Anthropic has secured a major legal victory in its escalating dispute with the U.S. military after a federal judge struck down one of the Pentagon measures that branded the Claude developer a national-security supply-chain risk.
U.S. District Judge Rita Lin ruled in a 59-page order that the Pentagon’s designation of Anthropic was “illegal and baseless”, rejecting the government’s attempt to justify broad procurement restrictions on national-security grounds.
CNBC reports that Lin also found the government’s actions violated Anthropic’s First Amendment rights and deprived the company of due process under the Fifth Amendment. The ruling vacated the supply-chain risk designation challenged in the California case.
The Fight Started With Claude’s Military Guardrails
The conflict stems from Anthropic’s refusal to remove restrictions governing how the Pentagon could use its AI models.
Reuters says Defense Secretary Pete Hegseth imposed the unprecedented designation after Anthropic refused to permit Claude to be used for U.S. surveillance or autonomous weapons. Anthropic has argued that current models are not reliable enough for safe use in autonomous weapons and that domestic surveillance conflicts with civil rights.
The Pentagon maintained that private companies should not be able to constrain military operations and argued that Anthropic’s restrictions could create uncertainty over how Claude could be deployed during missions. The government said its designation arose from contractual disagreements rather than Anthropic’s views on AI safety.
Lin rejected that justification. She wrote that the government could not use an “empty invocation of national security” as a “blank check” to punish its critics.
Billions in AI Business Were Potentially at Stake
The Pentagon’s action was unusually consequential because Anthropic became the first U.S. company publicly designated a supply-chain risk under the procurement statute, a framework originally intended to protect military systems from foreign sabotage. Anthropic warned the designation could cost it billions of dollars through lost contracts and reputational damage.
CNBC reports that the designation also affected defense contractors using Anthropic technology in work for the Pentagon. Anthropic sued in both San Francisco and Washington, D.C., because the Defense Department relied on two separate legal designations.
That distinction matters: Lin’s decision resolves the California challenge, but Anthropic’s separate Washington, D.C., litigation remains pending.
Anthropic welcomed the ruling, saying it remained focused on working productively with the government on AI and national security.
AI Safety Policies Are Becoming Procurement Architecture
For AI companies, the case exposes a growing strategic tension.
Frontier labs increasingly embed restrictions into models governing surveillance, autonomous weapons and other high-risk applications. Governments, meanwhile, may expect technologies purchased for national-security missions to remain available across a much broader range of lawful uses.
That means model guardrails are no longer only product-safety decisions. They can determine who wins government contracts, which systems defense contractors can integrate and whether an AI vendor remains viable inside sensitive public-sector infrastructure.
Anthropic’s victory therefore matters beyond Claude. As AI becomes part of military software stacks, the boundary between vendor safety policy and government control over deployed technology is becoming one of the industry’s most consequential procurement battles.
