In 2018, Google employees staged one of the most dramatic revolts in Silicon Valley history. Thousands signed a letter demanding the company drop Project Maven, a Pentagon drone-targeting program. Google listened. It pulled out, drew a bright line between itself and the military-industrial complex, and walked away from defense AI work.

That line just got erased.

According to The Information, confirmed by Reuters and Newsweek, Google is now in active negotiations with the Department of Defense to deploy its Gemini AI models in classified national security settings. The company reportedly wants a deal similar to the one OpenAI already secured with the Pentagon, but with contractual prohibitions on mass domestic surveillance and autonomous lethal weapons.

Read that again: Google is asking the Pentagon to let Gemini handle classified work. The same Google that once reminded everyone it had a "don't be evil" motto.

Why Now?

The timing tells the whole story. In February, the Trump administration ordered all federal agencies to stop using Anthropic products. That blacklisting created a vacuum. OpenAI moved in fast, securing Pentagon contracts and positioning itself as the government's preferred AI vendor. Now Google wants a seat at that table too.

But Google faces a massive catch-up problem. Amazon and Microsoft have been embedded in classified defense work for years. Their cloud systems are already trusted, tested, and integrated into Pentagon operations. Google's cloud services have never touched classified environments. Starting from zero against two entrenched players is brutal, no matter how good Gemini is.

A Pentagon official told Newsweek: "The Pentagon will continue to rapidly deploy frontier AI capabilities to the warfighter through strong industry partnerships across all classification levels."

The Guardrails Gambit

Google is trying to learn from the backlash its competitors faced. The proposed contract language explicitly prohibits Gemini from being used for mass domestic surveillance or autonomous weapons without human oversight. It is a clever positioning move: we will do defense work, but with ethical boundaries built into the contract itself.

Whether those boundaries survive contact with actual military procurement remains to be seen. Contract language is paper. Operational reality is something else entirely.

What This Actually Means

Here is the uncomfortable truth: the era of Big Tech keeping its distance from the Pentagon is over. Not winding down. Over. Every major AI company is now either already working with the military or actively trying to.

OpenAI has its Pentagon deal. Anthropic was in until it got politically blacklisted. Microsoft and Amazon have been there for years. And now Google, the company that once represented the moral high ground of Silicon Valley, wants in on classified work.

The defense AI market is becoming the most consequential contract race in technology. Whoever gets embedded in Pentagon workflows will have a structural advantage for decades. Google knows this. Its 2018 conscience was a luxury of a less competitive era.

In 2026, with AI becoming a genuine national security tool and hundreds of billions of dollars in government contracts at stake, principles have a price tag. Google just told us what theirs was.

There is also a complication Google cannot ignore. A report released this week found that AI search powered by Gemini returned roughly 9% incorrect results. For consumer search, that is a headache. For classified military intelligence, it could be catastrophic. The Pentagon will not miss that detail.

First reported by The Information. Confirmed by Reuters and Newsweek. Google and the Pentagon did not respond to requests for comment at time of publication.