Google in Talks With Pentagon for Confidential AI Deal, Rebuilding Military Ties

(Bloomberg) -- Google is in discussions with the Pentagon for a confidential artificial intelligence agreement, a move that signals a significant effort by the Alphabet Inc. unit to rebuild its relationship with the U.S. military after internal protests derailed a previous contract.

The talks, reported on April 16, 2026, represent a potential reversal of the company's 2018 decision to exit Project Maven, an AI initiative for analyzing drone footage, following widespread employee opposition. "Their Xeon roadmap gives us confidence that we can continue to meet the growing performance and efficiency demands of our workloads," Amin Vahdat, Google's chief technologist for AI infrastructure, said in a recent statement regarding a separate partnership, highlighting the company's focus on scalable and efficient AI systems.

A renewed partnership would place Google in a stronger position to compete for a slice of the Pentagon's massive technology budget, which includes billions for cloud computing and AI modernization. The discussions come as Google also deepens its collaboration with chipmaker Intel Corp., committing to using its latest Xeon 6 CPUs for AI workloads, even as it develops its own custom processors like the Arm-based Axion CPU and Tensor Processing Units (TPUs).

A confirmed Pentagon contract would likely be bullish for Google's stock (GOOGL), as it would open a significant, long-term revenue stream from defense spending. The deal could unlock access to the government cloud market, estimated to be worth tens of billions, and enhance Google’s competitive footing against rivals like Microsoft and Amazon, who have established strong ties with the military.

From Canceled Contracts to Renewed Collaboration

Google's relationship with the U.S. military has been fraught since 2018 when thousands of employees signed a petition protesting the company's involvement in Project Maven. The internal backlash led Google to abandon the project and publish a set of ethical principles for its AI, which included a pledge not to develop AI for use in weapons. This stance put Google at a disadvantage compared to competitors like Microsoft, which secured the $10 billion JEDI cloud contract with the Pentagon.

The current discussions suggest a strategic shift within Google, potentially driven by the immense market opportunity in government AI contracts. A successful agreement would demonstrate Google's ability to navigate the ethical complexities of military work while capitalizing on its advanced AI research and infrastructure.

A Multi-Pronged AI Hardware Strategy

The Pentagon discussions coincide with Google's broader strategy of diversifying its AI infrastructure. While the company has long relied on Intel processors, it recently expanded this partnership to include Intel's newest Xeon 6 CPUs for AI training and inference. This move is notable as the industry has been dominated by Nvidia Corp.'s GPUs for AI tasks.

Simultaneously, Google continues to invest heavily in its own custom silicon. For over a decade, it has developed its Tensor Processor Unit (TPU) specifically for AI workloads. In 2024, Google also introduced Axion, its first custom Arm-based CPU, signaling a desire to control its hardware stack and optimize performance from the chip level up. This multi-pronged approach—using chips from partners like Intel, its own TPUs, and custom CPUs—provides Google with a balanced and resilient infrastructure to handle the diverse demands of large-scale AI, a capability that would be critical for servicing a major government contract.

This article is for informational purposes only and does not constitute investment advice.