On Friday evening, Washington time, Anthropic received a phone call that put its flagship product on ice. The US government issued an export control directive ordering the company to block access to its Claude Fable 5 and Mythos 5 models from all "foreign nationals" — both inside and outside the US, including Anthropic's own employees.
The implication was clear: the only way to comply was to completely cut off access to the models, then fly to Washington and try to convince President Donald Trump to change course.
Why it matters
This is the first time a US administration has ordered a leading AI company to disable a publicly released product on national security grounds. The impact could reshape AI regulation across the entire industry. If the administration follows through, US AI companies could find themselves under export controls similar to those governing the semiconductor industry.
What the experts say
A source familiar with the negotiations told The Verge that the government gave Anthropic a 90-minute ultimatum: either shut down access voluntarily, or face Commerce Department export controls imposed by force. Within 15 minutes of the initial call, Anthropic executives were talking to the White House. CEO Dario Amodei spoke directly with Treasury Secretary Scott Bessent, Commerce Secretary Howard Lutnick, and National Cyber Director Sean Cairncross — in some cases more than once.
Anthropic says the government provided no specific details about the national security concern behind the order. The company claims the only information it received was about a potential "narrow, non-universal" jailbreak, shared verbally, and that the vulnerabilities found also exist in other models, including OpenAI's GPT-5.5.
David Sacks, the former White House AI and crypto czar who stepped down in March, noted on X that the entity which alerted the government was "a highly credible trusted partner of both Anthropic and the USG." Sacks did not mention China in his post, but other reports linked the move to US concerns about Chinese access to the technology via a global telecom company that had initially been cleared for Mythos Preview access.
Fable 5 was the first model from the Mythos class released to the general public, priced at $10 per million input tokens and $50 per million output tokens — double the cost of its predecessor Opus 4.8. Anthropic said that in testing, 95% of Fable sessions ran entirely on the model without falling back to a weaker model.
More from AI and Tech
Hot IPO summer. The IPO market is back with a new face: FAANG is giving way to MANGOS — Meta (or Microsoft), Anthropic, Nvidia, Google, OpenAI, and SpaceX. Half of these companies are heading to public markets in the same window, testing investors, valuations, and the very concept of a public tech company in 2026.
xAI's gas-powered data center. The Justice Department intervened in the NAACP's lawsuit against xAI, arguing that xAI's use of gas turbines in Mississippi is necessary for national security. The reason: "Grok provides critical support for the Department of War's military operations."
Disney embraces AI. Walt Disney Imagineering has adopted Adobe's Firefly Foundry platform for theme park design. The team will use models trained on Disney's own IP to turn sketches into fully rendered concept art and create 3D prototypes. Adobe, which first announced Firefly Foundry in January, is also collaborating with major Hollywood talent agencies including CAA, UTA, and WME.
Epic Games goes generative. Epic detailed how it uses GenAI tools within Fortnite, including Google's Nano Banana and its own internal GenMedia Bridge, claiming the tools keep "creative control in the hands of the creator."
Another court loss for Musk. Elon Musk lost again in his lawsuit against OpenAI, according to reports. Musk continues his legal battle against the company he co-founded with Sam Altman.
The bottom line
The shutdown of Fable 5 is a watershed moment for AI regulation in the US. The key question now is whether the administration goes further — and how OpenAI, Google, and other AI companies that may be next in line will respond.
