Nvidia Anchors CoreWeave Expansion

This makes Nvidia less like a supplier and more like CoreWeave's demand anchor, capital provider, and roadmap partner all at once. The $2B equity purchase gives CoreWeave fresh balance sheet support, while the $6.3B order form means Nvidia is also standing behind utilization on reserved capacity that CoreWeave is building out through 2032. That is unusually deep coupling for a cloud provider, because the same company is funding the GPU buildout, supplying the chips, and agreeing to absorb unsold capacity.

• The capacity backstop is not a normal customer contract. CoreWeave first sells reserved cloud capacity to end customers, and Nvidia gets access to residual unsold capacity under the September 9, 2025 order form. In practice, that helps CoreWeave finance infrastructure earlier because some utilization risk is shifted to Nvidia.
  3 sec 4 sec
• The strategic logic is straightforward. Nvidia wants scaled cloud partners that are fully committed to Nvidia hardware, especially as AWS, Google, and Microsoft push custom AI chips. CoreWeave fits because it buys huge volumes of Nvidia GPUs and is not trying to replace them with in house silicon.
  2 nvidia 5 sacra 6 sacra
• This also explains why CoreWeave can expand faster than smaller GPU clouds. It ended 2025 with $66.8B of backlog, more than 850 MW of active power, about 3.1 GW of contracted power, and a plan with Nvidia to build more than 5 GW of AI factories by 2030. Few peers have that mix of chip access, guaranteed demand, and financing support.
  1 sacra 2 nvidia 4 sec 7 sec

Going forward, the relationship is likely to make CoreWeave one of Nvidia's main vehicles for turning future GPU generations into dedicated AI factory capacity. That should keep CoreWeave near the front of the allocation line, but it also means the company's expansion path will remain tightly tied to Nvidia's product cadence, sales priorities, and ecosystem strategy.