Altman vowed limits on military AI, then signed a Pentagon deal
Mar 6th 2026
Sam Altman said OpenAI would draw the same red lines as rival Anthropic on mass surveillance and autonomous weapons, but within hours OpenAI signed a roughly $200 million Department of Defense deal that did not include those contractual limits, sparking industry backlash and changing market dynamics.
- Sam Altman publicly said OpenAI would not support mass surveillance or autonomous lethal weapons, but OpenAI signed a Department of Defense agreement that did not include those contractual protections, according to reporting.
- The DoD deal is reported to be worth about $200 million and gives the military broad rights to use OpenAI technology within legal limits.
- Anthropic lost government contracts after declining language that would have allowed the military wide latitude to use its models, according to multiple news reports.
- Market data and industry reports show Anthropic's enterprise share rose to about 40 percent while OpenAI's enterprise share fell to roughly 27 percent in 2025.
- The deal prompted public and internal pushback, with employees and online communities criticizing OpenAI and Altman saying publicly he would try to improve the agreement while internal messages suggested limited staff input.