The U.S. government's order to disable Anthropic's Fable model for foreign users exposed a critical vulnerability: Europe and other regions depend entirely on American AI systems they cannot control, and building sovereign alternatives is becoming urgent.

What happened: The United States government instructed Anthropic to switch off its most capable model, Fable, for all foreign nationals—a commercial product disabled four days after release based on a claim, which Anthropic disputed, that its safeguards could be jailbroken. The action illustrated how advanced AI infrastructure is concentrated in American hands under American law, with access subject to sudden policy changes.

Why it matters: European hospitals, banks, and defense ministries increasingly depend on American AI systems whose operation can be switched off by a single directive from Washington. As AI systems begin to improve themselves and become the input to all economic sectors, countries without frontier models of their own lose control over the inputs to everything else. Europe's regulatory efforts like the EU AI Act are no substitute for actually building and owning AI infrastructure on home soil.

What to watch: Europe faces a company worth almost $1tn and American tech giants spending an estimated $450bn a year on AI infrastructure, while a single American site, xAI's Colossus in Memphis, runs more than half a million GPUs. Sovereignty requires a credible domestic model good enough that being cut off is survivable rather than catastrophic—a capability gap that compounds monthly as frontier systems improve faster than alternatives can chase.