Anthropic has just been told to jog on by a federal appeals court. On Wednesday, judges refused to lift the Defence Department’s supply-chain risk label, leaving one of the US’s best-known AI shops stuck outside the Pentagon gates.
The court accepted that the company is taking a financial hit, yet it would not second-guess a national security call. That leaves Anthropic fighting on two fronts after a California court recently handed it a preliminary injunction against President Trump’s wider ban on federal agencies using its Claude models.
With Wednesday’s ruling, the Defence Department’s “security threat” tag stays put. Anthropic remains frozen out of new contracts and Pentagon systems, while the government continues to treat Claude as if it were carrying malware.
In late February, Trump gave the Defence Department six months to transition away from Claude, which is being used in the war in Iran.
The appeals court did not mince its words. “On one side is a relatively contained risk of financial harm to a single private company. On the other side is judicial management of how, and through whom, the Department of War [sic] secures vital AI technology during an active military conflict.”
The case now runs in parallel with the government’s appeal of the California ruling, which is the sort of legal fog contractors hate. Firms working with both Anthropic and the government are already talking about dropping Claude from public-sector projects to avoid getting caught in the blast radius.
Computer & Communications Industry Association chief executive Matt Schruers said, “The D.C. Circuit’s denial will prolong ambiguities regarding whether political considerations can drive federal procurement.”
The Pentagon used two different statutes to slap Anthropic with the supply-chain risk designation, and one of them sits squarely with the Washington appeals court.
Defence Secretary Pete Hegseth made the designation after contract renegotiations collapsed. Anthropic wanted assurances its models would not be used in fully autonomous weapons or for domestic surveillance, while the Pentagon insisted policies and laws already cover that and pushed for broader permitted use.
Department of Justice acting attorney general Todd Blanche claimed a win straight away. “Today’s D.C. Circuit stay allowing the government to designate Anthropic as a supply chain risk is a resounding victory for military readiness.”
The supply-chain risk tool is normally pointed at firms tied to US adversaries such as China, which makes turning it on a domestic AI supplier look close to unprecedented. It reads like the Pentagon has decided it wants maximum flexibility and minimum moral lecturing.
Some of Anthropic’s competitors, including OpenAI and Elon Musk’s xAI, have since accepted the Pentagon’s terms.
An Anthropic spokeswoman said, “We’re grateful the court recognised these issues need to be resolved quickly and remain confident the courts will ultimately agree that these supply-chain designations were unlawful.”
“While this case was necessary to protect Anthropic, our customers, and our partners, our focus remains on working productively with the government to ensure all Americans benefit from safe, reliable AI.”
On Tuesday, Anthropic said it is offering a preview of a new model to about 50 critical infrastructure organisations, including US tech giants, ahead of a public release. The company said it feared putting it out too soon could trigger widespread disruption online.
