Anthropic to Challenge DOD Supply-Chain Risk Label in Court
SAN FRANCISCO (AP) — Anthropic CEO Dario Amodei announced Thursday that the AI company will sue the Department of Defense after the Pentagon formally designated it a supply-chain risk, a label Amodei called “legally unsound.” The move comes despite Anthropic’s AI models, including its Claude system, reportedly being used by some government entities and even in Iran, according to multiple reports.
The Department of Defense notified Anthropic of the designation effective immediately, according to statements from Amodei and reporting by TechCrunch, CNBC and the BBC. Amodei said the company has “no choice” but to challenge the decision in court, arguing that the designation lacks legal grounding. He added that most of Anthropic’s customers remain unaffected by the label.
Anthropic, the creator of the Claude family of large language models, has previously resisted providing defense agencies with unrestricted access to its technology. The company has cited concerns over potential use in mass surveillance and autonomous weapons systems, a stance that appears to have contributed to the Pentagon’s decision. Last week, Anthropic issued a statement saying it would challenge “any supply chain risk designation in court,” and Amodei reiterated that position Thursday.
The designation as a supply-chain risk typically restricts federal agencies from using or procuring the company’s technology in certain sensitive contexts, though the precise scope of restrictions for Anthropic has not been fully detailed in public announcements. The label is often applied to companies viewed as potential security vulnerabilities due to foreign ownership, data-handling practices or other risk factors. Anthropic is an American company founded by former OpenAI executives, including Dario and Daniela Amodei.
Background and Competitive Context
The dispute highlights growing tensions between frontier AI labs and the U.S. government over how advanced AI systems should be controlled and deployed for national security purposes. While competitors such as OpenAI and Google have pursued deeper partnerships with defense and intelligence agencies, Anthropic has maintained a more cautious approach regarding military applications.
The timing of the DOD action is notable given reports that Anthropic’s Claude model has been used in Iran, raising questions about export controls and the effectiveness of current AI governance mechanisms. Anthropic has not publicly commented on the Iran usage reports.
Impact on Developers, Users and the Industry
For enterprise and government customers, the designation creates immediate uncertainty. Organizations subject to federal compliance requirements may need to evaluate alternatives or seek waivers, potentially slowing adoption of Claude models in regulated sectors. However, Amodei’s statement that “most Anthropic customers are unaffected” suggests the commercial impact outside of direct DOD-related contracts may be limited in the near term.
The case could set an important precedent for how the U.S. government classifies AI companies as supply-chain risks. A successful legal challenge by Anthropic might make the Pentagon more cautious about applying such labels without detailed public justification, while an unsuccessful challenge could encourage other AI firms to align more closely with defense priorities to avoid similar designations.
What’s Next
Anthropic has not yet filed its lawsuit, and specific legal arguments or a timeline for court proceedings have not been disclosed. The company is expected to provide further details as the case progresses.
The DOD has not issued a detailed public explanation for the designation beyond confirming the action. It remains unclear whether the label applies only to specific Anthropic products or the company as a whole.
Industry observers anticipate that resolution of the dispute could influence broader U.S. policy on AI security reviews, particularly as additional large language model providers come under similar scrutiny. No timeline has been given for when a court challenge might conclude or whether negotiations between Anthropic and the Pentagon could still occur in parallel with litigation.
This article is based on statements from Anthropic CEO Dario Amodei and reporting from TechCrunch, CNBC and the BBC published March 5-6, 2026.