Judge Rita Lin ruled Thursday that the Pentagon’s move to label Anthropic a “supply-chain risk” — a designation typically applied to hostile foreign actors — constituted unlawful retaliation against the company. The judge pushed back firmly on the government’s justification, stating that national security claims cannot be used as cover to target critics.
The conflict began earlier this year when Anthropic declined to sign an open-ended defense contract. CEO Dario Amodei maintained the company’s commitment to two firm boundaries: barring use of its AI in autonomous lethal weapons and domestic surveillance programs. That stance reportedly angered both Defense Secretary Pete Hegseth and President Trump.
In response, the administration imposed the risk designation in late February, effectively cutting off federal defense contractors from using Anthropic’s products, though a transition period allowed continued Pentagon use in the interim.
Thursday’s ruling means other federal agencies remain free to use Anthropic’s tools, including its advanced Mythos model, which several agencies have been piloting for cybersecurity and national security analysis.
Despite the legal fight, ties between Anthropic and parts of the administration have shown signs of warming, particularly around the rollout of Mythos. The company still faces a separate appeal in Washington, and the government may challenge Thursday’s decision.
Anthropic said it remains committed to working constructively with federal officials on national security applications of its technology.
