Palantir CEO First Responds to Pentagon Ban, Says Will Integrate Other Large Language Models Besides Claude

Gate News, March 13 — During the AIPcon 9 conference today, Palantir CEO Alex Karp publicly responded for the first time to the Pentagon’s designation of Anthropic as a supply chain risk. Karp stated, “Our products have integrated Anthropic, and we may incorporate other large language models in the future.” He also said that Palantir is committed to providing the military with “the best and most lethal technology,” but believes there are “many reasons” to restrict the use of these technologies in domestic law enforcement and other scenarios, “because they are too powerful.” Although the Pentagon officially listed Anthropic as a supply chain risk and banned its participation in government contracts last Friday, the Department of Defense is still using the Claude model to support operations related to Iran. Department of Defense CTO Emil Michael today said, “It’s impossible to remove a deeply embedded system overnight.” A six-month transition plan is in place, but if the conflict persists and sensitive operations are involved, “obviously exceptions will be made to avoid jeopardizing current military actions.” An internal memo permits “mission-critical activities” to continue using Anthropic products beyond the six-month period. Anthropic has filed a lawsuit against the U.S. government this week, seeking to revoke the supply chain risk designation and citing billions of dollars in damages. Palantir partnered with Anthropic and AWS in 2024 to provide the Claude model to U.S. defense and intelligence agencies. Other defense contractors like Lockheed Martin have also instructed employees to stop using Claude.