The US government has applied export controls to a commercially available AI model for the first time, cutting off foreign nationals from Anthropic's Fable 5 and Mythos 5.
The US Department of Commerce issued a directive on June 12 ordering Anthropic to suspend access to its Fable 5 and Mythos 5 AI models for all foreign nationals worldwide. It is the first time export controls have been applied to a commercially available AI model already in widespread use.
How the directive came together #
The chain of events that led to the order moved remarkably fast. Amazon CEO Andy Jassy raised concerns about a potential jailbreak vulnerability in Fable 5, flagging what he characterized as a national security risk. Within 24 hours of those communications, the Commerce Department had issued its directive.
White House officials had reportedly urged Anthropic to address the issues voluntarily before the formal order came down. Anthropic CEO Dario Amodei has engaged in discussions with senior White House officials multiple times throughout 2026 on matters related to access restrictions and security designations, suggesting this confrontation had been building for some time.
Over 100 cybersecurity professionals have called on the administration to rescind the restrictions. Their argument is counterintuitive but worth understanding: cutting off allied nations from these tools doesn’t just inconvenience them, it actively weakens the collective security posture of the US and its partners while potentially strengthening adversaries who will develop or acquire similar capabilities regardless.
The collateral damage extends to allies #
The directive doesn’t just target adversarial nations. It affects all foreign nationals, which means close US allies who had integrated Fable 5 and Mythos 5 into their own cybersecurity operations are now locked out.
The order also impacts Anthropic’s own foreign employee base, creating operational challenges for a company that, like most major AI firms, draws talent from around the world.
Amazon is a major investor in Anthropic and also Anthropic’s primary cloud infrastructure provider through AWS. Jassy’s role in triggering the export controls adds a complicated corporate dimension to an already complex geopolitical situation.
What this means for the AI market and investors #
The Commerce Department has now demonstrated that it can and will apply export controls retroactively to AI models that are already commercially deployed. Every AI company building frontier models, from OpenAI to Google DeepMind to Meta, is now operating in a world where their products can be globally restricted after launch.
For Anthropic specifically, the path forward likely involves navigating between compliance with government directives and preserving the international partnerships and user base that make its models commercially viable. Amodei’s repeated meetings with White House officials suggest the company is trying to find a middle ground, but the Commerce Department’s willingness to act unilaterally and swiftly means that middle ground may not exist in any stable form. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our