Over the last twenty years, artificial intelligence (“AI”), defined as the development of advanced computer systems able to perform tasks that normally require human intelligence, such as visual perception, communication and decision-making, has been developing at an increasing pace around the globe1. AI is not necessarily a technology in and of itself. It is an ambition, a quest for the intelligent machine, where AI-labelled innovations considered cutting edge in the past are now general applications common enough to question whether they should still be labelled AI.
As is often the case with technology, some locations have taken up positions at the forefront of AI research and development (“R&D”). As of 2018, the United States (“US”) and China are leading AI development, but other countries around the globe have emerged to produce high-quality AI applications and R&D, notably in cities such as Montreal, London, Paris and Tel Aviv. Alphabet CEO, Eric Schmidt, has compared the momentum of AI development to the moon race2. When analysing global AI development and initiative to ensure competitiveness, the AI race is even more heated as it involves a broad spectrum of industrial applications and is reinforced by national interests as well as growing capital investment.
AI has quickly progressed with the development of IT infrastructure, computing power and data availability. In addition, AI powerhouses have emerged in clusters because of an intertwinement of the availability and proximity to qualified talent and academic institutions, financial capital, a culture of innovation sharing, and a growing involvement of public and private entities3. Clustering is common in science as innovation rarely happens in a single, hidden place. As explained during an MIT interview with Yoshua Bengio, a deep learning specialist based in Montreal, science moves in small steps thanks to the collaboration of diverse communities where actors interact and share information in a spirit of collaboration leading to orthogonal research directions and exploration paths4. This article provides an overview of leading AI hubs and their particular dynamics around the globe.
AI Ecosystem Drivers
1. KEY HUBS OVERVIEW
United States
The US, as the birthplace of some of the largest digital players such as Alphabet, Apple, Amazon, Facebook, IBM, and Microsoft, has been driving a large share of innovation. These companies have had the necessary ingredients to move AI forward through access to large volumes of proprietary data, technology and capital, as well as the ability to attract highly skilled labour. With over 850,000 people working in AI, the US has one of the biggest pools of qualified professionals and has the capacity to train students in some of the best science, technology, engineering and mathematics (“STEM”) academic institutions5. Between 2013 and 2016, the US had the lion’s share of global private AI investments with over 60% of investments (valued at US$30–40bn). However, with the race heating up, US companies received 38% of global funds amounting to US$15.2bn in 20176. The US has an active merger and acquisition (“M&A”) market led by large strategic players, with over 40 acquisitions by Facebook, Amazon, Apple and Google between 2012 and 20177. These conditions, combined with the American entrepreneurial spirit, have led to the growing emergence of over 2,040 AI start-ups, representing 40% of the AI start-up ecosystem worldwide8. The government has been supportive with a strategic plan, flexible regulation and financial support amounting to over US$1.1bn annually in 2015–2017 and over US$2bn in 2018 from the US Defense Advanced Research Projects Agency (“DARPA”)9.
Silicon Valley has been at the forefront of technological innovation with a symbiotic ecosystem comprising universities, start-ups, tech companies and venture capitalists (“VC”). With over two million tech workers employed at the headquarters of some of the biggest tech players in the world as well as in 15,000 start-ups, Silicon Valley boasts the highest number of entrepreneurs globally10. The pipeline of talent comes from all over the world and from the best local universities such as Stanford, UC Berkley and UC San Diego. These institutions have been in the vanguard of AI development through their positioning as some of the best academic and corporate AI labs. In terms of financing, the Valley has historically received up to 40% of global capital investments in AI11.
The East Coast Boston-New York area has been driving innovation due to the presence of top tier academic institutions such as NYU, Cornell and Boston-based MIT, which developed early natural language processing programs in the 1960s and has been an innovation driver ever since. It is also where the term artificial intelligence was first coined in 1956 at Dartmouth University. In 2018, the MIT announced an investment of US$1bn to create a new college combining AI, machine learning, and data science with other academic disciplines12. This is the largest financial investment in AI by any US academic institution to date. New York is also a major hub where 11% of US AI job postings are located and where labs such as NYU Tandon Future Labs bridge academia, start-ups and industry collaborations13. The area is also a leading financial hub and the second-strongest funding ecosystem after Silicon Valley in the number of early-stage VC investments. Many global banks including Goldman Sachs, JP Morgan and Credit Suisse have set up machine learning teams to apply AI to investment and retail banking14.
Global AI start-up equity funding
China
In less than a decade, China has demonstrated its ambition to become the global leader in AI with an expected GDP impact going as far as 0.8–1.4 per cent per year by 203015. In 2017, the Ministry of Industry and Information communicated its vision via the Next Generation AI Development Plan where it set forth the following goals:
– reach globally advanced level in AI technology, models and methods by 2020;
– make AI a major economic driving force and be the world’s premier AI innovation centre by 2025;
– build an AI core industry exceeding RMB 150bn (US$21.5bn), and exceed RMB 1tr (US$143bn) in related industries by 203016.
China is walking the walk and has given itself the means to achieve its goal with major research centres in Beijing (US$2.3bn pledged), Tianjin (US$5bn pledged and US$16bn planned until 2025) and Shenzhen (US$5bn)17. Chinese technology giants, who have the resources to move AI forward, have agreed to organise innovation streams, with Alibaba leading smart cities, Baidu covering autonomous vehicles, Tencent responsible for medical imaging and IFlyTek managing smart voice. Due to the collection of large amounts of data and internet oversight, data access is facilitated for firms like IFlyTek to have access to data such as biometric information from the government more easily. Beyond the Chinese technological giants, China ranks second for the number of AI enterprises with over 1,000 firms in mainland China only18.
In regards to talent, even if mainland China produces more graduates in science and engineering every year than the United States, Japan, South Korea as well as Taiwan combined and forms professionals in strong STEM universities such as Peking and Tsinghua, it still currently lacks sufficient AI talent to fully achieve its aspiration. A growing number of job postings, rising salaries and desire to hire the best talent worldwide demonstrate the race to hire more AI experts19. Nonetheless, China’s R&D efforts are reflected in the rise of Chinese unicorns and in AI patent applications, which were for example six times more than in the US for deep learning related keywords in 201720.
WIPO Patents in Artificial Intelligence Technology Related Sectors as at 2018
Europe
If considered as a unified entity, Europe has significant innovation mass through its AI start-ups (circa 22% of global share), highly reputable academic institutions and a willingness to develop AI as illustrated by the EU Commission’s innovation policy21. The UK (London), France (Paris) and Germany (Cyber Valley) lead the competition in the European AI space. However, even if EU members agree to compete efficiently on a global scale, Europe needs an ambitious and rapid deployment strategy, covering both business and public administration, to create an ecosystem where ideas and research translate into pragmatic socioeconomic opportunities.
United Kingdom
The UK, and more specifically London, is perceived as the leading European AI hub with over 760 enterprises of which around 650 are in London22. The ecosystem benefits from expertise, collaboration and a talent pipeline from universities such as Cambridge, Imperial College London and Oxford. As a global financial centre, London has a strong AI position with applications in finance, insurance and law23. London has government support with a strategy created to identify actions supporting AI growth across industries to drive innovation and productivity. An example of governmental initiative is the Tech Nation Visa, where applicants with exceptional talent can work without sponsorship requirements24. The programme proves beneficial as 43% of London AI start-ups have been founded by non-UK nationals25. Start-ups in London tend to raise lower funding compared to American and Chinese hubs, but the city is leading European VC investment with over £200m in private funding and £500m in public funding invested in 2017. London also benefits from the presence of top AI players such as Google Deepmind, which focusses on deep-learning applications for positive impacts, and OakNorth, which concentrates on fintech26.
France
France became a place of AI renaissance when President Macron made AI development a priority with his national strategy stating that AI is not only a technological, but also an economic, social, ethical and political revolution27. President Macron has pledged investments to stimulate innovation and turn France into a country of unicorns. As part of this plan, over €1.5bn are devoted towards AI development28. Through the AI for Humanity platform, France is among the countries giving most thought to regulation, diversity and ethics to ensure AI development is in line with the best standards of acceptability for citizens29. Amongst these initiatives are tech visas and the ambition to share governmental data to allow anyone to build AI services. The ecosystem is mainly clustered in Paris where there are strong STEM universities including CNRS, Paris Saclay and the National Institute for IT research and automation (INRIA). Over the last few years, Paris has benefitted from material foreign direct investment with notably IBM, Samsung and Facebook establishing labs. More recently, Google has indicated that it will add 1,000 individuals to its fundamental research lab. The French capital is home to a vibrant start-up community with over 20 incubators including the biggest in the world (Station F, a 34,000 m2 campus). The ecosystem bolsters over 120 companies and includes notable start-ups such as Dataiku, which develops machine learning on “dirty” data, and Prophesee, which develops computer vision sensors and systems in all fields of artificial vision30.
Germany
Germany is making increasing efforts to be at the forefront of AI development. The 2018 government coalition plan looks to attract talent, respond to the changing nature of work, integrate AI into government services, make public data more accessible, and further the development of ethical AI31. As part of the plan, the government has pledged €3bn until 2025 and expects funding to be matched by the private sector. It has also laid out a vision for German-built AI solutions to have a “Made in Germany” seal of quality32. Germany prefers to focus on increasing productivity in factories and supply chains around the world by leveraging industrial rather than consumer data, which is harder to access given public concerns about data privacy33. The German private sector has been showing increased activity since 2015 with the creation of over 100 AI related start-ups in a variety of sectors including ADA Health in healthcare and Arago in process automation. Germany is also looking to establish 12 research centres to train talent and conduct R&D in collaboration with industrial players34. R&D efforts are diffuse, but players such as Porsche, Daimler and Bosch have concentrated their efforts in the Stuttgart Cyber Valley to provide an industry-backed push to create a stimulating AI ecosystem conducive to technology transfers between academic laboratories (including University of Tubingen as well as University of Stuttgart) and industry35. With its Industry 4.0 efforts, Germany is perceived as taking a leading position in industrial applications including autonomous vehicles and robotics36.
Canada
With a significant AI talent base, Canada has pioneered several advances in AI, robotics and deep learning since the 1980s. In 2017, it was the first country to release a national AI strategy distinctly focussed on R&D and talent across the country. The strategic plan reaffirms support to the cities of Toronto and Montreal, which have become some of the most advanced global hubs, by pledging CA$1.3bn (US$1bn) in financing and creating three new AI institutes in Edmonton (AMII), Toronto (Vector Institute), and Montreal (MILA)37. The ecosystem has been rapidly expanding and now has over 280 enterprises mainly located in Toronto and in Montreal.
Toronto is North America’s second-largest financial services hub and is a major manufacturing centre, making it fertile ground for industry and start-ups. It is one of the 20 strongest start-up ecosystems in the world and benefits from leading research conducted at its 16 academic institutions, notably at the University of Toronto and the University of Waterloo38. The city is a fintech leader with over 500 ventures, which raised US$400m in 201739. The city has one of the highest concentrations of AI start-ups in the world with, amongst others, the Vector Institute, NextAI, and the Creative Destruction Lab. Toronto is also home to the R&D labs of Uber, Thomson Reuters, the Royal Bank of Canada, Shopify, Amazon and Google. Google is also looking to apply AI in a new kind of project with the Google Sidewalk Toronto initiative combining forward-thinking urban design and new digital technology to create people-centred neighbourhoods40.
In Montreal, firms can find a hospitable environment offering the cheapest operating costs in North America and a pool of over 90,000 skilled information and communication technology (IcT) workers41. The city has six universities including McGill University and Université de Montreal, which established labs collaborating with the ecosystem by providing talent and knowledge-sharing. The city boasts one of the largest concentrations of AI scientists in the world and one of most respected deep learning research groups led by Yoshua Bengio. Since 2010, the ecosystem has become a global AI hub by attracting the likes of Facebook, Google, Samsung as well as Microsoft and having them invest to take advantage of Montreal’s expertise. Notable start-ups include Element AI, which raised US$102m in Series A funding in 2017, making it one of the first large-scale funding rounds in the world.
Israel
Israel, and particularly Tel Aviv, is an active innovation hub led by four universities and a start-up ecosystem with over 950 active start-ups, 50% of which raised one or more funding rounds42. Over the last four years, there were over 140 new AI centred start-ups created per year across a variety of industries. Israeli start-ups raised over US$7.5bn cumulatively from private and public sources43. AI is part of the national innovation strategy, and the government is supportive, encouraging collaboration with industry and academia. This effort is yielding results with applications in some of the most advanced defence technology in the world; over 30% of border protection is ensured by AI-powered systems in 201844. Furthermore, with a US$275m public investment, there is keen interest to develop digital health AI-powered services.
Other Asian Hubs
Other Asian hubs such as Singapore, Seoul and Tokyo are active players in the AI race in machine learning, deep learning and robotics. Singapore’s AI.SG initiative is a pioneering model, backed by S$150m (US$109m) in investment over five years to attract more resources, talent, and institutional support45. The initiative focusses on applying AI to finance, smart cities, and healthcare. These are all priorities for Singapore, a financial centre with an ageing population and constrained by space. In South Korea, the government has announced investments totalling South Korean won (KRW) 2.2tn (US$2bn) to strengthen AI R&D and build a public–private AI research centre jointly with leading Korean conglomerates46. Meanwhile in Japan, AI development is integrated in an industrialisation road map aiming to increase the use of data-driven AI applications and build ecosystems connecting multiple domains. The strategy focusses in particular on three priority areas for Japan: productivity, health and mobility. Like in many other countries, the policy includes investments across the ecosystem including R&D, talent, start-ups and data47.
Estimated Number of Artificial Intelligence Start-up as at 2018
2. COMMON CHALLENGES
Whilst AI hubs are scattered across the globe, they face similar challenges. There is a growing global war for talent where qualified professionals are scarce, as demonstrated by the demand for AI-related roles that more than doubled over the last three years, and where 40% of companies are reporting difficulties in filling IcT vacancies48, 49. Another challenge AI firms face is the availability of quality data to train AI algorithms. The human mind can learn to make a logical decision based on a few examples, but AI needs millions of data points to learn a similar decision pattern50. The challenge is exacerbated by the difficulty in transferring learning across domains because most algorithms trained in one domain cannot be transferred to perform in another area, creating a need for large data sets51. In addition to data, current computational power can prove insufficient in complex applications such as deep learning where hundreds of overlaid iterations must be processed simultaneously52. Another challenge is that in many circumstances, AI is developed for medical, defence, legal, and financial applications where health, life and human well-being are at stake. Thus, the ethics, regulation and, more broadly, the acceptability and implementation of pragmatic AI-driven solutions might be the ultimate challenge for an AI-driven society53. The rapid pace of AI development and R&D investments creates opportunities to gain efficiency and increase productivity, but it also creates challenges for industries to understand and implement these technologies as well as for governments to regulate AI.
3. ASSESSING NEW OPPORTUNITIES
Accuracy has a global reach aligned with some of the most vibrant AI hubs, such as London, Paris, Montreal and Beijing. The firm benefits from multidisciplinary professionals across a broad spectrum including advanced analytics, strategy, valuation and data science. Accuracy is always exploring forward-looking avenues to apply open innovation technologies and continue providing bespoke advice to fulfil the strategic and financial needs of our clients. Positioned as a partner of choice to guide change, we can help build bridges to innovation ecosystems, leverage data science and assist in creating change with smart investments. In the next series of the perspectives articles, AI driven technology and its applications will be further explored.
AI Hubs and Accuracy Offices
1 Artificial Intelligence Definition. Oxford Dictionaries.
2 Interview at GS (2018). Eric Schmidt: “The Artificial Intelligence Revolution”.
3 British Council (2018). “Hubs, clusters and regions”.
4 MIT Artificial Intelligence Podcast (2018). Yoshua Bengio: “Deep Learning”.
5 QS World University Rankings 2018 (2018).
6 CB Insights (2017). “The 2016 AI Recap: Startups See Record High In Deals And Funding”.
7 CB Insights (2017). “Top acquirers AI startups M&A timeline”.
8 CAICT (2018). “World AI Industry Development Blue Book”.
9 US Office of the President, National Science and Technology Council (2016). “US National Artificial Intelligence Research and Development Strategic Plan”.
10 Startup Genome (2018). “Global Startup Ecosystem Report”.
11 Asgard.VC, Rolland Berger (2018). “A Strategy For European AI Startups”.
12 MIT News (2018). “ MIT reshapes itself to shape the future”.
13 NYU. Tandon Labs “Research and Innovation”.
14 Efinancial Careers (2018). “The top machine learning teams in investment banks”.
15 Statista (2018). “Impact of artificial intelligence (AI) on the gross domestic products worldwide in 2030”.
16 Department of International Cooperation Ministry of Science and Technology (MOST), P.R.China (2017). “Next Generation Artificial Intelligence Development Plan”.
17 Reuters (2018). “China’s city of Tianjin to set up $16-billion artificial intelligence fund”.
18 CAICT (2018). “World AI Industry Development Blue Book”.
19 CB Insights (2018). “Top Ai Trends to Watch in 2018”
20 CB Insights (2018). “Top Ai Trends to Watch in 2018”.
21 European Commission (2018). “ESPC The Age of Artificial Intelligence Towards a European Strategy for Human-Centric Machines”.
22 Mayor of London, CognitionX (2018). “London: The AI Growth Capital of Europe”.
23 IBID
24 IBID
25 Tech City UK (2017). Tech Nation Visa Scheme “High demand as scheme enters fourth year”.
26 Deepmind (2018).
27 French Government (2018). “Artificial Intelligence: Making France a leader”.
28 IBID
29 IBID
30 CAICT (2018). “World AI Industry Development Blue Book”.
31 German Government (2018). “Key points for a Federal Government Strategy on Artificial Intelligence”.
32 IBID
33 Politico, J. Delcker (2018). “Germany’s €3B plan to become an AI powerhouse”.
34 German Research Center for Artificial Intelligence (DFKI) (2018).
35 CyberValley (2018). ‘’ Artificial intelligence finds a home’’.
36 German Government (2018). “ Key points for a Federal Government Strategy on Artificial Intelligence”.
37 Government of Canada (2018). “Canada – A leader in Artificial Intelligence”.
38 Startup Genome (2018). “Global Startup Ecosystem Report.”
39 DMZ Tech Incubator (2018). “ How Canada became a hotspot for artificial intelligence research”.
40 Google Sidewalk Toronto (2018). ‘’Sidewalk Labs is reimagining cities to improve quality of life’’.
41 Montreal International (2018). “Greater Montreal: An Artificial Intelligence Hub”.
42 Startup Genome (2018). “Global Startup Ecosystem Report.”
43 StartupHub.ai, Daniel Singer (2018). ‘’ Israel’s Artificial Intelligence Landscape 2018 ‘’.
44 Asgard.VC, Roland Berger (2018). “A Strategy For European AI Startups”.
45 Infocom Media Development Authority (2018). ‘’Digital Economy Framework for Action’’
46 Korean Ministry of Science and ICT (2018). ‘’ Government Innovation Plan’’.
47 Government of Japan (2018). ‘’ Artificial Intelligence: A Rival for Humans, or a Partner?’’.
48 Indeed Hiring Lab (2018). ‘’ Demand for AI talent rises ‘’.
49 Eurostat (2016). “ICT specialists – statistics on hard-to-fill vacancies in enterprises”.
50 MIT Artificial Intelligence Podcast (2018). Yoshua Bengio: “Deep Learning”.
51 University of Wisconsin, L.Torrey, J.Shavlik (2018). “ DeepMind’s latest AI transfers its learning to new tasks “.
52 IBM, Advantage Reports (2018). “The new AI innovation equation”.
53 IBM, Building trust in AI (2018). “Building trust in AI”.