Same ‘prompt,’ different responses: Anthropic supply chain risk designation stands in D.C. Circuit, for now, splitting from California district court

Updated, June 2, 2026 

The Anthropic supply chain risk designation dispute just got messier. As we wrote about last week, in March a California federal court issued a preliminary injunction against the government’s designation of Anthropic as a supply chain risk under one federal law. This week, the D.C. Circuit went the other way, rejecting Anthropic’s request for an order temporarily staying Department of War (DoW) directives labeling the artificial intelligence (AI) company a supply chain risk pursuant to a different statute and set the hearing on the merits for mid-May.

The D.C. Circuit did not reach the question of whether Anthropic showed a likelihood of success on the merits, though it noted that the case involved difficult and novel questions for which there is little precedent and emphasized that though Anthropic couched its interests as constitutional they “seem primarily financial in nature.” Id. at 2. Instead, the D.C. Circuit focused on weighing the costs and benefits to each side of granting a stay. 

With respect to the harm to Anthropic of allowing the designation to remain in place pending a decision on the merits, the D.C. Circuit acknowledged that it would “suffer some degree of irreparable harm.” But it found the evidence of any financial harm to be unclear and noted that “some record evidence suggests that Anthropic has financially benefited from its refusal to accede to the” DoW’s requests. Order at 2-3. The D.C. Circuit also found that Anthropic had not shown that its speech had been chilled during the pendency of the litigation. Id. at 2. 

Turning to the harm of a stay for the government, the D.C. Circuit was reluctant to “force the United States military to prolong its dealings with an unwanted vendor of critical AI services in the middle of a significant ongoing military conflict.” Id. at 3. It highlighted the parties’ recent disagreements over the use of Anthropic’s AI for military operations under the existing usage policy as well as their publicly deteriorating relationship. Id. at 3-4. For these reasons, the D.C. Circuit concluded that granting a stay would be a “substantial judicial imposition on military operations” and noted that it would “not lightly override” the DoW’s “judgments on matters involving national security.” Id. Thus, it denied the stay of the Pentagon’s designation of Anthropic as a supply chain risk. 

This is inconsistent with the decision coming out of the U.S. District Court for the Northern District of California. That court held that Anthropic was entitled to a preliminary injunction against the government’s directives and designations, labeling it a national security supply chain risk at issue there. That injunction remains in place and bars enforcement of the designations that prohibited numerous federal agencies from contracting with Anthropic or using its AI models. See Anthropic blacklisting blocked, for now: What the Anthropic injunction means — and what it doesn’t — for AI businesses.

The disparate rulings create uncertainty as to whether the designations will stick and what that means from a business perspective for Anthropic and other AI companies, making it all the more important that such entities take proactive steps to protect themselves when contracting with government agencies. See HSF Kramer Insights, ‘All lawful uses’: Precautions AI businesses need to take after Anthropic v. US DoW.

Update: On May 19, 2026, the D.C. Circuit heard oral argument on Anthropic’s challenge to the supply-chain risk designation. The argument, heard before a three-judge panel of Judge Katsas, Judge Rao, and Judge Henderson, revealed a potentially meaningful divide in approach between the judges. Judge Henderson appeared skeptical of the government’s position, pressing the view that the challenged designation may have represented an unwarranted escalation from what is, at bottom, a contractual dispute and signaling concern that the statute requires some showing of subversive or malicious conduct not clearly present here. By contrast, Judges Katsas and Rao appeared more receptive to the government’s arguments, although Judge Katsas probed the government on the contours of any “maliciousness” requirement. Judge Rao focused heavily on threshold justiciability issues—particularly whether the court has jurisdiction under the statute authorizing review of a “covered procurement action” and what specific action was properly before the court. And, both Judge Rao and Judge Katsas posed questions suggesting possible concerns about whether there was a sufficiently final agency action to permit judicial review . In sum, the questioning indicated that at least some members of the panel were exploring whether the case could be resolved on jurisdictional or prudential grounds without reaching the merits.