Europe’s dependence on American AI models hit a breaking point this past summer, and outgoing European Central Bank President Christine Lagarde wants to make sure it never happens again.
Her warning comes at a critical moment. The continent already learned painful lessons from Russian energy coercion and China’s rare earth export controls. Now artificial intelligence joins that list of strategic vulnerabilities.
Speaking Wednesday at the Paris Peace Forum, Lagarde framed AI sovereignty as an economic security imperative. She pointed to a specific incident: the United States government ordering Anthropic to block foreign access to Claude Mythos 5 and Claude Fable 5 on June 12. Because Anthropic couldn’t verify user nationality in real time, the company suspended both models globally. The restriction lifted on June 30, but the message resonated across European capitals.
This wasn’t Lagarde’s first warning. In May, she told an audience in Spain that European firms face a competitive disadvantage because American companies can leverage these advanced models. The computing gap tells the story. According to figures she cited in September, the United States controls roughly 75% of global AI compute capacity. Europe holds about 5%.
Meanwhile, Europe races to build alternatives. Lagarde confirmed plans for a digital euro pilot launching in the second half of next year, with 36 payment providers already participating. She acknowledged the project lacks glamour, but argued that without independent payment infrastructure, an outside power could freeze Europe’s money flows. The timing matters: the pilot begins after Lagarde leaves office on February 7, 2027.
Cyber sabotage adds pressure. France blames Iran’s Revolutionary Guard for an October attack that compromised refrigeration systems at roughly 700 French supermarkets and retail locations.
Other European leaders caution against overcorrection. Parliament member Sandro Gozi warned that autonomy must not slide into nationalism or isolationism, pointing to Brexit as evidence of failure. France’s minister delegate for European Affairs argued that strategic independence should build on cooperation with partners.
The Mythos blackout lasted only three weeks. Resolving Europe’s reliance on American AI will take considerably longer.
