At 5:21 p.m. Eastern on Friday, Anthropic received a directive from Commerce Secretary Howard Lutnick ordering it to obtain government permission before exporting its Fable 5 and Mythos 5 models to any destination worldwide, or to any foreign national regardless of location. Within hours, the company disabled both globally. Claude Opus 4.8 and the rest of its lineup remained live.

The letter, a copy of which was seen by Bloomberg News, threatened criminal and civil penalties for noncompliance and reached, by Anthropic’s account, the company’s own foreign-national employees. It cited national security authorities without specifying the concern. Anthropic says the trigger was a jailbreak claim from another company, first reported by Axios; the company reviewed the demonstration and concluded it surfaced minor, previously known vulnerabilities also discoverable in publicly available models including OpenAI’s GPT-5.5.

“We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people,” Anthropic said in a statement, warning that such a standard would “essentially halt all new model deployments.”

The escalation didn’t arrive cold. In February, President Donald Trump issued an order barring federal agencies from using Anthropic’s models. In early March, the Pentagon designated the company a “supply chain risk,” a label Anthropic is contesting in federal court. Earlier this month, following a funding round valuing it at $965 billion, the company confidentially filed for an IPO.

That sequence is the structural story. An administration that has spent six months building a paper trail of national-security findings against a single frontier lab now has export-control authority pointed at that lab’s two most capable models, at the precise moment the lab is trying to price itself for public markets. Talks with the administration, Anthropic says, are continuing.

Sources