U.S. Blocks Foreign Access to Anthropic’s Most Advanced AI Models
In an unprecedented escalation of export controls targeting artificial intelligence, the U.S. Commerce Department has ordered Anthropic to suspend all access to its most advanced AI models — Fable 5 and Mythos 5 — by any foreign national, forcing the company to abruptly disable the models for all users globally. The directive, issued on the evening of June 12, marks the first time the U.S. government has used national security export controls to restrict access to AI models themselves rather than the hardware that powers them.
The Directive and Its Scope
The order, which Anthropic described in an official statement as lacking specific details about the government’s national security concern, applies to foreign nationals both inside and outside the United States — including Anthropic’s own non-citizen employees. The company said it had no choice but to disable both models for all customers to ensure compliance, while clarifying that access to its other Claude models, including Opus 4.8, Sonnet, and Haiku, was unaffected.
According to The Guardian, the order comes as a previous dispute between Trump administration officials and the IPO-bound company showed signs of easing across parts of the U.S. government. The New York Times described the directive as “unusually expansive,” noting it could prevent Anthropic’s employees in allied nations such as Canada or the United Kingdom from using the models.
The Alleged Jailbreak
The government’s concern centers on a potential “jailbreak” method that could bypass Fable 5’s safeguards to identify software vulnerabilities. However, Anthropic disputes the severity of the threat, characterizing it as a “narrow, non-universal jailbreak” that reveals only minor, previously known vulnerabilities. The company stated that other publicly available models, including OpenAI’s GPT-5.5, can discover the same vulnerabilities without requiring a bypass.
Katie Moussouris, CEO of cybersecurity firm Luta Security, who reviewed the research that prompted the directive, told Fortune that the technique was not a jailbreak at all. “I’ve seen the paper. It’s not a jailbreak. It was Defense Oriented Prompting (DOP), capabilities defenders need,” she said. “If Nat defense is the goal, this just scored an own goal against us.”
A Deteriorating Relationship
The export control directive represents the second major action by the Trump administration against Anthropic in just a few months. The conflict began in February 2026 when President Trump ordered all federal agencies to stop using Anthropic’s models after the company refused to agree to Pentagon contract terms that would allow its AI to be used “for any lawful purpose” — including autonomous weapons and mass surveillance.
In early March, the Pentagon declared Anthropic a “supply chain risk,” requiring the U.S. military to cease using its models and prohibiting defense contractors from using them for government contracts. Anthropic is challenging that designation in federal court. The Pentagon’s chief information officer, Kirsten Davies, publicly supported the latest directive, posting on X that “some things are simply more important than revenue cycles, clickbait and pre-IPO valuation. America First. Always.”
As Deutsche Welle reported, key Trump administration figures — including former AI and crypto czar David Sacks and Pentagon undersecretary Emil Michael — have publicly attacked Anthropic, with Sacks accusing the company of being “woke” and engaging in “regulatory capture.”
Industry and Expert Reaction
AI policy experts reacted with surprise and criticism. Dean Ball, an AI policy expert who briefly served in the Trump administration, called the directive “cartoonish” in a post on X, according to Fortune. “An administration whose posture is that we should export advanced AI chips to China, which also wants to ban… Britain (and every other non-American on Earth)… from using our best models? I have no words,” Ball wrote.
Gary Marcus, a prominent AI critic, warned that the move could drive Chinese-born AI researchers back to China, potentially undermining American AI leadership. Ben Murphy of the Institute for Progress cautioned about the “balkanization of technology,” noting that requiring proof of citizenship to access services is becoming increasingly common.
IPO Implications
The directive could significantly impact Anthropic’s financial future. The company confidentially filed for a U.S. IPO in June 2026, with a recent funding round valuing it at approximately $965 billion. The export control decision raises questions about regulatory risk and whether Anthropic can maintain its technological edge, potentially dampening investor enthusiasm.
A Watershed Moment for AI Regulation
This action sets a significant precedent: the U.S. government has demonstrated it can unilaterally shut down access to AI models based on national security export controls, without a statutory framework or transparent process. Anthropic has called for a regulatory process that is “transparent, fair, clear, and grounded in technical facts” — principles it says this directive does not follow.
Historically, U.S. export controls on AI have targeted hardware and chips, such as restrictions on exporting advanced semiconductors to China. This directive marks a sharp escalation by directly restricting access to AI models themselves — a move some experts compare to failed 1990s efforts to control encryption software exports.
What’s Next
Anthropic says it is working to restore access as soon as possible, but the timeline remains unclear. The company is already challenging the Pentagon’s supply chain risk designation in federal court, and a legal challenge to this directive may follow. Meanwhile, the international community — particularly allies like the UK and Canada whose citizens are affected — may respond diplomatically.
The incident underscores the urgent need for clear statutory frameworks for AI regulation, as opposed to ad hoc executive actions that critics say blur the line between legitimate national security concerns and political retaliation against a company that has resisted administration demands.