Trump Administration Blocks Foreign Access to Anthropic's Most Advanced AI Models

The Trump administration has imposed sweeping export controls on Anthropic's most advanced artificial intelligence models, effectively banning foreign governments, corporations, and individuals from accessing the technology. The unprecedented directive marks a severe escalation in Washington's effort to treat cutting-edge AI systems as tightly guarded national security assets rather than globally available commercial software.[1][2]

In a letter sent late Friday to Anthropic CEO Dario Amodei, U.S. Commerce Secretary Howard Lutnick outlined the new restrictions targeting the company's flagship Mythos 5 and Fable 5 models. The blockade applies to any location outside the United States and, notably, extends to all foreign nationals currently residing within the country.[1][2]

The sweeping export ban is the latest—and most commercially damaging—salvo in a bitter, months-long standoff between the White House and the San Francisco-based AI safety lab. The conflict centers on a fundamental dispute over who controls the deployment of frontier AI: the developers who build the ethical guardrails, or the military that seeks to harness the technology for national defense.[6][8]

The rift traces back to late 2025, when the Department of Defense sought to integrate Anthropic's systems into its GenAI.mil platform, a government-certified hub designed to provide AI tools across the military. While Anthropic had previously secured a $200 million Pentagon contract and allowed its Claude models to be used for analytical support—reportedly including "Operation Absolute Resolve" in Venezuela—the company drew a hard line on future deployments.[6]

During negotiations, defense officials demanded that the military be permitted to use the technology for "any lawful use," effectively asking Anthropic to bypass its standard terms of service. Amodei refused, insisting that the company's safety guardrails must remain intact to prevent the AI from being utilized for fully autonomous lethal weapons or mass domestic surveillance.[6][7]

The refusal infuriated the administration. In March 2026, Defense Secretary Pete Hegseth formally designated Anthropic a "supply chain risk"—a severe classification typically reserved for foreign adversaries or companies linked to the Chinese military. The designation effectively blacklisted Anthropic from federal contracting, with President Trump publicly boasting that he had fired the company "like dogs."[5][8]

The federal government defended the blacklisting, arguing that Anthropic's refusal to lift its restrictions constituted a genuine risk to national security systems. Pentagon officials asserted that the military could not rely on software that required them to call Anthropic for permission to shoot down enemy drone swarms that target Americans.[7]

Anthropic retaliated with high-stakes federal lawsuits in California and Washington, D.C., accusing the administration of unlawful retaliation and violating the company's First Amendment rights. In late March, a federal judge in San Francisco temporarily blocked the Pentagon's blacklist, setting the stage for a landmark legal battle over executive authority and corporate speech.[4][6]

The administration's response has been to double down. The new export blockade coincides with the issuance of National Security Presidential Memorandum 11, a sweeping directive that revokes Biden-era guardrails on military AI. The memo explicitly orders the termination of federal contracts with any company that attempts to restrict how the government uses its artificial intelligence.[3]

By cutting off Anthropic's international access, the White House is weaponizing market dynamics to enforce compliance. The ban on foreign users—including multinational corporations and individual subscribers overseas—poses a massive financial threat to the company, which recently filed confidentially for an initial public offering in the United States.[1][8]

The inclusion of foreign nationals living within the United States adds a layer of unprecedented complexity to the export controls. Legal experts and tech industry observers are already questioning how Anthropic will technically enforce a ban that requires verifying the citizenship of every individual user attempting to access Mythos 5 or Fable 5 without violating domestic privacy laws.[1][6]

Beyond the immediate financial damage to Anthropic, the blockade signals a profound balkanization of the global technology landscape. By restricting commercial software based on citizenship, the administration is effectively creating digital zones of thought, where access to superhuman analytical tools is strictly confined within American borders.[1][8]

The tech industry is watching the standoff with mounting anxiety. If the government can successfully leverage export controls and supply-chain risk designations to force a domestic company to abandon its ethical guardrails, other AI developers may find themselves forced to choose between their safety principles and their survival in the global market.[6][8]