Europe spent years warning that its dependence on American technology could one day be used against it. On the evening of Friday 12 June, it found out what that looks like.
The order arrived as a letter from US Commerce Secretary Howard Lutnick, imposing export controls on Anthropic's Fable 5 and Mythos 5 models with immediate effect for any foreign national — including the company's own non-American employees. Because it is practically impossible to police nationality in real time inside shared cloud infrastructure, Anthropic was left with one option: shut both models down for everyone, everywhere.
Washington cited national security, acting after another company claimed it had jailbroken Mythos, raising fears the model could be misused. Anthropic pushed back on that rationale, calling the concern narrow and arguing it did not justify pulling access so broadly. The US Justice Department is appealing a court order that paused the ban.
The European Commission moved quickly to register its discomfort. Spokesperson Thomas Regnier said the body was still examining the directive's implications for the EU, amid concern that the United States can simply switch off technology that allied partners may soon rely on heavily. The Commission also warned against discriminatory treatment of European users.
For years, the danger was hypothetical: what happens if Europe builds its economy, research and even its security on tools owned by companies subject to a foreign government's orders? This week it became concrete. A single directive in Washington reached into European offices and laboratories and turned off the lights — not through a cyberattack or a price hike, but with a stroke of an export-control pen.
The political reaction across Europe was immediate. French leaders called on Paris to accelerate its backing for Mistral, the Paris-based startup that is the EU's only serious competitor to Anthropic and OpenAI, and the episode has poured fuel on a wider debate about sovereign cloud capacity and a European technology stack. The problem is scale: Mistral is one company facing American giants with vastly deeper resources.
This is also not an isolated shock. Only weeks earlier, EU finance ministers had sought access to Mythos to help defend their banks against cyberattacks, and Washington said no. Now access has been cut for everyone. Taken together, the two episodes look less like a one-off and more like a pattern — American AI as a lever that can be tightened or released depending on US interests.
The lesson Europe has long discussed in white papers landed in a single weekend: a foreign government can switch off tools embedded in European businesses, universities and public institutions, and there is little Brussels can do about it after the fact. The harder truth is how far behind the alternatives are. Sovereignty is no longer an abstraction — but building a credible European answer to Fable and Mythos will take years, and the clock started long before this week.
