Banking industry investment into early stage AI companies is being dominated by five US banks, according to new research into industry-wide artificial intelligence (AI) innovation from AI benchmarking & intelligence platform Evident. The report focuses on six levers of AI innovation – in-house research, patents, strategic investments, ecosystem partnerships, open source engagement, and cross-business idea generation.
“AI innovation is critical for banks to gain and maintain competitive advantage, and our research shows that North American banks continue to draw ahead of their European counterparts,” said Alexandra Mousavizadeh, Evident co-founder and CEO, in a statement. “Pulling the different levers of AI innovation not only helps to drive efficiencies in day-to-day banking operations, but it offers a map to the future of the industry, and the opportunity to fundamentally reimagine what it means to be a bank. Crucially, this competitive advantage is likely to be self-sustaining – a big problem for any banks that fall too far behind in the AI innovation race today.”
Ten key takeaways:
1. Different strategies are emerging when it comes to AI innovation in banking. All banks are thinking about how AI can enable their bank to operate better, faster and more efficiently. But a handful of banks appear to be pursuing cutting-edge AI innovation. They are characterized by:
- A focus on pure and applied AI research. Scarcity and cost of research talent make this is an expensive strategy that few banks can follow
- A volume of highly cited AI-related patents. Once again, this is a strategy that not every bank can follow – especially in Europe where patent filing is more restricted
- A strong ecosystem: participation in the open source community, as well as a wide range of partnerships with universities, accelerators and vendors Χ Strategic investments into early stage companies at the cutting edge of AI – ideally with an internal venture rather than a VC returns focus alone
2. Concentration at the top is clearly emerging. In each area of this report the top five banks have:
- Published 67% of the AI research
- Filed 94% of the AI patents
- Made 51% of the AI investments
3. North American banks are accelerating away from their European competitors. They:
- Published 80% of all AI research in 2022
- Filed 99% of all the AI-related patents in 2021 (the most recent full year of filings)
- Made 60% of all AI-related investments in 2022
- While there are structural reasons for this (e.g. AI patenting is more restricted in Europe), these findings reinforce the message from our previous work on the Evident AI Index and the Evident AI Talent Report published earlier this year
4. The same names consistently appear: JPMorgan Chase, Capital One, Wells Fargo alongside RBC (and in some instances TD Bank) from Canada
- Combined with their strong showing in our Talent work, we can see leadership positions being taken that will become increasingly hard to overcome
- The long tail of North American banks will find that competition with the top tier banks is going to get even harder
5. We see comparatively little AI innovation across European banks
- BNP Paribas typically leads the French banks but is outside the top 10 in almost all categories
- The UK banks rarely feature amongst the leading banks in any areas (the exception is Barclays’ top 10 performance on AI-related ventures)
- It is possible that a Spanish bank like BBVA could emerge as a European champion, a space that awaits a serious claimant
- Whilst we only touch on the Rest Of World banks, institutions in APAC currently look to be the only significant potential AI innovation competitors to the North Americans
6. Publication of AI research across the banks and payment providers has grown at a CAGR of 70% from 2017 to 2022
- JPMorgan Chase leads the field in terms of volume of published AI research papers (nearly 300 papers published since 2017) and team size (more than 120 AI researchers at the bank)
- We find more than 650 individuals working in AI research positions across the banks – 40% have joined the banks since the start of 2022
- India is second only to the US in terms of the number of AI researchers, indicating a significant pool of high caliber AI talent in the country
- Research focuses on a mix of applied and theoretical research, with a focus on deep learning
- We see a high proportion of researchers in trading relative to published papers, suggesting that this is an area with significant unpublished research
7. Patent registration is a strategy being seriously pursued by only six of our 60 banks, with a 40% CAGR (2017-2021) and more than 1,400 patents filed in 2021 alone
- Bank of America has historic dominance in this area, but is being challenged by Capital One, which filed 477 patents in 2021 (more than any other bank)
- The escalating pace of patent filing suggests that this is an area where first-mover advantage, and protective lock-in, is being eagerly sought
- Nearly 15% of all patents filed by banks are focused on trading, followed closely by payments and compliance
8. Ecosystems are being built with multiple partners
- Universities provide recruitment, research and accelerator opportunities – and smart banks are broadening their access to global ideation by stepping beyond purely domestic relationships
- Capital One’s engagement with open source is a key part of building its brand and IP – similarly for JPMorgan Chase, BBVA and ING 9. Investment in AI companies is also growing – at 15% CAGR from 2017 to 2022
- The top five AI investors are all US banks, accounting for almost 50% of all investment deals from 2017 to 2022
- Wells Fargo is the clear leader in this space, followed by Goldman Sachs and First Citizens (after its acquisition of Silicon Valley Bank)
- European banks are challenging US dominance: 90% of AI-related investments were made by US banks in 2015; this has since decreased to 60% in 2022
- French banks lead the European charge – BNP Paribas and Credit Mutuel – followed by Barclays.
- Not all investments are strategic: around 70% of AI investments look to offer potential assistance to banks in their core mission or functions, while 30% are non-bank related
10. Does Generative AI change the rules of the game?
- To date, AI innovation has been a top-down game characterized by strong leadership, aggressive focus of resources and highly educated AI specialists
- Advancements in Generative AI have reinforced this – banks need a centralized AI innovation strategy, with group-wide orchestration and leadership
- That said, easy access to Generative AI tools like ChatGPT also allows for new levels of bottom-up innovation, creating new working practices and business tooling
- Those that manage to let their staff test and share their new AI hacks will have cracked possibly the biggest innovation opportunity on offer: how to genuinely become learning organizations.
“Banks are investing in startups at the cutting edge of AI innovation to accelerate adoption, feed their product development and generate future returns. And while the field has so far been dominated by a relatively small number of highly active investors, as the AI race accelerates, we expect to see many more banks scanning both domestic and global startup ecosystems in search of the brightest investment prospects,” said Mousavizadeh in a statement.