CoreWeave and Leidos to Build an AI Cloud Inside Classified US Intelligence Facilities

CoreWeave and defense contractor Leidos announced a strategic partnership on July 30, 2026. CoreWeave's specialized GPU platform will be deployed inside classified facilities of US intelligence agencies and military organizations. For a commercial GPU cloud, this is the first move into classified computing, the government-grade sector of secure computation.

Why an AI cloud needs a security clearance

CoreWeave, one of the largest specialized GPU cloud providers, makes its money by renting compute to leading AI companies. Its customers include OpenAI and Microsoft, both of which use CoreWeave infrastructure to train and run inference for large language models. The business model is proven in the commercial sector, and CoreWeave is now carrying it into a fundamentally different environment.

US intelligence and military organizations handle classified data in specially equipped, isolated facilities. Law prohibits processing classified data on a standard public cloud: the information must not leave a physically secured perimeter. Yet government agencies need GPUs for AI workloads such as intelligence analysis, computer vision and predictive modeling no less than commercial tech companies do.

Leidos is the key partner and gateway. One of the largest IT contractors to the US Department of Defense, with annual revenue above $15 billion, it already operates inside classified perimeters and holds all the clearances required to manage sensitive government infrastructure.

What changes for the defense AI market

Before this partnership, classified organizations had a limited set of options: build GPU infrastructure themselves, or use certified government clouds such as AWS GovCloud and Microsoft Azure Government. Both routes were expensive and slow to deploy, and performance did not always meet the demands of modern AI workloads.

"CoreWeave plans to run the same cloud inside the secured facilities where

America keeps its secrets, with Leidos as the gateway," is how TNW describes the partnership model.

Key facts from the companies' joint announcement:

• The agreement was announced on July 30, 2026
• CoreWeave is a specialized GPU cloud provider; its customers include OpenAI and Microsoft
• Leidos is a US defense IT contractor with revenue of more than $15 billion a year
• CoreWeave's AI platform will be deployed inside classified facilities operated by Leidos
• Target customers are US intelligence agencies and the Department of Defense

CoreWeave and Leidos are offering a third path: a ready-made platform, proven with commercial customers and built on the latest GPU clusters, deployed directly inside the secure perimeter. That is potentially faster than building from scratch and higher-performing than general-purpose government clouds.

Why the deal matters for the industry

CoreWeave's move into classified computing reflects a broader trend. GPU clusters that were, only a few years ago, the exclusive tool of large tech companies are now in demand inside classified government facilities.

For CoreWeave, this is a strategically attractive market, with large long-term contracts and a high barrier to entry for competitors. Leidos sharply lowers that barrier for CoreWeave: it has already cleared all the required regulatory checks and holds active contracts with leading US intelligence and defense agencies. According to TNW, CoreWeave relies on Leidos's expertise and existing clearance to operate inside secured perimeters. US intelligence agencies spend tens of billions of dollars on IT infrastructure every year, which makes the market critically important for any GPU provider.

What this means

The CoreWeave-Leidos partnership sets a precedent: a classified GPU cloud for AI workloads. If the model succeeds, it could become a template for other commercial AI providers looking to enter the market for classified government computing, and it would accelerate AI adoption at US intelligence agencies and the Department of Defense.