The Commerce Department dropped an export control directive on Anthropic’s two newest AI models, Mythos 5 and Fable 5, restricting foreign nationals from accessing them. Rather than try to figure out which users are American and which aren’t, Anthropic simply turned both models off for everyone worldwide the next day.
When the government tells a cloud-based AI company to stop serving foreigners, the company faces an almost impossible technical problem. So Anthropic, led by CEO Dario Amodei, took the nuclear option. Every user, domestic and international, lost access on June 13.
What triggered the shutdown
The directive landed on June 12, reportedly prompted by a jailbreak vulnerability that third parties flagged shortly after the models went public. In plain terms, researchers found a way to bypass the safety guardrails baked into both Mythos 5 and Fable 5, potentially allowing misuse of the systems’ most advanced capabilities.
This wasn’t exactly a bolt from the blue, either. Anthropic and the Trump administration have been circling each other since early 2026. The tension started in February with disputes over military applications of Anthropic’s technology. Things escalated in March when Anthropic received a “supply chain risk” designation. The export control directive represents the sharpest escalation yet in what has become one of the most consequential standoffs between Silicon Valley and Washington.
