Palantir's Claude AI Plans Strikes in Pentagon Demos
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic refused the Pentagon unconditional access to Claude in late February, insisting the models not be used for mass surveillance of Americans or fully autonomous weapons; the Pentagon labeled Anthropic's products a "supply-chain risk," prompting two lawsuits this week alleging illegal retaliation by the Trump administration.
- Palantir announced in November 2024 that it would integrate Claude into software sold to US intelligence and defense agencies, and Claude reportedly played an instrumental role in the US military operation that led to the capture of Venezuelan president Nicolás Maduro in January.
- Project Maven — the Maven Smart System Palantir has developed for the Pentagon since 2017 — can apply computer vision to satellite imagery, automatically detect "enemy systems," and "nominate" targets for ground or aerial bombardment, with an AI Asset Tasking Recommender proposing which bombers and munitions to assign.
- Maven is accessible across the Army, Air Force, Space Force, Navy, Marine Corps, and US Central Command (which is overseeing military operations in Iran), according to Pentagon chief digital and AI officer Cameron Stanley, who said the tool is being deployed "across the entire department."
- A 2023 Palantir demo showed an AIP Assistant powered by Claude helping a fictional military operator plan a ground attack on tanks — generating three courses of action (air asset, long-range artillery, tactical team), analyzing the battlefield, drawing a troop route, and assigning jammers to sabotage enemy communications.
- Palantir's AIP lets Pentagon customers choose among third-party LLMs from Anthropic, Google, and Meta and pull from classified data sets; a June 2025 demo by Anthropic public-sector lead Kunaal Sharma showed Claude generating an intelligence report on Ukraine's Operation Spider's Web in minutes rather than the five hours a human analyst would need.
Why it matters: The disclosures reveal how commercial AI chatbots are embedded in Pentagon targeting workflows — from detecting objects in satellite imagery to recommending specific bombers and munitions — while Palantir and Anthropic have disclosed almost nothing about which systems actually use Claude, leaving no public visibility into life-and-death decisions increasingly mediated by AI.
