US Judge Blocks Pentagon Action Against Anthropic

A US federal court has rejected the Trump administration’s action against artificial intelligence company Anthropic, ruling that the Pentagon went too far when it classified the firm as a national security supply chain risk.

The decision came after Anthropic challenged the government’s move in court. The company had been caught in a dispute with the US military over how its Claude AI systems could be used, particularly in areas such as surveillance and autonomous weapons.

Judge Rita Lin, in a 59 page ruling issued on Thursday, found that the government’s actions were intended in part to retaliate against Anthropic for publicly expressing concerns about its policies. She said national security concerns could not be used as a justification for punishing a company for criticizing the government.

The judge also ruled that the Department of Defense had violated the due process protections of the Fifth Amendment. As part of the ruling, the court removed Anthropic’s supply chain risk designation and ordered the Pentagon to withdraw related guidance, directives, communications and instructions issued against the company.

The dispute began after Anthropic refused to give the US military unrestricted access to Claude for certain applications. The company has argued that current AI systems are not dependable enough to be safely deployed in autonomous weapons.

Anthropic has also raised concerns about the use of AI for domestic surveillance, saying such applications could threaten individual rights. Those positions brought the company into direct conflict with Pentagon officials, who argued that private technology companies should not be in a position to place limits on military operations.

The Pentagon’s decision to classify Anthropic as a national security supply chain risk had serious consequences for the company. Such a designation is generally used when the government believes a company or its technology could create risks involving infiltration, sabotage or other threats to sensitive systems.

Following the designation, Anthropic was prevented from participating in certain military contracts. The company argued that the move could cost it billions of dollars in potential business while also damaging its reputation in the rapidly growing AI industry.

Anthropic subsequently accused Defense Secretary Pete Hegseth of exceeding his authority by imposing the restrictions. The company took the dispute to a federal court in California, challenging the Pentagon’s decision and the restrictions that followed.

The latest ruling represents a significant legal setback for the Defense Department. The court’s decision means the Pentagon cannot simply rely on a broad reference to national security when taking action against a company that has publicly disagreed with government policy.

Judge Lin’s ruling also focused on the constitutional protections available to companies facing government action. By finding a violation of the Fifth Amendment’s due process clause, the court made clear that government agencies must follow legal procedures even when national security is cited as the reason for their decisions.

The case has broader implications for the growing relationship between the US government and private AI companies. Washington is increasingly looking to technology firms for advanced AI systems, while companies are facing difficult questions about how far their products should be used in military operations.

Anthropic has maintained that there should be clear limits on the use of AI in areas where mistakes could have serious consequences. The company’s position on autonomous weapons has become one of the central issues in its disagreement with the Pentagon.

The legal battle is not completely over, however. Anthropic has another lawsuit pending in Washington, DC, involving a separate Pentagon decision related to the same supply chain risk designation. That case could determine whether similar restrictions affect the company’s ability to work with civilian government agencies.

For now, the California ruling gives Anthropic an important victory. The Pentagon has been ordered to revoke the measures directed at the company, while the court has rejected the administration’s attempt to use the supply chain risk designation against Anthropic.

The decision also puts renewed attention on how governments respond when private technology companies disagree with policies involving national security and military technology. As AI becomes more deeply integrated into defense systems, disputes over control, safety and acceptable use are likely to become increasingly important.