Google Signs Classified Pentagon AI Deal Over Internal Revolt

Google moved forward with a Defense Department contract for classified AI work despite an open letter from over 580 employees demanding Pichai refuse — formalizing a values reversal Project Maven previewed in 2018.

Google Signs Classified Pentagon AI Deal Over Internal Revolt

Google has signed a Pentagon contract to provide AI for classified military work, the company confirmed this week, finalizing a deal that more than 580 employees publicly opposed in an open letter to CEO Sundar Pichai. The contract effectively closes the chapter Google opened in 2018, when employee pressure killed Project Maven and the company published its "AI Principles" pledging not to build weapons.

From Maven Walkout to Cleared-Workforce Bid

The 2018 Maven cancellation cost Google a foothold inside DoD that Microsoft, Amazon, and Palantir have spent eight years entrenching. The classified contract changes that — putting Google's models, infrastructure, and a security-cleared engineering staff inside the same procurement perimeter as JEDI's successor cloud and the IL6 workloads that have historically been Microsoft Azure's domain. Decrypt reports the work covers classified analysis pipelines but stops short of weapons-targeting systems.

The Letter Is the New Normal — and It Doesn't Matter

The 580-signature letter argued that classified AI work for the Pentagon is incompatible with Google's stated values and exposes employees to liability for downstream uses they cannot audit. Leadership disagreed in writing. The shift signals that the post-ChatGPT competitive logic — AI revenue must scale to justify the $40B+ Google has committed to compute infrastructure — has overwhelmed the cultural veto employees held over military contracting between 2018 and 2024.

BlockAI News' View

This isn't just a Google story. Anthropic took FedRAMP Moderate this month; OpenAI just landed on AWS GovCloud; Meta is bidding for similar work. The frontier labs are converging on a single posture: no major US lab will sit out classified federal contracting. The 2018-2024 abstinence window is closing, and the hiring pipeline for AI-cleared engineers is now the most important talent market most outsiders aren't watching.

Sources

https://decrypt.co/365920/google-signs-ai-deal-pentagon-classified-work-employees-object

How we report: This article cites primary sources, regulatory filings, and on-chain data where available. BlockAI News uses AI tools to assist with research and first-draft generation; every article is reviewed and edited by a human editor before publication. Read our full How We Report page, Editorial Policy, AI Use Policy, and Corrections Policy.

Keep Reading

Codex Security Blueprint: How OpenAI Cages Its Autonomous Coding Agents

Codex Security Blueprint: How OpenAI Cages Its Autonomous Coding Agents

TL;DR

  • OpenAI published its production Codex security architecture on May 8, 2026, detailing sandbox modes, network egress rules, and telemetry for enterprise deployments.
  • Codex Security scanned 1.2 million commits in 30 days, surfacing 792 critical and 10,561 high-severity findings with a false-positive rate below 6%.
  • The blueprint targets enterprise security teams and CISOs who must govern agents writing, reviewing, and merging code autonomously at scale.

On May 8, 2026, the team behind Codex — now boasting more than 2 million weekly active users

Read full story →

Stay Ahead of the Market

Daily AI & crypto briefings — straight to your inbox, your phone, and your timeline.