (US)

New York-Newark-Bridgeport (US)Tokyo (JPN)

San Diego-Carlsbad-San Marcos (US)Washington-Baltimore-N.Virginia (US)

Los Angeles-Long Beach-Riverside (US)Seoul (KOR)

Philadelphia-Camden-Vineland (US)Kanagawa (JPN)

Seattle-Tacoma-Olympia (US)Chicago-Naperville-Michigan City (US)

Beijing (CHN)Raleigh-Durham-Cary (US)

Osaka (JPN)Gyeonggi-do (KOR)

Houston-Baytown-Huntsville (US)Paris (FRA)

Shanghai (CHN)Shenzhen - Guangdong (CHN)

Total No. of Biotech Patents filed in 20130 1000 2000 3000 4000 5000 6000 7000 8000 9000

Shenzhen - Guangdong (CHN)Tokyo (JPN)


San Diego-Carlsbad-San Marcos (US)Seoul (KOR)

Gyeonggi-do (KOR)Kanagawa (JPN)

Seattle-Tacoma-Olympia (US)Boston-Worcester-Manchester (US) New York-Newark-Bridgeport (US)

Kyoto (JPN)Los Angeles-Long Beach-Riverside (US)

Aichi (JPN)Osaka (JPN)

Portland-Vancouver-Beaverton (US)Stockholms län (SWE)

Chicago-Naperville-Michigan City (US)Houston-Baytown-Huntsville (US)

Noord-Brabant (Einhoven)- NL

Total No. of ICT Patents filed in 2013


© 2016 Citigroup

12

The Life Sciences Cluster in Europe – A Case Study of Paris and Cambridge France has maintained a leading role in Europe in terms of medicine production volumes and pharmaceutical revenues. In total, France employs nearly 400,000 people in research & development (R&D) and approximately 11% of its tertiary graduates specialize in science. The country has developed many life sciences clusters- the most dominant being the cluster in the Paris region. All of the largest pharmaceutical companies (both domestic and international) have a presence in Paris. These companies have their headquarters based in the western and southern inner suburbs with companies taking advantage of the favorable local market conditions such as a large amount of project subsidies and with the opportunity to benefit from highly competitive sites in terms of rents and accessibility. In addition, there are roughly 300 public research institutions, three universities and 20 Grand Ecoles that make up the backbone of Paris’s research network.

Although Cambridge in the U.K. has not filed as many pharma and biotech patents as other areas, it is still an important region for innovation in life sciences. Cambridge has the presence of a world renowned university and other important research centers such as the Sanger Institute (a genomic research center), the Babraham Institute (for immunology research), the Laboratory for Molecular Biology, and Astra Zeneca has recently announced plans to build a new $515 million R&D facility in the area.

Cambridge has also established research parks such as the Cambridge Biomedical Campus which is adjacent to Addenbrookes Hospital and Cancer Research UK. Cambridge University has also set up the Cambridge Enterprise, which works with academics to commercialize their research work, provide seed funding, and liaise with VC’s and other angel investors. The region is becoming an important hub not only for large pharma and biotech firms but for also for start-up companies. The biggest issue with the Cambridge life sciences cluster is that there is a high demand for lab and office space and the availability of large units can be limited. The U.K. government is trying to rival the Boston regional cluster for life sciences by establishing a golden triangle between the universities in Oxford, Cambridge and London. This initiative is trying to bridge the gaps of lack of available funding, commercialization support, and improvements to infrastructure.

ICT Cluster – A Case Study of Stockholm It is not surprising to see that two of China’s regions (Shenzhen-Guangdong and Beijing) are in the top five places for filing patents (refer to Figure 8). China’s ICT sector has been growing rapidly helped by support from the Chinese government’s plans and policies and the increase in the number of Internet and mobile users (630 million Internet users in 2014, 1.29 billion mobile users in the same period).16 China’s government has given significant support to cluster development including favorable land / tax policies, an improvement of local infrastructure, and has also established important research centers and industrial parks, etc. For example in the Hangzhou Yuhang Economic and Technological Development area, 100% of taxes can be refunded to eligible companies in the first two years after starting a business.17 China is also investing heavily in its graduates. The country’s universities now produce the most life sciences graduates and post-graduate in the world and at least 80,000 new graduates trained in western universities are returning to China to work in companies or academic institutions.

16 EUSME, China-Britain Business Council (2015), The ICT market in China. 17 IBID.


© 2016 Citigroup

13

Figure 8. Top 20 Regions for ICT Patents in 2013

Source: OECD, Citi Research

Most of the other regions that filed a large number of patents in ICT are based in the U.S. and Japan. In Europe, Stockholm is ranked as being one of the most important tech hubs and has some of the fastest growing start-ups. In 2013, it filed a total of 745 patents in ICT. Many fast growing technology companies such as Skype, MySQL, Mojang and Spotify have started in this area. Kista Science City in the region is an area that has become an important cluster for ICT companies and research. It was founded started in 1986 as a place where industry, academics and the public sector got together and started the Electrum Foundation. The aim was to make the region a world-leading center for electronics research, but in the 1990s it extended its portfolio and transformed itself in a region important for ICT. There are close to 1200 ICT companies located in Kista with 90% of these export companies.18 Many companies have offices or headquarters based in the area including Siemens, Huawei, Ericsson, IBM and others.19. It is also home to two universities – the Royal Institute of Technology (KTH) and Stockholm University. Currently there are over 6,800 university students studying ICT courses at the Kista campuses2019 and a total of 24,000 employees that work in the ICT sector.

The Successful Recipe for Innovation Clusters There are a number of things the regional innovation clusters we looked at have in common. The first is the availability of high skilled workers – it is not a coincidence that most of the regional clusters mentioned above are located close to universities and research centers such the Boston Route 128 corridor which has MIT and Harvard on its doorstep. The second is the availability of funding – the majority of the VC funding is currently being invested in the SF Bay Area, however places like Beijing have increased their VC funding at a staggering pace. In Europe, the availability of VC funding is low compare to the U.S. and China, however other forms of funding are also important such as the EU Research & Innovation program “Horizon 2020” with nearly $80 billion of funding available for both universities and small and medium enterprises. Low rents especially for start-ups are also essential. Incubators such Level39 at Canary Wharf for FinTech companies are excellent 18 Switzerland Global Enterprise, Opportunities and challenges – Sweden’s ICT market. 19 kista.com 20 kista.com

0 1000 2000 3000 4000 5000 6000 7000 8000 9000

Shenzhen - Guangdong (CHN)Tokyo (JPN)


San Diego-Carlsbad-San Marcos (US)Seoul (KOR)

Gyeonggi-do (KOR)Kanagawa (JPN)

Seattle-Tacoma-Olympia (US)Boston-Worcester-Manchester (US) New York-Newark-Bridgeport (US)

Kyoto (JPN)Los Angeles-Long Beach-Riverside (US)

Aichi (JPN)Osaka (JPN)

Portland-Vancouver-Beaverton (US)Stockholms län (SWE)

Chicago-Naperville-Michigan City (US)Houston-Baytown-Huntsville (US)

Noord-Brabant (Einhoven)- NL

Total No. of ICT Patents filed in 2013


© 2016 Citigroup

14

examples of spaces that encourage innovation amongst companies and provide cheap desk space for start-up companies. Good government policies and incentives are also essential .Tax breaks in China for start-up firms allow companies to grow more quickly while good immigration policies allow access to talent that is needed. Other plans such as investment in infrastructure, education and connectivity, good livability, and cheap residential housing can also provide additional benefits for successful regional clusters to grow and develop. There doesn’t seem to be a winning recipe for a successful innovation cluster, however it is evident that the majority thrive in an open, global environment where knowledge sharing is at the core of their activities.


© 2016 Citigroup

15

1. Big Data Disruption The Big Data Revolution in Energy The disruptive forces of Big Data, software, and analytics are creating seismic shifts in the traditionally hardware-dominated energy sector and beyond. The transformation affects the entire energy value-chain, from finding energy faster and producing it more inexpensively in the upstream, to more customized energy at the downstream user level, almost as if energy becomes un-commoditized. With this shift, current business models in many sectors could become obsolete. Better market efficiency and transparency would not only come to well-established energy markets, but potentially spread also to the local level in the form of transactive energy. These advances are already taking place, with regulations are being implemented in early-adopting regions to make this energy transformation happen. What is going to make all this happen is greater automation and optimization through data and advanced analytics on software platforms. Here’s what the future could look like:

Upstream: Data-driven automation would lower the cost of oil and gas production while advanced analytics would help uncover and recover even more oil and gas in the ground. This could lead to capex costs on a per unit production basis potentially falling by 50% or more from the already deflated levels.

Downstream: The democratization of energy could see renewables and distributed energy resources (DERs) proliferate at the local level meaning fewer new power plants would be needed due to demand-side management and optimal capacity utilization of power plants. Consumers could eventually “trade” energy with others in the form of “transactive energy.” Utilities are already developing ways to value DERs at the neighborhood level and the value of electricity would differ depending on locations and also usage in order to value ancillary services. The ability to collect, process, and analyze more data for optimization and automation is key to these developments.

Business models: Utilities, for example, would become distribution service platform providers for DERs, as the State of New York is already envisioning. Technology companies could provide energy network optimizing software or even operate platforms and energy companies that transition to providing services could become asset-light, as they could control how energy is routed and optimized. Third-parties or homeowners would become energy providers through DERs and auto companies would become service and energy providers (e.g. through their battery technology).

Markets: Greater data availability and transparency should raise market efficiency not only in major energy markets, but potentially down to the neighborhood or more micro levels. As new builds in the power sector involve more near zero-variable cost sources, such as wind and solar, along with greater demand and storage optimization, the goal of dramatically lowering energy costs for all, with the possibility of free energy in some corners, may finally come to fruition. The notion of free energy came to prominence in the 1960s, as nuclear fusion was touted as a way to provide free energy but fusion was in the embryonic stage then. Now, Big Data and advanced analytics are developing rapidly to improve forecasting, automation, customization, and the democratization of energy. The end result is that we are producing more energy with fewer resources.

Anthony Yuen, Ph.D. Global Commodities Strategy


© 2016 Citigroup

16

Upstream: Saving Billions in Oil and Gas Production Costs The drive to squeeze cost and boost production in the upstream oil and gas industry has become relentless. Big data and advanced analytics, including machine learning, are helping to drive this cost decline and production surge: an entire discipline ”Digital Oilfield” aims to drive operational efficiency, better decision support in making real-time decisions, data integration, workflow automation, and production optimization. Unleashing this massive amount of low-cost oil and gas could transform not just the sector, but entire economies.

A recent survey of 900 senior oil and gas professionals by DNV-GL, a major certification body, showed that 45% expect “solid or high potential for data and analytics to make the industry more efficient”, 36% plan for “moderate or significant investment in Big Data and analytics,” and 20% believe that “their operations are ‘highly digitalized today.’”

The oil and gas industry is a major generator of data, but the lack of sophisticated analytical tools has hindered the ability of the industry to focus on analysis of the data. Many variables can affect drilling and production operations. Advanced analytics can help to make sense of the data, particularly in analyzing why production costs vary across plays and across competitors. After implementation of an analytical platform by one super-major, the cost of a multimillion dollar well fell by ~$700,000. Pioneer, one of the best run U.S. exploration and production companies, is also active in the analytics space, in areas including rig power optimization, real-time communications, analytics for improved team and rig performance, and improved performance of drilling tools.

The drilling process itself generates vast amounts of data. Automated drilling, by processing this drilling data concurrently, could decrease drilling time and cost by 30 to 50%, according to DNV-GL. Automated drill pipe handling could eliminate the need to have workers on the drill floor thereby improving safety while drilling process monitoring and diagnostics connect topside and downhole measurements together to allow instant analyses and automated control of drilling operation. Managed pressure drilling, by enhancing pressure control automatically, could reduce downtime during complicated drilling operations. The measurements collected in drilling, along with smart completion techniques, would also benefit the completion process by identifying optimal locations of oil and gas “drainage points” and raising the recovery factor of shale production. Machine learning can find model inconsistencies, narrow uncertainties and improve forecasts and option assessments, thereby increasing productivity and reducing downtime.

More broadly, process-automation is very useful in managing the substantial amounts of operational data that oftentimes come from different sources in different formats. The work needed to integrate these data appears daunting enough that useful data often go to waste. Indeed, for example, the integration of geological, geophysical, petrophysical, and geomechanical models allowed a producer in the Eagle Ford shale region to space hydraulic fracturing stages geologically, rather than geometrically. The first 8 laterals completed using this integrated process saw their perforation cluster efficiency rise from the low-60% to low-80%.

More important, production optimization, in combination of other areas of digital oilfield, could substantially raise recovery rates, thereby slashing the capex cost per unit of production. Current recovery factors for unconventional wells are in the 5% range; in other words, more than 90% of the oil and gas in a well stays underground. Conventional wells see recovery rates in the 15% range before


© 2016 Citigroup

17

secondary recovery or tertiary recovery, such as CO2 injection, is applied. Raising recovery rates of unconventional wells, through improved data analysis of geological and well data, could lower the per barrel capex cost of a well by 50% if the primary recovery factor rises from 5% to 10%, halfway to the 15% recovery factor of conventional wells. Higher initial production rates that are not driven by simply focusing on sweet spots of a field (i.e. high-grading) could be indications that recovery rates are rising. Together, advances in data usage and analytics should drive costs even lower.

Downstream: Seismic Shift in Electricity The electricity sector is in the process of integrating even more renewables than previously thought possible, expanding Distributed Energy Resources (DER) and ushering in a world of “Transactive Energy.” Big data analytics and forecasting are making the accommodation of even more renewables possible.

Professionals in the utilities sector seem clear what’s ahead for them, with software-dominated areas appearing high on the list of new revenue streams that utilities should pursue. Asked to choose “all that apply” in a recent survey by Utility Dive on which new and emerging revenue streams your regulated utility was likely to pursue, 66% of the respondents chose “offering energy management and efficiency services to customers” as a new way that firms should generate revenue in the regulated utility space. “Developing distributed system platform”, at 31%, and “offering microgrids-as-a-service to customers”, at 19%, were also key areas that industry participants highlighted as emerging revenue sources. Note that these new areas, in some ways, all involve the use of software and sophisticated data analytics. “Deploying electric vehicle charging infrastructure”, at 52%, and “deploying distributed energy storage”, at 40%, involve a mix of hardware and software solutions, with the latter depending on how sophisticated the system becomes in generating and managing two-way energy and information flows.

The bi-directional nature of energy and information flows fundamentally changes the way the grid would operate. No longer would energy flow one way. However, managing the power grid is not as easy as plug-and-play in the information world. Electricity flows require the balancing of not only the active power that consumers use, but also reactive power and other physical characteristics.

Innovations in data collection and analytics are at the heart of this transformation:

From the Upstream (Smart Renewable Forecasting and Integration)…: The fundamental technologies of solar panels, wind turbines, converters, and energy storage have been around for years, but having nearly half or more of total electricity supplied by wind or solar was previously thought to be impossible due to grid integration issues. But Big Data and advanced analytics are helping the electric grid to function more seamlessly, enabling wind and solar utilization and penetration rates to rise more sharply, and integrating more distributed generation. This is done through more precise measurements and predictions of very fickle solar and wind generation, based largely on complex weather data, derived from results from supercomputers. Having more precise estimates of renewable generation allows the grid to schedule in the appropriate amount of backup generation and deploy other measures, such as demand-side management.


© 2016 Citigroup

18

…To the Downstream (Distributed Energy Resources (DER)): With the increase in distributed generation resources (DER) and two-way flows of energy, a key enabler in the future is data analytics that allow the measurement of data, two-way communications and deployment of generation or demand resources. These are important in the siting, pricing (i.e., compensation), and operation of DERs in the future. California and New York are at the forefront of these developments globally. This goes beyond the question if DER will grow but to how to accommodate this growth.

Figure 9. Future Grid Systems and Smart Building Can Communicate in Ways that Improve Overall System Efficiency and Reliability

Source: U.S. Department of Energy, National Institute of Standards and Technology

From Inside Buildings (Smart Lighting as Sensors)…: Sensors are key to collecting data. Smart lighting could be a crucial part of the new energy network, using lighting nodes to collect data on electricity usage. The vast amounts of data collected would then be used for real-time analytics. Lighting is a large market because just about every room in every building has lights and sockets. The ubiquitous nature of lighting could let it leapfrog other innovations in terms of adoption rate, coverage, and ease of use. Generally, a lighting unit has three lines connected to it: in, out, and ground. Information to and from the unit, now equipped with sensors, could be passed through a fourth line for communication. Data collected could help with demand response (DR) and demand side-management (DSM) in optimizing energy consumption as part of the smart grid. On the regulatory front, California’s Title 24, part 6, for example, on residential lighting, could be one of those enabling regulations that accelerate the adoption of smart lighting. Companies like GE are active in this area.

…To Aggregation (A Virtual Power Plant (VPP)): Virtual power plants can become a reality by using software to integrate distributed generating resources (DERs), energy storage, and demand response resources from different locations together as if it is a single power plant. Groups working on VPPs include New York’s ConEd, Vermont’s Green Mountain Power, and Kentucky’s Glasgow Electric Plant Board. DERs and the broader VPPs


© 2016 Citigroup

19

are critical new developments in the electricity space. Historically, if electricity demand in an area looked to outstrip supply, a typical regulated utility would ask the regulators permission to build new power plants and/or transmission lines, and “rate-base” these new builds to earn a regulated return. In the case of New York, the regulator denied ConEd’s request for a $1billion upgrade in resolving equipment overloading. Instead, ConEd created the BQDM (Brooklyn Queens Demand Management) program that relied on both traditional utility and non-traditional customer/utility improvements at a total estimated cost of only ~$200 million. While using DERs to provide both energy and ancillary services is valuable, valuing DERs is not as straight-forward. It would require pricing energy and ancillary services at the neighborhood or an even more micro level.

Figure 10. How Distributed Energy Resources Become Virtual Power Plants

Source: U.S. Department of Energy, Ventyx

Improving Market Efficiency and Price Discovery Through Advanced Analytics on Big Data The use of big data and advanced analytics raises the transparency of supply-demand fundamentals, energy flows, and the price discovery process. Greater market transparency is already coming through advanced analytics on real-time pipeline, shipping, and electricity flow data. Real-time tracking of energy flow and generation enables better prediction and coordination of systems, more accurate prediction, more efficient market, and lower electricity and energy prices.

In trading, advanced analytics on large datasets, from flow and weather data to some unstructured data, already play a key role in the analyses of fundamentals. Using midstream data, such as natural gas pipeline flows and power transmission, traders and analysts are able to obtain a much more accurate picture of supply-demand of natural gas and electricity. For example, there are more than 40,000 locations in the U.S. that report the amount of gas flow at particular flow points daily. These are commonly known as the pipe scrape data. The arms-race involved has pitted various traders and analysts to develop ever more sophisticated models,


© 2016 Citigroup

20

some involving machine learning, to capture flow rates for an even more accurate picture of regional and national supply-demand. Combined with data on temperatures, wind speed, solar irradiance, and electricity obtained through measuring power lines’ electromagnetic waves forecasting becomes more accurate, diminishing uncertainty and narrowing price spreads. However, price volatility may not have declined as much as it should, as results from massive data analyses, using datasets described above and more have changed the market’s acceptable margins of errors. One major improvement is what’s considered the acceptable range of a U.S. weekly natural gas storage estimate, which went from 10-billion cubic feet (Bcf) in the mid-2000s to less than 5-Bcf currently. Results from weather models, released four times daily, have also enabled traders to make better decisions on how changes in weather forecasts affect demand.

In markets that involve seaborne transportation, the tracking of oil and to some extent LNG tankers using satellite images, radio signals and even the amount of draft of tankers, provide valuable information on near term supply and predict whether tightness in the market could come quickly. The tracking of oil pipeline flow by monitoring the electricity consumed at compressor stations along pipelines is another way of understanding how much oil is flowing in and out of certain key regions. Storage levels in certain oil storage tanks are also being measured: the real-time flow in and out of Cushing in the U.S. and its storage levels are the subject of intensive scrutiny by various suppliers, end-users, and traders. Instead of waiting for the U.S. government data release every Wednesday to discern market fundamentals, the market is now able to piece together the fundamental picture daily.

In addition, in markets where non-quantitative information can have substantial impacts on sentiment — such as geopolitical events in oil — quantifying sentiment data is finally providing metrics that can be incorporated into supply or price calculations. Natural language processing and text analytics can allow commodity traders to make use of the vast expanse of data available through online communication networks and online media to gauge sentiment on current issues affecting commodity markets. Even in markets that have more quantitative data available, such as electricity and natural gas, accurate indicators of market sentiment could prove helpful in discerning whether the market might be too bullish, bearish or reaching an inflection point.

Policy is Key to Implementation Proper policies can facilitate the development, deployment, and adoption of new technologies. These types of regulations are already at different stages of progress. For example, while the regulatory framework needed for self-driving cars to proliferate are not in place yet, regulations on electricity that transform utilities’ business models and encourage the use of DERs have emerged in a number of locations.

New York’s Reforming the Energy Vision, or REV, is one of the most ambitious regulations put forth by a regulatory agency in changing the business models of utilities. Regulators foresee that utilities would become Distribution Service Platform (DSP) providers that operate a DER marketplace, deploy DERs optimally, and generate revenue from DER origination, as well as system integration and operation. In the distributed, decentralized world, particularly in the electricity space where the supply and demand of electricity has to match instantly, ensuring a smooth and optimal operation of the grid necessarily requires advanced analytics to process the vast amount of data generated.


© 2016 Citigroup

21

Changing Roles of Energy Companies The Big Data revolution means energy companies could increasingly become asset-light. Numerous companies providing software or services thrive online, building on the physical infrastructure that firms construct. As utilities in the future increasingly serve as platform providers, once they have created efficiency in and have control over the distribution of electricity, they could divest some of their physical assets.

In a transactive energy world, having the right platform, similar to what eBay or Alibaba has built, could be extraordinarily lucrative, especially as the network effect increases when the platform contains more and more customers. Indeed, nearly half of the companies that won bids in the California utility space bidding for energy storage were software companies. On a larger scale, some industrial and technology companies are already designing operating systems for the grid, i.e. GE which introduced DER aggregation software recently.

Traditional utilities need to be physically close to their own service areas due to the need to operate and maintain physical infrastructure. Operating a Smart Grid requires the ability to manage big datasets, an intelligent infrastructure and sophisticated analytics for forecasting and operation which means they don’t necessarily have to be physically close to infrastructure.

The blurring of the sectoral boundary is evident in how technology companies are getting into the energy space, completely upending the status quo. As energy consumers, tech companies already have high demand for electricity, especially for their data centers, so much so that some have even acquired electricity trading licenses. Their desire to go green by drawing more power from renewable energy sources also brings along the need to resolve the intermittency issues and integrate energy storage solutions. These entities require complex calculations to manage their energy needs and have the analytical platforms to develop the necessary tools. As solution providers, the Big Data component of energy, naturally the forte of tech companies, gives tech companies the advantage as the electricity sector becomes more distributed with multi-way energy and information flows. For example, Google’s Project Sunroof uses Google Earth’s high-resolution aerial mapping to identify solar energy potential of roofs and possible installation techniques. Google’s Nest thermostat uses built-in sensors and algorithms to automatically detect people’s daily patterns. Combined with energy demand and weather data, these smart thermostats could estimate users’ demand patterns and optimize energy supply and demand. Microsoft developed its own software that collects 500 million data transactions daily from sensors and other energy devices to understand its own energy demand level and patterns. Both Google and Apple also have licenses to trade electricity. The list goes on and on.

What does the future of energy look like? Producers could tap previously stranded assets and do it quickly; utilities could be winners but only if they transform with the times; renewables, despite intermittencies, could operate as smoothly as traditional fossil energy; emissions should be limited as energy demand is optimized and renewables proliferate. Trillions of dollars are at stake. This is a story of how software will transform a hardware-dominated sector; it is the kind of creative destruction that demands fundamental changes in an entire sector.


© 2016 Citigroup

22

2. Contextual Commerce What is Contextual Commerce? Contextual commerce is a simple concept at heart – give a consumer what they want when the want it without requiring the consumer to take the initiative or expend much effort in any way.

More formally, we might define it as follows – Contextual Commerce is the concept, or technology / platform that enables a consumer to interact and transact with their chosen merchant /brands in the consumer’s preferred context or medium. The context can be a social networking conversation, an exchange on a messaging platform, an online session, a chat or a (voice) call – we have even seen a unique print media contextual commerce example.

For example, a consumer who searched for “funny birthday cards” might be prompted with an offer for a gift card or flowers to simultaneously send…or a Facebook user who is having a conversation about meeting friends for dinner or drinks is now able to order a Uber to go to the event – the key being that they can complete the Uber transaction without leaving Facebook. This is a global trend – for example, Chinese mobile messaging platform WeChat is quite advanced in its Contextual Commerce push as it lets its users hail cabs, customize branded products (such as Nike shoes), pre-order a latte at Starbucks, order movie tickets and pay bills, all without ever leaving the messaging platform.

Context and Intent A fundamental tenet of marketing has been to use demographics as a proxy for a potential consumer behavior. While demographic information is not without merit, we note that: (1) it might not capture important segments of the market; (2) it is prone to suggesting stereotypes; and (3) it definitely does not help capture intent or timing. However, context can be used to capture intent and timing. Going back to our definition…context is used to determine what the consumer wants (intent) as well as when they want it (timing).

Some examples make the point.

40% of baby product purchasers in the U.S. live in households without children, according to a Google/Ipsos 2015 study. These product purchasers could be friends, family or co-workers. If two friends or co-workers are having a Facebook conversation about baby products, that context may be a useful one to suggest a product on or provide a coupon for a baby products company to incentivize a transaction. It could be a crucial advantage in a large and competitive end-market like baby products (worth ~$67 billion by 2017, per Statista).

Home improvement companies might overwhelmingly target a male audience…but 45% of home improvement searchers on mobile phones are women. Combine this with the fact that 51% of mobile buyers purchased a brand other than the one they originally intended because the alternate brand provided a better experience (for example, a convenient how-to video) and it implies that appropriately using context to help a consumer at the right time can pay good dividends.

Ashwin Shirvaikar, CFA U.S. Payments, Processors & IT Services Analyst


© 2016 Citigroup

23

How Well Known is the Opportunity? Contextual Commerce is a new-enough concept that there are still alternate descriptors for it. We use “Contextual Commerce” because it captures the entirety of what we want to convey whereas many alternate terms are narrower in scope.

1. Social Commerce is an older term (Yahoo!, c.2005) that captures the notion of a consumer’s social engagement but not the context.

2. Conversational Commerce (Chris Messina, Uber, c.2015) involves “…utilizing chat, messaging, or other natural language interfaces (i.e. voice) to interact with people, brands, or services and bots…” This is essentially the same thing as contextual commerce, but “conversations”, while they are an important context, are not the only one.

3. Terms like F-commerce (or Facebook commerce) and T-commerce (using a smart digital TV-set, e.g., Apple TV, with two-way communication that incorporates interactive advertising) are clearly limited to a single communication medium and hence are a subset of the bigger theme.

Regardless of moniker, there are a lot of smart people working towards the same end…to use context to determine both the intent to carry out a particular action (generally the purchase of a product or service) and the timing of that action…and then facilitate the intended action by minimizing friction in the process.

The Path to Contextual Commerce Over the years, e-commerce has grown rapidly and made considerable inroads at the expense of traditional physical commerce because it eliminates many of the cumbersome elements of physical commerce (see Figure 11 and Figure 12 below). Of course, one area where we may have stepped backwards is the payment transaction itself – the physical payment experience is easy while the online experience may involve form-filling (poor consumer experience) and higher cost (poor merchant experience). In many (but not all) situations, Contextual Commerce can provide a better experience (see Figure 13).

Traditional “Physical” Commerce

Figure 11 illustrates the traditional “physical” shopping experience and we note that shopping for a specific product actually requires significant consumer initiative and commitment, all the way to the check-out process.

Figure 11. Traditional “Physical” Commerce is Cumbersome

Source: Citi Research

Pre-Shopping Post-ShoppingShopping• Prep Shopping List• Do Research, if needed

- Call friends- Research on-line

or off-line• Drive to one/more

shops• Possibly repeat above

process, esp. for big-ticket items

• Walk the aisles• Seek alternatives for

unavailable items• Put stuff in shopping

cart• Wait in line• Scan purchased items • Remember coupons, if

any• Pay

• Load / Unload car• Deal with traffic• Possible customer

service calls• Possible physical return

process


© 2016 Citigroup

24

E-Commerce / M-Commerce

Figure 12 illustrates the traditional online / mobile shopping experience. Here, pre-shopping is less cumbersome than the physical process, although the lack of “touch and feel” can make some consumers less comfortable especially for big-ticket items. Second, similar to the physical process, the consumer needs to show a high level of commitment to go through with the process. Third, the actual payment process in “physical” is often smoother than in an online or mobile environment and over 50% of shopping carts are abandoned as a result. Finally, we must consider how many apps a consumer might reasonably have on their mobile phone.

Figure 12. E-Commerce Solves Some “Physical” Issues But Raises New Complications


Contextual Commerce

Figure 13 illustrates Contextual Commerce. Note that at least in the initial stages of its evolution, Contextual Commerce may be more appropriate for social and/or discretionary situations – for example, food with friends, a sporting event, a movie, transportation to such events, fashion- and fitness-related shopping. This is probably not how a typical grocery list gets filled. However, there are interesting physical-to-digital interfaces (e.g., Amazon Dash) that can extend the scope of Contextual Commerce.

Figure 13. Contextual Commerce is Simplicity, Personified


Why Now? A Confluence of Factors There is clearly a need for a better (more seamless) way to conduct m-commerce driven by:

1. Consumer needs – Convenience; Desire for lack of transaction-related friction within the transaction; and

Pre-Shopping Post-ShoppingShopping• Prep Shopping List• Do Research, if needed

- On-line and Social

• Find Items on merchant website or app

• Add to Shopping cart• Select Checkout• View Cart to Confirm

Contents• Enter Billing & Shipping

Info• Confirm Info and Buy

• Wait for Physical Goods. Services can be consumed right away.

• Possible return process – increasingly this is being simplified to the point of being seamless enough that “no hassle” return is a feature.

Post-ShoppingShopping• Relevant product or

service is recommended within the context.

• Consumer choice is to proceed (or not).

• Wait for Physical Goods. Services can be consumed right away.

• Possible return process – increasingly this is being simplified to the point of being seamless enough that “no hassle” return is a feature.


© 2016 Citigroup

25

2. Merchant needs – Desire to improve M-Commerce conversion rates…despite the above factors, conversion rates using a mobile platform re-hash of an outdated “shopping cart” process are shockingly low…perhaps as low as ~1%.

Multiple technology and infrastructure advancements and demographic factors have led to the emergence of Contextual Commerce, which we view as a better way to transact on the mobile phone.

1. Mobility Penetration…high smartphone penetration; Improved mobile data infrastructure.

2. Social media adoption…integral to how younger demographic interacts.

3. Generational differences in quality and quantity of shared personal information…the younger generation is comfortable sharing information in exchange for convenience…better data can lead to better analytics.

4. Rising desire to experience rather than transact…increasingly, we live in an experience economy where the expectation is that factors other than the central experience should be as simple and frictionless as possible.

5. Computer Science advances…Artificial Intelligence; Growing use of application programming interfaces (APIs); Micro-services.

How Does Contextual Commerce Work? Marketing platforms – even newer web-based or mobile ones – have historically been built around the notion of driving consumers to a merchant with the ultimate goal of achieving a sale. The specific paradigm shift in Contextual Commerce is that the merchant is brought to the consumer by integrating the merchant’s offering(s) with the consumer’s preferred platform of choice (i.e., context). Further, by securely and seamlessly exchanging the consumer’s payment credentials in the background, the consumer is freed up to enjoy the purchased product or experience rather than spend time and effort completing a transaction.

Figure 14 and Figure 15 illustrate this fundamental and quite powerful change, which can help a merchant project themselves considerably more than the current “pull”-based system.

Figure 14. Traditional Commerce Process Requires Consumer Initiative and Has Multiple Breakage Points

Figure 15. Contextual Commerce Paradigm Shift Provides Consumer What They Want, When They Want, With Limited Effort

Source: Citi Research Source: Citi Research

Context• Blog• Email• Chat• Social Network• Banner or Other Ad• Virtual Reality• Gaming• Other Contexts

Purchase or Experience

User or Consumer

1

2

3

1. Consumer makes a conscious decision & shows initiative to visit a merchant site.

2. Consumer is taken out of their preferred context to visit a merchant site.3. If Consumer opts to proceed with the Purchase, the Transaction takes

place at the merchant site, with all the associated “shopping cart abandonment” risks.

User or Consumer

1. Consumer stays in preferred context; Merchant functionality is brought to the context. Micro-services, Payment APIs and Advanced Analytics techniques (perhaps AI-based in the future) are used to make this happen.

2. If Consumer opts to proceed, the Transaction takes place within the Context, with consumer information and payment credentials checked in the background.

Context

• Blog• Email• Chat• Social Network• Banner or Other Ad• Virtual Reality• Gaming• Other Contexts

12

Contextual Commerce Platform

Purchase or Experience


© 2016 Citigroup

26

The Contextual Commerce provider is a relatively new party in the ecosystem. It is too early at this point to determine which specific business and financial model might win. But the provider will likely have to provide the functionality below. This is not a complete list, obviously, since the industry likely evolves over time.

1. Integration between the merchant’s systems (e.g., CRM, supply-chain, order / bill) and the Context. This option likely requires bespoke deals for each Merchant-Context combination. Some of the work is likely to be automated using “middleware” tools that emerge.

2. Develop the user interface / user experience (UI/UX) for the merchant.

3. Not all contexts are created equal and so there may be development, analytics and security work needed to incorporate certain contexts.

4. The actual processing of underlying payments. This may be outsourced, or in certain cases, the Contextual Commerce provider also acts as the payment processor.

5. Context and outcome-specific expertise – for example, knowledge of carrier billing may be important for the gaming context but not for emails.

How Real is the Contextual Commerce Opportunity? The examples below show that the underlying technology is proven out. Many of the use-cases are new enough that there is no measurable track record yet. Also our conversations with some innovative companies that are involved with these launches show that there is no set financial model to monetize these capabilities.

WeChat

WeChat is China’s leading messaging app for sending text, voice, and photos to friends and family. However its differentiation lies not in this social capability but in the functionality it brings to its users, including making doctor appointments, paying utility bills, hailing a taxi, buying movie tickets, etc. WeChat approves third parties for an official account which gives them access to certain kinds of user information via APIs. The user’s payment credentials, which are stored in a WeChat wallet, are also used to facilitate interactions.

Facebook / Uber / PayPal

In December 2015, Facebook and Uber launched the ability to “Request a Ride” without leaving Facebook Messenger. For example, if two users decide to meet for a drink, one of them can send the bar location on Messenger, and they can request a ride by tapping on the address. There are other practical applications – for example, the ability to share the current Uber trip with co-workers through Messenger so they know when you might arrive for a meeting. PayPal provides the underlying functionality in this case.

Pinterest Buyable Pins

A May 2015 Millward Brown survey showed that Pinterest users use it as a planning tool during key “life change” moments. In other words, both intent and timing are likely. Similarly Shopify has indicated that 87% of Pinners have purchased an item due to Pinterest. So it is no surprise that Pinterest is one of the early adopters of Contextual Commerce and announced significant new functionality to assist its Pinners in the shopping process.


© 2016 Citigroup

27

Iine-Card

This is an interesting physical-to-digital example of contextual commerce (“ii, ne?” means “good, no?” in Japanese). Iine-Card is a Japan-focused entity that has worked with a range of Japanese fashion magazines to create a digital catalog of the products that are displayed in the magazines. If a reader likes a jacket or handbag or belt that a model is wearing, they can digitally scan the image from the Iine app and order the product from the catalog.

Contextual Commerce Market Size Market size estimates for mobile commerce vary – BI Intelligence predicts mobile commerce will reach $79 billion in 2016; meanwhile, Forrester Research estimates the 2016 market size at $142 billion in the U.S. Regardless, this is a large and fast-growing opportunity, especially when juxtaposed with a shopping cart abandonment rate that has stayed above 70% for years (~74% in 1Q16, per eMarketer).

One can debate whether or not innovations like Contextual Commerce add to the size of the pie and it is possible to make fair points on both sides of this debate. Anecdotally, we believe that higher transaction efficiency should lead to more sales. What is not up for debate is that if Contextual Commerce is properly implemented, it should dramatically cut down on shopping cart abandonment rates because it eliminates the source of friction. So, just the “sales capture” opportunity runs into the billions of dollars even when one acknowledges that not every product and service category is appropriate for Contextual Commerce. Our view is that if a Contextual Commerce platform can show a merchant that it is responsible for incremental “sales capture”, there is an opportunity to share in the incremental economics generated by this system.

Note also that the Contextual Commerce opportunity extends beyond enabling traditional m-commerce and into a number of emerging technology themes including: (1) Internet of Things applications; (2) Wearable device – especially micro-transactions; and (3) Augmented / Virtual Reality (AR/VR) applications. In our view we are only scratching the surface of this opportunity currently.

What Are the Barriers to Adoption? Because the underlying technology for Contextual Commerce has been proven out, the limiting factor is not how to deploy – rather it is how to deploy well. This may be stating the obvious but adoption will rely on consumers’ attitudes toward transacting in native contexts and the related factor of how well a vendor implements Contextual Commerce.

1. Properly interpreting the context is important. The contextual medium should not be used as a mass distribution mechanism – in other words “spam” is annoying to begin with and it is even more annoying in the middle of a conversation between friends.

2. Concerns about data privacy.

3. Concerns about the security of payment credentials.


© 2016 Citigroup

28

A Disruption Accelerator Contextual Commerce is a disruption accelerator, in our view. By eliminating the sources of mobile commerce friction, it can accelerate the pace of the already rapid transition away from traditional “physical” model. Consider this – this year’s ~$79 billion mobile commerce market (using the lower end of the range of estimates we found) would be ~$101 billion with a 10% lower abandonment rate; ~$153 billion if the shopping cart abandonment rate was ~50% instead of ~74% and ~$306 billion if all of the abandonment was eliminated. This assumes, of course, that the average shopping cart value stays consistent, which is not data that retailers provide. The bespoke nature of Contextual Commerce contracts implies a multi-year adoption curve – so one good way to think of this is that the pace of m-commerce adoption should possibly accelerate and then stay elevated for years to come.

There are some clear beneficiaries from the “channel shift” we mention above..

1. Content providers, i.e., merchants that adopt Contextual Commerce techniques;

2. Context (not Content) providers – for example, social media networks that are investing in this trend benefit because they succeed in keeping their users “in platform” and become more valuable to advertisers. Mobile marketing affiliate revenues tend to lag desktop, so Contextual Commerce is an opportunity to bridge the gap;

3. Content Carriers, i.e., Logistics companies that carry the elevated content loads;

4. Payments processors that have the end-to-end capability to participate in all or part of the emerging Contextual Commerce ecosystem – vaulting, gateway, mobile and e-commerce processing, payment API, emerging capabilities like IoT payments and AR/VR-related payments;

5. Software companies that provide payments security, analytics and AI capabilities; and

6. Newer (as yet unknown) companies that may emerge due to the contextual commerce paradigm shift.

Conversely, merchants that have lagged in digital adoption and are not considering investments in Contextual Commerce and payments processors inordinately skewed to a physical infrastructure (likely to be smaller independent sales organizations) may suffer.


© 2016 Citigroup

29

3. Direct-to-Consumer Marketplace Another Threat to Retail The global apparel market is worth $1.65 trillion annually, of which 12.1% is now serviced by online retailers, up from 4.4% five years ago having grown at +25% annually. While this change has been rapid and has resulted in some significant disruptions to existing business models it has primarily focused on the last stage of the apparel industry value chain – the consumer facing retail stage. The advent of direct-to-consumer marketplaces has the potential to disintermediate many retail outlets completely.

Figure 16. Global Apparel Market Size and Online Penetration Figure 17. Global Online Apparel Market Size

Source: Euromonitor International Source: Euromonitor International

Apparel Industry Structure In the traditional apparel model, manufacturers produce garments that are either designed in-house or made to specification by designers or retailers. These are then supplied to wholesalers e.g. Nike, Adidas or directly to retailers e.g. Next, Macy’s.

Figure 18. Apparel Industry Value Chain


0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

$bn

$200bn

$400bn

$600bn

$800bn

$1,000bn

$1,200bn

$1,400bn

$1,600bn

$1,800bn

2010 2011 2012 2013 2014 2015$bn

$20bn$40bn$60bn$80bn

$100bn$120bn$140bn$160bn$180bn$200bn

2010 2011 2012 2013 2014 2015

Consumer

Retail Stores / Online

Export Chains / WholesalersIncludes brands and exporters

Apparel ManufacturersUsually third parties

Textile Plants

Raw Materials

Design feedback

Dan Fox Homan European General Retail Analyst


© 2016 Citigroup

30

The majority of apparel manufacturing is carried out in Asia where local wages and competitive markets keep the prices of garments low. There is however a reasonable amount of product that is made closer to the major end-markets of Western Europe and North America

In the last decade, there has been a large shift with consumers using stores less and increasingly shopping through an online channel. While this has been highly disruptive to retailers, the innovation was introduced years ago with the advent of online selling.

Large Consumer Audience Creates New Opportunities As online-only retailers have grown they have moved from being a minor disruptive force to major drivers of the market (see Amazon in the U.S. or Alibaba in China) as the foundation and growth of online aggregation websites create a virtuous circle between the website, the customers, and the brands. A lot of companies have been using the market share gained from their online presence to drive into new categories, expand their own-label supply base and move into new geographies. However even as they expand, the majority of the disruption that online-only retailers have caused has been at the last stage of the apparel value chain, namely replacing stores with online for customer focus.

Figure 19. Online Clothing: Virtuous Circle Exists Between Customers, Brands and Aggregator Websites

Source: Company Reports and Citi Research Estimates

Direct-to-Consumer Marketplace There is now an emergence of direct-to-consumer marketplaces. These models allow consumers to buy directly from apparel manufacturers rather than via a wholesaler or retailer. This has been made possible by technology which is helping marketplaces to link up the available inventory of many manufacturers without the need to directly own or control the product.

The advantage for the consumer is that the products are cheaper, there is a broader range of products and the payment/check-out systems are the same as if they went to an online retailer.

CustomersBroad selection with localized

content and offers

BrandsClear fashion focus & platform to

present brands in an engaging way

More Customers = More Brands

More Assortment = More Customers


© 2016 Citigroup

31

The manufacturers benefit from additional revenue outside of their normal sales stream to established retailers/wholesalers while acquiring the ability to directly control their product, develop their own brand, and expand their customer base. This is all done with very little technology investment.

Figure 21. Direct-to-Consumer Marketplace Advantages

Consumer Manufacturer

Fashionable items at lower prices Extra revenue

Good quality Further control

Large selection Brand development

Own customer base

Limited technology investment


The benefit of this direct-to-consumer model is that manufacturers will generate considerably increased profit per garment (as shown below) and will be able to offer consumers better prices.

This would not be without risk for the manufacturers as all the inventory risk would move from the retailer to the manufacturer. It may also cause planning issues for the manufacturers as demand becomes uncertain and cash flow would become more volatile.

Figure 22. Illustrative P&L for Retail vs Marketplace

Supply to Retail Direct to Consumer Revenue 50 100 Commission - (10) Cost - (23) Net revenue 50 68 COGS (40) (40) Operating profit 10 28 Margin 20.0% 27.5%


The obvious loser from this transition would be the current retailers that source directly from manufacturers. These companies would no longer be the only customer for the manufacturers. The convenience that they currently offer to consumers by selecting and merchandising a selected edit of goods would be replaced by a technology enabled website that could potentially link into social media and use algorithms/automated editing to determine customer preference.

Figure 20. Global Apparel Industry Sales

Source: Euromonitor International

Asia Pacific

36%

North America

23%

Western Europe

22%

Latin America

7%

Middle East and

Africa6%

Eastern Europe

5%Australasia

1%


© 2016 Citigroup

32

Opportunity and Development The combined size of the North American and Western European apparel markets is $700 billion annually. Citi estimate that around 20-30% of apparel sold is sourced from within or close to these markets. If this entire proximity-sourced product moved to direct-to-consumer marketplaces then it would create a $200 billion annual revenue opportunity.

There would be scope for this to increase if a reduction in the retail channels led to better economics for manufacturers which allowed them to expand operations in more economically developed regions.

Although the technology required to implement this change already exists and there are a number of websites with a large enough active user base to generate the necessary traffic, it is unlikely that direct-to-consumer marketplaces will become a significant part of the market in the medium term. The amount of risk and lack of visibility that would need to be assumed by the manufacturers is likely to limit the appetite to push this channel at too great a speed.


© 2016 Citigroup

33

4. Epigenetics Immunotherapy’s New Best Friend Epigenetic breakthroughs are quietly creating a tool box of powerful drugs to treat cancer and many other potential indications. Importantly, epigenetic drugs, which can switch genes on and off, may materially increase the percentage of patients likely to experience profound increases in survival in response to immunotherapy. Several epigenetic drug classes may materially broaden and deepen the efficacy of cancer immunotherapies, such as PD1 antibodies which enhance the body’s immune response to tumors. Low dose inhibitors of the enzyme catalyst DNA methyltranferase (DNMT) may improve immune priming by triggering the transcription or copying of repressed viruses long embedded in the human genome. Separately, inhibitors of EZH2, HDAC and FAK improve the tumor microenvironment through suppression of inhibitory immune cells. While decades of immunotherapy research (The Beginning of the End for Cancer) materially pre-date epigenetics, powerful diagnostic assays for the genome and the epigenome will rapidly drive development of novel epigenetic drugs

Almost all the cells in our body contain the identical inherited DNA sequence. Despite this, some cells become pancreatic cells, some become heart cells, while others become hair cells. Although the DNA in our cells is not subject to reversible change, epigenetic modification allows the cell to selectively and reversibly turn on or off differential gene expression (which determines whether the cell is a heart or hair cell). Our DNA serves as a ‘hard drive’ of genetic information that requires the ‘software’ of epigenetic control to determine cell functions. It is now understood that the initiation and progression of cancer, traditionally seen as a genetic disease, involves epigenetic abnormalities along with genetic alterations. Importantly, reversible epigenetic changes have the potential to reverse cell abnormalities and reprogram tumor cells. Key components of the epigenetic machinery in cancer including DNA methylation, histone modifications, nucleosome positioning and non-coding RNAs, specifically microRNA expression can potentially translate into material efficacy benefits compared with current standards of care. Epigenetics induce wide ranging and non-specific changes in gene expression profiles. We estimate the market for epigenetic approaches in cancer treatment will likely exceed $10 billion by 2025, driven by novel agents, combination therapy, long treatment times, and the advancement in genome and epigenetic mapping.

Figure 23. Potential Mechanisms for Episensitization to Cancer Immunotherapy


Andrew Baum, MD Global Head of Healthcare Research

https://ir.citi.com/bMkWjteBW7m96SwoQA6KpxzXcrCkuKY6%2b5AvD7lNAJ%2fxRPFTa5%2boFh%2fajE8yc4%2b45BCCUvnQXds%3d


© 2016 Citigroup

34

Figure 24. Mechanisms of Epigenetics


Cancer is the developed world’s second most common cause of death. We believe epigenetic drugs will likely become an important part of next generation cancer management regimes in the developed world. While epigenetic agents will not form the backbone of 21st century cancer care like immunotherapy (Immunotherapy: Bigger, Broader, Bladder), we anticipate mono and combination epigenetic therapy will be used in multiple indications leveraging the multiple mechanisms of epigenetic drug activity. We highlight that compared immuno-oncology agents, epigenetic modifiers are likely to have different levels of activity in different tumor types given their promiscuous action. Consequently, we anticipate a more uneven news flow compared with the ongoing and growing tidal wave of news on cancer immunotherapy.

Future Development Outlook for Epigenetic Oncology Drug Development We anticipate the implantation of immunotherapy as the future treatment backbone for the majority of advanced cancers will intensify efforts to use epigenetic drugs such as hypomethylating agents, and potentially HDAC and EZH2 inhibitors as immunopotentiators, especially in patients with poorly immunogenic tumors. Potential mechanisms for epigenetic drugs to defeat cancer include:

https://ir.citi.com/f%2fDHDIM4p3xvsAGrzafMTjb10dBMfaSOebUUIZtT5h%2bE9MTFbUpGqD8v4gF93iVadq5bf7DcBrs%3d

https://ir.citi.com/f%2fDHDIM4p3xvsAGrzafMTjb10dBMfaSOebUUIZtT5h%2bE9MTFbUpGqD8v4gF93iVadq5bf7DcBrs%3d


© 2016 Citigroup

35

Cancer with activating mutations. Small subgroups of patients with cancers secondary to activating mutations. Examples include BRD activation triggering NUT carcinoma (an aggressive skin cancer), subgroups of non-Hodgkin’s lymphoma (NHL) and small cell lung cancer (SCLC). Epigenetic drugs can represses these activating pathways.

Cancers with inactivating mutations. Larger patient populations with common inactivating mutations of nucleosome remodeling proteins such as SWI-SNF, RSF, and NURD/CHD. Epigenetic approaches may induce cancer control in these patients by a process known as “synthetic lethality”21.

Reversing resistance. Epigenetic drugs such as BRD4i may reverse acquired resistance to multiple first line ‘standard of care’ therapies including small molecules TKI, hormone receptors antagonist, and chemotherapy through down regulation of the myc and bcl2 oncogenes.

Eliminating cancer stem cells to prevent recurrence. Stem cell targeting epigenetic therapies has the potential to eliminate minimal residual disease in both early and advanced cancer.

Enhancing immunotherapy such as the anti-PD1’s through enhanced immunogenic response (epigenetic priming) or diminished immunosuppression in the tumor microenvironment.

Figure 25. Five Potential Mechanisms for Epigenetic Drugs in Cancer


While exploration of the role of epigenetic drugs as immunopotentiators will likely dominate near term industry activity, epigenetic drugs also have potential utility in multiple solid (i.e. breast or prostate cancer) and blood cancers (i.e. leukemia) through multiple treatment pathways including cancer with activating mutations, cancers with inactivating mutations, reversing resistance, and eliminating cancer stem cells. In particular there is much excitement over the BET inhibitors, a class of drugs with anti-cancer immunosuppressive effects, especially selective BRD4i agents. The ability of BRD4 to inhibit selective commonly expressed oncogenes such as myc and bcl2 offers significant potential both as monotherapy (i.e. on its own) but also in combination with existing approved cancer therapies to reverse acquired resistance. Separately, BRDi are under active investigation for pulmonary arterial hypertension and other non-oncology disorders, including (1) autoimmune disease, (2) Alzheimer’s, (3) seizures, and (4) HIV.

21 Synthetic lethality arises when a combination of mutations in several genes result in cell death. It can be a combination of two non-lethal genetic mutation events (synthetic sicknesses), or apply when the combination of a mutation and the action of a chemical compound causes lethality.

I. Inhibiting cancer activating mutations

II. Opposing cancers with inactivating mutation

through synthetic lethality

III. Reversing resistance to TKI

inhibition

IV. Eliminating Cancer Stem Cells

V. Enhancing Immunotherapeutics


© 2016 Citigroup

36

We anticipate 5 waves of commercial success for epigenetics, including: (1) mono or combination epigenetic therapy in NHL (lymphoma) and SCLC (lung cancer); (2) overcoming acquired resistance to oral cancer drugs such as Sprycel and Tarceva with BRD4i to inhibit c-myc cancer oncogene; (3) eliminating cancer stem cells to prevent disease relapse; (4) broadening and deepening response to cancer immunotherapy through better epigenetic priming and enhanced tumor microenvironment; and (5) use in non-cancer indications, such as BRD4i in pulmonary hypertension and HIV.

Figure 26. Market Estimates for Epigenetic Opportunities are Largely Absent from Financial Models But Could Easily Exceed $10 billion p\Per Year in Cancer Indications Alone


-

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

Sale

s ($

bn)

US EUUS EU

Initial Indications Resistance reversal CSC elimination

IO Potentiation


© 2016 Citigroup

37

5. The Future Look of Devices Reshape, No Shape We believe consumer devices in 2021 could be reshaped as thin and flexible as a piece of paper or could even be shapeless, thanks to augmented reality and mixed reality.

Reshaping the Future with OLED From mainframe computers in 1960 to smartphones in 2007, computing devices have shrunk significantly in size. This is partially driven by Moore’s Law which predicts the number of transistors per square inch doubles every 18 months. Over the next 5 years, we expect the display industry to produce a substantial change in terms of form factor, with the development of organic electroluminescent display (OLED) technology. Currently, glass and metal casings together contribute a significant portion of the weight of a smartphone. With OLED, both can be scrapped, allowing devices to be as thin and light as a piece of paper – as shown in Figure 27.

OLED could become the next generation of displays for mobile applications. The form factor of future mobile devices will be different and In the future, OLED displays can be used for phones that are bendable. Compared with rigid OLED which uses glass substrates, flexible OLED uses plastic substrate (i.e. polyimide or PI) – these plastics are thin, light, power-efficient, and unbreakable. However, the flexibility of devices is just starting to evolve. The first generations of devices that use flexible OLEDs are not really flexible from the end-user perspective. The displays are bended/ curved by their manufactures, but the final end-users are not able to bend the devices themselves. When the technology is ready, we may see OLED panels that end-users can fold, bend or stretch.

In early June 2016, Samsung commented that its bendable phone screens are “relatively right around the corner.” The vendor once showed a gadget that unfolds like a wallet in one concept video – as shown in Figure 28. We believe this type of device is most likely to be released first, followed by paper-like devices. The wallet-like gadget has a flexible screen, but it still houses a number of rigid components and a casing. We believe this type of device may be released in 2017-18.

Figure 27. Flexible OLED panel made by LG Display Figure 28. A foldable device from a Samsung concept video

Source: SlashGear Source: Samsung, Business Insider

Arthur Lai Greater China Hardware Sector Analyst Dennis Chan Asian Handset Sector Analyst


© 2016 Citigroup

38

Augmented Reality In terms of function, we envision future devices will enable augmented reality functions which allow users to view and touch objects from different angles in an almost realistic fashion and to use a gyroscope to adjust views and aspects – as shown in Figure 29. With voice and gesture recognition functions, input/output (I/O) will be no longer be needed. Longer term, we also imagine future devices which are invisible – using regular glasses/ contact lenses as projectors to enable augmented reality. Moreover, with 5G connections, storage and computing power can be stored on a cloud, thereby saving components and enabling further power saving.

Figure 29. You Can Touch 3D Hologram Figure 30. AR: Computer Graphics Overlaid onto Real World Scenes

Source: 123RF Source: By Edward (via Wikimedia Commons)

OLED Obstacles

To enable the function of OLED devices, we imagine them as being virtual private network (VPN) tools controlling a computer miles away. It’s like using a notebook to control a server sitting in your office while you are traveling – all you need are the very basic components to show the progress of the machine (i.e. a screen) and to take the commands (i.e. voice/gesture control or touch screen). In this case, we can keep other key components like the computer processing unit (CPU) and the battery, on the cloud, while connecting to it via 5G networks.

For now, we believe the major obstacle preventing a bendable, flexible type of device from coming to reality would be the connectivity issue, i.e. the availability of 5G networks. We also view the user interface and ecosystem as key concerns. The user interface needs to be user-friendly and we believe the device will require a very high data transmission module (i.e. a 5G connection) and a very powerful cloud computing ecosystem to support the real time interaction. So far, the 5G standard is yet to be clearly set with market consensus forecasting 5G to be commercialized in 2020 at the earliest. Despite this, the speed of 5G should be much higher versus the speed of the current 4G connection, with speeds higher than 1Gbps, vs. 100Mbps with 4G. Based on these limitations, we don’t see elastic OLED devices being available in the commercial market within the next five years.

Market Growth We believe that the current market is only focusing on devices that will be available over the next 1-2 years and not on devices to be released several years later.

Given that consumer electronic devices of the future will be flexible and not restricted in terms of form-factors, we believe they could cannibalize the demand for smartphones, tablets, PCs, and even TVs. As a result, we believe the potential total


© 2016 Citigroup

39

addressable market for these devices will be big – potentially being the size of the current consumer electronics market. With 5G becoming availability in 2020 at the earliest, we take 2020 as a starting point for our market analysis. In the Disruptive Innovations III report published in July 2015, Citi’s Japanese Consumer Electronics analyst Kota Ezawa forecasted the market size of the Augmented Reality / Virtual Reality could reach $200 billion in the first five years after its launch. We envision a similar market size for the future device.

Figure 31. Potential TAM for Future Devices


Casing and Disruption We believe the metal casing segment will be negatively impacted by the new OLED technology, given the future of devices that we envision will be as thin and light as a sheet of paper. Moreover, our view is that metal casing is no longer a premium and differential factor any more in mobile device design. Apple has been a leader in terms of form factor design. We identify several reasons why Apple might want to switch to a different casing material. 1) Lack of differentiation with iPhone’s current metal casing design: iPhone has adopted metal backcover design for 3.5 years since iPhone 5. At MWC in Feb 2016, we observed that over 80% of the new smartphones adopted metal casing and/or metal frames. This makes it harder for Apple to differentiate itself as a high-end phone with unique form factor. 2) Easier design-in of wireless charging: Given the shielding effect of metal casing, it may be easier to adopt a different material at the back if Apple wants to adopt wireless charging. 3) Potential improvement in glass casing/sapphire technology could mean a unique value proposition: The current glass casing /sapphire casing is prone to scratches and breaking. A breakthrough in glass/sapphire technologies would mean Apple and other high-end brands could adopt a glass casing that was lighter/thinner and yet more scratch-/drop-resistant. This could offer a unique proposition versus past glass designs (i.e., iPhone 4S).

0

0.5

1

1.5

2

2.5

2015 … 2020 … … 2035

0.2

0.9

2.3

related market

Film, DVD, DL game

e-commerce

Mobilecommerce

Mobile device

addressable market

(US$, tn)


© 2016 Citigroup

40


© 2016 Citigroup

41

6. The Smart BoxBattle to Be the Heart of the Connected Home In previous GPS editions we reviewed the opportunity from the Internet of Things (IOT) and the move toward connected devices – smart phones, smart watches, smart cars, smart homes. In this section we look at the opportunity from ‘home networking’ and the potential for one device (a ‘Smart Box’) or one provider, to become the gateway to the connected home.

Pay TV/broadband/mobile markets suggest that if a single player can control the consumer device or operating system that’s used to engage with the digital world, it can capture the lion’s share of the economics. Within the home, although the ‘hinterland’ of potential participants may vary – from pay TV, to retail or even home security – the opportunity is significant. As such we see scope for significant disruption across media, telecoms, technology, retail, leisure and beyond.

The Home Use Case Broadband connectivity is enabling customers to take greater control as well as enabling the integration of a variety of applications within the home. We think the Smart Box, which we can also refer to as the Home Hub, can become a disruptive integrator of technology and services for customers that will provide a variety of functions:

Connectivity. The assumption with the Internet of Things is that devices willindividually be connected to the Internet. This is compelling in as far as it goes,but the thinking behind the ‘home network’ is that one device or platform will actas a hub within the home either providing simple connectivity to third partydevices or actively controlling them. We also see scope for the Home Hub todisintermediate whole categories of technology. Who needs an expensivespeaker system if a Home Hub can easily act as a speaker or transmit music toremote devices and speakers? Who needs an expensive home integrationsystem if smart lights can be controlled via your TV remote control? Are youfrustrated that the WiFi router from the broadband provider doesn’t provideadequate coverage and speed across your home? The Smart Box can increasethe ease at which customers can place WiFi extenders and hot spots into ahome to improve the overall coverage and connectivity experience.

Functionality. A smart lighting system, like Lutron, allows consumers to switchlights on and off, manage energy consumption, and change or review settingsremotely. Smart security systems allow remote surveillance. A connected fridge,meanwhile, may remind you when you are running low on milk and place anorder. These are valuable services. But, they represent only marginal upgradesin functionality with a significant uplift in hardware, installation and service costs.

Media Integration. The future of a connected media device, however,represents a significant break from the past. Instead of a dumb boxdecompressing and unencrypting video signals (maybe with some enhancedfunctionality like DVR) or simply playing music, the Home Hub of the future willact as: (1) a recommendation engine both for content; (2) a distributor of contentto other devices both inside and outside the home; (3) a gateway to onlinecontent, social media, and gaming; and (4) a platform for e-commerce; aplatform for non-traditional applications such as streaming music, showingphotos, even boosting WiFi signals.

Thomas A Singlehurst, CFA Head of European Media Research Team

Jason B Bazinet U.S. Entertainment, Cable & Satellite Analyst

Michael Rollins, CFAU.S. Telco & Telco Infrastructure Analyst

Jim Suva, CPAU.S. IT Hardware & Tech Supply Chain Analyst


© 2016 Citigroup

42

Figure 32. Putting One Device at the Heart of the Home Network


Why Will Media Devices Be the Launchpad for the Connected Home? The proliferation of smart devices – from whole lighting systems to individual light bulbs, from the connected thermostat to the connected refrigerator – is a well-established trend. Our core thesis, however, is that consumer media devices could be the focal point of the true home network, either set top boxes (STBs) connected to your TV or maybe, like the Amazon Echo, something as simple as a standalone speaker. Hence, a ‘Smart Box’ may emerge that integrates a variety of services and connectivity into a single location. While more and more applications are available on new televisions and cable STBs, we believe the use of a ‘Smart Box’ at the center provides much greater flexibility for consumers to install these devices onto their existing equipment.

Why do we think the home network will start with consumer media devices e as opposed with other potential focal points in the home, e.g. thermostats, lighting systems or connected refrigerators? The answer is these are areas where consumers still spend the majority of their time. Looking at data provided by Ofcom in the U.K., the average consumer still spends almost four hours per day watching TV and over three hours listening to the radio. This dwarves time spent fiddling with heating or lights, setting alarms or getting snacks from the fridge.

Figure 33. Time Spent Using Communications Services in the U.K. (Minutes per Day)

Note: Time spent may be duplicated, i.e. spent on more than one device / service at a time. Source: Ofcom

Internet

TV

PC

Tablet

Refrigerator

SecuritySystem

HiFi/SpeakerSystem

Thermostat

LightingSystem

InternetHom

eHub

TV

PC

Tablet

Refrigerator

SecuritySystem

HiFi/SpeakerSystem

Thermostat

LightingSystem

Internet of Things Home Network

219 216 218 225 225 242 242 241 232 239

185 182 177 172 170173 175 170 166 183

7 12 14 17 1313 29 29 28

87

15 1469 59

2728

27 68 68

65

15 14

14 1313

12 11

10 9

19

0

100

200

300

400

500

600

700

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Aver

age

Tim

e Sp

ent P

er D

ay b

y D

evic

e

TV Radio Mobile phone Internet on PC/ laptop Fixed phone


© 2016 Citigroup

43

Is This Happening? So, is this happening? Yes it is – but consumers may not have appreciated the significance of how transformative it could be. In the table below, we consider four devices that target different use cases in the home but which ultimately target a much bigger market opportunity beyond each companies ‘hinterland’, or as Enders Analysis put it, ‘vocation’. In this context, Sky Q in the UK is clearly trying to move beyond being a simple pay TV proposition into being a more holistic home technology provider. Amazon, meanwhile, is using media and content to extend the reach of its retail services. Similar ambitions appear to be behind Apple TV (now in its third generation) and Google Home (yet to be launched).

Figure 34. Four Devices That Could Change Your Home

Source: Company reports, Citi Research

Sky Q Amazon Echo

Google Home

Functionality:- Delivery of linear / non-linear content- Access to 3rd party web apps (YouTube/Vevo/Facebook)- Access to transactional services via Sky Store- Acts as WiFi hotspot/network for home via Sky Q mini boxes- Bluetooth touch remote with voice control- Music streaming via Airplay

Functionality:- Music streaming / Internet radio- Voice control via Alexa- Access to 3rd party web apps (e.g. Accuweather, Uber)- Ability to connect to/manage 3rd party smart-home devices- Access to transactional services

Apple TV

Functionality:- Music streaming / Internet radio- Voice-based control and access to search- Access to 3rd party web apps- Ability to connect to/manage 3rd party smart-home devices- Access to transactional services

Functionality:- Delivery of non-linear TV content via apps- Access to transactional services via iTunes- Access to 3rd party web apps (e.g. Netflix)- Video gaming capability with games via App Store- Touch remote with voice control by Siri


© 2016 Citigroup

44

Sizing the Market The ‘hinterland’ from where each player hails will have a key bearing on the size of the incremental opportunity. Reflecting this, we think about the opportunity in terms of concentric circles starting with the core addressable market and moving outward toward ancillary markets. There will inevitably be some dispute about what belongs at the center of the circle, but we think this will be for the most part home entertainment, and for many this will be the subscription pay TV market.

In Figure 36 we show our estimate of the total addressable market for Home Hubs in the U.K. which incorporates the core pay TV market with ancillary home leisure segments (transactional home entertainment, home delivery plus music streaming), advertising markets (both free-to-air (FTA) and Digital), broadband connectivity, home technology (home security as well as consumer expenditure on brown and grey goods), and retail (non-grocery e-commerce).

The upshot is that we estimate the total addressable market in the U.K. could be as much as £86 billion ($113 million) per year. For Sky this is over 10 times its core market. A similar analysis for the U.S. suggests the total market could be as much as $727 billion. (see Figure 35). The potential prize for being the gateway of the home is significant.

Figure 36. The Market Opportunity for the Connected Home in the UK Could be as Much as £86 Billion

Source: Company reports, GroupM, Euromonitor, Forrester, IFPI, UK Government, Citi Research

Figure 35. Market Opportunity in the U.S. Could Be As Much as $727 Billion

Category $ Billion U.S. E-commerce 342 Consumer Technology 75 Home Security 8 Music Streaming 5 Home Delivery/Takeaway

23

Home Entertainment 3 Digital Advertising 60 FTA TV Advertising 36 Fixed Line Broadband/Telephony

64

Pay TV 112 = Total Market 727



© 2016 Citigroup

45

Factors to Drive Success We think there are three high-level keys to success in this market:

First, we think it is important that a successful player has a strong line-up of devices. In the first instance this will be the central hub for the home and as we have argued it makes sense that this will be a device that firstly sits at the heart of the home – most likely the kitchen or the living room – and that it provides functionality associated with every day use. As above, music, television or gaming make a lot more sense than thermostats, home security systems and white goods all of which are likely to be located away from the main communal areas and probably only take up minutes of most users’ time. We actually think the point about devices extends beyond the central home hub to other complementary devices.

Second, we think it is essential that the devices are underpinned by a strong underlying software platform which includes an inter-operability strategy, machine learning capabilities, likely voice functionality, and an operating system that is prevalent in other environments. As Google has shown in mobile, it is entirely possible to end up dominating a platform without necessarily controlling all of the devices.

Third, we think it would be helpful to have a robust ecosystem around the central Home Hub. This is both a virtual ecosystem in the sense that we think it is important that there is access to a broad range of apps – i.e. the ability to order an Uber via the Amazon Echo – but also a physical ecosystem in the sense that the Home Hub can connect to other smart devices in the home whether they are provided by the same company or by third parties whether these be thermostats, fridges, washing machines, dishwashers, TVs, phones etc.

Figure 37. Considering the Home Hub Opportunity / Landscape vs. Mobile / Desktop

Desktop Mobile Home Hub

Devices Multiple Apple, Samsung Apple, Amazon, Google, Sky, Comcast,

???

Operating System Windows, Linux, iOS iOS, Android iOS, Android, Alexa, ???

Ecosystem Keyboards, mouse, monitor, speakers

Wireless speakers, battery packs

Wireless speakers, TVs, smartphones, games consoles,

thermostats, white goods, home security

systems, lighting systems


If we consider the companies and products we mention in Figure 37, one that appears to have a solid head start and is a good example/template for others is Amazon. What Amazon has done is, in our view, incredibly impressive. Specifically we note:

The work the company has done on Alexa, both the voice service/skills kit and the Alexa-powered home networking hub devices already in the market and gaining share


© 2016 Citigroup

46

The partnerships that Amazon has already struck with third party web-based applications. The partnerships with Uber, Dominos, Spotify, Pandora, ESPN as well as Amazon’s own hosted applications – in particular Prime Music and Audible – significantly enhance the functionality of the Amazon Echo platform.

The interoperability of the Amazon platform with third party hardware. Amazon itself highlights the deals it has done with Samsung SmartThings (TVs and other brown goods), Philips Hue (lighting systems/lightbulbs), Wink, Wemo and Insteon (plugs/switches/ centralized remote control systems) which allow Amazon to be the centralized point of (voice-based) control for the entire smart home.

Barriers to Adoption There are a number of reasons why some of the consumer technology companies may struggle to fulfil the opportunity from the connected home/home networking:

First, it may be difficult to unpick some of the components of the value chain. Take for example basic connectivity: incumbent telecoms/cable companies insist consumers use their devices – routers and STBs – to distribute WiFi signals around the home or de-compress and decrypt linear video signals. This may well change in part driven by regulation – we note the Federal Communications Commission’s (FCC’s) Notice of Proposed Rulemaking (NPRM) in the U.S. potentially presages exactly this – but it will likely be a while before cable/telco companies are relegated to being a ‘dumb pipe’.

Second. the cost of the devices may be prohibitive. If this is the case, as with the early days of mobile, it is possible that hardware will have to be subsidized / leased in order to drive mass take up, something we are already seeing with Sky in the U.K..

Third, firms may shun an open architecture with open standards. Already there are signs that the various players are trying to extend the walled garden to the home hub. Amazon has locked out Chromecast from the Amazon marketplace. In response, Amazon Prime Video is not available on Chromecast. Enders Analysis makes the point that there is a high likelihood Google will do the same on Home. These strategic choices by operators may make sense to the individual operators but is anathema to the consumer and could inhibit take up.

Fourth, privacy may matter. For the full transactional potential of products like Amazon Echo, Google Home or Sky Q to be fulfilled, consumers will need to surrender information to companies. Moreover, with always-on voice control, what consumers do and say in their own homes will be constantly monitored and recorded. Even if consumers are relaxed about this, regulators may not be.


© 2016 Citigroup

47

Winners and Losers We believe that the big potential players in the space may come from one of four areas:

Incumbent cable/pay TV players/telecoms companies are clearly well placed to benefit as they have scale relationships with consumers and content providers and a subscription-based model that would be a solid basis for a Home Hub offering. The question is whether they have the technological knowhow to deliver best-in-class consumer grade technology at scale.

Specialist hardware manufacturers, companies like Apple, Sony and Samsung, look well placed in terms of their ability to deliver attractive consumer technology into homes at scale, but potentially lack the content with which to prime the service and a subscription- or even transaction-based model with which to monetize the broader ecosystem.

The large scale internet players, with Google the most obvious example, may eventually be able to control the entire ecosystem – or at least take a decent proportion of the economics – by providing a standardized operating system. In short, if Android can dominate the Home Hub space in the same way it dominates mobile, Google could end in a very strong position even if it is not the main hardware provider.

E-commerce providers, and in particular Amazon, are clearly going to be actively involved in the space because, as we show above, the e-commerce opportunity is ultimately one of the biggest/most attractive potential markets.

While it is not clear who will win at this stage, there are a number of different markets that we think could be potentially disrupted if home networking gains scale and the connected home moves from concept to reality. We highlight the following areas:

Traditional linear content businesses within the media landscape. The move toward more home networking presages more non-linear viewing and with it, attendant pressure on ad based revenue.

Certain revenue lines for incumbent telecoms/cable providers, in particular charges for set-top boxes and STB-related functionality, could be at risk. As our U.S. research team point out, losing control of the UI could wipe out almost 100% of the U.S. cable companies’ free cash flow.

Providers of niche standalone consumer technology offerings. Anything from high-end wireless speaker systems, lighting systems, and even home security systems could be at risk of disintermediation from more sophisticated Home Hub solutions.

Retail, both traditional and online, could be impacted if providers of Home Hub services can exploit their direct relationship with consumers in the home.


© 2016 Citigroup

48


© 2016 Citigroup

49

7. Next Gen Ocular Delivery Sustained Drug Delivery – The Future of Ocular Medicine A new wave of technologies is leveraging the eye as a natural, self-contained drug delivery chamber. These new technologies will change the way physicians approach the treatment of ocular disease in the next 5-10+ years by moving toward sustained vs. intermittent approaches to therapy. The shift to sustained drug delivery in ocular medicine will have a significant disruptive impact on traditional therapeutic markets in ophthalmology where eye drops and injections are currently the mainstays of therapy.

The burden of eye disease is large and growing rapidly. Three big classes of eye disease (glaucoma, retina, dry eye) will affect ~21 million people in the US by 2020. The affected population is growing rapidly with ~34 million expected to be affected by 2050. Eye diseases increase in prevalence as people age.

The worldwide market for three major ophthalmic drug classes currently stands at $16 billion and is expected to grow to $29 billion by 2030. Treatment for ophthalmic conditions has and continues to be dominated by episodic intervention either via drops or injections.

Figure 38. Desirable Attributes for Next Gen Ocular Drug Delivery Figure 39. Ocular Disease Prevalence Increases Significantly With Age Attribute Advantage

4-12 month delivery Reduces office visits and injection burden

No adverse or minimal side effects Avoids causing secondary condition (glaucoma, cataract)

Ability to vary dose Customize dose by patient, stop delivering drug completely if needed

Minimize intraocular debris Prevent inflammation, floaters

High patient compliance Improved outcomes

Cost-effective manufacturing Commercial viability and scalability with the large ophthalmology market


The Problem What's wrong with today's drugs for ocular diseases? In the case of injections, treatment requires an office visit and an injection procedure that carries potential risks such as retinal detachment. Eye drops are significantly limited by natural obstacles to delivery including physical barriers in the anatomy of the eye and tears that wash away the majority of the delivered drug before it is absorbed.

Next generation (Next gen) delivery overcomes these limitations by reducing treatment frequency and increasing drug availability.

0%

2%

4%

6%

8%

10%

12%

14%

50-54 55-59 60-64 65-69 70-74 75-79 80+Glaucoma prevalence AMD prevalence

Yigal Nochomovitz, Ph.D. U.S. Mid/Small Cap Biotechnology Analyst


© 2016 Citigroup

50

Figure 40. Current vs. Next Generation Delivery Technologies


Next gen technologies address both anterior (front of the eye) and posterior (back of the eye) delivery. Implants or inserts can be loaded with a drug and released in a sustained or controlled manner to the back of the eye over a period of months or years. Eye drops that incorporate nanoparticles or liposomes can improve delivery over traditional eye drops, or contact lenses can provide a depot for sustained release to the front of the eye.

Figure 41. Global Market Opportunity: Including Glaucoma, Retina, and Dry Eye

Source: Citi Research, Allergan filing, Visiongain, Presentation by Cunningham at F.C.Cordes Eye Society Meeting 2015

Recent advances in ocular drugs have seasoned the market for next gen delivery. Although a handful of sustained delivery technologies have been approved in the past, their impact has been limited. However, the use of anti-VEGF drugs in wet AMD (age-related macular degeneration) has grown rapidly in recent years (i.e., Eylea, Lucentis, Avastin) and changed the landscape in ocular therapy. The new wet AMD therapies require monthly injections. However, the treatment burden could be mitigated by sustained delivery technologies. Thus, the time is now for development of new and effective ocular drug delivery technologies to complement the emerging treatment landscape.


© 2016 Citigroup

51

The underlying science and technology have advanced. Implants, inserts, and depots has been limited by technology hurdles in the past, but new chemistry and engineering has driven down size and driven up effectiveness of next gen delivery technologies. Thus, the foundation has been laid for meaningful advancements in the field in the next 5-10 years.

Figure 42. Affected U.S. Population with Glaucoma, Retina, and Dry Eye Figure 43. Perspectives on Ocular Drug Delivery

Source: Citi Research, National Eye Institute, Dry Eye: A practical approach, Springer-Verlag, 2015


Figure 44. Glossary of Key Terms

Disease Definition Current Treatment Glaucoma Buildup of pressure in the eye that damages the optic nerve.

Since the optic nerve sends vision signals to the brain, enough pressure/damage leads to vision loss

Eye drops to reduce fluid or increase outflow, laser surgery to increase fluid outflow from eye, microsurgery to drain fluid

Dry Eye Tear imbalance leading to pain, light sensitivity, itching, redness, blurry vision.

Eye drops/ointments to lubricate, temporary punctal occlusion to prevent tear drainage, punctal plugs for longer punctal occlusion

AMD Age Related Macular Degeneration. Leading cause of vision loss in elderly, caused by deterioration of the retina, which sense light.

Anti-angiogenic drugs, laser therapy, photodynamic therapy, vitamins (B, C, β-carotene, zinc, copper) all for wet AMD.

Neovascular (wet) AMD Advanced form of AMD when new blood vessels being to form in the choroid (beneath the macula). Leaky blood vessels cause

vision problems including vision loss as abnormal blood vessels begin to scar.

Same as above

Choroidal Neovascularization The formation of new blood vessels in the choroid during wet AMD progression.

Same as above

Geographic Atrophy Describes loss of photoreceptor cells in the macula in dry AMD. Photoreceptor cells cannot regenerate so once enough are lost,

vision begins to deteriorate.

None approved

Retinitis Collectively describes the group of diseases in which the retina is damaged, including retinitis pigmentosa (genetic cause) and

CMV retinitis (caused by viral infection).

Vitamin A, preventative care (avoidance of UV light), antiviral regimen for CMV retinitis (HAART)

Diabetic Macular Edema In diabetic patients, high blood sugar damages blood vessels in the eye which can cause leakage, leading to vision problems.

Anti-angiogenic drugs, laser therapy, vitrectomy (to clear out fluid in the eye), plus treatment of underlying diabetes

Uveitis Inflammation in the middle layers of the eye (uvea, which includes the iris, choroid, and ciliary body) caused by immune disorders, infection, or spontaneously. Can lead to irritation,

blurred vision, pain, light sensitivity, floaters.

Eye drops or pills (steroids to reduce swelling), antibiotics for infectious uveitis, sunglasses

Macular Telangiectasia (MacTel) Blood vessels become dilated in the macula, leading to deteriorating and scarring and eventually vision loss

laser therapy, intravitreal injection of steroids or anti-angiogenics

Source: WebMD, Citi Research


© 2016 Citigroup

52

Figure 45. Next Gen Ocular Drug Delivery Approval Figure 46. Next Gen Ocular Delivery Technologies Come in All Shapes and Sizes

Source: Company reports, Citi Research Source: Citi Research

The Opportunity Next gen ocular delivery companies haven’t quite reached prime time yet. The competitive landscape is fairly wide open. A handful of public companies play in this space but most remain private and venture capital-funded. Relative to the overall worldwide market opportunity (~$22 billion in 2020), investment in the sector is nascent. This creates investment opportunities as technologies begin to mature and more clinical data materialize.

Figure 47. Majority of Ocular Delivery Companies are Not Yet Known to the Public and are Supported by VC Funding


Ocular drug delivery is rich with pipeline assets across all stages of clinical development. The market is big enough to accommodate many winners. A diverse set of assets across many eye diseases ranges from pre-clinical to post Phase 3. See our full page table (Figure 49) highlighting the accelerating progress in the field. We expect that over the next decade we will see multiple new approvals for sustained delivery drugs to the eye. This progress should expand the current set of just four sustained release intraocular drugs approved over the last 40 years of drug development to many-fold that, and in one fourth the drug development time.

Ocusert (Glaucoma)

Vitrasert (CMV Retinitis)

Retisert (non-infectious uveitis)

Ozurdex (Macular edema, Uveitis, RVO, DME)

Iluvien (DME)

Next Gen Technology

Nano

Micro

DexNP (nanoparticle eye drops)ENV515 (nanoparticle for extended release)

DexNW (drug loaded nanoreservoirs)ProDex (liposome based)

RetisertVitrasertOzurdexIluvein

Durasert

NormalOcusert (ring insert)

Ocu-ject, Oct-mix, CLS-1001 (Needle Technology)EGP-437 (Eyegate)

Mydriasert (insert)DextenzaOTX-TP

ENV905 (implant) Micropump (implantable drug pump)

Implants Depots

$0

$50

$100

$150

$200

$250

$300

$350

$400

VC funding ($M) Mkt cap ($M) on 06/15

Graybug

Clearside Biomedical

Neurotech

IconBioscience

Envisia

ForsightpSivida

Ocular

Eyegate

Alimera

Replenish


© 2016 Citigroup

53

Figure 48. Sustained Ocular Drug Delivery Provides Smoother Drug Exposure


Bioa

vaila

bilit

y

Time

PeriodicTreatment

Suboptimalavailability in trough

Safety risk from peak availability

Sustained-releaseachieves constant, optimal availability


© 2016 Citigroup

54

Figure 49. Summary of Next Gen Ocular Drug Delivery Technologies in Development

Product Company/institute Technology description Drug delivered Target diseases Development stage Estimated number of

years to reach market

Dextenza Ocular Resorbable sustained release depot

Dexamethasone Post-surgical ocular pain & inflammation

Under FDA review 1

IBI-10090 Icon Bioscience Biodegradable, intravitreal injection, based on Verisome

technology

Dexamethasone Cataract surgery inflammation

Ph. 3 completed, NDA filing in process

1-2

Medidur pSivida Injectable micro-insert NA Posterior uveitis Ph. 3, NDA expected mid 2017

2

EGP-437 EyeGate Utilizes a low-level electrical current to deliver a specified

amount of drug (iontophoresis), based on EyeGate II technology

Dexamethasone Noninfectious anterior uveitis and macular edema

Ph. 3 2-3

Bimatoprost SR Allergan Biodegradable, intracameral implant

Bimatoprost Glaucoma Ph. 3 2-3

CLS-1001 Clearside Biomedical

SCS microinjector to deliver drug precisely

Zuprata Macular edema associated with non-infectious uveitis

Ph. 3 2-3

OTX-TP Ocular Resorbable sustained release depot

Travoprost Glaucoma and ocular hypertension

Ph. 2 4-5

Bimatoprost ring Forsight Non-invasive, rest on the eye Bimatoprost Ocular hypertension, glaucoma

Ph. 2 4-5

Punctal plug drug delivery system

Mati Punctal plug Latanoprost Glaucoma Ph. 2 4-5

ProDex Taiwan Liposome Company

A lipid-based technology that carries drugs, based on BioSeizer

technology

Dexamethasone Macular edema Ph. 2 4-5

Ranibizumab delivery port

Roche/ForSight Refillable drug port delivery system (PDS) designed to release Lucentis over a period of months

Ranibizumab AMD Ph. 2 4-5

NT-501 Neurotech Ocular removable implant for delivering therapeutics, based on encapsulated cell therapy (ECT)

Ciliary Neurotrophic

Factor (CNTF) from NTC-200

Macular telangiectasia (MacTel)

Ph. 2 4-5

ENV515 Envisia Biodegradable PRINT nanoparticle formulation for extended release

Travoprost Glaucoma Ph. 2 4-5

DexNP Oculis γ-cyclodextrin nanoparticle eyedrops

Dexamethasone Diabetic macular edema (DME) and

posterior/intermediate uveitis

Ph. 2 4-5

Durasert Pfizer/pSivida A bioerodible, sustained-release product based on Durasert

technology

Latanoprost Glaucoma Ph. 1/2 6-7

DSP-Visulex Aciont Noninvasive electroosmotic delivery method

Dexamethasone Anterior Uveitis Ph. 1/2 6-7

MicroPump Replenish Refillable, implantable drug pump releasing nanoliter-sized doses

Range of drugs Diabetic macular edema Pre-clinical 8-10

GB-102 Graybug Biodegradable polymer matrix, intravitreal injection

Sunitinib AMD Pre-clinical 8-10

ENV905 Envisia Difluprednate Implant Difluprednate Inflammation and pain associated with ocular

surgery

Pre-clinical 8-10

Nanowafer Baylor Dissolvable polymer hydrogel film with drug-loaded nanoreservoirs

TKI drugs such as axitinib

CNV Pre-clinical 8-10

Dex-NW Baylor Dissolvable polymer hydrogel film with drug-loaded nanoreservoirs

Dexamethasone Dry eye Pre-clinical 8-10

JDE-003 Jade Therapeutics (now EyeGate)

Hyaluronic acid (HA) polymer platform as liquid gel drop

Thiolated-HA topical drop

(0.75%)

Corneal repair Pre-clinical 8-10

SKS sustained release technology

Ohr Hydrogel template to make microparticles, where drugs load

NA Glaucoma Pre-clinical 8-10

Microrobots EZH Coated tube movable in electromagnetic field

NA NA Pre-clinical 8-10

Microparticles Ohr Biodegradable, injectable microparticles with consistent size

distribution

OHR1031 Glaucoma Pre-clinical 8-10



© 2016 Citigroup

55

8. Open Source Robotics Open Source to Drive Robotics Beyond Its Automotive Roots The modern robotics industry has its roots with automotive customers. Robots were first used by General Motors in the 1960s and are now ubiquitous in auto assembly plants globally – penetration of robots in automotive welding for example is now >90%. Historically, robots have focused on high volume and highly repetitive tasks, but the robot industry is changing – industrial robots are becoming more flexible (through the introduction of collaborative robots) and there is a complete new industry in service and other non-industrial robots.

The history of a narrow suite of robot applications – such as material handling and welding - and a relatively concentrated number of suppliers has meant that robots have historically had proprietary software in the controlling interface. The surge of robot demand outside the factory, and the much wider variety of uses however has led to another development – the use of open source software to control robots. Open source software is that where the source code is made available to users to amend as they see fit, and usually entitles unrestricted re-distribution to others. This is a benefit to developers in a variety of ways ranging from collaborative development to standardized robotic interfaces, but also potentially represents a lost revenue stream for any would-be provider of proprietary code.

With the market for robots in general industry expected to double by 2020, this ability to share development costs is arguably as important for the customer as payback period calculations on the hardware.

Figure 50. Worldwide Supply of Industrial Robots (‘000 units)

Source: IFR, Citi Research

0

50,000

100,000

150,000

200,000

250,000

2009 2010 2011 2012 2013 2014

Martin Wilkie European Capital Goods Analyst Graeme McDonald Japanese Machinery & Shipbuilding Analyst


© 2016 Citigroup

56

Open Source for Non-Industrial Looks Set to be the Norm Unlike their industrial brethren, service and consumer robots have no existing legacy infrastructure, are more likely to operate on a standalone basis, are more likely to be used by smaller customers (with smaller software budgets and less in-house expertise), and are likely to be used on a far wider range of applications. This all lends appeal to the potential for open source. As robots get more complex – and the number of hardware suppliers increases – the appeal of collaboration on software also increases. This is potentially hugely important – incumbent robotics companies have substantial domain expertise in automotive markets, but arguably far less in more diverse general industrial markets.

Figure 51. Around 90% of Robots are Currently Used for Industrial Applications

Figure 52. Service Robots Account for ~10% of Units

Source: Citi Research, IFR data (2014) Source: Citi Research, IFR data, 2014

One key benefit is that the software becomes portable across hardware systems. According to the IEEE, robotics software has historically contained algorithms which theoretically could be reusable, but have not been in practice as they have been so intertwined with the hardware specific operating system. Open source systems that separate the “messaging” to the robot from any hardware specific software solves this problem. Examples include collaborative platforms like ROS (Robot Operating System), part of the Open Source Robotics Foundation and which stemmed from Stanford University. In 2012, Rethink Robotics launched its Baxter two-armed robot, aimed at small and medium industrial customers, which runs on the open source ROS platform for academic applications, although is sold with its proprietary “Intera” software for industrial customers.

There are other open source platform possibilities – in May 2016, Softbank announced that its humanoid “Pepper” robot would be compatible with Google’s Android SDK (software developers kit), allowing certain parts to be open source. Pepper’s own operating systems is already based on Linux, an open-source operating system.

Sizing the Opportunity Putting a dollar value on a free product might seem counterintuitive, but we see the value of open source as coming from driving hardware penetration, as customers benefit from the shared development costs of robot applications.

Automotive43%

Electrical/ Electronics

21%

Other36%

Defense46%

Milking21%

Logistics11%

Mobile Platforms

7%

Medical5%

Other10%


© 2016 Citigroup

57

The World Robotics Federation (IFR) estimates that the world market for robots stood at 229,000 units in 2014, with a market value of $10.7 billion. Most robotics companies do not separately disclose their software revenues, but the IFR sees the value of software, peripherals, and systems engineering collectively to be about three times the level of hardware sales, or around $32 billion.

The U.S. based Robotics Industry Association forecasts growth of cobot units (collaborative robots) to reach 40,000 units and $1 billion revenue by 2020. We think that the unit forecast looks potentially too low, although with an average cost of closer to $15,000 per unit, the value figure could potentially be too high.

Arguably the bigger opportunity is for general industry more broadly; non-automotive end markets account for close to 60% of robot unit shipments, where industry players see the market doubling by 2020. This is the equivalent of ~$6 billion of incremental non-automotive robots annually.

Lessons From Other Industries – How Successful has Open Source Been? It’s hard to predict the precise potential success of open source for robotics, although we can glean some insights from open source successes in other areas. The penetration of open source software varies enormously by application – penetration of open source in desktop operating systems is quite low (Linux at <2%), but much higher on web server operating systems (collectively around 67%).

Closer to industrial markets, most incumbent automotive original equipment manufacturers (OEMs) currently use proprietary software inside cars (although this is changing with cars models such as Cadillac, which uses Linux), and the same is true for avionics in commercial aircraft. CAD (computer aided design) has only 2% penetration from open source. Open source software penetration in industrial applications is generally quite low, although General Electric’s Predix, a software platform for industrial asset management and operations optimization, was launched as open source in 2015.

In their 2012 book “Internet Success: A Study of Open-Source Software Commons”, MIT Professors Charles Schweik and Robert English cite several common success factors for open source projects. These include a clearly defined and clearly communicated vision by an active leader, well-articulated and clear project goals, and higher perceived utility of the resultant software to end users. ROS Industrial, the industrial variant of ROS (Robot Operating System) counts numerous industrial companies in its consortium, including robot manufacturers like ABB and Yaskawa, general industrial automation players like Siemens, and large industrial customers including Boeing, Caterpillar, BMW and Ford.


© 2016 Citigroup

58

Figure 53. Open Source Success Varies Substantially Across Different Technologies

Figure 54. Global Smartphone Sales By Platform (Millions of Units)


Strategic Implications for Robotic Players There are arguably two sets of suppliers that are impacted by the choice of robot software in factories – the suppliers of robots, and the suppliers of the broader industrial automation software that controls the overall factory.

We think that there is an important distinction to be made between industrial and non-industrial robots. Most industrial robots from the leading players have their roots in the automotive industry – this is largely the case for the “big 4” European and Japanese incumbents, although Universal Robots a relatively new entrant in collaborative robots in 2009, and now owned by Teradyne, is a notable exception.

For these players, open source software probably accelerates robot penetration through lower adoption costs for customers, but equally lowers entry barriers for new entrants. Lessons from tech hardware suggest that open-source shifts the balance of power away from hardware players.

From a Japanese perspective, the move towards open-source robotics appears to be relatively slow with most large OEMs typically more focused on hardware. There is however a drive to expand into general industry, with co-bots being one area of focus. Among the Japanese OEMs, Kawasaki Heavy Industries argues that it has been the most receptive to open source and believes that it will be a winner, helped in part by a relative broad customer base encompassing not just auto and general industry customers but also medical institutions and tech customers. However, the potential for growth from non-auto customers, coupled with the move towards open source, is prompting change even at Fanuc, a name that in the past epitomized the closed nature of proprietary systems and black box solutions to its users. It has long been said that one reason the largest robot supplier in Japan does very little business with Japan's largest car maker is the latter's historical reluctance to use proprietary systems from its suppliers. However, the launch of Fanuc's FIELD (Fanuc Intelligent Edge Link and Drive) system in collaboration with Rockwell Automation, Cisco and Preferred Networks (a deep learning start-up in Japan) leads us to believe that Fanuc recognizes the shift towards open-ness and the necessity of changing its whole strategy.

0%

10%

20%

30%

40%

50%

60%

70%

Desktop PCs(Linux)

Browsers(Firefox)

Web Servers(Linux)

HTTP servers(Apache)

Smartphones(Android)

0

200

400

600

800

1000

1200

1400

1600

2009 2010 2011 2012 2013 2014 2015

Non-Android Android


© 2016 Citigroup

59

Strategic Implications for Industrial Automation Players Most industrial automation in a factory hinges on the PLC – the “programmable logic controller” that is the digital brain that controls (amongst other things) factory machinery. PLCs are not part of the machinery per se, but rather dedicated industrial computers that can be programmed to direct machines as part of the broader factory process. Siemens is the market leader for PLCs in Europe, with Rockwell Automation the leader in the U.S. – other key suppliers globally include Mitsubishi, Schneider Electric, and Omron. Although PLCs are widely used across most industries, the automotive market is the biggest single end market, we estimate accounting for around a third of market demand. Automotive markets are therefore the biggest single customer base for both PLCs and industrial robots. Despite this commonality of customer base, the programming languages used to control robots has historically been different to that used to program PLCs – in part due to being supplied by different manufacturers.

For PLCs there is a standard programming language (IEC 61131), whereas Robot controllers have had separate proprietary software. Due to the complexities involved (mainly involving errors being longer to identify due to the existence of separate programs, and the cost of two sets of software engineers), a number of robot makers have now incorporated PLC software directly into their robot controllers, including Fanuc and Yaskawa. We see the ability to more closely integrate robots into the software systems of industrial automation players as a potential net positive for both robot adoption and revenues of industrial automation players.

Changes in the Competitive Landscape The shift in growth towards non-auto is just one trend in robotics; the other is the anticipation that China will remain the fastest growing market.

As part of the “Made in China 2025” plan, the Chinese government has previously highlighted robots as a focus sector, targeting to increase the "Made-in-China" robotics market share to >50% by 2020, from 30% in 2015, and to reach 800,000 units of operational stock of industrial robots by 2020. Any industry trend that lowers entry barriers – for example software development costs – might help this aspiration.

Winners and Losers We see two trends. First, open source software probably accelerates robot penetration through lower adoption costs for customers – a boost for hardware sales – but equally lowers entry barriers for newer (and smaller ) hardware players, meaning the continuing dominance of European and Japanese incumbents is not assured. Lessons from tech hardware also suggest that open-source shifts the balance of power away from hardware players. Considering the Chinese government's intention of nurturing domestic robot suppliers in particular, and the robotics industry in general, there is no guarantee that the top Japanese and European makers today will still be the same in five to ten years’ time. This is true not just for the OEMs but also key component suppliers, like precision speed reducers. Secondly, for broader suppliers of industrial automation systems, the ability to more closely integrate robots into their own software systems is a potential net positive. The increasing diffusion of open source robotics could easily accelerate the growth of co-bots in general industry, cementing the dominance of key component players.


© 2016 Citigroup

61

9. Thermoplastic Subsea Pipes Enabling Stranded Deepwater Assets Deepwater oilfields account for almost 20% of industry’s growth portfolios, with its material portion at risk of being left stranded in the current oil price environment. Costs have been a focus for the industry and we believe new technology will play a significant role in driving change. Combined with weaker pricing (driven by supply chain overcapacity) we expect offshore development costs to drop by 50% over the medium term. Transition to a new subsea pipe technology – bonded thermoplastic pipes (TCP) – looks one of the most exciting technological shifts in the deepwater (DW) industry in years.

A step change in development costs — The potential gains of TCP versus existing rigid/flexible steel pipe technology look significant. Less weight, longer life, and easier installation will drive lower upfront (installation) and life of field costs. We estimate that subsea costs could drop by 30-40% and total deepwater costs by 10%, enough to lower the breakeven oil price by $4/bbl.

Adoption within five years — Two independent players are already manufacturing plastic pipe for various deepwater applications and have attracted the interest of large operators (e.g. BP and Royal Dutch Shell). GE – one of the largest players in the space – has also recently announced plans to launch a similar product for use in ultra deepwater. We think full-scale adoption in production risers could occur within the next five years, as necessary safety and performance qualifications are completed (e.g. Magma/BP/Subsea7 program ending in 2017).

A threat to the established hierarchy — While a shift to bonded pipe is potentially good for kickstarting deepwater development, there are negative implications for existing SURF players. Flexible pipes (a $1.5bn/year business) stand to lose market share. The light weight of TCP will call into question the role of the heavy installation vessel fleet – industry has a total of $16bn of net fixed assets primarily in heavy boats.

Figure 55. TCP Pipe (Shown vs. Steel) Set to Challenge Steel/Flexibles in SURF Application…

Figure 56. …With Potential To Reduce Ultra Deepwater SURF Costs by 40%

Source: Magma Global Source: Citi Research

0%

-90%

-25%

-50%

0%

20%

40%

60%

80%

100%

Conventional Bonded composite pipe

Risers, Jumpers, flowlines Installation Engineering Other equipment and fabrication

Total cost saving potential of up to 40%

Mukhtar Garadaghi European Oilfield Services Analyst


© 2016 Citigroup

62

Disruptive Pipe Technology to Deliver a Step Change in SURF Costs Subsea infrastructure (“SURF” - Subsea, Umbilicals, Risers, and Flowlines) comprises about 25% of typical deepwater project costs (subsea equipment makes up another 10%). Disappointingly, it has seen limited cost deflation to date (e.g. vs wells) and remains a key area of focus in the industry’s cost cutting efforts.

Figure 57. Cost Breakdown of a Typical Offshore Project Figure 58. SURF Cost Breakdown


So far, savings in subsea have been driven by tighter integration with the production systems and lower vessel pricing. Beyond these incremental improvements in SURF, we also see potential for the downturn to accelerate the adoption of new technological solutions that could deliver a step change in cost. Importantly, the conventional riser and flowline materials (steel and steel/polymer combinations) are also approaching their physical limits as developments move into waters with depths of >2 kilometers, harsher environments, and reservoirs with corrosive liquids content. This creates a wide range of issues including prohibitively high riser assembly weight, material fatigue and wear when subjected to high hydrogen sulfide (H2S) or reservoir stimulation fluids.

With continued advancement in high performance materials, bonded TCP pipes appear to offer that step change solution: (1) on a 5-year view SURF costs could see a 35% reduction (i.e. up to 10% of total project cost, for water depth of 2 kilometer+); (2) previously inaccessible UDW resources could be developed; and (3) installation/exploitation risk could be reduced. Some polymers/plastics are already used in riser/flowline/jumper manufacturing (the flexible “non-bonded” solution), however bonded TCP surpass the properties of existing rigid and flexible pipe solutions in many respects:

90% lower weight is key – TCP density equals to 10-30% of that for equivalent steel or flexible pipe (in water) and has a tendency for natural buoyancy even when flooded. This in turn means that significantly lower top tension is required at the top of the top of the riser. (see Figure 60) An order of magnitude smaller (and cheaper) buoyancy equipment can be used, with fewer and smaller installation vessels required. This is especially pronounced in water depths of >2km, where TCP weight delivers the highest incremental benefit (delta far narrower in <1km).

Drilling rig18%

Well services20%

Facilities29%

Subsea production systems

10%

SURF23%

Risers, jumpers, flowlines

40%

Installation31%

Engineering5%

Other equipment,

fabrication, etc24%

Scope for materialcost reduction

Figure 59. Schematic Cross Section of a Bonded Composite Pipe

Source: Citi Research, DNV GL


© 2016 Citigroup

63

Figure 60. Composite Rise Require Significantly Lower Top Tension Due to a Natural Buoyancy – Reducing Equipment Demands

Figure 61. Composite Pipes Also Deliver Materially Improved Fatigue Performance All Depths Notably at the Weakest Points Close to the Upper Rise Assembly

Source: Citi Research Source: Citi Research, 2H Offshore The Benefits of Composite Materials in Deepwater

Rise Applications

Increased durability/longevity – Smooth, single solid wall TCP bore (as opposed to multiple thinner walls in flexible pipes or multiple weld junctions for rigid risers) appears to exhibit significantly less wear during exploitation. This is demonstrated by a significant improvement of fatigue life (see Figure 61– study by 2H Offshore) of composite vs. steel riser at all depths, notably at weak points near the pipe interface with Upper Riser Assembly. Combined with corrosion resistance, this translates into a projected life-span of 25 years for composite risers/flowlines (vs 7-10 years for flexible and 10-15 years for rigid) meaning that TCP is far less likely to require replacement (i.e. lower late life capex and less downtime).

Simpler spooling, deployment – High flexibility (vs. steel) allows for spooling the pipes on small/medium diameter reels (14m for a 6”), while manufacturing in continuous length of up to 4km saves precious time offshore otherwise wasted on welding individual pipes.

Figure 62. Bonded TCP Pipe vs. Rigid and Flexible Rises: Key Stats

Bonded TCP pipe Conventional/flexible pipes Material Thermoplastics, carbon fiber, glass Steel / Steel, polymer, plastics Weight/Density Up to 10x less -- Pressure rating Up to 20k psi Up to 15k psi Cost per kilometer $6k/m $4-7k/m for flexibles, fully coated rigid Wear/corrosion resistance High Variable Lifespan, years up to 25 years 7-10 years for flexibles, <15 years for rigid Type system deployment time* 3 months 7 months


Corrosion resistance and better flow characteristics – TCP’s high

resistance to corrosion by hydrocarbons (with high H2S, CO2 content) and injection chemicals is among its key benefits. Figure 63 shows a life cycle cost for three key types of flowlines, with TCP showing potential for c40% and 60% full cycle costs vs. equivalent flexible and rigid pipes, respectively.


© 2016 Citigroup

64

SURF Cost Reduction of Up to 40% Driven by Cheaper Installation A significant reduction in installation time and expenditure is the key benefit of TCP adoption for SURF, more than offsetting the marginally higher manufacturing costs. We estimate that bonded pipe’s lower weight, reel deployment and natural buoyancy have the potential to deliver up to a 90% reduction in installation cost and a greater than 50% reduction in installation time. High capacity (J-Lay), expensive (up to $1m/d) installation vessels could be replaced by mid-sized reel capable pipelays, delivering a 50-70% day rate saving, with significantly better availability.

Full Scale SURF Application of TCP Likely 5 Years Away But Limited Deployment/Tests Have Had Early Success TCP/bonded pipe solutions for subsea are being brought to market primarily by two specialized manufacturers:

Magma Global is a UK-based bonded pipe manufacturer targeting subsea applications in oil and gas. Its product (m-pipe®) has recently been certified for subsea application by DNV GL and is undergoing industry tests with m-pipe jumpers and injection risers already deployed in the North Sea. The key components of its product are carbon fiber and Victrex PEEK polymer (a high performance thermoplastic), which are combined through a fully automated 3D laser print process.

Airborne Oil and Gas is a Netherlands-based TCP manufacturer with similar technology but employing fiberglass. Importantly, Airborne has strong industry backers: Royal Dutch Shell, Chevron and Evonik, which demonstrates operator interest in TCP and is likely to accelerate the deployment as product matures. To date Airborne’s TCP has been deployed the 3”, 5kpsi downline in Brazil (2.1 kilometer water depth) as well as downlines and dynamic jumpers in a number of other geographies.

GE’s Oil & Gas division – one of the largest players in the oil and gas equipment space – has recently joined the TCP race in May 2017. It has announced (May 2016) plans to manufacture a hybrid composite riser based on TCP technology in a drive to reduce weight and costs in UDW. With momentum around cost reduction in offshore building and key industry players like BP, Royal Dutch Shell, and Chevron) supporting innovation by subsea TCP manufacturers, we believe full scale deployment of bonded pipes in SURF could be only 5 years away with the lack of offshore track record a key short term challenge for TCP. Manufacturers are pursuing qualification processes with operators to combat this, e.g. Magma targeting qualification of m-pipe® for riser technology BP/Subsea7 by end-2017, with Airborne successfully completed a 3 year qualification process for Petronas (for 6” TCP flowline), as well Shell and Total in various applications.

The most direct effect of a wide scale adoption of TCP would be the displacement of flexible and the rigid pipe volumes from subsea. And while subsea is a very small portion of rigid pipe market, flexible pipe manufacturers (e.g. Technip and National Oilwell Varco) have a more focused exposure. Longer term implications for DW installation names are also likely negative, with TCP having the potential to reduce the size of awards (project scope) and pressure demand for high-end installation vessels.

Figure 63. Significant Operating Expense and Late-Life Capital Expenditure Reductions Due to Reduced Corrosion Effects on TCP


0

5

10

15

20

25

30

35

40

45

0 5 10 15 20

$m

Carbon Steel, S-lay Flexible TCP Flowline

Years


© 2016 Citigroup

65

10. Wide Bandgap Semiconductors Powering the Future Semiconductors power the entire technology industry and are the building blocks for all electronic devices. While personal computers have historically been the dominant end market for semiconductors, the last 15 years have seen the emergence of handsets as a major end market. And over the next 10 years, we believe growth of semiconductors will in fact be driven by diverse applications targeting multiple end markets ranging from automotive to industrial and communication infrastructure.

Figure 64. Semi End Markets, 2000 Figure 65. Semi End Markets, 2015 Figure 66. Semi End Markets, 2025E

Source: WSTS, Garner, Citi Research Source: WSTS, Garner, Citi Research Source: WSTS, Garner, Citi Research

As devices evolve and applications diversify, existing semiconductors are losing their ability to keep pace with the increasing performance demands. This is particularly relevant when it comes to conversion, control, and processing of electric power (interchangeably referred to as ‘power electronics’ / ‘power semiconductors’). The traditional semiconductor manufacturing technology based on Silicon has matured to a point where it is pushing the fundamental limitations of physics.

Wide bandgap (WBG) semiconductors have shown the capability to meet the higher performance demands yet enable design of modules that are smaller, faster, and more reliable than their silicon-based counterparts, resulting in improved efficiency when it comes to existing applications such as power chargers, and fully meeting the needs of new emerging applications such as MRI scanning and automotive radar.

According to a 2014 study by Power America, the Innovation Institute backed by the U.S. Department of Energy, adoption of WBG could in the case of industrial motor systems, translate into savings of electricity of up to 5.9 million households. An older study in 2010 pegged that widespread adoption of new WBG-centric power electronics could trigger more efficient workloads and potentially save over 25% of worldwide annual energy consumption by 2025. Although the latter in particular might seem a bit optimistic given the time frame, we believe these studies demonstrate the long term disruptive potential and the wider socioeconomic and geographic (global warming) impact of WBG semiconductor technology.

In the subsequent sections, we first take a step back to explain the science behind semiconductors in general and wide bandgap semiconductors in particular. We follow this up with an overview of our medium term forecasts as well as our take on the likely path of adoption. We conclude the discussion with an overview of the challenges and explain why we still remain optimistic on widespread adoption.

Automotive7%

Industrial8%

Comm Infrastructure

11%

Handsets8%

Consumer16%

PC50%

Automotive9%

Industrial11%

Comm Infrastructure

9%

Handsets23%

Consumer11%

PC37%

Automotive15%

Industrial16%

Comm Infrastructure

10%Handsets17%

Consumer13%

PC29%

Amit B Harchandani Head of European Technology Research Team


© 2016 Citigroup

66

The Technological Perspective The main objective of a semiconductor is to transmit electronic signals. To make this controlled movement possible, a semiconducting material (i.e. one with electrical conductivity between an insulator and conductor) is “doped” with impurities to create "unpaired" electrons which, if induced, can move around and transmit signals. The two broad types of semiconductor components are called diodes and transistors.

Semiconducting materials can be two types – elemental or compound. Elemental semiconductors are silicon (Si) and germanium (Ge), while silicon carbide (SiC) and gallium nitride (GaN) are examples of the latter type. Silicon is cheaper than competitors and readily available (given that it is the second most common element in earth's crust), as well as easier to process. This has seen silicon emerge as the most popular material for production in the rapidly growing electronics industry.

However, silicon has its own fundamental limitations. To understand these limitations, we need to introduce the concept of “bandgap”: simply put this generally refers to the energy difference (measured in units called ‘electron volts’ or eV) between the valence (highest range of energy) and conduction band (lowest level of empty state where an electron can move) of any material.

Figure 67. Band Gap Representation Figure 68. Semiconducting Material Band Gap Property

Source: Citi Research Source: U.S. Department of Energy, Citi Research

Wide bandgap semiconducting materials such as silicon carbide (SiC) and gallium nitride (GaN) are those that have a bandgap higher than silicon. As a consequence, WBG semiconductors permit devices to operate at much higher temperatures, voltages, and frequencies – making the power electronic modules using these materials smaller (due to lower cooling requirements / energy loss), faster (on account of chemical composition) and more reliable (thanks to higher temperature tolerance) than those made from conventional semiconductor materials.

Sizing the Market for WBG Semiconductors The total size of the worldwide semiconductor market in 2015 stood at ~$335 billion according to data from semiconductor industry body WSTS. We estimate that the share of power semiconductors stood at ~$16 billion, based on contributions from four broad product groups – power diodes, power transistors, rectifiers, and thyristors.

Power semiconductors are used in a wide range of applications. According to data from industry research firm Gartner, automotive and industrial end markets each roughly accounted for a third of the sales of power transistors in 2015. German manufacturer Infineon was the leader in terms of market share, followed by Japanese firm Mitsubishi

Conduction Band

Valence Band

Band Gap

Ener

gy

Material Symbol Bandgap Energy (eV)

Maximum Temperature

Silicon Si 1.1 150C

Silicon Carbide SiC 3.3 300C+

Gallium Nitride GaN 3.4 300C+

Semiconductors transmit signals through controlled movement of electrons

Silicon has been the dominant material used in production of semiconductors

Bandgap refers to energy difference between valence and conduction band

Wide bandgap semiconductors such as SiC and GaN have a higher bandgap than Si

We estimate power semiconductors market was ~$16bn in 2015

Power semiconductors are used in a wide range of applications.


© 2016 Citigroup

67

Figure 69. Power Semis by Product (2015) Figure 70. Power Transistors by Market (2015) Figure 71. Power Transistors by Vendor (2015)

Source: WSTS, Citi Research Source: WSTS, Gartner, Citi Research Source: Gartner, WSTS, Citi Research

While growth of power diodes and transistors has been somewhat flattish in recent years, we believe a combination of a consolidating market, exposure to higher growth application areas such as electronic/hybrid vehicles and renewable energy, combined with the ever increasing emphasis on energy efficiency should see the growth improve to a compound annual growth rate (CAGR) of ~2% over the period 2015-2020.

Zeroing in on WBG semiconductors, we believe these are still early days, and we estimate that proportion of GaN and SiC-based components (both diodes and transistors) as proportion of the overall power diodes and transistors market stood at 0.9% in 2015. However, we see overall penetration rate improving to 6.7% by 2020, and hence we estimate WBG semiconductor revenue to grow at a robust CAGR of ~60% over the period 2015-2020, resulting in it surpassing $900 million by 2020. As SiC has wider range of applications relative to GaN and is also a more mature technology in our view, and we therefore see it being the more dominant driver of this growth.

Figure 72. Forecast for Power Transistors and Diodes (2015-2020) Figure 73. Forecast for WBG Semiconductor Revenue (2015-2020)

Source: WSTS, Citi Research Source: WSTS, Citi Research

Power transitors

69%

Power diodes

9%

Others22%

Automotive33%

Consumer14%

Wired Comm

4%

Wireless Comm

5%

Data Processing

12%

Industrial30%

Military2% Infineon

19.5%

Mitsubishi13.0%

Fairchild Semi8.4%Toshiba

7.8%STMicro

6.3%

Others45.0%

-

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

2015 2016 2017 2018 2019 2020

$bn

Power Transistors Power Diodes

- 100 200 300 400 500 600 700 800 900

1,000

2015 2016 2017 2018 2019 2020

$mn

SiC GaN

We see power transistors and diodes sales growing at a CAGR of ~2% in 2015-2020

We see WBG semiconductor sales growing at a CAGR of ~60% in 2015-2020


© 2016 Citigroup

68

Applications of WBG Semiconductors In the near-to-medium term, we see greater traction for WBG semiconductors across the following three end markets – Automotive, Industrial and Consumer.

Automotive – We see WBG materials used in building the charging infrastructure for electric and hybrid vehicles (EV/HEV) given their ability to cut electricity losses. We also see benefits from adoption in traction applications.

– A 2012 study by McKinsey showed that SiC devices in hybrid electric vehicles offer a better value proposition than traditional silicon devices after taking into account all cost savings across the entire value chain including fuel savings, reduction in cooling, and weight requirements.

Industrial – We see increasing uptake in case of renewable energy to convert direct current (DC) to alternating current (AC) electricity used in homes and businesses. However, despite the inherent benefits, we are more cautious on the near-term adoption of WBG materials in the case of variable speed drives used in industrial motors.

– Switching from Si-based devices to WBG-based devices could increase conversion efficiency from 90% to 98% within data centers, thereby amounting to a 8.3% reduction in energy usage by data center power electronics.

– From an analysis carried out by Infineon, a converter built on SiC devices is a third of the size and 25% of the weight compared to a current silicon based reference solution. Thanks to the significant reduction in volume and weight, the system costs can also be reduced by more than 2% as per Infineon.

Consumer – The ability to make power supplies smaller makes WBG-based devices ideal for high-efficiency data centers. We also see use cases across other consumer applications ranging from fast chargers to LED lighting.

– According to U.S. Department of Energy (DoE), LEDs based on WBG semiconductors produce 10x the light per watt and over 30x+ service life.

– In another case study the U.S. DoE found the use of GaN-based semiconductors in tablet and cell-phone adapters can result in annual savings on the scale of the annual output of a midsized coal power plant.

Challenges and Reasons for Optimism Based on our analysis of data from the U.S. DoE and other available literature on WBG semiconductors, we acknowledge that multiple challenges must be overcome to make them cost effective in more applications. However, we remain optimistic on their widespread adoption.

High substrate cost – Substrate materials account for 30-50% of the cost of a SiC device, while traditional silicon-based power device substrates account for only 5-7%. Nonetheless, it is possible to transmit a greater amount of electronic signals in a SiC device, making SiC devices competitive on a cost/area basis.

– Further in the case of SiC, cost reductions can be realized through the high-volume processing of wafers.


© 2016 Citigroup

69

– When it comes to GaN, we note the initiatives around GaN-on-Si devices – this is a more cost effective approach and involves applying a layer of GaN on a Silicon substrate that enables a substantial reduction in material costs.

Device design and cost – The device designs that effectively exploit the properties of WBG materials are different as compared to those for conventional semiconductors. These need to be implemented in order to achieve the voltage and current ratings required in certain applications. From a back-end perspective, alternative packaging materials are also needed to withstand the high temperatures in WBG devices.

– The key motivation for producing the new designs facilitated by WBG semiconductors is miniaturization, integration, and improvement in performance.

Systems integration – WBG devices are not always suitable drop-in replacements for Si-based devices. The larger, more complex systems must be redesigned to integrate the WBG devices in ways that deliver unique capabilities.

Citi Global Perspectives & Solutions (Citi GPS) is designed to help our clients navigate the global economy’s most demanding challenges, identify future themes and trends, and help our clients profit in a fast-changing and interconnected world. Citi GPS accesses the best elements of our global conversation and harvests the thought leadership of a wide range of senior professionals across the firm. All Citi GPS reports are available on our website www.citi.com/citigps

Digital Disruption How FinTech is Forcing Banking to a Tipping Point March 2016

The Coming Pensions Crisis Recommendations for Keeping the Global Pensions System Afloat March 2016

Technology at Work v2.0 The Future is Not What It Used To be January 2016

Global Political Risk The New Convergence between Geopolitical and Vox Populi Risks January 2016

Investment Themes in 2016 New Normal or No Normal January 2016

2016 Corporate Finance Priorities January 2016

Energy 2030 Financing A Greener Future November 2015

The Global Art Market Perspectives on Current Drivers & Future trends November 2015

The Curtain Falls How Silicon Valley is Challenging Hollywood October 2015

Energy Darwinism II Why a Low Carbon Future Doesn’t Have to Cost the Earth August 2015

Disruptive Innovations III Ten More Things to Stop and Think About July 2015

Public Wealth of Nations Unlocking the Value of Global Public Assets June 2015

Women in the Economy Global Growth Generators May 2015

Car of the Future v2.0 Mobility Transformation: Full Steam Ahead May 2015

Beyond China The Future of the Global Natural Resources Economy March 2015

Technology at Work The Future of Innovation and Employment February 2015

Investment Highlights in 2015 Dealing with Divergence January 2015

Corporate Finance Priorities 2015 Driving Corporate Growth in Divergent Markets January 2015

The Re-Birth of Telecom Monopoly Is the Industry Broken & Heading Back to its Monopolistic Roots November 2014

Energy 2020: Out of America The Rapid Rise of the US as a Global Energy Superpower November 2014

Asset Allocation for a New Era Diversification, Not Rotation, is the New Watchword October 2014

Future Opportunities, Future Shocks Key Trends Shaping the Global Economy and Society October 2014

Taking It To The Streets The New Vox Populi Risk May 2014

The Car of the Future Transforming Mobility As We Know It May 2014

Disruptive Innovations II Ten More Things to Stop and Think About May 2014

Upwardly Mobile III Mobility Unchained: From Mobile Commerce to IoT January 2014

2014 Year Ahead Investment Themes January 2014

Abenomics Four Arrows to Target Four Challenges October 2013

Energy Darwinism The Evolution of the Energy Industry October 2013

Call of the Frontier II On the Right Track to be Tomorrow’s EMs September 2013

Energy 2020 Trucks Trains & Automobiles: Start your Natural Gas Engines June 2013

The Global Search for Yield How Today’s Markets Shape Tomorrow’s Companies May 2013

Disruptive Innovation Ten Things to Stop and Think About April 2013

Energy 2020 Independence Day: Global Ripple Effects of the N. American Energy Revolution February 2013

2013 Year Ahead Investment Themes January 2013

2013 Year Ahead Corporate Finance Priorities January 2013

Upwardly Mobile II A Long and Winding Road for Mobile Payments November 2012

China in Transition What We Know, What We Don’t Know November 2012

Global Debt Mr. Macawber’s Vindication November 2012

Sub-Saharan Africa The Route to Transformative Growth September 2012


© 2016 Citigroup

72

IMPORTANT DISCLOSURES This communication has been prepared by Citigroup Global Markets Inc. and is distributed by or through its locally authorised affiliates (collectively, the "Firm") [E6GYB6412478]. This communication is not intended to constitute "research" as that term is defined by applicable regulations. Unless otherwise indicated, any reference to a research report or research recommendation is not intended to represent the whole report and is not in itself considered a recommendation or research report. The views expressed by each author herein are his/ her personal views and do not necessarily reflect the views of his/ her employer or any affiliated entity or the other authors, may differ from the views of other personnel at such entities, and may change without notice. You should assume the following: The Firm may be the issuer of, or may trade as principal in, the financial instruments referred to in this communication or other related financial instruments. The author of this communication may have discussed the information contained herein with others within the Firm and the author and such other Firm personnel may have already acted on the basis of this information (including by trading for the Firm's proprietary accounts or communicating the information contained herein to other customers of the Firm). The Firm performs or seeks to perform investment banking and other services for the issuer of any such financial instruments. The Firm, the Firm's personnel (including those with whom the author may have consulted in the preparation of this communication), and other customers of the Firm may be long or short the financial instruments referred to herein, may have acquired such positions at prices and market conditions that are no longer available, and may have interests different or adverse to your interests. This communication is provided for information and discussion purposes only. It does not constitute an offer or solicitation to purchase or sell any financial instruments. The information contained in this communication is based on generally available information and, although obtained from sources believed by the Firm to be reliable, its accuracy and completeness is not guaranteed. Certain personnel or business areas of the Firm may have access to or have acquired material non-public information that may have an impact (positive or negative) on the information contained herein, but that is not available to or known by the author of this communication. The Firm shall have no liability to the user or to third parties, for the quality, accuracy, timeliness, continued availability or completeness of the data nor for any special, direct, indirect, incidental or consequential loss or damage which may be sustained because of the use of the information in this communication or otherwise arising in connection with this communication, provided that this exclusion of liability shall not exclude or limit any liability under any law or regulation applicable to the Firm that may not be excluded or restricted. The provision of information is not based on your individual circumstances and should not be relied upon as an assessment of suitability for you of a particular product or transaction. Even if we possess information as to your objectives in relation to any transaction, series of transactions or trading strategy, this will not be deemed sufficient for any assessment of suitability for you of any transaction, series of transactions or trading strategy. The Firm is not acting as your advisor, fiduciary or agent and is not managing your account. The information herein does not constitute investment advice and the Firm makes no recommendation as to the suitability of any of the products or transactions mentioned. Any trading or investment decisions you take are in reliance on your own analysis and judgment and/or that of your advisors and not in reliance on us. Therefore, prior to entering into any transaction, you should determine, without reliance on the Firm, the economic risks or merits, as well as the legal, tax and accounting characteristics and consequences of the transaction and that you are able to assume these risks. Financial instruments denominated in a foreign currency are subject to exchange rate fluctuations, which may have an adverse effect on the price or value of an investment in such products. Investments in financial instruments carry significant risk, including the possible loss of the principal amount invested. Investors should obtain advice from their own tax, financial, legal and other advisors, and only make investment decisions on the basis of the investor's own objectives, experience and resources. This communication is not intended to forecast or predict future events. Past performance is not a guarantee or indication of future results. Any prices provided herein (other than those that are identified as being historical) are indicative only and do not represent firm quotes as to either price or size. You should contact your local representative directly if you are interested in buying or selling any financial instrument, or pursuing any trading strategy, mentioned herein. No liability is accepted by the Firm for any loss (whether direct, indirect or consequential) that may arise from any use of the information contained herein or derived herefrom. Although the Firm is affiliated with Citibank, N.A. (together with its subsidiaries and branches worldwide, "Citibank"), you should be aware that none of the other financial instruments mentioned in this communication (unless expressly stated otherwise) are (i) insured by the Federal Deposit Insurance Corporation or any other governmental authority, or (ii) deposits or other obligations of, or guaranteed by, Citibank or any other insured depository institution. This communication contains data compilations, writings and information that are proprietary to the Firm and protected under copyright and other intellectual property laws, and may not be redistributed or otherwise transmitted by you to any other person for any purpose. IRS Circular 230 Disclosure: Citi and its employees are not in the business of providing, and do not provide, tax or legal advice to any taxpayer outside of Citi. Any statements in this Communication to tax matters were not intended or written to be used, and cannot be used or relied upon, by any taxpayer for the purpose of avoiding tax penalties. Any such taxpayer should seek advice based on the taxpayer’s particular circumstances from an independent tax advisor. © 2016 Citigroup Global Markets Inc. Member SIPC. All rights reserved. Citi and Citi and Arc Design are trademarks and service marks of Citigroup Inc. or its affiliates and are used and registered throughout the world.


© 2016 Citigroup

73

NOW / NEXT Key Insights regarding the future of Disruptive Innovation

COMMODITIES The oil and gas industry is a major generator of data, but the lack of sophisticated

analytical tools has hindered the ability of the industry to focus on analysis of the data. / The use of big data and advanced analytics raise the transparency of supply-demand fundamentals, energy flows and the price discovery process. Real-time tracking of energy flow and generation enables better prediction and coordination of systems and lowers electricity and energy prices.

INNOVATION Drug delivery via eye injections require an office visit and an injection procedure that carries potential risk such as retinal detachment. Eye drops are limited by natural obstacles to delivery including physical barriers in the anatomy of the eye and tears. / Next gen ocular delivery technologies will address both the front of the eye and back of the eye delivery. Implants and inserts can be loaded with drugs and released to the back of the eye over an extended period. Eye drops that incorporate nanoparticles or liposomes can improve delivery over traditional eye drops.

TECHNOLOGY Software to control industrial robots has historically been proprietary and hardware-specific. / Open source software in robots boosts collaborative development. As robots get more complex – and the number of hardware suppliers increases – the appeal of collaboration on software also increases.

Disruptive technologies review citi july 2016

Business

Transcript of Disruptive technologies review citi july 2016