Pentagon to adopt Palantir AI as core US military system after dispute with Anthropic

Soon after the Pentagon-Anthropic rivalry, the Pentagon announced it will deal with Palantir's AI for US military systems.

Palantir’s Maven artificial intelligence system will become an official program of record, Deputy Secretary of Defense Steve Feinberg said in a letter to Pentagon leaders, a move that locks in long-term use of Palantir’s weapons-targeting technology across ‌the U.S. military.

In the March 9 letter to senior Pentagon leaders and U.S. military commanders, Feinberg said embedding Palantir’s Maven Smart System would provide warfighters “with the latest tools necessary to detect, deter, and dominate our adversaries in all domains." ”.

The decision is expected to go into effect by the close of the current fiscal year, which ends in September, according to the letter, which was reviewed by Reuters and has not been previously reported.

Maven is a command-and-control software platform that analyzes battlefield data and identifies targets. It is already the primary AI operating system for the U.S. military, which has carried out thousands of targeted strikes against Iran over the last three weeks.

Designating Maven as a program of record will streamline its adoption across all arms of the military and provide stable, long-term funding, Feinberg said.

The memo ordered oversight of Maven be moved from the National Geospatial Intelligence Agency to the Pentagon’s Chief Digital Artificial Intelligence Office within 30 days. Future contracting with Palantir will be handled by the Army, the letter said.

“It is imperative that we invest now and with focus to deepen the integration of artificial intelligence (AI) across the Joint Force and establish AI-enabled decision-making as the cornerstone of our strategy,” Feinberg wrote.

During a presentation at a Palantir event earlier this month, Pentagon official Cameron Stanley, who leads its AI office, demonstrated how the company’s Maven platform could be used for weapons targeting in the Middle East, and he showed heat map screenshots from the Maven platform.

United Nations expert panels have warned AI weapons targeting without human intervention raises ethical, legal and security risks since AI picks up inadvertent biases from the data sets used to train it.

Palantir says its software does not make lethal decisions and humans remain responsible for selecting and approving targets.

Palantir developed its AI system to serve the Pentagon’s Project Maven, which began as a drone-imagery labeling program in 2017. In 2024, the Pentagon awarded Palantir a contract worth up to $480 million. 

That year, Palantir’s Chief Technology Officer Shyam Sankar told the House Armed Services Committee, opens new tab that Maven had “tens of thousands” of users and urged Congress to provide more funding. 

In May 2025, the Pentagon increased the contract ceiling to $1.3 billion.

One potential complication in deeper Maven adoption is the software’s use of the Anthropic-made Claude AI tool, Reuters previously reported. 

Anthropic was recently deemed a supply chain risk by the Pentagon amid a months-long spat over safety guardrails surrounding the AI.