It was over a swanky lunch in Mayfair in 2013 that a local venture capital investor reality-checked the ambitions of what was once Europe’s most important AI company. In between bites of Cantonese cuisine, the founders of DeepMind were telling their backer about their plans to change the world with AI models that were smarter than humans. Chief Executive Officer Demis Hassabis and co-founder Mustafa Suleyman mentioned they’d just come back from Mountain View, California, where they’d rejected a takeover offer in the hundreds of millions of dollars. They wanted to stay independent and reach for the stars.
“Mountain View?” the investor spluttered. “You mean Google?”
The founders nodded. Their investor leaned forward to give them advice oft repeated to British and European startups who dreamed big – advice that would keep holding back the region’s tech ecosystem.
“Take the offer,” he said.
Hassabis was passing up an $80 million nest egg for his young family, the VC investor pointed out. Did the entrepreneur really want to keep living in an apartment in north London? Google would also poach his talent and crush his fledgling business.
Months later, DeepMind sold itself to Google for $650 million. It was a steal for the search giant: Although Hassabis had lacked the oceans of capital that Silicon Valley entrepreneurs had on their doorstep, his base in Europe had enabled him to tap a pool of cheap academic talent.
The UK and the continent are home to nine of the world’s top 25 research institutes, more than any other region including the US and China. Google, Facebook and Microsoft Corp. were only just starting to realize that a branch of artificial intelligence called deep learning was primed to jump-start the field. Yet Hassabis had already hired many of the world’s top deep-learning researchers. Now Google could use their expertise to enhance its services, and DeepMind’s investor could get a tidy return. Europe, meanwhile, lost another technology champion to Silicon Valley.
Europe is the world’s largest producer of scientific publications, spearheading research into deep-tech fields like artificial intelligence. Yet it has just 14% of the world’s unicorns, or startups worth more than $1 billion. It was researchers at Britain’s University of Cambridge who first mapped the structure of DNA, split the atom and developed gene therapies; Londoner Tim Berners-Lee invented the world wide web in Switzerland, and Alan Turing laid the foundations of artificial intelligence at Bletchley Park in England. All those breakthroughs were commercialized and scaled by American firms.
Europe’s focus on democratizing technology isn’t always a drawback. Berners-Lee famously insisted that the world wide web be made freely available to everyone, a decision that unleashed global creativity and innovation. But sharing — and not selling — technology comes at a price. Europe is a distant third behind China and the US in the race to build the most advanced AI models.
Tech influencers have sneered at the region, posting images of the iPhone next to single-use plastic bottles with tethered caps to show what US and European “innovation” look like side by side. Many have blamed the continent’s heavy focus on regulation and laws like the European Union’s Artificial Intelligence Act, a sweeping, world-first effort to govern AI responsibly. But regulation is no villain in this story. To say so is to embrace the spin from large tech companies pushing back against sensible, necessary laws that could disrupt their business models.
Europe’s real AI problem is money, specifically its ongoing failure to invest large amounts in its most innovative firms. And the solution, at its most basic level, is simply more of it. Take Wayve Technologies Ltd., an autonomous-driving startup that’s also the latest jewel in the British AI crown. The company has developed systems that let modern cars drive themselves using just cameras and data, and with a more flexible approach to navigation than Tesla Inc. In 2024, it raised $1 billion in a funding round led by SoftBank Group Corp., with Microsoft and Nvidia Corp. making up the other largest investors. Hardly any of the money came from the UK or Europe.
“We had to look offshore for that support,” says Wayve CEO Alex Kendall from his office in Kings Cross, London, a short walk from the grander headquarters of Google DeepMind. He is pragmatic about why: “Silicon Valley has been at this since the 1960s.” The Bay Area has a decades-long head start in building a thriving tech ecosystem through cycles of boom and bust, cultivating generations of entrepreneurs and investors. Less than 10% of European venture capitalists have previously founded companies, compared with 60% of VCs in Silicon Valley.
European startups that want to raise early-stage capital in the tens of millions of dollars have little trouble doing so. The problem is when those like Wayve need hundreds of millions to start building an empire.
The AI gold rush has made this bottleneck in late-stage funding worse. Europe had been growing its global share of VC funding until 2022, when it reached an all-time high of 22%. But then came ChatGPT, and that share started to decline as US tech companies outspent everyone on the capital-intensive business of building AI models. OpenAI raised a record $40 billion in March. Its San Francisco-based rival Anthropic PBC is in talks to receive $10 billion. Four of the biggest technology firms are expected to spend $344 billion this year, with much of that going to the data centers needed to power AI. The more they spend, the more the funding gap with European tech firms has widened.
There’s a solution for all this, and it has nothing to do with relaxing tech regulations. It lies in an industry that’s a world away from the futuristic vision of AI and hides in plain sight: pension funds. European pension funds invest, on average, just 0.018% of their assets in venture capital, according to research by London-based venture capital firm Atomico, compared with roughly 2% for US pension funds. That difference is significant. European pension funds have more than $10 trillion euros ($11.7 trillion) in total assets under management, but for every $1 they invest in the venture capital firms that are the lifeblood for tech startups, a US fund puts in more than $100. Investing 2% of Europe’s pension assets, or $236 billion, could transform its tech sector.
Their extreme caution is partly cultural, partly due to policy blunders. In Britain, for instance, the government under former Prime Minister Gordon Brown abolished a major tax relief plan for local pension funds in the late 1990s, pushing them away from backing tech startups, or from building the expertise they needed to invest in them.
The opposite happened in the US, with the 1979 reform of the “prudent man rule.” For more than a century, American pension fund managers had been guided by the standard that they should invest only as a reasonably prudent person would, with their beneficiaries’ interests in mind. When the US Department of Labor clarified the law to consider the whole portfolio’s safety — not just each individual investment — pension funds could suddenly back riskier assets like venture capital. Over the next eight years, the share of US pension funds in VC firms rose from 15% to more than 50%, unlocking huge new pools of money for startups and feeding the unstoppable growth of Silicon Valley.
Europe can’t make that same kind of sweeping change. Pension policy is decided nation by nation, and culturally, fund managers still prefer to protect their retirees’ money over maximizing growth.
*Bloomberg
