OpenAI Sweeps In With Pentagon Deal as Anthropic Faces ‘Supply Chain Risk’ Designation

When the Pentagon moved to designate Anthropic as a supply chain risk late Friday, the AI industry braced for impact. What happened next was swift and unexpected: within hours, OpenAI announced its own deal with the Department of Defense—one that appeared to accomplish what Anthropic had been fighting for, but through a different path entirely.

“We are asking the Department of War to offer these same terms to all AI companies, which in our opinion we think everyone should be willing to accept.” — Sam Altman, OpenAI CEO

The Deal That Changed Everything

OpenAI’s agreement with the Pentagon, announced late Friday, allows the military to use its AI models in classified systems. But the announcement carried a twist: Altman revealed the deal contains the same two limitations Anthropic had been insisting on—restrictions against mass surveillance of Americans and autonomous weapons.

The difference lies in how those limitations are enshrined. While Anthropic sought explicit contractual language spelling out the restrictions, OpenAI agreed the Pentagon could use its technology for “any lawful purpose” while also stating the company “put them into our agreement.” The contract appears to rely on existing U.S. law that prohibits mass surveillance and military policy requiring human judgment over lethal force.

OpenAI went further, announcing the Pentagon agreed to let the company build technical solutions into its models to prevent misuse for mass surveillance or autonomous weapons. The deal also includes a third red line banning “high-stakes automated decisions” such as social credit systems.

The Anthropic Fallout

The supply chain risk designation represents uncharted territory for AI companies. Secretary of War Pete Hegseth interpreted the designation broadly, stating that “effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic.”

The business implications could be catastrophic if Hegseth’s interpretation stands. Many enterprises adopting Claude models also have Pentagon contracts. Legal experts question whether the designation can survive court challenge—former Biden NSC official Peter Harrell noted DoW cannot legally tell contractors “don’t use Anthropic even in your private contracts.”

The funding threat adds another layer of complexity. Anthropic’s recent $30 billion funding round valued the company at $380 billion. But if major investors like Amazon, Google, and Nvidia face pressure to divest due to their own Pentagon relationships, Anthropic’s IPO prospects could be in jeopardy.

“It will take years to resolve in court. And in the meantime, every general counsel at every Fortune 500 company with any Pentagon exposure is going to ask one question: is using Claude worth the risk?” — Shenaka Anslem Perera, Independent Analyst

Legal Challenges and Industry Response

Several legal experts have questioned whether the supply chain risk designation can withstand judicial scrutiny. The statute requires the government to prove risk of sabotage, subversion, or manipulation by an adversary—something Amos Toh of the Brennan Center noted is unclear in Anthropic’s case. The government must also have exhausted less intrusive alternatives and notified Congress, steps that appear to have been skipped.

The situation has drawn unusual allies. Employees at both Google and OpenAI signed an open letter supporting Anthropic CEO Dario Amodei’s position on usage restrictions. Altman himself had previously publicly supported Anthropic’s stance, making Friday’s announcement a notable pivot.

What emerges from this standoff could reshape how AI companies engage with government contracts. If OpenAI’s approach becomes the template, other labs may follow suit. If Anthropic prevails in court, it could establish stronger contractual protections for the entire industry.

For now, the AI industry watches and waits. The Pentagon has its OpenAI deal. Anthropic has its legal battle. And every other AI lab is recalculating what it means to work with the world’s largest customer.

This article was reported by the ArtificialDaily editorial team. For more information, visit Fortune and The New York Times.