Google Signed a Classified Pentagon AI Deal. 600 Employees Are in Revolt — and Losing.
Google has signed a classified contract giving the U.S. Department of Defense access to its Gemini AI models for "any lawful governmental purpose" — and the 600-plus employees opposing it are discovering they have far less power than their 2018 counterparts did.
What Happened
According to Fortune's May 4 analysis, the contract grants Gemini access across approximately three million DoD personnel. The classified nature of the deal means specifics — including what tasks Gemini will be used for — are not publicly disclosed. CEO Sundar Pichai approved the deployment without publicly discussing usage restrictions.
More than 600 employees at Google DeepMind and Google Cloud signed an open letter to Pichai opposing the deal, citing concerns about deploying AI for military purposes without clear ethical guardrails.
The 2018 Comparison That Explains the Stakes
In 2018, employee dissent killed Project Maven — a DoD contract for AI-assisted drone imagery analysis. Roughly 4,000 employees petitioned. Google did not renew. That outcome created a durable belief inside Google that organized internal opposition could constrain military AI decisions.
That leverage is gone. The classified defense AI market has grown dramatically since Maven. Microsoft Azure, Amazon Web Services, and Palantir have all signed major DoD AI contracts in the intervening years, normalizing the space and establishing it as a standard line of business for cloud providers. By 2026, Google entering defense AI is catching up to competitors, not pioneering new ethical territory — which changes the calculus for executives weighing employee sentiment against competitive positioning.