NVIDIA Invests $2 Billion in Nebius as AI Cloud Buildout Targets More Than 5 GW

NVIDIA announced a $2 billion investment in Nebius and a partnership to scale full-stack AI cloud infrastructure.

NVIDIA and Nebius outline a 5 GW AI infrastructure plan

NVIDIA said in its investor-relations press release that the agreement covers AI factory design, inference, infrastructure deployment and fleet management. The company said the partnership is intended to help Nebius deploy more than 5 GW of NVIDIA systems by the end of 2030.

Nebius, an AI cloud company, is building capacity for customers that need large-scale GPU infrastructure. In its Q1 2026 shareholder letter, Nebius said its contracted capacity had already exceeded 3.5 GW and that it had raised its 2026 contracted-power guidance to more than 4 GW.

The same Nebius shareholder letter said the company had secured a 1.2 GW AI factory site in Pennsylvania. Data Center Dynamics separately reported the Pennsylvania plan and cited Nebius’s disclosure that it had more than 3.5 GW of secured capacity and a year-end contracted-power target above 4 GW.

Capital spending guidance moves higher

Reuters, in a report carried by Investing.com, said Nebius raised its annual capital-expenditure forecast to $20 billion to $25 billion, up from a previous range of $16 billion to $20 billion. Reuters attributed the increase to Nebius’s effort to secure AI infrastructure capacity and power.

Nebius’s Q1 2026 shareholder letter said the company spent about $2.5 billion in capital expenditure during the quarter, including GPUs, GPU-related hardware and data-center expansion. Those figures show that Nebius’s growth plan is not only a software or services expansion, but also a physical infrastructure buildout involving power access, facilities and hardware procurement.

Data Center Dynamics also reported the increase in 2026 capital-expenditure guidance to $20 billion to $25 billion, alongside the Pennsylvania campus plans and the company’s contracted-power targets.

Why the partnership matters

NVIDIA’s announcement positions Nebius as a customer and infrastructure partner for large deployments of NVIDIA systems. The stated scope — design, inference, deployment and fleet management — suggests the relationship extends beyond chip supply into operational support for AI data-center infrastructure.

For Nebius, the partnership comes as it seeks to convert power access and capital spending into available AI cloud capacity. The company’s own disclosures indicate that contracted power is a central metric for its growth plan, with more than 3.5 GW already contracted and more than 4 GW targeted by the end of 2026.

For NVIDIA, the $2 billion investment aligns the chipmaker with a cloud provider planning multi-gigawatt deployments of NVIDIA systems. NVIDIA did not describe the investment as a full acquisition, and the company’s announcement framed the transaction as part of a broader partnership to scale AI cloud capacity.

The constraint is physical as much as financial

The available source documents point to a simple conclusion: Nebius’s expansion depends on access to power, data-center sites and hardware, not just customer demand. The Pennsylvania 1.2 GW site cited by Nebius and reported by Data Center Dynamics is one example of how AI cloud providers are pursuing gigawatt-scale campuses to support future computing demand.

At the same time, the higher 2026 capital-expenditure guidance reported by Reuters and Data Center Dynamics shows the cost of moving quickly. Nebius is guiding toward $20 billion to $25 billion in annual capital spending while pursuing more than 4 GW of contracted power by year-end 2026 and a longer-term deployment path that NVIDIA says could exceed 5 GW of NVIDIA systems by the end of 2030.

The partnership therefore reflects a broader shift in AI infrastructure: leading AI cloud providers are being evaluated not only on model access or software platforms, but also on their ability to finance and operate very large power-backed computing facilities.