The D.C. Circuit Court of Appeals ruled 2-1 on Friday to uphold the Pentagon's designation of Anthropic as a supply chain risk, handing the Trump administration a significant win in its escalating campaign against AI companies that attach conditions to military use of their models. The two judges in the majority, Gregory Katsas and Neomi Rao, were both Trump appointees. Judge Karen LeCraft Henderson, a George H.W. Bush appointee, dissented. The underlying dispute is straightforward: Anthropic signed a $200 million Pentagon contract in July 2025 and began negotiating Claude's deployment on the DOD's GenAI.mil platform that September. The DOD demanded unfettered access across all lawful purposes. Anthropic wanted assurances against fully autonomous weapons and domestic mass surveillance. Talks collapsed. Defense Secretary Pete Hegseth accused Anthropic of attempting "to seize veto power over the operational decisions of the United States military," and in March the Pentagon designated the company a supply chain risk — a tool designed for foreign adversaries like Huawei, now turned inward against a domestic AI lab. The legal picture is split. Anthropic sued in two courts because the DOD relied on two distinct designations. A San Francisco federal judge already ruled one designation illegal. Friday's D.C. ruling upholds the second. Anthropic now faces a procedural labyrinth: it can petition for rehearing by the full D.C. Circuit or escalate to the Supreme Court. The appellate panel delayed immediate effect to preserve those options. Judge Katsas's reasoning is where the precedent bites. He acknowledged Hegseth's claims that Claude might shut down unexpectedly or be "subject to manipulation" — speculative assertions Anthropic disputed. But Katsas ruled that the decision-making authority rests with the President and the Secretary of Defense, writing that Hegseth "did not transgress any limits on his authority under the Supply Chain Security Act or the Constitution." This is judicial deference stretched to cover what began as a failed vendor negotiation. The political dimension is unmistakable. Trump has publicly attacked CEO Dario Amodei on Truth Social, writing that his administration has "stopped AI 'people' from doing bad, or potentially bad, 'things,' like Dario (Anthropic!)." Amodei was conspicuously excluded from a White House state dinner for Chinese President Xi Jinping. The message to the AI industry is legible: attach safety conditions to military contracts and you become a national security threat. The Supply Chain Security Act was built to prevent foreign adversaries from embedding compromised components in defense infrastructure. Its application to a domestic company over contractual terms about use restrictions is novel and alarming. The ruling establishes that the executive branch can designate any vendor a supply chain risk based on the Secretary's judgment about "competing risks" — with minimal judicial scrutiny of the underlying factual claims. What is being decided here is not whether Anthropic's safety conditions were reasonable. It is whether AI companies have any leverage to negotiate the terms under which their models are deployed by the military. Friday's ruling says they do not — or more precisely, that exercising such leverage can be recast as a national security threat, triggering exclusion from the entire defense ecosystem.
