The investment is not simply a financial stake. Under the terms of the partnership, Nvidia will collaborate with Nebius on AI factory design, inference and agentic AI software, fleet management, and the deployment of multiple generations of Nvidia hardware, including the forthcoming Rubin platform, Vera CPUs, and BlueField storage systems. Nebius will also receive early access to Nvidia’s latest accelerated computing architecture as it works toward deploying more than 5 gigawatts of data center capacity by the end of 2030. To put that figure in context, 5 gigawatts is roughly equivalent to the power needs of more than 4 million U.S. households.
Nebius and the neocloud moment
Nebius sits within a growing category of infrastructure companies sometimes called neoclouds. Unlike the major hyperscalers, which serve a broad range of commercial customers, neoclouds focus on building compute capacity specifically tailored for AI workloads. Nebius, along with CoreWeave, has gained significant traction through deals to supply that infrastructure to some of the largest names in technology, including a reported $17 billion deal with Microsoft and a $3 billion deal with Meta Platforms.
Nebius CEO Arkady Volozh described the company as having been built for AI from its founding rather than adapted from a general-purpose cloud model. The Nvidia partnership, he said, extends that commitment from gigawatt-scale AI factories down through inference and software.
The company’s capital spending reflected the pace of that build-out. In the December quarter, Nebius reported $2.1 billion in capital expenditures, up sharply from $416 million in the same period a year earlier.
Nvidia’s expanding investment portfolio
Today’s announcement is the latest in a string of $2 billion investments Nvidia has made across the AI ecosystem. The company announced similar stakes in optics firms Lumentum and Coherent earlier this month, took a $2 billion position in chip design firm Synopsys in December, and announced a $2 billion stake in CoreWeave in January. On Tuesday, Nvidia also announced a significant investment in Thinking Machines Lab, the AI startup founded by former OpenAI chief technology officer Mira Murati, as part of a multiyear strategic partnership.
The chipmaker contributed $30 billion to the $110 billion funding round OpenAI announced late last month and said in November that it planned to invest up to $10 billion in Anthropic. Nvidia CEO Jensen Huang noted at a conference earlier this month that those investments may be among the last Nvidia makes in those companies before they go public.
The pattern has drawn attention from analysts who note that many of these companies are also Nvidia customers, raising questions about the circular nature of the deals. Nvidia has so far framed the investments as a way to ensure the infrastructure supporting AI development keeps pace with demand for its chips.
Nvidia produces the graphics processing units that underpin nearly all large-scale AI model training and deployment, making it one of the primary financial beneficiaries of the current AI build-out. The Nebius partnership extends that position further down the stack, from the chips themselves into the cloud architecture that runs on top of them.
