Nvidia deepens CoreWeave AI data center alliance

Nvidia commits $2 billion equity investment to CoreWeave

Nvidia and CoreWeave have expanded their strategic relationship with a new $2 billion equity investment by the chipmaker into the neocloud provider. Announced on Jan. 26, the deal prices CoreWeave shares at $87.20 and follows an earlier arrangement under which Nvidia agreed to purchase up to $6.3 billion of any unsold CoreWeave data center capacity through April 2032. The latest agreement moves Nvidia from being solely a major customer and offtake backstop into a direct financial backer of CoreWeave’s growth plans.

CoreWeave specializes in renting dedicated AI data center capacity to large technology customers, including Microsoft, Meta Platforms and OpenAI, as well as other AI-focused enterprises. Its infrastructure is built primarily around Nvidia graphics processing units, and the closer tie-up underscores the companies’ shared interest in expanding access to high-performance computing tailored to artificial intelligence workloads.

Funding 5 GW of AI ‘factories’ and expanding infrastructure

According to the joint announcement, Nvidia’s investment is intended to help CoreWeave accelerate the buildout of more than 5 gigawatts of AI data centers, described as “AI factories,” by 2030. Those facilities will be powered by Nvidia’s advanced data center chip systems, including its upcoming Vera Rubin offerings. The companies say the expanded capacity is designed to support AI adoption at a global scale, reflecting what they describe as insatiable demand for AI data center resources from hyperscale cloud and AI customers.

Beyond hardware, Nvidia’s financial strength is expected to aid CoreWeave in securing land and electricity to support new data centers. The capital will also underpin operational planning and infrastructure deployment as CoreWeave scales to meet larger, longer-duration contracts with major technology platforms. The investment complements Nvidia’s existing commitment to purchase unused CoreWeave capacity, which provides additional revenue visibility for the neocloud provider.

Use of proceeds and avoidance of ‘circular’ financing

CoreWeave has indicated that the $2 billion in new funds will be directed primarily toward research and development, workforce expansion and other data center-related expenses, according to comments cited from a company spokesperson. The money will not be used to purchase Nvidia’s hardware, a point highlighted to address potential concerns that the arrangement might constitute circular financing, in which vendor-supplied capital is recycled back to that same vendor via equipment purchases.

In addition to physical infrastructure, Nvidia plans to help strengthen CoreWeave’s software stack. The companies intend to develop CoreWeave’s platform into a more comprehensive offering that not only provides raw computing power but also supports customers across the full AI lifecycle, from model building through deployment and scaling. That focus on software and services is positioned as a way to differentiate CoreWeave from providers that offer only commodity compute.

Backlog surge underscores demand from AI hyperscalers

CoreWeave enters this new phase of growth with a rapidly expanding revenue backlog that reached nearly $56 billion at the end of the third quarter of 2025. That compares with about $15 billion a year earlier, reflecting a near fourfold increase in contracted future business within twelve months. The backlog growth has been driven by large deals with Meta Platforms, OpenAI and other hyperscale customers seeking long-term access to dedicated AI infrastructure.

The company and Nvidia suggest that the fresh capital infusion should help convert a greater portion of that backlog into realized revenue more quickly. With significant contracts already in place and additional demand expected as AI adoption widens, the investment is framed as a way to align CoreWeave’s buildout schedule with customer deployment timelines while reinforcing Nvidia’s position at the center of AI computing supply chains.