Story perspectives

Pentagon Signs AI Deals with Tech Giants, Excludes Anthropic

5/1/2026

Summary

Pentagon Signs AI

The U.S. Department of Defense signed AI agreements with OpenAI, Google, Microsoft, Amazon Web Services, Nvidia, SpaceX, and Reflection for classified networks.
The contracts allow AI to run at Impact Level 6 and 7 while the Pentagon’s multi-vendor strategy, with SpaceX’s xAI acquisition as a defense-stack provider, seeks to avoid lock-in.
The Pentagon excluded Anthropic after it refused to relax safety guardrails on weapons and surveillance, deeming it a supply-chain risk.

Summary

Google Employees Protest

Over 600 Google employees signed a letter urging CEO Sundar Pichai to block the Pentagon AI contract, citing “inhumane” applications.
The request references the 2018 Project Maven protests, when Google ended a drone-analysis contract.
Alphabet removed the AI-principles clause barring weapons and surveillance work in early 2024.
After the deal’s disclosure, Alphabet shares rose about 1.7% while the contract runs on air-gapped Google Distributed Cloud TPU systems.

History

Historical Context

Project Maven, launched in 2017 to add machine-learning to U.S. intelligence, drew a $9 million contract with Google that sparked a 2018 walkout by more than 600 Alphabet engineers. After Google withdrew, the Pentagon broadened its AI supply chain, signing 2024 agreements with OpenAI, Microsoft, Amazon Web Services, Nvidia, SpaceX and Reflection while keeping Anthropic out.

1 of 3

Pentagon Signs AI

The U.S. Department of Defense signed AI agreements with OpenAI, Google, Microsoft, Amazon Web Services, Nvidia, SpaceX, and Reflection for classified networks.
The contracts allow AI to run at Impact Level 6 and 7 while the Pentagon’s multi-vendor strategy, with SpaceX’s xAI acquisition as a defense-stack provider, seeks to avoid lock-in.
The Pentagon excluded Anthropic after it refused to relax safety guardrails on weapons and surveillance, deeming it a supply-chain risk.

1 / 3