The Race to Build the DeepSeek of Europe Is On
Against that background, Europe's reliance on American-made AI is starting to look more and more like a liability. In a worst-case scenario, although experts consider the possibility remote, the US could choose to withhold access to AI services and crucial digital infrastructure. More plausibly, the Trump administration could use its dependence on Europe as leverage if the two sides move on iron out of a trade deal. “That dependence is a liability in any negotiation – and we will negotiate more and more with the US,” says Taddeo.
The European Commission, the White House and the UK Department for Science, Innovation and Technology did not respond to requests for comment.
To protect against these risks, European nations have tried to spend the production of AI on land funding programs, targeted deregulationand partnerships with academic institutions. Some efforts have focused on building competitive large language models for native European languages, such as Aperture and GPT-NL.
However, as long as ChatGPT or Claude continues to outperform European-made chatbots, America's lead in AI will only grow. “These domains are very often winner-takes-all. If you have a very good platform, everyone goes there,” says Nejdl. “Not being able to produce state-of-the-art technology in this field means that you will not catch up. You will always just feed the bigger players with your input, so that they will be even better and you will be more behind.”
Mind the gap
It is unclear how far the UK or the EU intend to take the pressure for “digital sovereigntylobbyists claim. Does sovereignty require total self-sufficiency across the extended AI supply chain, or only an enhanced capability in a narrow set of disciplines? Does it require the exclusion of US-based providers, or only the availability of domestic alternatives? “It's rather vague,” says Boniface de Champris, a senior policy manager at the Computers & Technology Association for communications companies. be more of a narrative at this stage. “
There is also no broad agreement on which policy levers to pull to create the conditions for Europe to become self-sufficient. Some European suppliers are advocating a strategy where European companies are required, or at least encouraged, to buy from their own AI companies – similar to China reported approach to their domestic processor market. Unlike grants and subsidies, such an approach would help seed demand, says Ying Cao, CTO at Magics Technologies, a Belgium-based outfit that develops AI-specific processors for use in space. “That's more important than just access to capital,” Cao says. “The most important thing is that you can sell your products.” But those who advocate for open markets and deregulation argue that attempts to eliminate US-based AI companies risk putting domestic companies at a disadvantage to global peers, left to choose which AI products suit them best. “From our perspective, sovereignty means having a choice,” says de Champris.
But for all the disagreement over policy minutiae, there is a broad belief that bridging the performance gap to US leaders remains eminently possible for even budget- and resource-constrained labs, as DeepSeek illustrated. “If I thought we wouldn't catch up, I wouldn't [try]”, Nejdl says. SOOFI, the open-source model development project Nejdl is involved in, aims to deploy a competitive general-purpose language model with around 100 billion parameters within the next year.
“Progress in this field will largely no longer depend on the largest GPU clusters,” claims Nejdl. “We will be the European DeepSeek.”
