Like it or hate it, artificial intelligence – especially generative AI – is the technology story of 2024.
OpenAI, with its rollouts of viral services like ChatGPT and billions in funding, may have gobbled up the lion’s share of attention and money so far. But according to a new report out from top VC Accel and the analysts at Dealroom, in Europe and Israel there is a wave of hopefuls emerging to make their mark.
Together, Europe and Israel typically make up some 45% of all venture funding annually, yet when you translate that to the specific sphere of AI, the proportion drops to less than half of that — and generative AI even less. You can take that as a signal that Europe and Israel are lagging in the market. Or more optimistically, it means that we will see a number of interesting developments in the months and years ahead as the region catches up.
Investors are now on the hunt for the next big thing, potentially at prices that are less inflated than in the U.S. Interestingly, Accel partner Harry Nelis tells me that one of the reasons this report materialized is because of his firm’s efforts to evaluate all the generative AI startups emerging across the region. Here’s what they learned while trying to determine what to fund:
London has generated the most GenAI startups.
Of the 221 startups Dealroom and Accel analyzed, some 27%, nearly one-third of the group, were created in London.
Tel Aviv took the number-two slot at 13%; Berlin 12%; and Amsterdam 5%. Interestingly, although Paris is the city everyone has been talking about for a while as a hotbed for AI development, it found itself very much in the middle of the city rankings, at 10%.
But those Parisian startups are packing a punch.
French-founded GenAI startups are raking in the most money.
Collectively, French startups that self-describe themselves as working in the field of generative AI have raised $2.29 billion to date, the most of any country across Europe, and more than Israel. Recent rounds have included Mistral AI raising $640 million earlier this month, atop more than $500 million before that; and “H” raising a $220 million seed round, astonishingly, a few weeks ago. Meanwhile, the foundational player Poolside, which moved its headquarters from the U.S. to Paris last August, is reportedly also in the midst of raising a hefty round.
Other notable AI startup activity in Paris includes Hugging Face, the open-source repository for machine learning models, which raised $235 million in August 2023, and a new research-focused organization called Kyutai, which itself is armed with hundreds of millions of euros to make some waves in open-source AI models.
Why do some places perform so much better than others?
Altogether, France’s $2.29 billion is nearly as much as the next three countries have raised combined. The U.K. has seen $1.15 billion in generative AI startup funding (Stable Diffusion maker Stability AI, Synthesia, and PolyAI are among the bigger players in the region); Israel has seen $1.04 billion, owing to startups including AI21 and Run:ai, which Nvidia recently acquired; and Germany’s haul has hit $636 million, with Aleph Alpha’s $500 million round last year accounting for the bulk of that amount.
Beyond that, other countries in the region have pulled in less than $160 million each — sometimes significantly less, with some countries in Europe seeing overall funding in the very lowest realms of seven digits.
Nelis believes that the numbers reflect where some of the strongest educational institutions are situated and that produce lots of technical talent, as well as large tech companies that build out their own operations to tap that talent.
“You can see the importance of real, long-term investment in education yielding a lot of founders in Paris,” Nelis said. “The same goes for London feeding from schools like Cambridge, Oxford and UCL.”
The step between universities and founders, however, isn’t immediate: the middle stage has been, for many, working in big tech companies, which set up shop to improve recruitment.
“Universities are clearly very important for attracting hyperscalers,” Nelis said, citing Facebook/Meta establishing its AI research labs in Paris early on, as well as Google eventually establishing a similar set-up there, having already built out an operation with DeepMind both in London and Paris.
“Founder factories” – hyperscaler tech companies – are a big part of the story.
Indeed, while startups may feel like the crucible of AI development, big tech has a major role to play in feeding the flames.
Looking at the long tail of GenAI startups, some 25% of them have founders who previously worked at Alphabet (DeepMind or Google), Apple, Amazon, Meta, or Microsoft (let’s call the group MAAMA). It gets even more clubby the higher up you go. Among the top 10 of these startups, a full 60% of the founders come from one of the MAAMAs.
In fact, one company in particular — Google — stands out as a clear AI founder feeder, exceeding even some of the most prestigious universities in the world once you count people up.
MAAMA mia! It’s not a brilliant message for outsiders and outliers — although that, too, is likely to evolve and expand as the field matures and grows.