Palantir Retains Anthropic AI for Military Operations Despite Pentagon Ban
Denver, Thursday, 12 March 2026.
Despite a strict Pentagon blacklist, Palantir CEO Alex Karp confirmed the company still uses Anthropic’s Claude AI to power precise military operations in the ongoing U.S. conflict with Iran.
Navigating the Pentagon’s Blacklist
Speaking at Palantir’s AIPcon 9 in Maryland on Thursday, March 12, 2026, CEO Alex Karp clarified the immediate reality of defense technology integration: unwinding a deeply embedded artificial intelligence system takes time [1]. Despite the Department of War designating Anthropic a “supply-chain risk” in early March 2026, Palantir Technologies (NYSE: PLTR) continues to utilize Anthropic’s Claude large language model [1][4]. Karp noted that while the department plans to phase out Anthropic, the technology remains active in current operations [1].
The Ideological Clash Over AI Guardrails
The rift between Anthropic and the U.S. government stems from a fundamental disagreement over the ethical deployment of artificial intelligence [5]. During a February 24, 2026, meeting at the Pentagon, Anthropic CEO Dario Amodei and Secretary of War Pete Hegseth clashed over the military’s demand for unrestricted access to Claude for “all lawful use cases” [3][5]. Anthropic firmly maintained two red lines: prohibiting the use of its models in fully autonomous kinetic weaponry and banning its application in mass domestic surveillance [3][5]. Following an ultimatum, the Trump administration officially ordered all federal agencies to immediately cease using Anthropic’s technology, with President Trump publicly condemning the firm [5].
Market Shifts and the Future of Defense Tech
The blacklisting of Anthropic has created a massive vacuum in the defense sector, rapidly shifting economic and technological power to its competitors [GPT]. Anthropic had previously held the unique distinction of being the only AI firm offering services for classified use cases to U.S. defense agencies, largely due to its partnership with Palantir [5]. However, within days of the fallout, rivals quickly moved to capture the market share. On February 28, 2026, OpenAI announced a new agreement to deploy its models in classified Pentagon environments, while the Department of War also advanced negotiations with Google and xAI [3][5].