CoreWeave and Leidos build classified AI cloud for US intelligence and defense agencies

CoreWeave Enters the Classified Computing Arena

CoreWeave, known for renting high-end GPUs to the biggest AI labs, announced a collaboration with government contractor Leidos to deliver a classified AI cloud for US spies and military operations1

1

. The Leidos CoreWeave partnership will place CoreWeave's AI-native cloud platform inside Sensitive Compartmented Information Facilities, the sealed rooms known as SCIF-accredited data centers where classified intelligence is stored and analyzed2

2

. This marks a shift for CoreWeave from purely commercial operations into the highly regulated world of the US Intelligence Community and Department of Defense.

The partnership addresses a critical constraint in classified computing: not chips, but clearance. While commercial clouds are built for shared regions and elastic capacity over the open internet, classified work demands hardware sitting in cleared facilities, operated by cleared staff, and completely isolated from public networks1

1

. Leidos will handle the complex aspects a commercial provider cannot manage alone: accreditation, secure architecture, cyber operations, and mission integration for national security organizations.

What Secure Sovereign AI Cloud Services Will Deliver

The companies outlined five core offerings designed to provide AI capabilities for defense and intelligence missions. First is a classified environment for model training, fine-tuning, evaluation, deployment, and continuous monitoring3

3

. Second, AI-powered workflows will support intelligence fusion from multiple sources, imagery analysis and exploitation, and decision support for analysts.

The remaining features lean into cybersecurity war-gamin g and operational readiness. Cyber AI ranges will enable simulation, autonomous defense testing, threat modeling, and adversarial AI evaluation. Synthetic data generation and digital twins will support training and operational planning in secure environments. Finally, edge-to-cloud AI orchestration will connect centralized platforms with forward-deployed, disconnected tactical units in the field2

2

.

Jason O'Connor, president of Leidos Intelligence, described this as "the next evolution of mission technology," delivering "sovereign AI compute at scale, secure by design"1

1

. CoreWeave will provide purpose-built infrastructure, advanced networking, AI-optimized storage, and cloud-native orchestration for training and inference workloads.

Why CoreWeave Needs Federal AI Priorities

For CoreWeave, the AI cloud for defense agencies solves a pressing business problem: customer concentration. The company reported $2.08bn in first-quarter revenue, up 112% year-over-year, against a backlog worth $99.4bn1

1

. However, much of that revenue rests on a handful of giants, including a $21bn commitment from Meta and a multi-year deal with Anthropic. Selling to intelligence and military customers diversifies that risk.

Leidos brings the scale and credentials necessary for this work. The company reported approximately $17.2bn in revenue for the fiscal year ended January 2, 2026, and employs more than 50,000 people2

2

. Sachin Jain, CoreWeave's chief operating officer, said federal teams need "secure, scalable platforms" to operationalize AI, and the company intends to extend the capabilities it provides to top AI labs "to highly secure government environments"3

3

.

The collaboration builds on CoreWeave Federal, the company's dedicated business launched last year for delivering AI cloud services to US government agencies and the Defense Industrial Base. CoreWeave passed 1GW of active power in the first quarter and has contracted more than 3.5GW, with plans to reach more than 8GW by 20301

1

.

What Actually Gets Built Depends on Washington

Despite the announcement, definitive agreements have not yet been signed. CoreWeave's own filing notes that deployment will follow "mission requirements and federal appropriation priorities"1

1

. In practical terms, this moves at the speed of Washington procurement, not commercial contracts. The classified slice will be tiny compared to CoreWeave's overall capacity, limited less by megawatts than by where an accredited rack can legally sit.

This push reflects a wider trend. Washington is consolidating software contracts, and the hunt for sovereign AI now extends beyond US borders, with Europe pursuing its own military-AI initiatives. Rivals including military cyber startups and chipmakers are chasing the same budgets, creating a competitive landscape for federal contracts. For readers watching this space, the key question is whether CoreWeave can navigate government accreditation processes faster than competitors and whether federal budgets will prioritize AI infrastructure at the scale these partnerships envision.