Trump gives Anthropic 90 minutes to pull down Fable 5

The Trump administration gave Anthropic 90 minutes on Friday to pull down its most powerful models before a licensing regime took effect, a move that turned a technical product decision into a national-security confrontation. The clash centered on Fable 5 and Mythos 5, and on how quickly the White House was prepared to act once officials concluded the models should not be broadly available.

Anthropic said the order forced an immediate shutdown. “The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,” the company said Friday evening. Axios reported that the administration was blocking foreign governments, companies and individuals from accessing Anthropic’s most advanced models, a restriction that left non-U.S. customers and entities, including European Union institutions, scrambling for access.

The speed of the intervention underscored how rapidly the policy frame around frontier AI has hardened in Washington. The White House had already issued an executive order on June 2, 2026, titled “Promoting Advanced Artificial Intelligence Innovation and Security,” signaling that it intended to pair AI acceleration with tighter national-security controls. That followed a July 23, 2025, order promoting the export of the American AI technology stack, showing that AI had already been treated as both an economic and strategic asset before the Anthropic fight escalated.

The dispute also hit users at the worst possible moment for Anthropic’s rollout. The company had just made Fable 5 generally available days earlier, and Forbes reported that the model was priced at twice the standard rate of Anthropic’s prior flagship. Free access on paid subscriptions was set to expire on June 23, 2026, adding another deadline into a rollout already defined by scarcity, cost and limits. Forbes also reported that Fable 5 was designed to handle long-running tasks and agents, but with higher costs, tighter safeguards and usage restrictions.

Customers were immediately asking whether Anthropic would offer refunds after the withdrawal. Forbes reported mixed results, with some EU customers finding better success through Anthropic’s in-app support chatbot and the EU Consumer Rights Directive’s 14-day cooling-off period becoming a key factor in refund requests.

The broader political warning had been building for months. In March 2026, Politico reported that lobbyists, investors and a former Trump adviser warned the Anthropic fight could damage the administration’s AI agenda. Axios reported on June 9 that Anthropic had unveiled a safer general-use Mythos-class model, but the latest confrontation made clear that frontier AI is now being judged less like ordinary software and more like a strategically sensitive system subject to emergency state power.