Sam Altman has stepped into the growing clash between Anthropic and the Pentagon — as OpenAI pushes forward with its own military deal.
Speaking on CNBC, Sam Altman said the US government “needs AI partners” and that tech companies should work with the military. However, he added that he does not believe companies should be forced to cooperate under pressure.
The dispute centers on Anthropic and its negotiations with the US Department of Defense. Anthropic CEO Dario Amodei has said the company will not cross two “red lines”: mass domestic surveillance and fully autonomous weapons.
Anthropic claims the Pentagon has not made enough progress in restricting how its AI model, Claude, would be used. Reports say officials have warned Anthropic it could be blacklisted from future contracts if it refuses to agree.
A senior Pentagon official has reportedly suggested the government could use the Defense Production Act to compel cooperation. Altman said he does not think that step is appropriate, but stressed that AI firms should support the government if legal safeguards are in place.
Meanwhile, OpenAI is pursuing its own agreement with the Defense Department. According to reports, the deal would allow its models to operate in classified environments while staying aligned with company principles.
Competition is heating up among AI firms. OpenAI, Anthropic, and Google have been cleared to work with government information, while xAI’s Grok model has been approved to handle classified Pentagon data.
OpenAI’s stance on military work has shifted in recent years. In 2024, it removed language from its usage policies that had banned military and weapons-related applications. The company also added former National Security Agency director Paul Nakasone to its board.
As AI becomes central to national defense, the debate highlights a growing tension between innovation, ethics, and government power.
