Pentagon signs AI deals with OpenAI, Google, Microsoft, and Amazon for classified networks, leaves out Anthropic

The Pentagon is opening its doors wider to artificial intelligence—and closing one, at least for now. On Friday, the Department of Defense said it has reached agreements with a group of major AI providers to bring their tools into its most sensitive systems. The list includes OpenAI, Google, Microsoft, Amazon Web Services, NVIDIA, SpaceX, and Reflection. One name is missing: Anthropic.

“The War Department has entered into agreements with eight of the world’s leading frontier artificial intelligence companies, SpaceX, OpenAI, Google, NVIDIA, Reflection, Microsoft, Amazon Web Services, and Oracle to deploy their advanced AI capabilities on the Department’s classified networks for lawful operational use. These agreements accelerate the transformation toward establishing the United States military as an AI-first fighting force and will strengthen our warfighters’ ability to maintain decision superiority across all domains of warfare,” The Pentagon said in a news release.

The timing stands out. The report lands just two months after Anthropic pushed back against the Pentagon over restrictions on how its AI tools could be used.

The move pushes more AI capability into the Pentagon’s classified networks, known as Impact Levels 6 and 7, where sensitive military operations are handled. For the Defense Department, the goal is simple: give troops broader access to AI tools used in planning, logistics, and targeting, and speed up decision-making across large-scale operations.

At the same time, the decision signals a clear break with Anthropic. Earlier this year, the Pentagon labeled the startup a supply-chain risk, blocking its tools from being used across the department and by contractors. That stance has been difficult for some inside the system to accept. Pentagon staff, former officials, and contractors told Reuters they have grown used to Anthropic’s tools and see them as stronger than many alternatives. Orders are now in place to phase them out over the next six months.

Inside Washington, the shift is about control as much as it is about capability. By bringing in multiple vendors, the Pentagon is trying to avoid dependence on a single provider. In its own words, the expansion will help prevent “vendor lock”—a sign of how central AI has become to military infrastructure.

The scale is already significant. The Pentagon’s internal platform, GenAI.mil, has reached more than 1.3 million Defense Department personnel within five months of launch. That kind of adoption is rare for enterprise software, let alone within a classified environment.

There are signs that the door is not fully closed on Anthropic. Defense Department Chief Technology Officer Emil Michael said the company still presents a supply-chain risk, but pointed to its Mythos model as a separate issue—one that has raised alarms across both government and industry for its potential to accelerate cyberattacks. Access to Mythos has been extended to select organizations to help strengthen defenses, though it’s unclear whether the Pentagon is among them.

The political tone has shifted slightly. President Donald Trump said last week that Anthropic was “shaping up” in the eyes of his administration, hinting that the current standoff may not last forever.

For now, the Pentagon is moving ahead with a broader bench of AI partners. That creates new opportunities for smaller defense tech startups seeking to break into government contracts, a space long dominated by a handful of incumbents. The message is clear: the military wants more options, not fewer, as AI becomes embedded in its operations.