US Judge Blocks Pentagon's Attempt to Label AI Firm 'Supply Chain Risk' Over Military Use Dispute

In a ruling that could reshape how Washington handles artificial intelligence vendors, a federal judge has temporarily blocked the Pentagon’s attempt to label Anthropic a “supply chain risk.” The decision hands an early win to the AI firm in its legal standoff with the Pentagon, while raising deeper questions about free speech, national security, and the government’s growing dependence on private AI. At the center of the dispute is whether the government can penalize a company for refusing to align with its preferred military uses of AI. For now, the court says no. U.S. District Judge Rita Lin issued a 43-page ruling siding with Anthropic, blocking both the “supply chain risk” designation and a directive from President Donald Trump to cut federal contracts with the company. Judge Lin called the government’s actions “classic First Amendment retaliation,” arguing they were not clearly tied to legitimate national security concerns. The ruling hinges on a critical idea: companies have the right to hold and express viewpoints, even when those views conflict with government priorities. Anthropic, led by CEO Dario Amodei, refused to allow its Claude AI model to be used for: The Pentagon, under Defense Secretary Pete Hegseth, pushed for broader usage, including “all lawful applications.” The court found that punishing Anthropic for resisting these uses could violate constitutional protections. Judge Lin’s reasoning was blunt: if the Pentagon had concerns about operational integrity, it could simply stop using the technology. Instead, it imposed sweeping restrictions that appeared punitive. The dispute didn’t emerge overnight. It grew out of increasing tension between AI developers and government agencies over how far these tools should go in military contexts. Anthropic’s Claude model reportedly saw limited use during the early phase of the Iran conflict, including support roles like target determination. That alone placed the company at the intersection of cutting-edge tech and real-world warfare. But Anthropic drew a line. The company argued that certain uses of AI, especially autonomous weapons, cross ethical boundaries. That position clashed directly with the Pentagon’s desire for flexible deployment. Typically, this label is reserved for foreign adversaries or entities that could compromise national security. Applying it to a U.S.-based firm was highly unusual. The designation triggered: Judge Lin questioned whether the government had overstepped by using a national security tool to settle a policy disagreement. This case is bigger than one company. It sits at the crossroads of three powerful forces: constitutional law, military policy, and the future of AI. If upheld, the ruling could restrict how the government pressures private companies to align with its priorities. That matters in an era where: A good place for an infographic here would be a flowchart showing how AI companies interact with federal agencies across defense, healthcare, and intelligence sectors. The Pentagon increasingly depends on private-sector AI for: But companies like Anthropic are asserting boundaries. This ruling signals that participation in defense contracts may not require full alignment with military objectives. That could embolden other firms to set similar limits. Critics of the ruling may argue it complicates national defense by limiting the government’s ability to vet and control suppliers. Supporters counter that: The decision is temporary. Judge Lin paused the ruling for one week to allow the Trump administration to appeal. Even with the ruling in place, the Pentagon retains significant flexibility: What it cannot do, at least for now, is impose broad punitive measures tied to the “supply chain risk” label. Despite the legal win, Anthropic faces practical challenges. Some federal agencies, including the Department of Health and Human Services and the General Services Administration, have already removed its products. Rebuilding those relationships may prove difficult even if the company ultimately prevails in court. This case is quickly becoming a litmus test for how the U.S. governs artificial intelligence. These questions are not theoretical. They will influence: