Archive · 26 Sep 2026
Appeals Court Upholds Pentagon’s Anthropic Supply Chain Risk Designation
Anthropic makes an artificial intelligence model called Claude.
The Pentagon wanted Anthropic to remove safety rules for certain military uses.
These uses included fully autonomous lethal weapons and mass surveillance of Americans.
Anthropic refused to remove the safeguards.
The Pentagon then labeled Anthropic a supply chain risk and canceled its military contracts.
A federal appeals court voted 2-1 to let that designation continue.
This means the military can remove Claude from its systems and limit Anthropic’s defense work.
Anthropic says the decision is wrong and is considering further review.
Another federal court in California reached a different conclusion about a broader government ban.
A federal appeals court ruled 2-1 that the Pentagon may continue designating Anthropic a supply chain risk.
The dispute began after Anthropic refused to remove Claude safeguards covering fully autonomous lethal weapons and mass domestic surveillance.
The March designation canceled Anthropic’s military contracts and barred other Pentagon contractors from using its technology.
The ruling allows the Defense Department to remove Claude from its systems and restrict Anthropic’s products for defense work.
The decision conflicts with a separate California ruling that found a broader Trump administration ban against Anthropic unlawful.
- Who
- Anthropic, the United States Department of Defense, the federal appeals court, and Pentagon officials including Pete Hegseth and Emil Michael.
- What
- The appeals court upheld the Pentagon’s supply chain risk designation against Anthropic.
- When
- The dispute began in February, the designation was made in March, and the appeals court ruling was issued on Friday.
- Where
- The ruling came from a US federal appeals court; a separate related case was decided in California.
- Why
- The Pentagon acted after Anthropic refused to remove safeguards on Claude’s use in autonomous lethal weapons and mass domestic surveillance, while officials also raised concerns about military control of the software.
Anthropic’s Position
Pentagon’s Position
Supply chain designation
Anthropic’s Position
Anthropic says the designation is unlawful, points to the separate California ruling, and is considering further review.
Pentagon’s Position
The Pentagon maintains that it can exclude Anthropic under public procurement rules, and the appeals court left the designation in place.
AI safeguards
Anthropic’s Position
Anthropic refused to remove safeguards limiting Claude’s use in fully autonomous lethal weapons and mass domestic surveillance of Americans.
Pentagon’s Position
Pentagon officials argue that private companies should not insert their preferences into the military chain of command.
Military software control
Anthropic’s Position
Anthropic has maintained that it cannot alter its models after they have been delivered to the military.
Pentagon’s Position
The appeals court said Anthropic determines the behavior of new model versions, while the military must update its systems to use them, creating concerns about operational risks.
- Court ruling
- A federal appeals court ruled 2-1 against Anthropic.
- Designation
- The Pentagon may continue designating Anthropic as a supply chain risk.
- Military consequences
- The Defense Department can remove Claude from its systems and restrict Anthropic’s products for defense work.
- Contracts
- The March designation canceled Anthropic’s military contracts and barred other Pentagon contractors from using its technology.
- Safeguards
- Anthropic refused to remove restrictions on fully autonomous lethal weapons and mass domestic surveillance of Americans.
- Competing companies
- OpenAI, xAI, Google, and Microsoft later signed agreements with the Pentagon at different times.
- Separate case
- A California federal court found a broader Trump administration ban against Anthropic unlawful last month.
