🔍 Read the full analysis: How The Court Ruled On The Pentagon Blacklisting Anthropic Over Claude Features on ThorstenMeyerAI.com
Get the latest gadgets delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
TL;DR
The U.S. Court of Appeals for the D.C. Circuit ruled 2-1 that the Pentagon may designate Anthropic a supply-chain risk and exclude Claude from Defense Department work. The majority said the procurement law it reviewed does not require proof of malicious intent; a separate California ruling found the government’s parallel designation unlawful under a different law.
The U.S. Court of Appeals for the D.C. Circuit ruled 2-1 on September 25 that the Pentagon may designate Anthropic a supply-chain risk and exclude its Claude AI products from Defense Department work, as described in the original analysis. The ruling rejects the company’s challenge under a federal procurement law, leaving the Pentagon able to continue removing Claude from its systems and barring defense contractors from using Anthropic products on department work, according to the Associated Press.
The majority concluded that the law it reviewed, 41 U.S.C. § 4713, allows a supply-chain-risk designation without a finding that a supplier acted with malicious intent. The judges said the statute covers the risk that a person could manipulate a technology product in ways that deny or disrupt its function. They accepted the Pentagon’s concern that Anthropic could use Claude’s built-in restrictions to prevent it from carrying out military functions the department considered contractually authorized and necessary.
The dispute centers on Anthropic’s refusal to agree to the Pentagon’s requested terms for Claude’s use. The company has maintained restrictions against use for mass domestic surveillance and lethal autonomous warfare. The court’s opinion also discusses prior instances in which Claude refused some government prompts and a disagreement over whether its terms permitted use in an overseas military operation. The judges acknowledged that Anthropic said some earlier refusals were glitches that were resolved with government stakeholders, and that the company described the operational dispute as a misunderstanding.
Judge Gregory Katsas wrote the majority opinion, joined by Judge Neomi Rao. Judge Karen Henderson dissented. The majority denied Anthropic’s petitions for review. The ruling does not mean the company’s products are barred from all federal agencies: the AP reports that the decision concerns Pentagon work and does not impose a government-wide ban. The department had already begun removing Claude from its workflows earlier this year.
The Reach of Pentagon Procurement Power
The decision gives the Defense Department judicial backing to exclude a supplier over concerns about whether its AI system will perform as the department expects. That matters to companies selling software to the military because product rules, model behavior and contract terms can become part of a national-security procurement dispute. The ruling’s reasoning could also shape how agencies apply the law when a supplier’s product controls limit particular uses.
The judges did not hold that Anthropic acted maliciously. The majority said the relevant statute turns on what a supplier’s product controls may do to system function, rather than the supplier’s motive. That distinction is central: Anthropic says its restrictions reflect safety and privacy commitments, while the Pentagon argued that it needs confidence Claude will carry out lawful functions covered by its contract.
The immediate effect described by the AP is confined to Defense Department work, including contractors doing work for the department. The ruling does not settle every dispute over Anthropic’s federal business. A separate case in California addressed another legal basis for the government’s actions and reached a different conclusion. The coexistence of those rulings means the practical scope depends on which statute and government action are involved.
Two Laws, Separate Court Cases
The Pentagon designated Anthropic a supply-chain risk after the company refused to authorize all uses the department sought for Claude. Anthropic challenged the designation in separate proceedings. The D.C. Circuit case reviewed the department’s action under 41 U.S.C. § 4713, a procurement statute that gives the appeals court exclusive jurisdiction over certain challenges to these determinations.
A federal district judge in California separately reviewed the designation under 10 U.S.C. § 3252. Ars Technica reported that the California court found that law’s definition of a supply-chain risk required a bad motive, and found no such motive by Anthropic. The D.C. Circuit majority said it did not dispute that conclusion about the California statute; it held that the separate procurement law before it had a broader definition and did not require malicious intent.
The contrast between the cases helps explain why the outcomes differ. The California ruling addressed whether the administration could apply one statutory definition and impose broader restrictions. The D.C. Circuit ruling concerned the Pentagon’s authority under a different procurement provision. The two decisions should not be read as a single court resolving the same legal question in opposite ways.
““We respectfully disagree with the court’s decision” and are “considering all options, including further review.””
— An Anthropic spokesperson, in a statement reported by the Associated Press
What the Ruling Leaves Open
The ruling does not resolve the separate California case or determine the full effect of that court’s restrictions on the government. The legal reach of each decision, and how they interact in practice, remains dependent on further proceedings and the specific actions the government takes. The D.C. Circuit decision concerns the Defense Department’s procurement authority under the statute before it.
Anthropic has said it is considering further review, but it has not confirmed in the statement reported by AP which route it will take. Ars Technica reported that the company could seek review by the full D.C. Circuit or petition the Supreme Court. Whether it will pursue either option, and whether a reviewing court would take up the case, remain unknown.
The majority accepted the Pentagon’s concerns about possible effects of Claude’s restrictions, but the opinion does not establish that Anthropic remotely controls models already deployed on classified systems. It discusses the company’s ability to shape model behavior through updates and new versions; Anthropic told the court it cannot access, alter or shut down a model after delivery for use on those systems. The immediate deployment and contract consequences beyond the existing removal process are not fully described in the news reports.
Further Review and Pentagon Implementation
The next major legal step may be a request from Anthropic for review by the full D.C. Circuit or a petition to the Supreme Court. The company has said it is considering further review, but has not announced a filing. Any such request would determine whether another court considers the majority’s interpretation of the procurement statute.
In the meantime, the Pentagon can continue its removal of Claude from department systems and keep Anthropic products out of work performed for the Defense Department, as described by the AP. Further clarification may come from how the department applies the designation to ongoing contracts and contractor work. The separate California case also remains relevant to the broader government restrictions, but this appeals court ruling does not resolve that case.
For now, the confirmed development is narrow: a divided appeals court upheld the Pentagon’s authority under one procurement law to designate Anthropic a supply-chain risk. The final reach of that decision, any further appellate review, and the practical effects for Anthropic’s government business remain unsettled.
Key Questions
What did the appeals court decide?
The D.C. Circuit ruled 2-1 that the Pentagon could designate Anthropic a supply-chain risk under the procurement statute before the court and reject the company’s challenge.
Does the ruling ban Anthropic from all federal work?
No. The Associated Press reports that the decision concerns Pentagon work and does not bar Anthropic products from the federal government as a whole.
Why did the majority uphold the designation?
The majority read 41 U.S.C. § 4713 to cover risks that a supplier’s controls could deny or disrupt a product’s function, without requiring proof of malicious intent.
How does the California case differ?
It reviewed the government’s action under a different statute, 10 U.S.C. § 3252. The D.C. Circuit said the California court’s interpretation of that law did not control the separate procurement-law question before it.
Will Anthropic appeal?
Anthropic said it is considering further review. It has not confirmed whether it will ask the full D.C. Circuit or the Supreme Court to hear the case.
Primary source: Anthropic · via ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
