Lead: The US Department of Defense has formally designated Anthropic as a supply-chain risk, a move announced as effective immediately and prompting the AI company to say it will sue. The designation bars use of Anthropic technology for defense contracts while leaving non-defense commercial relationships intact, according to company statements. CEO Dario Amodei said Anthropic received a formal letter from the department and called the action legally unsound. The decision follows recent, fraught talks between Anthropic and the US government and public statements from senior administration figures.
Key Takeaways
- The Pentagon placed Anthropic on a supply-chain risk list and stated the designation is effective immediately, restricting DoD use of its services.
- Anthropic announced it intends to challenge the designation in court, with CEO Dario Amodei calling the action legally unsound.
- Microsoft said it will continue integrating Anthropic technology for commercial clients but will exclude the US Department of Defense.
- Anthropic has supplied tools to US government agencies since 2024 and was among the first advanced AI firms used on classified work.
- Public criticism from President Donald Trump and administration officials preceded the designation and were cited in reporting on stalled negotiations.
- Rival OpenAI has secured a defense contract described by its CEO as having extensive guardrails for classified deployments.
- Anthropic reports more than a million new Claude signups per day and that Claude remains among the most downloaded AI apps in several countries.
Background
Anthropic, co-founded by former OpenAI researchers, built the AI system Claude and has been a prominent commercial provider of large language model services since 2023. The company began working with US government agencies in 2024, including deployments that involved classified tasks, which positioned Anthropic as an early vendor for sensitive uses. Over recent weeks and months, Anthropic and the Department of Defense engaged in negotiations about access and contractual terms for military use; sources described discussions as active but ultimately unresolved.
Tension grew after public comments from President Donald Trump and administration officials criticized the company, and an administration post used the colloquial term Department of War when discussing policy toward Anthropic. Anthropic has resisted proposals that it says would grant unfettered access to its models, citing concerns about mass surveillance and autonomous weapons. Those safety and policy stances, combined with political pressure, help explain the unusually public trajectory that led to the formal supply-chain designation.
Main Event
The Pentagon issued the supply-chain risk designation in a letter Anthropic says it received the day before its public response, and the department described the step as aimed at ensuring the military can use technology for lawful purposes without vendor-imposed restrictions. Anthropic framed the designation as narrow in scope and emphasized that it does not, and legally cannot, constrain the company from providing Claude or other services in contexts unrelated to specific DoD contracts.
CEO Dario Amodei said the company does not consider the designation legally well founded and announced plans to challenge it in court. Company spokespeople and filings indicate Anthropic intends to argue that the department did not use the least restrictive means required by law when applying the supply-chain risk label. The company also disputed public claims that followed from administration posts and said it received no prior notice from the White House that such statements were imminent.
Microsoft, a commercial partner of Anthropic, informed the BBC that it will keep Anthropic-based features for its customers but will not use those products for the US Department of Defense. The company said its legal review concluded Anthropic products can remain available to commercial customers and that non-defense collaborations can continue. Meanwhile, the department characterized the step as necessary to prevent vendors from restricting lawful use by the chain of command and to avoid putting warfighters at risk.
Analysis & Implications
The designation is unprecedented in modern US technology policy because it explicitly marks a domestic AI company as a supply-chain security concern. Practically, the immediate effect is to bar Department of Defense contracts that rely on Anthropic services, but the broader implications touch on procurement norms, vendor governance, and the balance between national security and commercial technology stewardship.
Legally, Anthropic will likely base its challenge on statutory constraints that require the secretary to adopt the least restrictive measures to protect the supply chain. That argument will test how courts interpret executive discretion in labeling vendors and whether the department met procedural and evidentiary thresholds. A successful challenge could limit future uses of the supply-chain risk tool; an adverse ruling would underscore expansive executive authority over defense procurement decisions.
Politically, the episode underscores how public rhetoric from senior officials can affect commercial negotiations and regulatory outcomes. The case may accelerate Congress and agencies to clarify standards for AI vendors in the defense ecosystem, including contractual guardrails, data access terms, and certification processes. Internationally, adversaries and partners will watch whether the US approach emphasizes access and control or instead protects vendor autonomy and safety constraints.
Comparison & Data
| Item | Anthropic | OpenAI |
|---|---|---|
| Government use since | 2024 | 2024 |
| Reported daily signups | Over 1,000,000 | Not disclosed here |
| Defense contract status | Designated supply-chain risk, barred for DoD | Contracted with DoD with stated guardrails |
The table highlights core, publicly reported contrasts: Anthropic has been used by US agencies and reports rapid consumer uptake, but the new designation limits its defense work. OpenAI has secured at least one DoD agreement described by its leadership as having strong safeguards. These numbers and statuses will be central to court filings and policy discussions moving forward.
Reactions & Quotes
Voices across government, industry, and advocacy groups responded quickly to the announcement, ranging from support for the Pentagon’s emphasis on operational flexibility to criticism that the move punishes a company for pursuing safety limits.
The military will not allow a vendor to insert itself into the chain of command by restricting lawful use of a critical capability and put our warfighters at risk.
Pentagon official, official statement
The Pentagon framed the designation as a matter of ensuring lawful military use, arguing vendors cannot bind the chain of command. That framing focuses on operational necessity rather than broader commercial policy.
We do not believe this action is legally sound and we see no choice but to challenge it in court.
Dario Amodei, Anthropic chief executive
Anthropic emphasized the narrow legal scope of the designation and signaled immediate litigation. The company also underscored that non-DoD commercial work is unaffected by the label, according to its public statements.
Designating an American company for refusing to compromise its safety measures is shortsighted and a gift to our adversaries.
Senator Kirsten Gillibrand, public statement
Some lawmakers criticized the department’s move as counterproductive and potentially harmful to national competitiveness, framing the dispute as a clash between safety-conscious corporate policy and force-readiness arguments.
Unconfirmed
- Reports that Anthropic is disliked by some administration officials due to limited campaign donations or public praise come from anonymous sources and remain unverified.
- Details about the internal negotiating positions of both Anthropic and the Department of Defense were reported by sources familiar with discussions and have not been fully disclosed by either party.
Bottom Line
The Pentagon’s action against Anthropic marks a rare escalation between a US technology vendor and the federal government, with immediate effects on defense contracting and potentially long-term consequences for AI procurement policy. Anthropic’s announced legal challenge will test the boundaries of executive procurement authority and statutory requirements to use the least restrictive measures when labeling vendors.
Outside the courtroom, the episode is likely to prompt legislative and administrative clarifications about how safety-driven vendor policies interact with national security needs, and to influence how large cloud and software providers negotiate access and guardrails for sensitive government work. For readers, the most important near-term developments will be the companys court filings, any judicial rulings on the designation, and whether Congress or the administration moves to codify clearer standards for AI vendors working with defense agencies.
Sources
- BBC News (media reporting)
- U.S. Department of Defense (official site)