Google Expands Pentagon AI Access After Anthropic Refuses Defense Contract

Apr 29, 2026

Google signed a contract granting the U.S. Department of Defense access to its AI models for classified networks, allowing essentially all lawful uses despite including non-binding language about not intending the technology for domestic mass surveillance or autonomous weapons. The deal came after Anthropic refused to grant the Pentagon unrestricted AI access and proposed binding guardrails, leading the Department of Defense to brand Anthropic a supply-chain risk—a designation normally reserved for foreign adversaries that is now subject to a judge-granted injunction while a lawsuit proceeds. Over 950 Google employees signed an open letter protesting the contract and asking Google to adopt Anthropic-style restrictions, while OpenAI and xAI also signed Pentagon deals with similar non-binding language. The controversy highlights tensions between AI companies, employees, and military agencies over appropriate use restrictions for advanced AI systems in defense applications.

Google's Pentagon AI Deal and Contract Terms

▪The provisions in Google's U.S. Department of Defense contract are unclear whether they are legally binding or enforceable.
▪Google's deal with the U.S. Department of Defense essentially allows all lawful uses of Google's AI.
▪Google granted the U.S. Department of Defense access to Google's AI for classified networks.
▪Google's agreement with the U.S. Department of Defense includes language saying Google doesn't intend for its AI to be used for domestic mass surveillance.
▪950 Google employees signed an open letter asking Google to follow Anthropic's lead and not sell AI to the U.S. Department of Defense without similar guardrails.

Anthropic's Refusal and Legal Battle with the DoD

▪The supply-chain risk designation is normally reserved for foreign adversaries.
▪The U.S. Department of Defense wanted unrestricted use of AI from Anthropic.
▪A judge granted Anthropic an injunction against the supply-chain risk designation while the case proceeds.
▪Anthropic refused to grant the U.S. Department of Defense unrestricted access to Anthropic's AI.
▪Anthropic wanted guardrails to prevent Anthropic's AI from being used for domestic mass surveillance.
▪The U.S. Department of Defense branded Anthropic a supply-chain risk after Anthropic refused certain use cases.
▪The judge granted Anthropic an injunction last month as of April 28, 2026.
▪Anthropic and the U.S. Department of Defense are embroiled in a lawsuit.

Competitive Response from Other AI Companies

▪Google marks the third AI company to sign a deal with the U.S. Department of Defense following Anthropic's refusal.
▪OpenAI immediately signed a deal with the U.S. Department of Defense after Anthropic refused.