The Pentagon has stopped using Anthropic’s artificial intelligence tools after previously designating the company a national security supply-chain risk. The move ends the U.S. military’s use of Anthropic AI at the Pentagon, including Claude models that had been integrated into intelligence and data-analysis systems.
A U.S. Department of Defense official confirmed that the department had ceased using Anthropic products. The decision follows months of tension between the Pentagon and the AI company over safeguards governing military use of its technology.
Defense Secretary Pete Hegseth designated Anthropic a supply-chain risk in February and ordered the department to phase out its technology. Anthropic has challenged that designation in court, calling the government’s action unlawful.
The company had worked with the U.S. government and military since 2024. Its AI models became some of the first advanced commercial AI systems deployed in classified government environments.
Dispute Over Military Use of Claude
The conflict intensified after the Pentagon sought broader access to Anthropic’s Claude models. According to the report, defense officials wanted the company to remove some safety restrictions governing how the technology could be used.
Anthropic resisted the request, citing concerns about applications including mass surveillance and autonomous weapons.
Despite the dispute, Claude reportedly remained in use inside the Defense Department after the February designation. People familiar with the Pentagon’s AI operations said the models supported research, intelligence gathering and analysis.
Claude was also integrated into the Maven Smart System, a major Pentagon data platform operated by Palantir. The system helps organize large volumes of intelligence, including satellite imagery and drone footage.
Analysts could then use large language models to process the information for tasks such as intelligence analysis. The deep integration reportedly made replacing Claude more complicated than simply switching to another AI model.
Pentagon Expands Use of Other AI Systems
As the dispute with Anthropic developed, the Pentagon expanded its relationships with other major artificial intelligence companies.
The department has signed agreements involving Google, xAI and OpenAI, while OpenAI tools have reportedly gained wider adoption in some parts of the Defense Department.
The Pentagon’s decision comes as Anthropic continues to emphasize safety controls around advanced AI. The company has warned that increasingly capable systems could create serious security and societal risks.
At the same time, Anthropic’s relationship with the White House appears less confrontational. President Donald Trump recently met Anthropic CEO Dario Amodei at the White House alongside other technology executives.
The dispute nevertheless highlights a growing challenge for governments and AI developers: how to balance access to increasingly powerful artificial intelligence systems with restrictions on their military use.
The end of Anthropic AI at the Pentagon also shows how difficult replacing AI systems can become once they are deeply integrated into military data and intelligence infrastructure.
