In 2024 the Maven Smart System, the operational AI tool that grew out of the US military’s Project Maven, moved firmly into fielded use. In May 2024 the US Army awarded Palantir a five-year contract to expand the system to thousands of users across several combatant commands, a milestone in turning what began in 2017 as an experimental computer-vision effort into a standing piece of military software.
The Maven Smart System fuses data from many intelligence, surveillance, and reconnaissance sources, such as satellite imagery and sensor feeds, to detect and track potential targets and to support commanders’ decisions. It is built on Palantir’s broader analytics technology; the company describes its Artificial Intelligence Platform as software that “connects AI with your data and operations” and is designed to “drive automation across operational processes.” Maven Smart System applies that approach to the battlefield, where the core challenge is making sense of overwhelming volumes of sensor data faster than human analysts can on their own.
The system’s expansion was significant because Project Maven had been the flashpoint for the 2018 Google employee revolt over military AI. Six years later, the same program had not only survived but become a flagship deployment, with its prime contractor work moving to companies that embraced defense work. Its growth tracked a broader US push to put AI-driven decision support at the center of military operations.
For a general reader, the Maven Smart System is a concrete answer to the question of what military AI actually looks like in practice today. It is not an autonomous weapon but a decision-support tool, which is exactly why it sits at the heart of debates over automation bias and how much weight human operators should give to machine-generated targets and analysis.