A federal appeals court upheld the Defense Department's designation of Anthropic as a supply-chain risk on September 25, 2026, keeping in place restrictions that bar Pentagon employees and contractors from using the company's Claude AI tools for defense work. The ruling ends Anthropic's legal challenge to a designation typically reserved for firms tied to adversarial governments.
What's happening
- The D.C. Circuit Court of Appeals upheld DOD's supply-chain risk designation for Anthropic in a 2-1 decision issued September 25, with judges Gregory Katsas and Neomi Rao forming the majority.
- DOD imposed the designation on March 3, 2026, after Anthropic refused to permit its AI tools for all lawful Pentagon uses, including surveillance and autonomous weapons applications.
- The ruling means Pentagon employees and contractors remain prohibited from using Anthropic's products for defense business.
- Anthropic won one earlier legal challenge in the Northern District of California in August but failed in the D.C. Circuit appeal.
Why it matters
- The designation blocks a major AI vendor from the defense industrial base during rapid military adoption of large language models across the Pentagon and its contractors.
- The ruling creates precedent for the government to restrict commercial AI tools from defense procurement based on policy disagreements rather than traditional supply-chain security concerns.
- Pentagon contractors and program offices must now identify alternative AI solutions or remove Anthropic tools from classified and unclassified defense workflows.
- The label affects adoption rates of Claude across the defense-industrial base at a moment when DOD is accelerating AI integration for operational and administrative functions.
Going deeper
- The original DOD-Anthropic contract dispute centered on use restrictions for domestic mass surveillance and fully autonomous weapons; DOD sought to remove these prohibitions on February 24, 2026.
- Anthropic rejected DOD's ultimatum to permit all lawful uses by February 27, 2026, leading Secretary of Defense Pete Hegseth to direct the supply-chain risk designation.
- Judge Karen LeCraft Henderson dissented, suggesting that Anthropic's legal position was strong and that panel composition influenced the outcome.
- Legal experts note that appeal prospects at the en banc D.C. Circuit or Supreme Court level exist but are discretionary and not guaranteed.
Financial impact
- The ruling does not immediately affect Anthropic's revenue, which is reported to be growing tenfold annually from non-defense customers.
- Anthropic's planned November 2026 IPO valuation expectations of approximately $2 trillion remain based primarily on commercial and non-defense markets.
- Defense revenue will remain blocked or severely curtailed for Anthropic as long as the supply-chain risk designation holds, removing a segment that competing AI vendors may capture.
- DOD contractors must evaluate capex for alternative AI platforms or internal development to replace Anthropic tools currently deployed in defense systems and workflows.
The intrigue
- The ruling may reflect judicial deference to the Trump administration's assertions of national security authority rather than the substantive merits of the supply-chain risk claim.
- INDOPACOM had been using Anthropic solutions for defense AI applications, and adjustment of those systems creates hidden costs and schedule risk across the command.
The fine print
- Anthropic stated it is considering further review options, including potential appeal to the en banc D.C. Circuit or Supreme Court, though such appeals are discretionary.
- A parallel federal court ruling in the Northern District of California previously held one of DOD's two statutory justifications for the designation unlawful, creating legal ambiguity.
- The designation uses statutory language typically applied to companies with ties to adversarial governments, not commercial policy disputes, setting an unconventional precedent for application.