US Judge Blocks Pentagon Blacklisting of AI Developer Anthropic

A United States federal judge has issued a ruling barring the Department of Defense from penalizing artificial intelligence firm Anthropic, declaring that the Pentagon acted illegally when it designated the technology enterprise as a national security risk. District Judge Rita Lin delivered the 59-page decision on Thursday evening in a California federal court, halting a military ban that had prevented the artificial intelligence vendor from competing for lucrative government contracts.

The legal battle originated from decisions made by Defense Secretary Pete Hegseth, who designated the developer of the Claude AI system a supply-chain hazard. The blacklisting occurred after Anthropic refused to permit military authorities to utilize its advanced artificial intelligence models for domestic surveillance programs or autonomous weapons systems. Anthropic has maintained that existing artificial intelligence models are not reliable enough to be trusted with autonomous weapons, while the Department of Defense has argued that private technology companies should not impose operational restrictions on military activities.

In her written opinion, Judge Lin concluded that military officials had unlawfully targeted and penalized the technology company because of its public opposition to the Pentagon’s approach to battlefield AI deployment. The judge emphasized that relying on broad claims of national security was “not a blank check to punish and retaliate against government critics.” During a preliminary court hearing on July 30, Judge Lin had characterized the administration’s stance as “really troubling” and “at odds … with the First Amendment,” observing that the government’s evidentiary record had grown weaker over time.

Defending the military’s administrative action, Department of Justice lawyers argued in court that artificial intelligence models are “so staggeringly enormous and opaque” that military assessors cannot evaluate them in the same manner as physical hardware. Anthropic countered in its federal lawsuit that the government’s retaliation violated its constitutional rights to due process and free speech, noting that defense officials took adverse action against the company without providing a formal mechanism to dispute the security claims.

Company executives previously warned that exclusion from government procurement could cost the firm billions of dollars in potential revenue. The Pentagon’s decision represented the first instance in which a domestic American corporation was publicly labeled a supply-chain risk under a federal procurement law originally intended to safeguard defense systems against foreign sabotage. Anthropic welcomed the court’s order, stating it remains intent on working productively with federal authorities to deploy AI for national defense, though government lawyers are expected to appeal the decision.

Background

The dispute reflects growing friction between Silicon Valley technology firms and the U.S. military over ethical boundaries for artificial intelligence. As the Department of Defense seeks to integrate modern AI software into strategic hardware and logistics, several commercial software developers have attempted to place contractual limits preventing their products from being used in lethal operations or domestic spying. U.S. defense procurement rules give military leadership broad discretion to disqualify vendors considered supply-chain risks, an administrative mechanism traditionally used to prevent foreign state-sponsored espionage or sabotage in defense infrastructure.