By: Margaret M. Cassidy & Jelena Tasic
Background
The Anthropic-Department of Defense (DoD) battle started earlier this year when, over social media, DoD designated Anthropic a “supply chain risk” and along with the White House restricted the use of Anthropic products, particularly its Claude AI models, initially across the whole federal government. DoD and the White House also issued directives that federal government contractors could not use Claude AI even for commercial business. The ban was a result of Anthropic refusing to permit DoD to use its AI for mass domestic surveillance and for use in autonomous weapons systems
In addition to the notice of the ban on social media, on March 4th, DoD sent Anthropic two different letters referencing a different law in each to support its action labeling Anthropic a supply chain risk to national security and banning its technology from the government.
One letter explained DoD was using the Federal Acquisition Supply Chain Security Act (FASCSA) (41 U.S.C. §§ 1321–1328; 41 U.S.C. § 4713) to ban Anthropic. This statute is designed to apply across all federal agencies and among other things, it allows a banned contractor to challenge the decision by filing a petition in the U.S. Court of Appeals for the D.C. Circuit. Anthropic filed a petition challenging the ban in this court.
The other letter explained that DoD was using the Requirements for Information Relating to Supply Chain Risk (10 U.S.C. § 3252) to ban Anthropic. This law applies only to DoD and does not mandate where a case challenging a decision must be filed. Anthropic filed its challenge to DoD’s decision that it was a supply chain risk in federal district court in California. [Click here for The Defense Salon article with more details on the background.]
Each court issued a ruling on Anthropic’s request:
- The federal district court in California stopped the ban
- The D.C. appellate court allowed the ban to continue.
The conflicting rulings are not the only unusual occurrence in this case, DoD’s decision to send two separate letters rather than one citing both statutes is as well and the irregularities compound from there. The letters were dated March 3 after Secretary Hegseth had already publicly announced on February 27 that DoD had determined Anthropic was a supply chain risk indicating that they were drafted retroactively to paper over a decision already made. Neither letter explains what less intrusive measures were considered before imposing the outright ban, as both statutes require.
National Security vs. Contractor Risk: Competing Judicial Priorities
The two courts arrived at contradictory decisions in part for how each viewed the dispute with the California court focusing on the contracting risks of DoD’s decision while the D.C. court focused on national security implications.
California Court: On March 26, 2026, the California court issued a preliminary injunction halting the DoD designation of Anthropic as a supply chain risk under the DoD specific law by concluding that key aspects of the government’s actions are likely unlawful.
The California court focused on Anthropic’s rights as a contractor and systemic risks to the procurement ecosystem, including reputational harm, potential de facto debarment, and the broader impact on private-sector participation in defense innovation if the ban stood. The court questioned whether the government’s stated national security rationale was adequately supported; raised concerns that the designation could be applied in a manner inconsistent with statutory requirements; and considered whether DoD trampled on Anthropic’s due process and First Amendment rights. The court concluded that Anthropic is likely to succeed on the merits because the government’s actions likely constitute unlawful retaliation, due process violations, and improper use of procurement authorities.
Based on this reasoning, it granted injunctive relief halting key elements of the DoD’s ban.
D.C. Federal Appellate Court: In contrast, April 8, 2026, the D.C. Circuit court denied Anthropic’s request to suspend DoD’s directive labeling Anthropic a supply chain risk, allowing the designation and related restrictions to remain in effect while the case proceeds on an expedited basis.
The D.C. Circuit court did not address the legality of the supply chain risk designation, deferring at least for now to DoD’s decision that Anthropic may be a national security risk. It applied a stringent standard that courts must use when determining if the court should stop some type of harm. Using this standard, the court determined that Anthropic failed to show that the balance of equities favored immediate judicial intervention to stop the ban, particularly given national security considerations and given it was unclear whether Anthropic was financially harmed. The court said in its order that did not want to mandate that DoD continue using a contractor it did not want to use especially when there are ongoing military conflicts. As a result, the designation remains operative pending further review into DoD’s judgment on matters of national security even as the appellate court acknowledged that the case presents novel and complex legal questions
Implications for Defense Contractors
For now, DoD’s determination remains in effect.
Contractors and subcontractors should anticipate continued scrutiny of their technology supply chains and contracting risks, particularly where Anthropic AI tools are integrated with their operations. Government agencies, primes and higher-tier subcontractors may interpret and operationalize the ban differently and may continue to act on the designation by discouraging or restricting the use of Anthropic products in performance of covered contracts.
What’s Next?
Short answer – who knows, especially since it is not only the courts that signaled inconsistent approaches. The government itself made contradictory statements on engaging with Anthropic. While Secretary of Defense Hegseth has hunkered down telling the Senate that Anthropic is run by “an ideological lunatic who shouldn’t have sole decisionmaking authority” on DoD operations, President Trump has recently stated that Anthropic employees are “high-IQ people” who can “be of great use” so “we’ll get along with them just fine.”
The final decision will be with the courts.
The D.C. appellate court has established an expedited briefing schedule, with oral argument set for May 19, 2026, indicating that a more definitive appellate ruling may emerge soon. At the same time, the California court’s injunction stopping the ban remains in place, though its potential long-term effect remains subject to further appellate review since DoD appealed the decision to the federal appellate court in California.
Until the courts resolve each case on the merits, contractors should expect continued unpredictability in enforcement and guidance, particularly across different agencies and with prime and subcontractors.
The conflicting cases present material compliance and business risk that requires close monitoring of agency direction, contract requirements, and supply chain exposure. How appellate courts ultimately resolve these issues will likely shape the future scope of government authority over technology contractors in the defense industrial base.
