Federal judge blocks Trump administration's attempt to bar AI firm Anthropic from government contracts

A federal judge in San Francisco issued a permanent injunction preventing the Trump administration from enforcing rules that would have cut off artificial-intelligence company Anthropic from federal contracts. Judge Rita Lin found that the administration’s actions, including a presidential order and a Defense Department declaration labeling Anthropic a "supply chain risk," constituted unlawful retaliation for the firm’s protected speech and violated due-process protections.

Anthropic had sought to limit the use of its AI model Claude in mass-surveillance and fully autonomous weapons, prompting the military to argue for unrestricted use. After negotiations failed, the administration ordered agencies to stop using Claude and declared the company a security risk, citing concerns that Anthropic might sabotage its own technology. Lin rejected those claims, stating the evidence showed Anthropic could not maintain back-door access and that the national-security rationale was a pretext for punishment.

The 59-page opinion affirmed that the government may still contract with other AI providers, provided actions comply with applicable laws. Lin also nullified a February social-media post by Defense Secretary Pete Hegseth that barred all military contractors from working with Anthropic, describing it as an unlawful boycott. The ruling emphasized that national-security considerations do not grant a blank check to punish critics of the government.

Anthropic’s spokesperson welcomed the decision, saying the company remains committed to collaborating with the government on AI that benefits national security. The case underscores ongoing tensions between federal agencies and private AI developers over the balance between innovation, security, and regulatory oversight.