A federal judge has ruled against the Trump administration in its fight with artificial intelligence company Anthropic, finding that the Pentagon’s decision to designate the company a national security supply-chain risk was unlawful.
The ruling represents a significant setback for the administration’s attempt to pressure Anthropic over restrictions the company placed on how its artificial intelligence systems could be used by the US military.
US District Judge Rita Lin found that the government’s action was not supported by a legitimate national security justification and amounted to unlawful retaliation against Anthropic. The ruling also found violations of the company’s constitutional protections.
The dispute began with a disagreement over the military’s proposed use of Anthropic’s Claude artificial intelligence system.
Anthropic had resisted demands for unrestricted military use of its technology, maintaining limits on applications including mass domestic surveillance and autonomous weapons. The company argued that sufficiently powerful AI systems should not be deployed for lethal decisions without meaningful human control.
The Pentagon took a different position.
Defense Secretary Pete Hegseth subsequently designated Anthropic a national security supply-chain risk, a classification with potentially far-reaching consequences for the company’s ability to do business with the federal government and defence contractors.
The administration argued that its decision was justified by national security concerns and the practical problems that could arise if the military depended on a company capable of imposing restrictions on the use of its technology.
Anthropic challenged that reasoning in court.
A dispute that became a constitutional question
What began as a disagreement over the terms of a government technology contract eventually became a broader dispute over the limits of executive power.
Anthropic argued that the administration had punished the company because it refused to accept the Pentagon’s demands and publicly defended its position on the use of AI.
The government rejected that characterization, arguing that its actions concerned national security and contractual policy rather than protected speech. In court filings earlier this year, the administration maintained that Anthropic’s refusal to remove its restrictions constituted conduct rather than speech protected by the First Amendment.
Judge Lin did not accept that explanation.
Her ruling concluded that the government’s designation was unlawful and described the action as retaliation. The decision also found that the administration had acted outside the statutory framework governing the supply-chain designation.
The judgment places the administration’s use of national-security authority under renewed scrutiny.
AI and the battlefield
Behind the legal dispute is a much larger question: how far should governments be able to direct the development and deployment of increasingly powerful artificial intelligence?
The Pentagon wants access to advanced AI systems for military planning, intelligence, cybersecurity and other national-security applications.
Technology companies, meanwhile, are increasingly being forced to decide what uses of their systems they are willing to permit.
Anthropic has drawn a line around certain military applications, particularly autonomous weapons and mass surveillance. The company has argued that the technology remains insufficiently reliable and controllable for some of these uses.
That position has placed it at odds with a Pentagon increasingly determined to incorporate commercial AI into military operations.
The disagreement is not unique to Anthropic. Other major AI companies have entered defence partnerships, intensifying competition over access to government contracts and raising questions about whether companies that refuse particular applications risk being excluded from one of the world’s largest technology customers.
The consequences could extend beyond one company
The ruling matters beyond Anthropic.
If the government can classify a domestic technology company as a national-security risk after a dispute over how its products may be used, the case could establish an important precedent for the relationship between Washington and the technology industry.
Judge Lin’s decision suggests that national security cannot simply be invoked as a blanket justification for government action against a company whose policies officials dislike.
That does not mean the government lacks authority to protect sensitive military systems or exclude companies that genuinely pose security risks.
It means that such powers must still operate within the law.
The distinction is particularly important as governments become more dependent on privately developed technologies.
Artificial intelligence is increasingly becoming part of the infrastructure of modern government, from cybersecurity to intelligence analysis. That dependence gives technology companies considerable influence, but it also gives governments powerful incentives to demand access on their own terms.
The Anthropic dispute demonstrates what can happen when those interests collide.
A legal battle that may not be over
The ruling is a major victory for Anthropic, but it may not end the dispute.
The administration could appeal the decision, and a separate legal battle concerning another Pentagon supply-chain designation remains active in Washington.
The practical consequences for Anthropic’s relationship with the US government therefore remain uncertain.
The company has nevertheless welcomed the ruling and said it remains willing to work with the government on national-security applications of AI.
For the Trump administration, the case represents another reminder that presidential power has limits, even when national security is invoked.
For the technology industry, it offers a different lesson.
The companies developing the systems that governments increasingly want to use are no longer merely software suppliers. Their decisions about how those systems may be deployed can become matters of public policy, constitutional law and national security.
The Anthropic case has brought those questions into the courtroom.
And as artificial intelligence moves deeper into the machinery of government and warfare, they are unlikely to be the last.
