On September 25, 2026, the U.S. Court of Appeals for the D.C. Circuit (D.C. Circuit) ruled that the Department of War (DoW) reasonably concluded Anthropic poses a supply chain risk under the Federal Acquisition Supply Chain Security Act (FASCSA).

As discussed in our previous articles on this topic, the DoW designated Anthropic a supply chain risk under two separate statutes, 10 U.S.C. § 3252 and 41 U.S.C. § 4713.  Anthropic challenged the DoW’s designation in the U.S. District Court for the Northern District of California (regarding the § 3252 designation) and in the D.C. Circuit (regarding the § 4713 designation).

The U.S. District Court for the Northern District of California (California District Court) ruled in favor of Anthropic in an August 27, 2026 decision.  In that case, the DoW’s reliance on 10 U.S.C. § 3252 to find Anthropic is a supply chain risk was at issue:

The term “supply chain risk” means the risk that an adversary may sabotage, maliciously introduce unwanted function, or otherwise subvert the design, integrity, manufacturing, production, distribution, installation, operation, or maintenance of a covered system so as to surveil, deny, disrupt, or otherwise degrade the function, use, or operation of such system.

10 U.S.C. § 3252 (d)(4).

The California District Court set aside the designation. It found that § 3252, which requires “an adversary” to engage in sabotage or subversion, does not reach Anthropic’s good faith negotiation of contractual usage restrictions, and that Anthropic acted without any “bad motive.”

In reaching its decision, the D.C. Circuit distinguished the “supply chain risk” definition relevant to FASCSA from the narrower 10 U.S.C. § 3252 definition at issue in the California District Court case. The FASCSA “supply chain risk” definition is set forth in 41 U.S.C. § 4713 (k)(6) and states:

The term “supply chain risk” means the risk that any person may sabotage, maliciously introduce unwanted function, extract data, or otherwise manipulate the design, integrity, manufacturing, production, distribution, installation, operation, maintenance, disposition, or retirement of covered articles so as to surveil, deny, disrupt, or otherwise manipulate the function, use, or operation of the covered articles or information stored or transmitted on the covered articles.

Unlike § 3252, which is limited to threats from an adversary, §4713 reaches “any person” and expressly includes conduct that may “extract data” or “otherwise manipulate”   covered articles. Although the D.C. Circuit stated it had “no quarrel” with the California District Court’s conclusion, it determined that a “bad motive” is not required to support a supply chain risk designation under the much broader definition in § 4713.

Applying that definition, the D.C. Circuit agreed with the DoW that Anthropic’s ability to “manipulate” its Claude model through training “in such a manner as to inhibit the DoW’s use thereof” created a supply chain risk, and the Secretary of War therefore acted within statutory authority in removing Anthropic from the supply chain under FASCSA.

The D.C. Circuit also held that:

• The DoW reasonably concluded that less intrusive measures were not available to reduce the supply chain risk given the pervasive integration of Claude into the DoW’s systems.
• Any lack of advance notice to Anthropic of the supply chain risk designation was harmless error because Anthropic was allowed to respond to the decision and its response did not change the DoW’s decision.
• Anthropic’s due process rights were not violated, in part because the “unprecedented velocity” of AI technology justified the DoW’s expedited process.
• Anthropic’s First Amendment rights were not violated because it was not removed from the supply chain due to its advocacy, but instead because it refused to accept the DoW’s essential contract terms.

What does this mean for Government Contractors?

Although the broader supply chain risk designation issued under 10 U.S.C. § 3252 remains subject to a permanent injunction entered by the California District Court, the D.C. Circuit’s ruling allows the DoW to use FASCSA to direct contractors to stop using Anthropic products and services in the performance of DoW contracts.

For now, contractors may still use Anthropic products commercially but there should be a strict separation from any DoW work. Although there is no current obligation or direction under FASCSA that prohibits contractors from using Anthropic, the DoW could do so at any time. Contractors should wait for written direction from a DoW contracting officer through a contract modification which identifies Anthropic as a covered source[1] and continue checking SAM.gov for Supply Chain Security Orders to see if Anthropic is added.

The Bottom Line

The two decisions are not in conflict; they were decided  on different statutes.  Section 3252 requires a threat from “an adversary,” while FASCSA reaches “any person” capable of manipulating a covered article, regardless of motive.  The DoW therefore now has a judicially approved path to exclude the use of Anthropic from DoW contract performance.  Other AI providers that retain post-deployment control over their models could face similar designations.  Contractors should inventory their use of Anthropic products, keep that use segregated from DoW work, prepare transition plans, and monitor SAM.gov and contracting officer communications for any FASCSA order.

[1] Under FAR 52.204-30(b)(4), a FASCSA order issued after the date of solicitation applies to a contract only if it is added by an amendment to the solicitation or through a modification to the contract.