On August 27, 2026, the U.S. District Court for the Northern District of California filed a 59-page opinion addressing the U.S. Department of Defense’s actions against Anthropic. The court found that designating Anthropic a supply chain risk violated 10 U.S.C. § 3252 and constituted arbitrary and capricious agency action. It also found unlawful retaliation under the First Amendment and a lack of the pre-deprivation process required by the Fifth Amendment.

The key point is that the ruling should not be expanded into either “a complete Anthropic victory” or “the removal of all U.S. government procurement restrictions.” The court recognized the Defense Department’s authority to choose the AI vendors it wants. What it found problematic was the manner and scope of designating Anthropic a supply chain risk because of its positions on AI safety, then restricting dealings with federal agencies as well as defense contractors and suppliers.

Key Changes

The measures challenged by the court were broad restrictions imposed under presidential and Defense Department directives, requiring federal agencies to stop using Anthropic technology and barring defense contractors and suppliers from doing business with Anthropic. The ruling found that these measures did not comply with the relevant law and lacked an adequate administrative basis.

The ruling’s legal conclusions can be summarized in three parts:

  • The supply chain risk designation violated 10 U.S.C. § 3252.
  • The action was arbitrary and capricious.
  • The action against Anthropic constituted retaliation under the First Amendment and lacked the pre-deprivation process required by the Fifth Amendment.

Accordingly, the central issue was not whether the Defense Department could select a particular AI company, but whether it could impose restrictions approaching a government-wide procurement ban because of that company’s policy positions.

Current Status

The dispute centers on two restrictions Anthropic sought to maintain: mass surveillance of people in the United States and the use of fully autonomous lethal weapons. Anthropic sought to preserve those limits, while the Defense Department demanded that AI be available for all lawful purposes. The broader government and military debate surrounding Claude arose from this conflict.

Anthropic welcomed the ruling and said it would work productively with the government in support of national security. However, the ruling did not declare Anthropic’s entire AI safety policy correct. The court addressed not every aspect of that policy, but the legal validity of the Defense Department’s supply chain risk designation and broad procurement restrictions arising from the dispute.

Anthropic is also pursuing a separate supply chain risk designation case in the U.S. Court of Appeals for the District of Columbia Circuit. That case concerns the possibility of exclusion from civilian government contracts. As a result, the California federal court’s ruling alone does not establish that the entire procurement dispute between the government and Anthropic has ended.

Impact

The most direct significance of the ruling is its distinction between the government’s discretion to choose AI suppliers and the manner in which that discretion is exercised. The Defense Department may choose the AI supplier it wants, but designating Anthropic a supply chain risk and broadly excluding contractors and suppliers from doing business with it requires a separate legal basis and appropriate procedures.

The case has become a test of whether procurement restrictions can be justified solely by government demands when AI safeguards conflict with national security uses of technology. Anthropic’s restrictions focused on mass surveillance and fully autonomous lethal weapons, while the Defense Department’s demand covered all lawful purposes. Rather than resolving the policy conflict itself, the court reviewed the legal basis and procedures for the government’s response.

Therefore, it would be difficult to interpret this ruling alone as meaning that Claude’s government and military contracts will immediately resume in full—or that they have all been suspended. The confirmed change is the court’s finding that the specific procurement restrictions and supply chain risk designation were unlawful. The actual scope of contracts and use may still depend on the remaining litigation and subsequent government action.

The benchmark set by this ruling is less “Who can choose AI?” than “On what grounds and through what procedures may that choice be restricted?”

Next Checkpoints

There are three developments to watch:

  1. How the U.S. Department of Defense responds to the Northern District of California’s ruling
  2. How the separate supply chain risk designation case before the U.S. Court of Appeals for the District of Columbia Circuit affects civilian government contracts
  3. What changes, if any, cooperation between Anthropic and the government brings to the actual scope of procurement and use

Based on the facts confirmed so far, it cannot be said that this ruling removed all restrictions on Anthropic’s U.S. government procurement. However, the court addressed the supply chain risk designation, broad transaction restrictions, constitutional rights, and pre-deprivation process together, placing major aspects of the Defense Department’s action on hold. The subsequent legal proceedings and government response will be the next benchmarks for determining the scope of Claude’s government use.