Anthropic Wins in Court as Pentagon’s AI War Hits a Constitutional Wall

Anthropic Wins in Court as Pentagon’s AI War Hits a Constitutional Wall
The fight over Anthropic’s AI models has become a test of whether the Pentagon can punish a contractor for drawing limits around military technology. A federal judge has now sided with the company, blocking sanctions that could have reshaped both government procurement and the boundaries of corporate speech.

Judge Rita Lin ruled that the Trump administration’s designation of Anthropic as a “supply chain risk” was unlawful, barring federal agencies from enforcing an order that restricted their use of the company’s Claude models. The decision makes permanent an earlier suspension of the sanctions and could cost the government its preferred leverage over a company that rejected military use of Claude for mass surveillance and autonomous weapons.

The ruling’s central distinction is between a contract dispute and political retaliation. The Justice Department argued that Anthropic’s refusal to accept the Pentagon’s terms created operational uncertainty and could threaten military systems. The government also maintained that the blacklist reflected contractual concerns, not disagreement with Anthropic’s views on AI safety. But Lin rejected what she called the “empty invocation of national security” as a justification for punishing critics, describing the designation as “illegal and baseless.”

Anthropic presented the ruling as confirmation that its safeguards were lawful, saying it welcomed the finding that the designation was unlawful. Its position rests on reliability and civil-liberties concerns: the company says current AI systems are not safe enough for autonomous weapons and that domestic surveillance violates basic rights.

The conservative framing points in the opposite direction, presenting the judgment as a strike against Defense Secretary Pete Hegseth’s authority rather than primarily as a speech case: its headline calls the decision a ruling that “strikes down Hegseth’s ban” and labels the retaliation theory unconstitutional. Both perspectives recognize the Pentagon’s desire for control over critical military technology. They diverge sharply over whether that control permits the government to brand a U.S. company a security risk after it refuses particular uses of its product. The administration may appeal, while a separate case could still threaten Anthropic’s access to civilian government contracts.