Google has granted the US Department of Defense access to its AI for classified networks. The deal allows essentially all lawful uses. Google becomes the third major AI company to capitalize on Anthropic's refusal to give the Pentagon unrestricted access following OpenAI and xAI despite nearly 1,000 of its own employees signing an open letter demanding it follow Anthropic's lead instead.
What Google Agreed To
The agreement gives the Pentagon access to Google's AI tools on classified military networks. The Wall Street Journal reported that Google's contract includes language saying the company does not intend for its AI to be used for domestic mass surveillance or autonomous weapons. OpenAI's Pentagon contract contains similar language.
But the operative word is "intend." It is unclear whether such provisions are legally binding or enforceable. The Pentagon wanted unrestricted access. Anthropic refused. Google, OpenAI, and xAI said yes with soft guardrails that may or may not hold up in practice.
Google did not respond to requests for comment.
Anthropic Drew the Line
The deal traces directly back to Anthropic's public stand against the Trump administration earlier this year. The Pentagon wanted AI without restrictions. Anthropic refused to allow its technology for fully autonomous weapons and domestic mass surveillance. It wanted enforceable guardrails. The Pentagon said no.
In response, the Department of Defense branded Anthropic a supply-chain risk a designation normally reserved for foreign adversaries. The label threatened to block Anthropic's technology across the entire federal government. Anthropic sued. A judge granted an injunction against the designation while the case proceeds.
The dispute became a defining moment in the AI industry. It forced every major AI company to declare where it stands on military use. OpenAI signed a Pentagon deal immediately. xAI followed. Now Google has joined them.
950 Employees Objected
Google's decision came despite significant internal opposition. Nearly 950 Google employees signed an open letter asking the company to follow Anthropic's example and refuse to sell AI to the Defense Department without guardrails against surveillance and autonomous weapons.
The internal dissent echoes Google's 2018 Project Maven controversy, when thousands of employees protested the company's work on Pentagon drone imagery analysis. Google ultimately withdrew from that contract. This time, the company appears to have made a different calculation.
The difference may come down to competitive pressure. In 2018, Google could walk away from a military AI contract without losing ground. In 2026, with OpenAI and xAI already inside the Pentagon and Anthropic locked out, refusing the deal would have meant ceding one of the largest potential AI markets to its competitors.
The Palantir Connection
Google's Pentagon deal aligns with the worldview expressed in Palantir's recent manifesto, which argued that AI companies have a moral obligation to support national defense. Palantir, which has long sold surveillance and analytics software to defense and intelligence agencies, criticized companies that refuse military contracts as failing in their civic duty.
The AI industry is now split into two camps. On one side: Anthropic, which drew red lines around military applications and is paying a significant price for it. On the other: Google, OpenAI, xAI, and Palantir, which have accepted Pentagon contracts with varying degrees of stated limitations.
Why It Matters
Google's deal is significant because of the company's scale. Google Cloud serves thousands of enterprises and government agencies worldwide. Its AI models, including Gemini, are embedded in products used by billions of people. Granting the Pentagon access to those tools on classified networks represents a major expansion of military AI capabilities.
The deal also puts Google in an awkward position with Anthropic. Google is simultaneously investing up to $40 billion in Anthropic while competing directly with it for Pentagon contracts. Google is funding the company that refused the military and then taking the deal Anthropic turned down.
For the broader AI industry, the pattern is clear. Anthropic's principled stand cost it access to the Pentagon. Every major competitor moved to fill the gap. Whether Anthropic's approach ultimately proves wise or whether it simply handed its rivals a lucrative market will depend on how the lawsuit, the court of public opinion, and the enterprise market respond.
