US orders Anthropic to halt foreigners access to its advanced models

Anthropic PBC abruptly pulled the plug on its two most powerful artificial intelligence models on Friday, yielding to an unprecedented United States government export control directive that bars any foreign national from accessing the systems.

The mandate, issued by the Commerce Department, legally prohibits foreign nationals—including Anthropic’s own overseas employees and clients from allied nations—from utilizing the newly launched Claude Fable 5 and Mythos 5 models. Because verifying the physical nationality of every digital user account globally is functionally impossible, the San Francisco-based startup responded by disabling the flagship models entirely.

“The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,” Anthropic said in a public statement. The shutdown occurs just three days after Fable 5 was launched to the public as the first of Anthropic’s next-generation “Mythos-class” tier.

Bypassing Guardrails and the “Jailbreak” Dispute

While the Commerce Department did not provide detailed evidentiary files, U.S. national security officials indicated the emergency intervention stems from an alleged “jailbreak” method. Federal investigators reportedly discovered a narrow prompt technique capable of bypassing the model’s core safety guardrails, potentially allowing malicious actors to use Fable 5 to autonomously analyze codebases and map out critical software vulnerabilities for cyber warfare.

Anthropic pushed back heavily against the government’s technical assessment, calling the suspension a severe “misunderstanding” of how frontier models function.

“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 stated, adding that the targeted flaw was minor and also present in rival models. “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”

The Backstory: Weapons Systems and the Pentagon Blacklist

The administrative whiplash marks a catastrophic escalation in a year-long feud between the Trump administration and Anthropic’s leadership team, led by CEO Dario Amodei.

The relationship originally fractured after Anthropic corporate leadership refused to grant the U.S. military licensing rights to use its neural networks for mass domestic surveillance and fully autonomous weapons platforms. In retaliation, the Pentagon placed Anthropic on a restrictive commercial supply chain blacklist slated to take effect later this year.

National security hardliners inside Washington quickly praised the embargo, signaling that software code is now explicitly viewed as a high-stakes geopolitical asset.

“Some things are simply more important than revenue cycles, clickbait, and pre-IPO valuation. America First. Always,” Kirsten Davies, the Chief Information Officer for the Department of Defense, wrote in a public post on X.

Chilling Effects on European Allies and Wall Street

The sudden deletion of Anthropic’s premier systems has sent shockwaves through international corporate circles and global tech hubs.

Former White House tech official Dean Ball noted that the unprecedented scope of the export ban alters basic user authentication paradigms. “This means you should expect to have to prove your citizenship to use Anthropic models,” Ball warned. The internal operational fallout is equally severe, as key Anthropic personnel—including world-renowned researchers Andrej Karpathy and co-founder Chris Olah—were born outside the United States and are legally barred from logging into their own firm’s creations.

Across the Atlantic, European leaders reacted with fury as western hospitals, logistics conglomerates, and defense institutions found their advanced research infrastructure turned off without warning.

“Disabling Fable 5 and other models for foreigners is not a misunderstanding or a mistake,” British MP Tom Tugendhat said. “It’s the inevitable result of technology shaping warfare so that sovereignty is more about code than cannons.”

The timing is financially brutal for Anthropic. The regulatory freeze lands just days after the company submitted its preliminary registration documents with the Securities and Exchange Commission (SEC) ahead of a multi-billion dollar initial public offering planned for late summer. Market analysts expect the regulatory standoff to deeply spook prospective institutional investors, potentially clouding Anthropic’s target valuation as rival tech firms march forward uninhibited.us-government-forces-anthropic-disable-fable-mythos-ai-models