Newsletter

EU can compete with US AI giants, says Brussels’ frontier-model champion

The FT's Europe Express newsletter of 29 Oct leads with Barbara Moens' interview with Uljan Sharka, founder and CEO of Italian company Domyn, which leads the 'Europa consortium' building an open-source frontier AI model under the European Commission's Apply AI strategy. Domyn has begun training the model on European supercomputing infrastructure and expects to release it ahead of its one-year deadline in June. Sharka says 'Europe has all the means and opportunities to actually build its own proposition in the market... The only limit is Europe itself in the sense that Europe is not investing enough', argues that Europe's linguistic diversity and industrial expertise are data advantages, and warns that 'if we build only software, we're going to have great products, but then someone can cut off electricity'. He calls risk-averse technical decision makers, and Europe's private sector's reluctance to buy homegrown tech, 'the real bottleneck'. Counterweights in the piece: it notes concerns that Europe is falling irreversibly behind OpenAI, Google and Anthropic and that Brussels' focus has shifted to AI applications. Caveats: it is a short newsletter item based on a single interview with the company leading the effort, and gives no euro amounts, GPU counts or AI gigafactory details. A second item covers a Copenhagen Economics study for the Managed Funds Association (a hedge fund trade body) linking public short-position disclosure to +2.9% daily volatility; the Commission says it has no plans for a review.

Why it matters

Domyn is the EU's chosen champion for a sovereign open-source frontier model, with a release targeted before June, so it is a European AI company to track closely in Dealroom. Sharka's argument that the bottleneck is European corporate demand, not talent or capital alone, is a useful framing for the sovereignty and scaleup-financing debate. The lack of any budget or compute numbers means the ambition should be tested against what is actually funded.

Read the full article: Financial Times

More top stories