Europeans have produced a lot of AI big hitters. From DeepMind, to Hugging Face and the tech behind Amazon Alexa, entrepreneurs from the continent have produced some major success stories in the industry.
The issue? That all of the names on that list are either now headquartered in the US, or have been bought by Silicon Valley-based companies. So how can Europe scale meaningful AI businesses on home soil, and what kind of infrastructure do they need to help them?
That was the subject of discussion for a special podcast recorded at Slush in Helsinki late last year, when Sifted’s Europe Editor Mimi Billing sat down with some of the continent’s leading AI builders.
Silicon Valley may still dominate when it comes to building the world’s most powerful models but, when it comes to developing the products that large enterprises actually use, there are many reasons to be bullish on Europe — even if there’s not so much VC cash sloshing around.
“Capital, compared to the US, is tighter [in Europe],” said Mona Li, VP of startup ecosystem at AI cloud compute provider Nebius. “That forces a different muscle. They need to build more efficient architectures.”
This lean company-building muscle, combined with entrepreneurial skill in navigating regulations and a growing supply of sovereign compute infrastructure to smooth that process, is fomenting a new generation of resilient, enterprise-ready startups in Europe.
And, with many warning of an AI bubble set to burst in the wake of a US-led funding frenzy, it’s not a bad time to pay attention to these kinds of businesses.
Enterprise ready
Enterprise adoption has been one of the biggest sticking points since the GenAI boom began in 2022. While tools like large language models (LLMs) are very useful for everyday users who might want to write better emails, large companies are still moving cautiously with these new technologies.
“Adoption from enterprises is pretty much in early-stage, pilot mode,” said Li. “It's really hard for startups to move the enterprise from pilot mode to full deployment at this point.”
And, on the face of it, that’s even harder when selling to European companies, compared to their risk-hungrier US counterparts,
We’re already winning deals with enterprises that we wouldn't think we would win, because we have this data advantage over American competitors.
“It’s somehow easier to start in the US just because there's this willingness, to take the risk, burn some cash, invest in the relationship. And see where it takes them ROI-wise,” according to Claire Nouet, chief operations officer of Pathway, who joined the ”Europe’s AI Opportunity” podcast in June last year.
But that doesn’t mean that European AI startups are out of the race. Many US enterprises are operating, or selling products here, and that can give European startups an edge, due to expertise in operating in a complex regulatory landscape.
“There are different levels of data compliance [in Europe]… It's very hard to serve,” Hakob Astabatsyan, CEO and cofounder of voice AI startup Synthflow, explained.
“Being a two-year-old company, we’ve done that and we’re already winning deals with enterprises that we wouldn't think we would win, because we have this data advantage over American competitors.“
The need for homegrown compute
Astabatsyan added that building up this kind of advantage isn’t just about dedicated in-house work: it’s about having the right infrastructure to support you.
“You cannot solve this problem alone, these are complex setups,” he said, adding that companies like his need compute providers who understand these challenges.
“That’s what we expect from our cloud provider.”
You have to make sure that the infrastructure you have is 100% compliance-ready with local policy.
Much of this is about understanding and building for the myriad of different regulations found in Europe to give enterprise clients confidence in your product, Li explained.
“You have to be compliance-ready in each region. All enterprises care about this: where’s your data and under what jurisdictions,” she said. “You have to make sure that the infrastructure you have is 100% compliance-ready with local policy. That’s what Nebius does.”
It’s for reasons like this that, back in June on the Europe’s AI Opportunity podcast, Thomas Wolf, cofounder and chief science officer at Hugging Face emphasised that it was “extremely important” to try and build more compute capacity in Europe.
That urgency isn’t going anywhere, Luca Bocchio, partner at global VC firm Accel, told Billing at Slush: “Infrastructure-wise, governments and different countries need to make sure that we do build things fast instead of procrastinating.”
Thinking bigger
If infrastructure providers can give European startups the tools they need to build complex AI products that are also compliant, could be well set to thrive in the coming years, even if the industry bubble does burst.
Bocchio, who is also an investor in Synthflow, described why the company caught the attention of his Silicon Valley-headquartered VC firm.
The challenge of building in a complex market like Europe actually inspires entrepreneurial DNA.
“For us, the exception with Synthflow was the ability to build a lot, with very limited resources,” he said.
“This is a competitive space, but a lot of players have been investing a lot of money to get relatively not far… It’s important to demonstrate that you can do a lot, quickly, iterate, move without having to invest too much too soon.”
This demonstrates how challenges faced by European startups can actually create great companies, a point echoed in July on the Scaling Europe podcast, by Steffen Tjerrild, cofounder of video AI scaleup Synthesia, which counts 90% of Fortune 500 companies as clients.
It's about setting the mission level high enough and not being complacent.
“The challenge of building in a complex market like Europe actually inspires entrepreneurial DNA,” he argued.
Li, Astabatsyan and Bocchio rounded off the conversation at Slush with their best advice for European entrepreneurs. The main takeaways: choose your partners wisely early on, and don’t let the narrative that Europe lags behind the US dampen your ambition.
“Don’t just think about the first market, I think we have that mindset limitation in Europe sometimes,” said Astabatsyan. “It's about setting the mission level high enough and not being complacent.”
