US Bans Anthropic, Taps OpenAI for Pentagon AI
The Trump administration has banned all federal agencies from using AI from Anthropic, citing “supply chain risk.” Just hours later, rival OpenAI signed a major contract to supply the Pentagon with its advanced models, cementing its position as the US government's preferred AI vendor in a dramatic reshaping of the defense AI landscape.
The standoff culminated after months of negotiations over military guardrails for Anthropic's flagship model, Claude. Anthropic insisted on "red lines" preventing its AI from being used for mass domestic surveillance or in fully autonomous weapons systems that operate without human oversight. The Pentagon argued these constraints were too restrictive for operational realities. Defense Secretary Pete Hegseth framed the issue as needing "full, unrestricted access" for any "lawful use," a stance that ultimately led to the impasse. The "supply chain risk" designation applied to Anthropic is an unprecedented move, historically reserved for foreign adversaries, not American tech companies. Anthropic has vowed to challenge the designation in court, calling it "legally unsound." The OpenAI agreement, announced shortly after the Anthropic ban, includes contractual "red lines" similar to those Anthropic advocated for: prohibitions on mass domestic surveillance, directing autonomous weapons, and high-stakes automated decisions. OpenAI states its deal has stronger, multi-layered protections, including retaining full control over its safety systems and having cleared OpenAI personnel involved in deployment. This conflict highlights a growing tension between Silicon Valley's safety-focused culture and the military's push for rapid AI integration. The Department of Defense's AI spending has surged, growing from $269 million to $4.323 billion and now accounting for 95% of all federal AI funding. This reliance on private tech firms for critical defense capabilities sets the stage for future clashes over ethics and control.
