US Government Shuts Down Anthropic's Most Powerful AI Models Over National Security Fears

The U.S. government just pulled the plug on two of Anthropic's most advanced AI models — Claude Fable 5 and Claude Mythos 5 — and the company says it had no choice but to comply.

The directive came Friday at 5:21 pm ET, according to Anthropic's announcement on X. The order forces the company to shut off access for all users globally, not just the foreign nationals the export control order was supposedly aimed at. Other Anthropic models remain unaffected.

Anthropic made clear it disagrees with the decision. The company called the move a mistake, but it didn't detail the government's specific national security concerns. The White House and the Commerce Department, which oversees export controls, haven't commented publicly.

This is a major escalation in the U.S. government's push to regulate powerful AI. Anthropic, co-founded by former OpenAI employees Dario and Daniela Amodei, has positioned itself as the safety-first AI company. Its models are known for being less likely to generate harmful content than competitors like OpenAI's GPT series. But that safety-first reputation may have backfired — the government's action suggests that even cautious AI can be seen as a threat.

The models affected are among the most powerful ever created. Claude Fable 5 is Anthropic's creative writing model, capable of generating long-form fiction and poetry. Claude Mythos 5 is a general-purpose reasoning model that excels at complex analysis. Both were released in early 2026.

The shutdown is unprecedented. While the U.S. has imposed export controls on AI chips and restricted some AI software, this is the first time it has ordered a company to disable existing models for all users worldwide.

What happens next is unclear. Anthropic could challenge the order in court, but national security directives are hard to overturn. The company may also negotiate with the government to allow limited access for research or safety testing.

For now, millions of users who relied on these models for work, creativity, and research are locked out. Developers who built applications on top of Claude Fable 5 and Claude Mythos 5 are scrambling to find alternatives.

This isn't just about Anthropic. The move signals that any AI company, no matter how safety-conscious, could face similar shutdowns. The U.S. government is drawing new lines around what it considers too dangerous to run, even for domestic users.

For now, the AI arms race just hit a new speed bump — and it's Anthropic, the company that built its brand on caution, that got stopped first.