California judge invalidates Pentagon label targeting Anthropic
A federal judge in California has ruled that the Trump administration illegally designated Anthropic a supply-chain risk. US District Judge Rita Lin found that Defense Secretary Pete Hegseth’s national-security label, followed by an order directing all federal agencies to stop working with the maker of Claude, amounted to unlawful retaliation and was arbitrary and capricious.
Lin also held that Anthropic had been denied due process under the Fifth Amendment. The decision is the company’s first court victory in its challenge to the Department of Defense, while a separate case filed in Washington, DC, remains ongoing.
Safety limits at the centre of the dispute
The conflict arose after Anthropic set firm guardrails on certain military uses of its AI models. The company objected to uses involving fully autonomous weapons and mass surveillance of American citizens. The Pentagon said it intended to use Anthropic’s models only for lawful purposes and argued that the company could seek to control how the military used systems it had purchased.
In her ruling, Lin wrote that the government’s actions and language showed a desire to make a public example of Anthropic for what officials described as its “arrogance” in criticising the government. She concluded that invoking national security did not justify punishment or retaliation against a government critic.
Government conduct undercut the risk claim
The court identified several actions that conflicted with the supply-chain designation. Hegseth had proposed applying the Defense Production Act to Anthropic, a step the judge said would treat the company as essential to national security rather than as a threat. The Department of Defense also continued pursuing a contract with the company.
Lin further noted government collaboration with Anthropic’s new Mythos model for cybersecurity. She said Anthropic “undisputedly lacks” backdoor access to its technology once it has been provided to the Department of Defense. Those details weakened the factual basis for broad restrictions across federal agencies.
The policy dispute is also framed by restrictions on Anthropic models and security arguments because it examines the tension between restrictions on Anthropic models and the security arguments made around them.
What the ruling means for AI suppliers
The judge stressed that the Department of War remains free to choose its AI vendors. Her ruling instead addresses the legality and basis of the measures imposed on Anthropic. Anthropic welcomed the decision and said it remained focused on working productively with government on national-security uses of AI.
For businesses selling AI systems to public-sector customers, the case reinforces the value of documenting model-use limits, procurement commitments and technical access boundaries. Those records can become central when safety requirements, national-security claims and vendor eligibility come into conflict.