After the U.S. Department of Defense designated AI corporations Anthropic as a "supply chain risk" corporation, moves to halt the use of Anthropic's AI models are spreading in the U.S. defense industry.
According to Reuters and CNBC on the 4th (local time), Lockheed Martin hinted it may stop using Anthropic's AI model "Claude," saying it would "follow the president and Ministry of National Defense guidance." Lockheed Martin said the impact would be limited because it does not rely on a single AI supplier.
Ten defense startups backed by defense investor J2 Ventures are also reportedly in the process of stopping the use of Claude and switching to other services. J2 Ventures said corporations with large defense contracts strictly comply with Ministry of National Defense requirements.
Reuters reported that Palantir, which holds contracts worth more than $1 billion with the Ministry of National Defense, has likewise faced a situation in which it must replace Claude on its platform. Palantir's "Maven Smart System" is known to be built on Anthropic's "Claude Code."
Anthropic has signaled legal action over the Ministry of National Defense designation, but it has not launched a full response because formal procedures, such as publication in the Federal Register, have not yet been completed.
Major Silicon Valley corporations have voiced concern over the moves. The Information Technology Industry Council (ITIC), which includes Nvidia, Amazon and Apple, sent a letter to Defense Secretary Pete Hegseth saying emergency powers such as supply chain risk designations should be limited to genuine emergencies, such as foreign adversaries. It also warned that if the designation is applied more broadly, U.S. corporations that contract with the federal government could find it harder to access top-tier technology.
The letter did not mention Anthropic by name, but Anthropic CEO Dario Amodei was said to have discussed the matter with Amazon, an investor in the corporation. Other investors are also reportedly in contact with members of the Trump administration.
Meanwhile, Amodei was reported to have said in an internal memo sent to employees on the 27th of last month that the designation is not unrelated to political considerations. He also strongly criticized OpenAI's contract with the Ministry of National Defense and reportedly questioned the effectiveness of the safeguards included in that contract.