Anthropic CEO Dario Amodei will attend the World Economic Forum in Davos in January 2025.Fabrice Coffrini/AFP/Getty Images
The Pentagon on Thursday gave the artificial intelligence research institute Anthropic a formal supply chain risk designation, restricting its use of technology that officials said is used in military operations in Iran.
The “supply chain risk” label confirmed in Anthropic’s statement is effective immediately and prohibits government contractors from using Anthropic’s technology in U.S. military operations.
However, companies can still use Anthropic’s Claude in other projects unrelated to the Department of Defense, CEO Dario Amodei said in a statement, adding that this restriction applies only to the use of Anthropic AI in contracts with the Department of Defense.
The risk designation comes after months of controversy over the company’s claims for safeguards that the Pentagon, which the Trump administration calls the Department of the Army, went too far. In a statement, Amodei reiterated that the company will challenge this designation in court.
Anthropic CEO says AI companies cannot ‘in good conscience’ accept Pentagon demands
Amodei said in a statement Thursday that in recent days Antropic and the Pentagon have discussed a possible plan for the Pentagon to decommission Claude. The two countries have been discussing how Anthropic can continue to cooperate with the military without dismantling its safety equipment, he added.
Amodei also apologized for an internal memo published Wednesday by technology news site The Information. In the memo, originally written last Friday, Amodei said one reason Pentagon officials don’t like the company is that “we don’t admire Trump like a dictator.”
The release of the internal memo comes as Anthropic’s investors race to stem the damage caused by the company’s conflict with the Department of Defense.
The Department of Defense did not respond to requests for comment.
The action marked an unusual rebuke from the United States for a U.S. technology company that cooperated with the Pentagon earlier than its rivals. The move comes as the department continues to rely on Anthropic’s technology to provide support for military operations, including in Iran, according to people familiar with the matter.
Claude is likely being used to analyze intelligence and assist in operational planning.
A Microsoft spokesperson said the company’s lawyers reviewed the designation and concluded that “Humanity products, including Claude, will continue to be available to our customers, other than the Department of the Army, through platforms such as M365, GitHub, and Microsoft’s AI Foundry.”
Microsoft can continue to work with Anthropic on non-defense-related projects, the spokesperson added.
Amazon, an investor in Anthropic and a major customer of the company’s Claude models, did not immediately respond to a request for comment outside of normal business hours.
Palantir’s Maven Smart Systems, a software platform that provides intelligence analysis and weapons targeting for the military, uses multiple prompts and workflows built using Anthropic’s Claude code, Reuters previously reported.
Anthropic has been the most aggressive of its competitors in courting U.S. national security officials. But the company and the Pentagon have been at odds for months over how the military can use its technology on the battlefield. The conflict erupted in public earlier this year.
Anthropic has refused to lift its ban on autonomous weapons and Claude AI to enhance US mass surveillance. The Pentagon pushed back, saying it should be able to use the technology as needed, as long as it complies with U.S. law.
The “supply chain risk” label gives Anthropic the status previously typically used by the U.S. government against foreign adversaries. The United States has taken similar steps to remove Chinese technology giant Huawei from the Pentagon’s supply chain.
