Love it or hate it, artificial intelligence, and generative AI in particular, is the tech story of 2024.
OpenAI may have monopolized most of the attention and funding so far, thanks to its viral services like ChatGPT and its billions of dollars in funding, but a new report from analysts at top VC firms Accel and Dealroom says there's a wave of up-and-coming companies in Europe and Israel looking to make a name for themselves right now.
Europe and Israel combined typically account for around 45% of all venture funding annually, but when you scope this out to the specific field of AI, the percentage drops to less than half that amount (and even lower for generative AI). This could be taken as a sign that Europe and Israel are falling behind in the market, or, more optimistically, it means that we'll see a lot of interesting developments in the coming months and years as the region catches up.
Investors are now chasing the next big thing, which may not be as priced out of the U.S. Interestingly, Accel partner Harry Nellis said one of the reasons this report came to fruition is because his firm has been hard at work evaluating all the generative AI startups popping up across the region to determine which ones to invest in. Keep an eye on this space.
For now, here are some of the most interesting data points from the report:
London is the city that has produced the most GenAI startups.
Of the 221 startups analyzed by Dealroom and Accel, roughly 27% were founded in London, nearly a third of the total. Tel Aviv came in second with 13%, followed by Berlin at 12% and Amsterdam at 5%. Interestingly, Paris, a city that has been talked about for some time as a hotbed of AI development, only made it into the middle of the city rankings with 10%.
But these startups are having a big impact.
GenAI startups founded in France are the most highly funded
French startups describing themselves as working in the field of generative AI have raised a total of $2.29 billion to date, more than any other country in Europe and more than Israel. Recent rounds include Mistral AI, which raised $640 million earlier this month (it had raised over $500 million prior to that), “H”, which raised $220 million a few weeks ago, and Poolside, which is reportedly in the midst of a larger fundraising round.
Other notable AI startup activity in Paris includes Hugging Face, an open-source repository of machine learning models that raised $235 million in August 2023, and Kyutai, a new research-focused organization that has raised hundreds of millions of euros in funding to make waves in open-source AI models.
Why are some areas so much better than others?
France's $2.29 billion is roughly equal to the amount raised by the next three countries combined: the UK, with $1.15 billion in funding for generative AI startups (big names include Stable Diffusion makers Stability AI, Synthesia, and PolyAI), Israel with $1.04 billion (startups like AI21 and Nvidia's recently acquired Run:ai), and Germany with $636 million (mostly driven by Aleph Alpha's $500 million raise last year). Outside of that, other countries in the region have raised less than $160 million each, sometimes significantly less, with amounts in the low seven-figure range.
Niels believes the main reason for this is the perfect combination of strong educational institutions that not only produce a large number of talented engineers but also attract major tech companies to set up their own operations to take advantage of that talent.
“The fact that Paris produces so many entrepreneurs speaks to the importance of real long-term investment in education,” says Nélis. “The same is true in London, where universities like Cambridge, Oxford, and UCL produce talent.” But the intermediate step between academia and entrepreneurship isn't immediate. The intermediate step is working for a big tech company that was founded to improve recruitment.
“Universities are obviously very important in attracting hyperscalers,” Nellis said, citing as examples Facebook/Meta setting up an AI research lab in Paris early on, and Google eventually setting up a similar facility in Paris, and already building operations with DeepMind in both London and Paris.
“Founder factories,” or hyperscale tech companies, are a big part of this story.
Sure, startups may feel like the melting pot of AI development, but big tech companies play a big role in fueling the fire. If you look at the long tail of GenAI startups, roughly 25% of their founders previously worked at Alphabet (DeepMind or Google), Apple, Amazon, Meta, or Microsoft (let's call them MAAMAs). The higher you go, the more clubby it gets: of the top 10 of these startups, 60% of the founders are from one of the MAAMAs.
In fact, one company in particular stands out as a clear feeder for AI founders.
As the field matures and grows, that's not a very good message for people coming from outside that group, but it's likely that group will evolve and expand as well.