OpenAI lands Pentagon classified network deal hours after White House bans Anthropic

OpenAI has secured access to classified Pentagon networks just hours after the White House ordered federal agencies to halt use of Anthropic’s AI tools and designated the company a national security supply-chain risk — handing OpenAI a dominant position in US military AI almost overnight.

Just hours after the White House ordered federal agencies to halt the use of AI tools from rival Anthropic, the Department of Defense confirmed it had reached an agreement with OpenAI to deploy its models across classified Pentagon networks. The timing underscored how quickly geopolitical risk, national security, and AI governance are now colliding.

In a late-night post on X, Sam Altman, chief executive of OpenAI, disclosed the agreement, stating that the company would make its models available within the Pentagon’s “classified network.”

According to Altman, the Defense Department demonstrated “deep respect for safety” and agreed to operate within OpenAI’s existing guardrails—language that has become central to public debate around the OpenAI Defense Contract.

The deal landed during a volatile week for the AI industry. Earlier that same day, Defense Secretary Pete Hegseth formally designated Anthropic as a “Supply-Chain Risk to National Security,” a label most often associated with foreign-linked technology threats.

The designation obliges defense contractors to certify that they are not using Anthropic’s AI models, effectively freezing the company out of US military procurement.

Soon after, Donald Trump issued a directive requiring all federal agencies to immediately stop using Anthropic technology, granting a six-month transition window for departments already dependent on its systems. The abrupt policy shift cleared the runway for the OpenAI Defense Contract to move forward at speed.

Anthropic’s Pentagon Talks Collapse

The reversal was especially striking given Anthropic’s earlier position. The company had been the first AI lab to roll out models across the Pentagon’s classified environment under a $200 million contract signed in July. That relationship unraveled after negotiations stalled over how the technology could be used.

Anthropic reportedly sought binding assurances that its models would not be applied to autonomous weapons systems or domestic mass surveillance programs. Defense officials, however, insisted that any AI deployed under military contracts must remain available for all lawful national defense purposes.

In a public statement, Anthropic said it was “deeply saddened” by the national security designation and confirmed plans to challenge the decision in court. The company warned that the move could establish a precedent that reshapes how American technology firms negotiate with government agencies, particularly as political oversight of AI partnerships intensifies.

That vacuum was quickly filled by the OpenAI Defense Contract, which now positions OpenAI as the Pentagon’s primary AI partner for classified operations.

Safety Promises Under the OpenAI Defense Contra

Altman has emphasized that OpenAI’s internal restrictions were written directly into the OpenAI Defense Contract. According to him, the company explicitly prohibits domestic mass surveillance and requires meaningful human accountability in any decision involving the use of force, including automated weapons systems.

“These principles are non-negotiable,” Altman has said in prior remarks, framing the OpenAI Defense Contract as a controlled deployment rather than an unrestricted handover of AI capabilities.

Defense officials, for their part, have argued that advanced AI is essential for logistics, intelligence analysis, cybersecurity, and threat detection—areas where speed and scale can be decisive. The OpenAI Defense Contract, they say, provides access to cutting-edge models while keeping them inside secure government infrastructure.

Political and Public Backlash Grows

Despite those assurances, the OpenAI Defense Contract has sparked vocal backlash online, particularly among technologists, civil liberties advocates, and crypto-native communities already skeptical of government–AI alliances.

“I just canceled ChatGPT and bought Claude Pro Max,” wrote Christopher Hale on X. “One stands up for the God-given rights of the American people. The other folds to tyrants.”

Other critics have framed the deal as a reversal of OpenAI’s early ethos.

One widely shared post read: “2019 OpenAI: we will never help build weapons or surveillance tools. 2026 OpenAI: Department of War, hold my classified cloud instance.”

Such reactions highlight the reputational risk embedded in the OpenAI Defense Contract, even as the company expands its influence inside government.

A Defining Moment for Military AI

Beyond the immediate controversy, the OpenAI Defense Contract marks a broader turning point. It illustrates how quickly trust, compliance, and political alignment can outweigh technical leadership in determining who supplies AI to the US military.

For the Pentagon, the OpenAI Defense Contract ensures continuity after the abrupt removal of Anthropic. For OpenAI, it cements a powerful—if contentious—role at the heart of US defense infrastructure.

And for the wider AI industry, the episode sends a clear message: national security considerations now sit at the center of commercial AI strategy. Companies unwilling or unable to align with government expectations risk being sidelined overnight.

As legal challenges loom and public scrutiny intensifies, the OpenAI Defense Contract will remain a focal point in debates over ethics, power, and the future of artificial intelligence in warfare. Whether it becomes a template for responsible military AI or a cautionary tale may define the next phase of the global AI race.