Google CEO Sundar Pichai authorized the company to sign a classified Pentagon AI agreement on Friday, moving forward despite an open letter from approximately 600 employees urging him to reject military work entirely. The deal grants Google access to Impact Level 6 and IL7 networks, the government's highest-tier classified environments used for secret intelligence analysis and sensitive national security data.

Why Did Google Employees Object to This Pentagon Deal?

Just days before the Pentagon's announcement, around 600 Google employees sent an open letter to Pichai expressing serious concerns about the military contract. The employees flagged two specific worries: lethal autonomous weapons and mass domestic surveillance. In their letter, they wrote that providing AI technology for these purposes could cause irreparable damage to Google's reputation and role in the world.

The timing of the employee pushback echoes a similar moment in 2018, when Google withdrew from Project Maven, a Pentagon computer vision initiative, after significant internal backlash. This time, however, leadership proceeded with the deal. A Google spokesperson responded by stating the company "remains committed to the consensus that AI should not be used for domestic mass surveillance or autonomous weaponry without appropriate human oversight," and characterized providing API access to its models as "a responsible approach to supporting national security".

What Exactly Did Google and Six Other Companies Agree To?

The Pentagon signed agreements with seven leading artificial intelligence companies on Friday: SpaceX, OpenAI, Google, Nvidia, Reflection, Microsoft, and Amazon Web Services. The War Department stated that "together, the War Department and these strategic partners share the conviction that American leadership in AI is indispensable to national security".

The agreements authorize these companies to deploy AI on the Pentagon's highest-classification networks for what the government calls "lawful operational use." This includes battlefield decision support, data synthesis, situational awareness, and intelligence work. Each company's artificial intelligence models will be integrated directly into the Pentagon's classified IL6 and IL7 environments.

• Scope of Deployment: The contracts cover battlefield decision support, data synthesis, situational awareness, and intelligence analysis on classified military networks
• Safeguard Language: Agreements include language stating AI should not be used for domestic mass surveillance or autonomous weapons without appropriate human oversight
• Pentagon Authority: None of the companies receive veto power over how the Pentagon ultimately deploys their technology in military operations

The Pentagon's GenAI.mil platform, which has been running since December with Google's Gemini model, has already demonstrated significant adoption. Over 1.3 million Department of Defense personnel have used the platform, generating tens of millions of prompts and hundreds of thousands of AI agents in just five months.

How to Understand the Anthropic Situation and Why It Matters

One notable absence from the Pentagon's list of seven companies is Anthropic, the maker of Claude, an advanced AI model. Until recently, Anthropic had been the only AI model cleared for Pentagon classified networks. The company's exclusion reveals a fundamental disagreement over how much control companies should have over military AI deployment.

• Anthropic's Red Lines: The company drew two firm boundaries: no fully autonomous weapons and no mass domestic surveillance, refusing to accept the Pentagon's "any lawful use" language
• Pentagon's Response: The War Department designated Anthropic a "supply chain risk," a designation historically reserved for foreign adversaries and never before applied to a U.S. company
• Legal Battle Status: Anthropic sued to challenge the designation; a federal judge blocked it in March, though an appeals court later declined to fully lift it, leaving the legal battle ongoing

Defense Department Chief Technology Officer Emil Michael confirmed Friday that Anthropic remains blacklisted from the current agreement. However, he separately flagged the company's powerful new cybersecurity model, called Mythos, as a "separate national security moment" that the entire government needs to address. The National Security Agency is reportedly already using Mythos despite the company's official designation as a supply chain risk.

Behind the scenes, the White House appears to be exploring a potential path forward. A draft executive action is reportedly in development that could give government agencies a way to work with Anthropic again, specifically to access the Mythos cybersecurity model, which has demonstrated the ability to find vulnerabilities in well-tested software. President Trump told CNBC it is "possible" a deal could happen, describing Anthropic as "very smart" and potentially of "great use." Anthropic CEO Dario Amodei met with senior White House officials earlier this month, with both sides characterizing the conversation as productive.

What Does This Mean for the Future of AI in Defense?

The Pentagon's announcement sends a clear strategic message: the department is not waiting for consensus on Anthropic. Instead, it is diversifying its AI stack and locking in competitors, making Anthropic's absence increasingly costly. With significant federal funding earmarked for AI and offensive cyber operations under the One Big Beautiful Bill Act, the seven companies now at the table have a substantial head start in securing defense contracts and government resources.

Google's decision to sign despite employee objections reflects a broader tension in the tech industry between worker values and corporate strategy. The company's willingness to move forward, unlike its 2018 withdrawal from Project Maven, suggests that leadership views national security applications as strategically important enough to override internal dissent. For employees concerned about AI ethics and military applications, the outcome represents a significant setback in efforts to shape corporate policy from within.