OpenAI Wins Pentagon AI Contract Hours After Anthropic Ban

OpenAI has reached an agreement with the Pentagon, according to CNBC, in a move that comes hours after President Donald Trump ordered U.S. agencies to stop using rival AI company Anthropic for government work.

The deal places OpenAI in a newly elevated position with the Defense Department at a moment of heightened scrutiny over how advanced AI systems can be used by the military and the federal government. The timing also follows a high-profile dispute involving Anthropic and the Pentagon that led to the administration’s ban.

Axios reported that Trump moved to blacklist Anthropic AI from all government work. NPR and Houston Public Media reported that OpenAI said it shares Anthropic’s “red lines” over military AI use, signaling that OpenAI is attempting to draw clear boundaries even as it expands its government footprint.

The New York Times reported that Trump ordered U.S. agencies to stop using Anthropic AI technology after a Pentagon standoff. AOL.com separately reported the same action, describing a clash with the Pentagon that preceded the order. Together, the reports underscore a rapid shift in which AI vendors are deemed acceptable for federal use, and how quickly access to government contracts can be granted or cut off.

This development matters because the Pentagon is a major customer with significant influence over standards for security, procurement, and deployment. An OpenAI agreement with the Defense Department suggests the government is moving forward with at least one leading AI provider even as it restricts another. That creates immediate operational consequences for agencies and contractors already using or evaluating AI tools, and it raises the stakes for how companies articulate and enforce policies around military applications.

It also matters for the broader AI industry, where public commitments about responsible use are increasingly being tested against real-world demands from governments. OpenAI’s statement that it shares Anthropic’s “red lines,” as reported by NPR, points to a growing emphasis on defining limits. At the same time, the administration’s decision to ban Anthropic indicates those limits and the relationships behind them can become politically and operationally consequential.

Next, federal agencies will need to implement the president’s order regarding Anthropic, including transitioning work where Anthropic tools were in use. The Pentagon and OpenAI will face questions about what the agreement covers, how it will be executed, and how safeguards will be applied in practice. Companies across the sector are likely to watch closely for further directives that could reshape which AI systems are permitted across the government.

For now, the U.S. government has signaled a sharper divide between approved and barred AI providers, and OpenAI has secured a major foothold with the Pentagon as that divide takes effect.