Defense tech companies are dropping Claude after Pentagon's Anthropic blacklist

In a swift reaction to the Trump administration's recent decision to blacklist Anthropic Technologies, defense contractors are instructing their employees to discontinue the use of the AI model known as Claude. This move comes in light of the designation of Anthropic's technology as a supply chain risk, prompting companies deeply involved in defense contracts to seek alternative AI solutions. Alexander Harstrick, managing partner at J2 Ventures, noted that many of his firm's ten portfolio companies collaborating with the Department of Defense (DoD) are already taking steps to replace Claude. In an email to CNBC, he stated that these companies are prioritizing compliance with the stringent regulations set forth by the DoD. Lockheed Martin and other defense contractors are expected to remove Anthropic's technology from their operations, as reported by Reuters. This marks a significant downturn for Anthropic, which relies on enterprise customers for approximately 80% of its revenue. CEO Dario Amodei revealed earlier this year that the company entered the DoD's ecosystem through a collaboration with Palantir, making Claude the first AI model integrated into the government's classified networks via a $200 million contract. Despite the Pentagon's announcement, Claude continues to support military operations in regions like Iran, illustrating the complexities surrounding the situation. Defense Secretary Pete Hegseth's declaration that contractors engaged with the U.S. military must cease all business with Anthropic highlights the urgency of the matter. This directive followed Anthropic's refusal to comply with government requests concerning the use of its AI systems in autonomous weaponry or domestic surveillance. While Anthropic has the option to appeal the government's decision, the company has yet to take formal legal action, largely voicing its concerns through social media. In a recent blog post, Anthropic asserted that the restrictions imposed by Hegseth may lack legal authority and cannot hinder contractors from using Claude for non-defense purposes. Defense executives, remaining cautious, have already begun transitioning their teams away from Claude to avoid any potential fallout. Some have set a timeline of one to two weeks to switch to alternative models, including open-source options, in anticipation of the impending deadline. Harstrick emphasized that the decision to move away from Claude is not indicative of any shortcomings in the AI's capabilities but rather a precautionary measure amid regulatory uncertainty. Meanwhile, OpenAI's CEO Sam Altman announced on social media that his company has successfully negotiated terms with the DoD for the use of their AI models, although he later acknowledged that the timing of the announcement was poorly executed. As federal agencies, including the Treasury and Health and Human Services, also pivot away from Claude, industry insiders suggest that the transition may not be as disruptive as it seems. Investors in defense tech believe that companies should avoid over-reliance on any single AI provider, easing the switch to alternative solutions. However, some experts caution that cutting ties with Anthropic could have serious implications. Tara Chklovski, CEO of Technovation, argued that Anthropic's meticulous approach to AI development for military applications sets it apart from potential competitors, warning that any new suppliers may not prioritize safety and efficacy in the same manner. As the situation unfolds, defense tech companies are left to navigate this complex landscape, balancing compliance with the need for advanced AI capabilities.