U.S. Military "Formalizes" AI Systems of American Tech Companies, Expert Analysis

[Global Times Report, Reporter Liu Yang] According to recent reports from Reuters and other foreign media, the Pentagon will designate the artificial intelligence (AI) system Maven from the American technology company Palantir as an "officially established project," allowing multiple branches of the U.S. military to utilize the company's related technology in the military field. The Pentagon emphasized that future strategies will focus on AI-driven decision-making to accelerate the comprehensive integration of various military branches. What does it mean for Palantir to receive the Pentagon's "official AI designation"? Why did the Pentagon choose Palantir?

According to reports, U.S. Deputy Secretary of Defense Steve Feinberg stated in a letter sent on March 9 to senior department officials and U.S. military commanders that the introduction of Palantir's Maven intelligent system will provide frontline combat personnel with "the latest tools needed to detect, deter, and suppress adversaries across all domains." The letter noted that this decision is expected to officially take effect before the end of the current fiscal year (by the end of September).

The Maven project developed by Palantir was launched in 2017 as a drone image labeling project. In 2024, the Pentagon awarded Palantir a contract worth up to $480 million. In the same year, Palantir's Chief Technology Officer testified before the U.S. House Armed Services Committee, stating that the Maven project has "tens of thousands" of users and urged Congress to provide more funding. In May 2025, the Pentagon raised the contract ceiling to $1.3 billion.

Feinberg stated that designating Maven as an "officially established project" will accelerate its deployment across the U.S. military branches and provide stable long-term funding support. In the U.S. military's defense procurement system, an "officially established project" refers to a procurement plan that has been included in future annual defense plans and has received approval and budget allocation.

Liu Wei, director of the Human-Computer Interaction and Cognitive Engineering Laboratory at Beijing University of Posts and Telecommunications, told Global Times reporters on the 24th that Maven is not a simple AI system, but a command and control platform that integrates various AI functions, capable of quickly analyzing vast amounts of data from satellites, drones, radars, sensors, and intelligence reports, and automatically identifying potential threats or targets through AI, such as enemy military vehicles, buildings, and weapons storage facilities. It has been revealed that the U.S. military has conducted thousands of precision strikes against Iranian targets using systems like Maven in the past three weeks.

According to reports, in early March, at an event hosted by Palantir, Cameron Stanley, head of the Pentagon's AI office, demonstrated how Maven is used for weapon targeting in the Middle East. After the demonstration, Stanley stated, "When we first started doing this, it took hours to accomplish what you just saw." Now, it only takes a few seconds.

Liu Wei believes that Palantir obtaining a Pentagon license and incorporating its AI system into the U.S. military's joint core operating system goes far beyond the scope of a technical contract and is a key step in the U.S. military's transformation towards intelligence. This cooperation may even influence the nature of future warfare to some extent. Its significance lies not only in enhancing the U.S. military's combat effectiveness, such as shortening kill chain times and improving target strike accuracy, but also in solidifying AI technology as a core competitive advantage of the U.S. military. With further improvements to the Maven system, such as integrating more AI models and enhancing data fusion capabilities, AI will become the "core weapon" in future U.S. military warfare.      

Liu Wei emphasized that the acceleration of AI militarization by the U.S. also brings ethical and legal risks, such as civilian casualties caused by target identification errors and accountability issues regarding AI decision-making. However, the U.S. military clearly places these risks behind "technological advantages." Although Palantir has stated that its software will not make lethal decisions, the selection and approval of targets remain the responsibility of humans. Nevertheless, how the AI systems that have deeply integrated into the U.S. military's intelligence analysis, decision-making, and strike chains will evolve remains uncertain.