US Agencies Phase Out Anthropic AI
Several U.S. government departments, including State, Treasury, and HHS, are phasing out their use of Anthropic's AI products. The shift to OpenAI and other providers follows a new White House directive, signaling a consolidation of preferred AI vendors within the federal government.
The rift with Anthropic stems from the company's refusal to remove "red lines" in its contract that would prohibit the U.S. military from using its AI for mass surveillance of American citizens or for fully autonomous weapons systems without human oversight. The Department of Defense (DoD) insisted on terms allowing for "all lawful purposes," leading to the impasse and President Trump's directive for all federal agencies to cease using Anthropic's technology. In a swift realignment, OpenAI announced a deal with the Defense Department to provide its AI models for use on classified systems just hours after the President's order against Anthropic. This move followed months of contentious negotiations between Anthropic and the Pentagon, which had previously awarded a $200 million contract to the company in July 2025 to develop AI capabilities for national security. The White House's move is part of a broader strategy to streamline AI procurement across the federal government. The General Services Administration's (GSA) "OneGov" initiative aims to consolidate purchasing, reduce duplicative contracts, and negotiate better terms with AI vendors as a single entity. Prior to the fallout, the DoD had awarded contracts of up to $200 million each to Anthropic, Google, OpenAI, and xAI, indicating an initial multi-vendor approach to fostering AI innovation. This vendor shift has significant implications for the use of AI in defense and robotics. The Pentagon is actively pursuing autonomous systems, including AI-piloted fighter jets and drone swarms. OpenAI's technology is already being used in a competition to develop voice-controlled drone swarm software for the military, translating a commander's verbal orders into digital instructions for unmanned systems. The focus on AI for autonomous systems extends to ground operations as well. The U.S. Army has initiated the "Generative Unwanted Activity Recognition and Defense" (GUARD) project, a $6.3 million contract to develop software that can identify and analyze unpredictable behavior in AI-enabled autonomous systems, aiming to create comprehensive risk profiles before deployment. For those in robotics, OpenAI's deepening relationship with the military coincides with its renewed focus on embodied AI. The company is reportedly reviving its robotics division with an emphasis on developing humanoid robots. This signals a strategic push to give its AI models physical interaction capabilities, which could accelerate advancements in both commercial and military robotics applications. Defense Secretary Pete Hegseth has designated Anthropic a "supply chain risk to national security," a move that could prevent any company that contracts with the military from also doing business with Anthropic. Anthropic has vowed to challenge this designation in court, calling it a "dangerous precedent." Meanwhile, other AI leaders like Google and Elon Musk's xAI also have DoD contracts, suggesting the government will continue to work with multiple major players in the AI field.
