Federal Judge Strikes Down Trump Administration’s Anthropic Blacklist as Illegal Retaliation

A federal judge has ruled that the Trump administration illegally designated Anthropic a national security supply-chain risk and ordered the government to rescind directives blocking the company’s AI products. U.S. District Judge Rita Lin said the measures amounted to unlawful retaliation against Anthropic for criticizing the Pentagon’s preferred uses of artificial intelligence.

Lin granted key portions of Anthropic’s motion for summary judgment in a 59-page order. She found violations of the First Amendment and said the government’s decision was arbitrary, capricious, and procedurally defective under the Fifth Amendment. The ruling vacates directives that barred federal agencies from using Anthropic products and prohibited defense contractors from doing business with the company, including on unrelated projects.

The dispute grew from Anthropic’s refusal to remove restrictions on military applications involving fully autonomous weapons and mass surveillance of Americans. President Donald Trump and Defense Secretary Pete Hegseth responded by labeling Anthropic a supply-chain risk and directing federal agencies to stop working with the Claude developer.

The Pentagon argued that Anthropic was attempting to control how the military used models it had purchased. Anthropic maintained that its safeguards addressed high-risk applications rather than ordinary government work. Lin wrote that the record showed the administration wanted to make a public example of the company for its “arrogance” in challenging government views, rather than acting on evidence that Anthropic would sabotage its model.

The decision establishes that national security cannot serve as a blanket justification for punishing a government critic. Lin wrote that the Constitution and the federal statute cited by the administration did not permit sweeping penalties based principally on Anthropic’s criticism of the government’s position on AI deployment.

The judge also highlighted contradictions in the government’s treatment of Anthropic. Hegseth had proposed applying the Defense Production Act to the company, a step that would frame Anthropic as essential to national security, while the Pentagon continued pursuing a contract with the firm. Those actions weakened the rationale for calling Anthropic a threat and reinforced the court’s finding that the blacklist lacked a coherent factual basis.

The government must rescind the directives that barred federal use of Anthropic’s products and restricted defense contractors from working with the company. The order therefore reaches beyond the Pentagon, affecting agencies and contractors that had been told to sever ties with Anthropic even when their work had no direct connection to military operations.

Anthropic welcomed the ruling and said it remained focused on working productively with the government on national security applications. The White House did not immediately respond to a request for comment, and the administration is expected to challenge the decision. The ruling addresses the broad blacklist case in California, while a separate, narrower Anthropic lawsuit remains pending in the federal appeals court in Washington, D.C.

The case gives AI companies a significant legal reference point as governments negotiate access to advanced models and providers set limits on dangerous uses. Anthropic’s position was that safeguards against autonomous lethal force and mass surveillance were conditions of responsible deployment, while the Pentagon treated those limits as interference with government authority.

The judgment separates a government’s ability to choose its technology vendors from its ability to impose industry-wide penalties on a company for expressing a policy position. Agencies remain free to select other AI providers, but the ruling rejects retaliation disguised as a procurement or national security decision. That distinction will matter as OpenAI, Google, Microsoft, Meta, and other firms negotiate contracts that define how their systems can be used.

The immediate next step is an expected government appeal, alongside the separate case proceeding in Washington. The administration must also determine how quickly it will withdraw the directives and whether it will pursue narrower restrictions tied to specific military or intelligence uses.

For Anthropic, the ruling removes the broadest federal barrier to government business but leaves the underlying argument over AI safeguards unresolved. The company still faces pressure to support national security work, while officials continue seeking systems for surveillance, weapons development, and other sensitive operations. The case now stands as an early test of whether model providers can set firm use restrictions without exposing themselves to government retaliation.