Earlier this week, Google officially signed an approximately $200 million AI contract with the U.S. Department of Defense, giving the Pentagon direct API access to its commercial models. On the same day, more than 600 Google employees — including researchers from DeepMind — sent a letter of protest to CEO Sundar Pichai, urging the company to reject the deal. Management moved forward regardless.
Contract Details: API Access, No Custom Development
Bloomberg confirmed the contract’s value and scope: the Pentagon may use Google’s AI systems for “any lawful government purpose.” Google’s spokesperson stressed this is an API-access agreement, not a custom development project. “We believe providing API access to our commercial models under industry-standard terms, including execution on Google Cloud infrastructure, is a responsible way to support national security,” the spokesperson said in a statement.
“We remain committed to the public-private consensus that AI should not be used for large-scale domestic surveillance or for autonomous weapons without appropriate human oversight.”
The contract does not involve building bespoke tools for the military; instead, it allows the Pentagon to call Google’s AI models directly for analysis. Google also reiterated its stance against deploying AI for mass surveillance or fully autonomous weapons.
A Gap Left by Anthropic’s Departure
The timing of this contract coincides with the Pentagon’s falling-out with Anthropic earlier this year. The Department of Defense is now pushing to remove Anthropic and its Claude AI tools from its supply chain, actively seeking a new big-tech partner. Google’s entry is widely seen as filling that void.
For Google, this marks a strategic shift. In 2018, employee protests forced the company to withdraw from Project Maven, a military program using AI to analyze drone footage, which Bloomberg called “the heart of modern American warfare.” Google then pledged not to build weapons or harmful technologies. But in 2025, the company quietly removed some restrictive language from its AI ethics policy. This new military contract is the first major outcome of that policy relaxation.
Notably, Google also quietly withdrew from a separate $100 million Pentagon competition to develop voice-controlled autonomous drone swarms. Internal documents seen by Bloomberg cite an ethics-review decision; Google publicly cited “resource constraints” — a pattern of ambiguity in its military AI dealings.
600 Signatures: Can Protest Change Anything?
On the day the contract was signed, more than 600 Google employees, including DeepMind researchers, delivered an open letter to Pichai. “We are Google employees deeply concerned about the ongoing negotiations between Google and the U.S. Department of Defense,” the letter read. “As people working on AI, we know these systems can concentrate power, and they do make mistakes.” The letter focused on two fears: the risk of AI errors in military contexts and the irreversible consequences of power concentration. The demand was clear: reject the contract. But management pressed ahead.
This is the second large-scale employee protest against Google’s military AI work — and the outcome is the opposite of 2018. Back then, protests drove Google out of Project Maven. This time, the letter and the contract arrived almost simultaneously, and 600 signatures changed nothing.