Key Takeaways
- Alphabet’s Google is currently in negotiations with the Pentagon regarding Gemini AI deployment in secure, classified environments.
- The arrangement would permit military applications of Google’s artificial intelligence for any legally permissible purpose.
- Google seeks contractual language preventing use in mass domestic surveillance operations and fully autonomous weapons systems.
- The restrictions mirror an agreement OpenAI previously secured with the Defense Department.
- Neither Google nor Pentagon officials have publicly acknowledged or discussed the ongoing negotiations.
Alphabet’s Google has entered serious negotiations with the Pentagon concerning the deployment of its Gemini artificial intelligence models within classified military systems, The Information reported, based on information from two sources directly familiar with the discussions.
The Information: Google is in talks with the Pentagon on a deal to use Gemini in classified settings, expanding its defense ties as it pushes deeper into military AI. The report says the discussions could allow use of $GOOGL AI for all lawful uses, if an agreement is reached. pic.twitter.com/iuU7YCXdWo
— Wall St Engine (@wallstengine) April 16, 2026
The negotiations represent a significant pivot for Google, which previously maintained distance from defense-related agreements. In 2018, the tech giant encountered substantial internal resistance when employees objected to its participation in Project Maven, an AI initiative for military drone operations. Following the controversy, Google withdrew from the program.
The current situation presents a markedly different scenario.
Under the terms being discussed, the Defense Department would gain access to Google’s artificial intelligence technology for various legitimate military applications. However, Google is advocating for specific contractual safeguards that would explicitly prohibit the technology’s use in widespread domestic surveillance programs or autonomous weapons platforms lacking adequate human oversight.
Proposed Restrictions Echo OpenAI Agreement
The safeguards Google is requesting closely resemble provisions OpenAI negotiated with the Pentagon in a deal finalized earlier this year. OpenAI CEO Sam Altman allegedly urged the Defense Department to extend identical contractual terms to all artificial intelligence vendors, seeking to establish uniform standards throughout the sector.
It remains uncertain whether Google’s desired protective clauses will appear in any finalized agreement. Discussions continue, and representatives from both Alphabet and the Pentagon have declined to provide statements.
The Defense Department’s artificial intelligence ambitions are growing rapidly. The current administration has directed military leadership to integrate AI capabilities throughout operations to reduce expenses and enhance both administrative efficiency and tactical decision-making speed.
A successful agreement with Google would strengthen Alphabet’s presence in the government contracting space, where AI companies are increasingly competing for lucrative federal partnerships.
Anthropic Dispute Provides Background
These negotiations unfold amid an unrelated conflict between the Pentagon and Anthropic. Earlier this year, Anthropic declined Pentagon requests to relax safety protocols governing its AI systems. In response, the Defense Department designated Anthropic as a supply chain security concern, threatening the company’s existing government business.
That confrontation exposed fundamental disagreements between AI safety standards and military requirements for flexible, unrestricted technological deployment.
Google’s emphasis on preemptive contractual language regarding surveillance and weaponry indicates an attempt to address these concerns proactively, establishing boundaries early rather than encountering the challenges Anthropic experienced.
President Trump has also directed the Defense Department to revert to its historical name, the Department of War, though this modification requires Congressional approval and has not been implemented.
Alphabet’s Google (GOOGL) traded marginally lower, declining 0.08% at the time of this report.
