Judge Signals Skepticism in Anthropic’s Pentagon Challenge

A federal judge signaled continued skepticism toward the Pentagon’s effort to designate Anthropic a supply-chain risk during a July 30, 2026 hearing, but she has not issued a final ruling.

U.S. District Judge Rita Lin heard cross-motions for summary judgment in Anthropic PBC v. U.S. Department of War, a Northern District of California case challenging federal directives and a designation made under 10 U.S.C. § 3252. Both Anthropic and the government asked Lin to decide the case on the existing written record.

According to hearing coverage from Axios and Lawfare, Lin said she did not see additional government evidence that improved its position. She reportedly said the record appeared, in some respects, to have become worse for the government and questioned whether officials had evidence that Anthropic could alter a model after delivery or activate a remote “kill switch.”

Those were preliminary observations during oral argument, not a written judicial finding. Lin took the matter under submission and said she would issue a written order; she did not rule from the bench.

What led to the dispute

The conflict centers on how the government wanted to use Anthropic’s Claude artificial-intelligence models. The Pentagon sought access for lawful military and intelligence purposes, including sensitive applications. Anthropic insisted that two uses remain off-limits: fully autonomous lethal weapons and mass surveillance of Americans.

Anthropic argues that the government retaliated after the company publicly defended those limits. Its legal theories include First Amendment retaliation, due-process violations, claims under the Administrative Procedure Act and challenges to the government’s procurement authority.

The government has argued that national-security officials had legitimate reasons to distrust Anthropic’s technology and its safety restrictions. Government lawyers said frontier AI is unusually opaque and that agencies need confidence that vendors will not impose undisclosed or changing limits that interfere with military operations.

What the March injunction blocks

Lin’s March 26 preliminary injunction remains the key interim protection in the Northern District of California case. The operative order blocked federal agencies from enforcing a February 27 presidential directive ordering them to stop using Anthropic’s technology.

It separately restrained the Department of War and Secretary Pete Hegseth from enforcing a February 27 Hegseth directive that designated Anthropic a “Supply-Chain Risk to National Security” and directed that contractors, suppliers and partners doing business with the U.S. military not conduct commercial activity with Anthropic. The order also blocked enforcement of the March 3 notice and formal designation under 10 U.S.C. § 3252, along with related guidance or actions intended to carry out those measures.

The order is a preliminary injunction, not a final judgment. It restores the prior status quo while the case proceeds. It does not require the government to buy Claude, continue using Anthropic products or award the company a contract. It also allows the Department of War to transition to other AI providers if that transition complies with applicable laws, regulations and constitutional requirements.

Why the separate D.C. Circuit case matters

Anthropic is also challenging a related designation under a different statute, 41 U.S.C. § 4713, in litigation before the U.S. Court of Appeals for the D.C. Circuit. That proceeding has produced a different interim result: the appeals court declined to issue an order shielding Anthropic from the consequences of the designation while the case develops.

The two proceedings do not yet amount to a final merits ruling or a definitive circuit split. But the differing interim postures create uncertainty for contractors and AI companies trying to determine which government restrictions remain operative while the litigation continues.

What to watch next

A final ruling in Lin’s case could help define how far an executive agency may go when a contractor disputes how its technology can be used. It could also influence how AI companies write safety restrictions into government contracts and how agencies document national-security justifications for excluding a vendor.

For now, the March injunction protects Anthropic from the challenged federal directives and the § 3252 designation, while leaving the government free to make lawful procurement decisions, including a transition to other providers. The next concrete development to watch is Lin’s written ruling, followed by possible appeals.

Sources

Look for updates to this story

Discover more from Interactive News

Subscribe to get the latest posts sent to your email.