Pentagon Pushes Deeper Into Google’s AI, But Warns Against Putting All Eggs in One Basket

The Pentagon is scaling up its use of Google’s Gemini AI model for classified work, marking a significant shift in the Department of Defense’s approach to artificial intelligence. Cameron Stanley, the DOD’s AI chief, confirmed to CNBC that the agency is expanding its relationship with Google roughly two months after cutting ties with Anthropic, which was designated a supply chain risk.

A source familiar with the arrangement—who requested anonymity because the details are not public—said Gemini is now being deployed on classified projects. The move comes as the DOD also works with OpenAI and other vendors to modernize wartime capabilities. Stanley stressed that relying on a single AI provider creates vulnerabilities. “Overreliance on one vendor is never a good thing,” he said. “We’re seeing that, especially in software.”

The Pentagon’s embrace of Google unfolds against a backdrop of legal battles with Anthropic. A federal appeals court recently denied Anthropic’s request to block the DOD’s blacklisting, even as a separate court granted the company a preliminary injunction allowing it to continue working with other government agencies. President Trump has hinted that a deal to bring Anthropic back into the DOD fold is possible.

Stanley said Gemini is already saving the military thousands of man-hours weekly. But the partnership faces internal pushback at Google, where over 700 employees signed a letter urging CEO Sundar Pichai to reject classified workloads, warning the technology could be used in harmful ways.

Stanley dismissed concerns with a homespun analogy: “You don’t cook a Thanksgiving turkey in the microwave. You need the right technology for the right use case.” He cited Anthropic’s recent Mythos rollout as a wakeup call, given its advanced cyber capabilities. The DOD, he said, is “taking this very seriously” to prepare for a wave of AI-enabled threats.

Source: CNBC