Anthropic was forced to take its two most advanced AI models offline globally on Friday after the Trump administration issued what analysts are calling an unprecedented export control order, one that has raised immediate questions about government authority over commercial AI systems and the reliability of cloud-based AI for enterprise users.
On June 12, 2026, the US Department of Commerce directed Anthropic to immediately cut off access to Claude Fable 5 and Claude Mythos 5 for all foreign nationals, a category that includes non-US citizens living inside America, overseas customers, and even Anthropic's own employees who are not US citizens. Commerce Secretary Howard Lutnick communicated the order directly to Anthropic CEO Dario Amodei.
Rather than attempt a selective geographic or user-based block, which would have been technically complex and legally risky, Anthropic chose to shut down both models entirely for all users worldwide to ensure full compliance. Paying enterprise customers, individual subscribers, and API developers lost access simultaneously.
Anthropic said the government did not provide specific details about its national security concerns, but the company believes the order was triggered by a publicly circulated jailbreak technique. The timeline is telling: Fable 5 and Mythos 5 were released to the public on June 9. Within 24 hours, a user on X published what was described as a working method to bypass the models' safety filters. Two days later, the Commerce Department's directive arrived.
Anthropic examined the technique and concluded it exploited a limited set of vulnerabilities that were already known, relatively minor in nature, and not unique to its models, meaning other AI systems could exhibit similar weaknesses. The company characterised the government's response as disproportionate, arguing that applying this standard across the industry would make it impossible to release any frontier AI model publicly. Anthropic called the situation a "misunderstanding" and said it was actively working toward restoring access.
Earlier this year, the US Department of Defense designated Anthropic a supply chain risk, limiting its ability to pursue government contracts. The June 12 order went further, granting authorities direct control over access to the company's AI tools. The move has raised concerns among businesses that rely heavily on centralised AI platforms and highlighted the risks of dependence on a single vendor.
What Remains Available
The order applies exclusively to Fable 5 and Mythos 5. All other Claude models remain fully operational. Both Anthropic and government officials have indicated they want the matter resolved quickly, though no timeline for restoring access has been given.
The shutdown arrived at a particularly sensitive moment for Anthropic, which had recently filed confidentially for a US IPO at a valuation approaching $1 trillion, making the episode one of the most significant regulatory interventions in the company's history, days before it was expected to enter public markets.
