NVIDIA has invested $2 billion in AI cloud provider CoreWeave, purchasing Class A shares at $87.20 each to support the buildout of new data center infrastructure. The companies announced the agreement along with plans to expand technical collaboration, including early adoption of NVIDIA’s upcoming compute, CPU, and storage platforms.
Following the news, CoreWeave’s stock rose 12%.
The deal comes as CoreWeave ramps up its delivery capacity to serve long-term infrastructure agreements with major AI clients.
In addition to integrating new NVIDIA hardware, the company will work with NVIDIA to test internal software tools for potential use across other cloud and enterprise environments.
Why It Matters: AI compute requirements are growing faster than existing infrastructure can handle. In response, providers are reworking hardware design and exploring new financing approaches. The agreement between NVIDIA and CoreWeave shows how newer entrants are addressing these demands through custom-built systems and closer coordination between data center design and AI software deployment.
- $2 Billion Equity Investment to Support Expansion: NVIDIA acquired CoreWeave stock at $87.20 per share, below the previous close of $92.98. The $2 billion in funding will support the buildout of new data centers and add to the billions already raised by CoreWeave since its March IPO. The investment strengthens NVIDIA’s role in a company that supports many of its AI workloads.
- 5-Gigawatt Data Center Capacity Target by 2030: CoreWeave and NVIDIA plan to build more than 5 gigawatts of AI-focused infrastructure over the next several years, an amount of electricity comparable to the annual usage of about 4 million U.S. homes. This capacity is intended to support long-term infrastructure contracts with OpenAI, Meta, and others that depend on continuous access to GPU resources for training and running large-scale AI systems.
- Deployment of NVIDIA’s Next-Generation Hardware: CoreWeave will be among the first providers to adopt NVIDIA’s Rubin architecture, Vera CPUs, and BlueField storage systems. These platforms are designed to support high-throughput AI workloads that place sustained demands on compute and memory. CoreWeave plans to operate multiple hardware generations within the same environment to serve clients with varying performance requirements.
- Evaluation of CoreWeave’s Software for Wider Use: CoreWeave’s internal platforms, SUNK and Mission Control, will be evaluated by NVIDIA for possible inclusion in reference architectures made available to cloud providers and enterprise customers. SUNK allocates GPUs for large AI jobs, while Mission Control monitors and coordinates infrastructure across locations. If incorporated, these tools could guide how NVIDIA’s customers manage compute resources.
- Customer Commitments and Capacity Guarantees: CoreWeave is under contract to provide $22.4 billion in infrastructure to OpenAI and $14.2 billion to Meta through 2032. NVIDIA’s investment supplies additional capital to help meet these delivery commitments and maintain the pace of new data center construction. A separate $6.3 billion agreement requires NVIDIA to purchase any unused compute capacity through April 2032, providing more predictable utilization and revenue across CoreWeave’s network.
CoreWeave stock jumps 12% as Nvidia invests $2 billion to expand AI data center capacity – CNBC
Trusted insights for technology leaders
Our readers are CIOs, CTOs, and senior IT executives who rely on The National CIO Review for smart, curated takes on the trends shaping the enterprise, from GenAI to cybersecurity and beyond.
Subscribe to our 4x a week newsletter to keep up with the insights that matter.
