Europe should accelerate the development of its own cutting-edge AI models, politicians across the continent said.
BRUSSELS — The U.S. order to suspend foreigners’ access to Anthropic’s latest models is a stark reminder for Europe that it’s heavily reliant on American AI models — and that the window to change that is closing.
The U.S. Commerce Department on Friday ordered Anthropic to cut non-U.S. citizens’ access to the company’s latest models. Anthropic suspended access for everyone, including U.S. citizens.
Europe got the message: for the first time, Washington had pulled the plug on Europeans’ access to U.S. cutting-edge technology.
The use of a “kill switch,” long seen as a hypothetical threat, is a wake-up call for Europe.
Over the past year, Trump’s erratic behavior over Greenland and the EU-U.S. trade deal created momentum to reduce Europe’s reliance on U.S. technology in areas such as cloud, chips and AI models.
But barely a week after the European Commission presented plans for the EU to become more technologically sovereign, the U.S. move shows that time is running out.
Europe should accelerate the development of its own cutting-edge AI models to avoid being left behind, European politicians said in unison on Saturday.
The week ahead offers ample opportunities to discuss that effort. From Monday to Wednesday, G7 leaders convene in France. On Tuesday, European Parliament lawmakers will discuss the Commission’s tech sovereignty plans. At the end of the week, EU leaders will huddle in Brussels.
European Commission spokesperson Thomas Regnier said the Anthropic case “further underlines Europe’s need for technological sovereignty.”
That sentiment was echoed across the continent.
“This sudden decision comes to remind us that artificial intelligence is already a major issue of national sovereignty,” said Jordan Bardella, leader of France’s far-right National Rally and front-runner in the polls for the French presidential election. “Nations that do not quickly develop their own [models] will always depend more and more on the choices of other powers.”
Several fellow European Parliament lawmakers joined Bardella in underscoring that Europe needs to develop its own frontier artificial intelligence models.
“Europe cannot keep building its tech stack on access that can be switched off overnight by a foreign government,” said Finnish conservative MEP Aura Salla. “We must take action to reserve our data and our market preliminary for European tech to scale it and build our own frontier AI.”
Bulgarian conservative lawmaker Eva Maydell said Europe needs “to identify which of our own offenses and defenses we wish to build up, how to do so quickly, and how to work with trusted partners.”
Many reactions focused on France-based Mistral, the EU’s most credible AI contender, which is reportedly in talks to raise €3 billion in funding at a €20 billion company valuation.
The topic united politicians across the French political spectrum. Edouard Philippe, former French prime minister under President Emmanuel Macron and a contender in next year’s presidential elections, urged Europe to “avoid the vassalization.”
He offered several policy options, such as a Buy European Tech Act, re-upping the French push to leverage government contracts to boost Europe’s tech sector, and preferential access to decarbonized electricity for European players.
French Digital Minister Anne Le Hénanff focused on AI’s significant energy needs, highlighting France’s advantage in that space.
“Thanks to abundant, decarbonized and competitive electricity, we are a natural home for the data centers and infrastructure that will drive the AI of tomorrow,” she said.
Other politicians focused on the need for Europe to quickly build up an offering of computing power, needed to train the most complex and advanced AI models.
German Greens MEP Sergey Lagodinsky pushed for the EU to team up with “middle powers,” such as Canada, Australia, Singapore and the U.K., “to pull together compute capacity through a consortium approach.”
He also said the Commission “must get serious” about its plan to build several AI gigafactories, massive AI compute hubs to train the models.
Océane Herrero contributed reporting.