Google Said Yes When Anthropic Said No

When Anthropic told the Pentagon “no,” it got blacklisted. When Google told the Pentagon “yes,” it got a classified contract worth billions and an employee revolt. The story of how two AI giants took opposite paths on military AI reveals a fault line that will define the industry for years to come.

The Deal That Divided Silicon Valley

Google Goes Classified

On April 28, 2026, Google signed a deal granting the U.S. Department of Defense access to its Gemini AI models on classified networks, essentially permitting “any lawful use.” According to Reuters, the deal moves Google’s AI from lower-sensitivity government settings into fully classified military workloads.

Pentagon chief digital and artificial intelligence officer Cameron Stanley confirmed the expansion to CNBC, stating: “Overreliance on one vendor is never a good thing.” The remark was a pointed reference to the DoD’s previous dependence on Anthropic’s Claude models before the relationship collapsed.

Anthropic’s Principled Stand

The backstory is critical. Earlier this year, Anthropic CEO Dario Amodei refused to grant the Pentagon unrestricted access to Claude. The company insisted on guardrails preventing two specific use cases: domestic mass surveillance of American citizens and autonomous weapons systems. Defense Secretary Pete Hegseth responded by invoking the Defense Production Act and designating Anthropic a “supply chain risk” – a label normally reserved for foreign adversaries like Huawei.

The designation effectively froze Anthropic out of all federal contracts. A federal judge later granted Anthropic an injunction against the designation while the lawsuit proceeds, but the damage to the company’s government business was immediate.

The Employee Revolt Inside Google

580+ Signatures and Counting

Before the ink dried, more than 580 Google employees – including researchers from DeepMind – signed an open letter to CEO Sundar Pichai urging the company to follow Anthropic’s lead and refuse classified military AI work without guardrails. “We feel that our proximity to this technology creates a responsibility to highlight and prevent its most unethical and dangerous uses,” the letter read, according to The Guardian.

The internal backlash echoes the 2018 Project Maven controversy, when thousands of Google employees protested the company’s work on Pentagon drone-targeting AI. Google eventually withdrew from Project Maven, but in 2026 the political and business calculus has shifted dramatically.

The $40 Billion Paradox

Adding an extraordinary twist, Google is simultaneously investing up to $40 billion in Anthropic itself. The initial $10 billion commitment, with an additional $30 billion contingent on milestones, values Anthropic at $380 billion. Google Cloud is providing 5 gigawatts of computing capacity to power Anthropic’s models.

The paradox is hard to miss: Google is bankrolling the company that refused the Pentagon deal while simultaneously taking that same Pentagon business for itself. As The Next Web noted, Google is funding both sides of the AI ethics debate.

The White House Reversal

Trump Signals a Thaw

In a surprising turn, the Trump administration is now backing off its hardline stance against Anthropic. On April 21, Trump told CNBC that Anthropic was “shaping up” and that a deal with the Pentagon was “possible.” On April 17, the White House held what was described as a “productive” meeting with Anthropic leadership, according to The New York Times.

Most significantly, Axios reported on April 29 that White House officials are drafting executive guidance that would allow federal agencies to bypass Anthropic’s supply chain risk designation and onboard new models – including Claude Mythos, Anthropic’s most powerful model yet, described internally as a “step change in capabilities.”

Why the Reversal?

Multiple factors are driving the shift. First, Anthropic’s Claude models are widely regarded as among the most capable available, and losing access to them puts federal agencies at a competitive disadvantage. Second, the legal battle is not going the Pentagon’s way – the injunction against the supply chain risk designation suggests courts may side with Anthropic. Third, the political optics of blacklisting an American AI company while China’s DeepSeek races ahead create an uncomfortable narrative.

What This Means for the AI Industry

The Google-Pentagon deal, the Anthropic blacklisting, and the White House reversal together mark a turning point in the relationship between AI companies and government power. Three lessons are emerging.

First, saying “no” to the Pentagon carries real costs. Anthropic lost federal contracts, faced a supply chain risk designation, and watched three competitors rush in to fill the gap. But the company’s principled stand also generated enormous public goodwill and may ultimately result in a better deal than blanket compliance would have produced.

Second, employee resistance still matters – but less than it used to. Google’s 580 signatories represent a fraction of the company’s workforce, and unlike 2018, leadership appears unwilling to retreat from military contracts in 2026’s geopolitical climate.

Third, the government needs AI companies more than the reverse. The speed with which the White House moved from blacklisting Anthropic to courting it back suggests that frontier AI capabilities are too strategically important to leave on the table over a policy disagreement.

Related

• Google Expands Pentagon’s Access to Its AI After Anthropic’s Refusal – TechCrunch
• Pentagon AI Chief Confirms DoD’s Expanded Use of Google – CNBC
• Trump Officials Draft Plan to Bring Anthropic Back – Axios
• 580+ Google Employees Urge Pichai to Refuse Pentagon AI Deal – TNW
• Google to Invest Up to $40B in Anthropic – TechCrunch

Sources

1. TechCrunch – Google expands Pentagon’s access to its AI after Anthropic’s refusal (Apr 28, 2026)
2. CNBC – Pentagon AI chief confirms DoD’s expanded use of Google (Apr 28, 2026)
3. Reuters – Google signs classified AI deal with Pentagon (Apr 28, 2026)
4. Axios – Trump officials draft plan to bring Anthropic back (Apr 29, 2026)
5. The Next Web – 580+ Google employees urge Pichai to refuse Pentagon deal (Apr 2026)
6. The Guardian – Google reportedly signs classified AI deal with US Pentagon (Apr 28, 2026)