Nvidia has committed $2 billion to AI cloud provider Nebius Group as part of a strategic partnership to build the next generation of hyperscale AI data centers, the companies announced. The investment underscores the chipmaker’s determination to ensure that the infrastructure needed to train and run advanced AI models keeps pace with the explosive demand for its processors.
Nebius, which was spun out of the Russian technology company Yandex and is now headquartered in the Netherlands, has emerged as one of the fastest-growing “neocloud” providers — companies that build cloud infrastructure specifically optimized for AI workloads rather than general-purpose computing. The partnership with Nvidia will see Nebius deploy more than 5 gigawatts of Nvidia systems by 2030, a scale of infrastructure that rivals the largest cloud providers.
Why Nvidia Is Investing in Cloud
The investment is part of a broader pattern. Nvidia has been steadily building relationships with a network of cloud providers — including CoreWeave, Lambda Labs, and now Nebius — that serve as alternative channels for companies that want access to Nvidia GPUs without going through AWS, Microsoft Azure, or Google Cloud. These neoclouds offer specialized AI infrastructure with fewer layers of abstraction and, in many cases, lower costs.
For Nvidia, the strategy serves multiple purposes. It diversifies the company’s customer base beyond the hyperscale cloud providers, reduces the risk that any single customer becomes too powerful, and ensures that AI compute capacity continues to grow fast enough to absorb the chips Nvidia is producing. The company is on track to ship more than $100 billion worth of AI chips this year, and every one of those chips needs a data center to live in.
The Nebius Opportunity
Nebius has positioned itself as a specialist in “AI factories” — data centers purpose-built for the unique demands of training and running large AI models. These facilities require different power, cooling, and networking configurations than traditional cloud data centers, and Nebius has been building them at a pace that has attracted attention from investors and customers alike.
The company went public on the Nasdaq through a SPAC merger and has seen its stock rise sharply on the back of the AI infrastructure boom. The Nvidia investment — which sent Nebius shares up 10% in premarket trading — provides both capital and a powerful endorsement. Nvidia is effectively betting that Nebius will become one of the primary destinations for the AI chips it manufactures.
The Bigger Picture
The partnership also reflects the changing economics of AI infrastructure. Training a single frontier AI model can cost hundreds of millions of dollars in compute, and the largest models now require clusters of tens of thousands of GPUs running for months. The companies that can provide that compute at scale — and at competitive prices — are positioned to capture an enormous share of the value being created by the AI boom.
Nvidia’s $2 billion investment is a bet that Nebius will be one of those companies. If the bet pays off, it will mean more demand for Nvidia chips, more AI models trained, and more of the infrastructure that the AI industry needs to sustain its extraordinary growth trajectory.