Technologie · US kill switch · Archives18 jours, un ordre : comment un « kill switch » américain a coupé l’Europe de l’IA de pointeRédigé 7 septembre 2026 à 21:22 UTCRegarderTranscription0:00For weeks last month, a US government0:02order cut off Europe from Anthropic's0:03most advanced AI, and the continent is0:05still absorbing what that means. The US0:07Commerce Department ordered Anthropic to0:09block foreign access to its latest0:11models in June. Although the ban only0:13lasted 18 days before it was lifted, the0:15episode exposed a painful truth.0:17Europe's AI future currently runs0:18through Washington. When the US0:20government flipped the switch, European0:22users and businesses lost access0:24overnight to Anthropic's frontier0:25models. No European regulator had a say.0:28No European court could stop it. The0:30decision came from Washington. The cost0:32was paid across the EU. German State0:34Secretary Thomas Jarzombek called it a0:36wake-up call for European tech0:37sovereignty. In an interview with DW, he0:40said the episode laid bare just how0:42dependent Europe has become on American0:44artificial intelligence. Now Germany is0:46racing to build its own path. The0:48government is pursuing tech sovereignty0:50through procurement policy, favoring0:52European AI providers, and by expanding0:54domestic data center capacity to host0:56and train advanced models on home soil.0:58The question for Europe is whether this1:00moment becomes a turning point or just1:02another warning. The ban lasted 18 days,1:05but the signal was permanent. The United1:07States can unplug its allies from1:08frontier AI at any time. Europe is now1:11deciding how to build a future where1:12that vulnerability no longer exists.Lire l’articleChronologie de l’histoireHeyDexLe format court ↗
