Anthropic has asked the full federal appeals court in Washington to reconsider its defeat against Trump's Pentagon, seeking another hearing after a three-judge panel upheld the government's decision to blacklist the Claude developer from its military supply chain. The October 13 request gives the company another route to challenge a designation it has fought since March.
The dispute centers on who controls the boundaries of military AI. Anthropic retained restrictions on mass surveillance of Americans and lethal autonomous warfare. The Pentagon wanted permission for all lawful uses and argued that supplier-imposed restrictions could prevent an AI system from performing national-security tasks when needed.
Anthropic seeks a larger hearing
CNBC reported the request for the full D.C. Circuit to rehear the case. The procedure, known as rehearing en banc, allows the court to reconsider a panel decision with a larger bench. It requires the court's agreement, rather than following automatically from a company's disagreement with the result.
Federal appellate rules reserve that procedure for circumstances such as conflicts with binding decisions or questions of exceptional importance. A majority of eligible active circuit judges can order it. Anthropic is therefore seeking another judicial review of the Pentagon's authority, rather than announcing that the restrictions have been overturned.
Why the Pentagon won
The September 25 decision rejected Anthropic's statutory and constitutional challenges by a 2-to-1 vote. Judges Gregory Katsas and Neomi Rao formed the majority; Judge Karen LeCraft Henderson dissented.
The majority accepted that Anthropic could encode restrictions into Claude that would prevent government uses the department considered contractually authorized. Katsas wrote:
In sum, the Department reasonably feared that Anthropic might manipulate Claude’s design to prevent it from performing national-security functions that the Department deems contractually authorized and necessary.
The court did not require a finding that Anthropic was acting maliciously. It interpreted the procurement law broadly enough to cover risks created by restrictions imposed with good intentions. It also rejected the company's claim that the exclusion punished protected speech, concluding that the department acted because Anthropic refused an essential contract term.
The safeguards Anthropic refused to abandon
In a March 5 statement, CEO Dario Amodei said the company accepted the military's responsibility for operational decisions while maintaining two limits:
Our only concerns have been our exceptions on fully autonomous weapons and mass domestic surveillance, which relate to high-level usage areas, and not operational decision-making.
Anthropic had already adapted Claude for government work, including intelligence analysis and cyber operations. The court's account describes a government-specific model and relaxed restrictions, while confirming that the two contested exceptions remained.
Henderson disagreed with the majority's interpretation of the supply-chain statute. She warned that its reasoning could force other vendors to choose between changing their policies and risking a national-security designation:
I cannot agree that this is the scenario the Congress had in mind when it enacted FASCSA.
A separate court victory remains relevant
Anthropic's September defeat did not settle every part of its legal battle. An August 27 California judgment had set aside a parallel designation under a different law, 10 U.S.C. § 3252.
The D.C. Circuit considered the Federal Acquisition Supply Chain Security Act, under 41 U.S.C. § 4713. Its majority said that statute defined supply-chain risk more broadly and did not require the bad motive relevant to the California case. One ruling therefore did not control the other.
The distinction matters for contractors: the Washington case concerns the Pentagon's exclusion of Claude from its systems and from contractor work for the department. It is not a ruling prohibiting every private-sector use of Claude.
Anthropic's next challenge is persuading the appeals court to revisit that interpretation. The dispute now tests how much control commercial AI suppliers can retain over military deployments when their safety conditions clash with procurement demands.