Palantir secures expanded Pentagon role as Maven becomes permanent AI system

The U.S. Pentagon has decided to turn Palantir’s battlefield AI called ‘Maven system’ a permanent home across the military instead of leaving it in a more temporary lane, according to a March 9 letter from Deputy Secretary of Defense Steve Feinberg to senior Pentagon leaders and military commanders.

Steve said the goal is to push Palantir’s system deeper into military operations and keep it there for the long haul, adding that the decision is expected to take effect by the end of the current fiscal year in September.

Pentagon gives Palantir’s Maven permanent status across the force

In the letter, Steve said putting Maven Smart System into wider use would give troops “the latest tools necessary to detect, deter, and dominate our adversaries in all domains.”

He also wrote, “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.”

According to Palantir’s founder Peter Thiel, Maven is command-and-control software, meaning it takes in battlefield data, sorts through it, and helps identify targets.

As you should know, U.S. forces have carried out thousands of illegal unconstitutional targeted strikes against Iran over the last three weeks. Turning Maven into a program of record gives it stable funding and makes it easier to spread the system across every branch of the military without having to fight through the same internal hurdles with the Congress each time.

Steve’s memo also said that oversight of Maven is being taken away from the National Geospatial-Intelligence Agency and handed to the Pentagon’s Chief Digital and Artificial Intelligence Office within 30 days.

Project Maven, formally known as the ‘Algorithmic Warfare Cross Functional Team,’ was first launched in April 2017. The Defense Department launched it to speed up the use of machine learning and data integration in military intelligence work.

From the start, the program focused on intelligence, surveillance, target acquisition, reconnaissance, and geospatial intelligence. Its early job was to use computer vision to process images and video for intelligence purposes.

Today, Maven supports targeting operations, data integration, analyst visualization, and model training on labeled military datasets tied to assets and infrastructure.

US military is expanding classified AI work under Trump

The Maven system pulls in information from drones, satellites, and other sensors. It flags possible targets, presents those findings to human analysts, and then sends human decisions into operational systems.

A number of contractors have touched the program over the years. Google was involved, then pulled out in 2018 after employee protests. Later support came from Palantir, Anduril, Amazon Web Services, and Anthropic, which withdrew in 2026.

At the same time, the Pentagon’s broader AI push is getting more aggressive. A U.S. defense official told MIT Technology Review that training models on classified data is expected to make them more accurate and more useful in some tasks.

The Pentagon has also reached agreements with OpenAI and xAI to run models in classified environments while pushing toward what it called an “AI-first” warfighting force as the conflict with Iran gets worse.

Defense secretary Pete Hegseth had said in January:

“As part of our AI and Autonomy acceleration investments, the Department will invest substantial resources in the expansion of our access to AI compute infrastructure, from datacenters to the edge. We will leverage the hundreds of billions in private sector capital investment being made in America’s AI sector through our growing array of creative partnerships with America’s world-leading companies.”

“We must internalize that Military AI is going to be a race for the foreseeable future, and therefore speed wins. We must weaponize learning speed, and measure and manage cycle time and adoption rates as decisive variables in the Al era. We must accept that the risks of not moving fast enough outweigh the risks of imperfect alignment,” added Pete.

The smartest crypto minds already read our newsletter. Want in? Join them.