On August 27, 2026, the U.S. District Court for the Northern District of California issued summary judgment on Anthropic’s principal claims in Anthropic PBC v. U.S. Department of War. The court’s conclusion went beyond a simple contract dispute between a company and the government. It ruled unlawful the U.S. Department of Defense’s designation of Anthropic as a supply chain risk, along with measures preventing contractors, suppliers, and partners doing business with the military from engaging in commercial activity with Anthropic.
The key point is not simply that “an AI company won.” The court examined the legal basis and procedures for the government’s actions, as well as whether they constituted retaliation for protected expression and policy choices.
Key changes
According to the court’s decision, Secretary of Defense Pete Hegseth’s supply chain risk designation constituted unlawful retaliation under the First Amendment and violated the pre-deprivation procedures required by the Fifth Amendment. The court also found that the designation violated 10 U.S.C. § 3252 and was arbitrary and capricious under the Administrative Procedure Act.
The challenged measures went beyond removing Anthropic from a government transaction list. The Department of Defense imposed broad restrictions preventing contractors, suppliers, and partners that contract or cooperate with the military from engaging in commercial activity with Anthropic. The court described the measures as “unlawful and unsupported.”
Current status
The dispute centered on conflicting positions over the permissible uses of Claude. Anthropic maintained limits on mass domestic surveillance of Americans and the use of lethal autonomous weapons without human control. The Department of Defense, by contrast, demanded the ability to use Claude for all lawful purposes.
According to the decision, President Donald Trump directed all federal agencies on February 27, 2026, to stop using Anthropic technology, and Secretary Hegseth ordered the Department of Defense’s supply chain risk designation that same day. Anthropic officially announced on March 4 that it had received confirmation of the designation from the Department of Defense and warned that it would pursue legal action.
The ruling does not mean that Anthropic’s AI safety restrictions themselves were found to be technically perfect. The issue before the court was how the government legally implemented the supply chain risk designation and transaction restrictions in response to a company’s policy positions and conditions of use.
Impact
The scope of the ruling must be clearly distinguished. The court did not find that the Department of Defense lacked authority to select other AI providers through methods consistent with the law and Constitution. Nor does the decision mean that the Department of Defense must use Anthropic products.
However, the government’s designation of Anthropic as a supply chain risk and its broad restrictions on the commercial activities of related private contractors and partners will be difficult to maintain as issued. As a result, a significant issue remains: what procedures and legal authority are required when a company’s safety standards and contract terms conflict with the federal government’s intended uses in AI procurement.
Next points to watch
The government is expected to appeal the ruling. The outcome of any appeal has not been predetermined, and the final scope of the current judgment may change depending on the appellate proceedings. Anthropic is also challenging a separate supply chain risk designation case under different laws in the D.C. Circuit.
Readers should watch three developments:
- Whether the government actually appeals the Northern District of California’s ruling
- If an appeal is filed, how the scope of the ruling concerning the supply chain risk designation and transaction restrictions is addressed
- How the separate D.C. Circuit case relates to this ruling
The conclusion currently established is that the U.S. Department of Defense’s designation of Anthropic as a supply chain risk and its broad transaction restrictions were found unlawful by a federal court. However, the ruling did not automatically restore all government contracts or require the government to use Claude. Because further proceedings remain, the government’s appeal and the separate case must be followed together.