NVIDIA has deepened its push into hyperscale AI infrastructure through a new strategic alliance with Nebius Group N.V.. The partnership centers on building a next-generation AI cloud platform designed to serve both AI-native companies and enterprise developers operating at global scale.

As part of the agreement, NVIDIA will invest $2 billion in Nebius, reinforcing confidence in the company’s engineering capabilities and its vertically integrated approach to AI infrastructure. The move also positions Nebius to scale capacity aggressively as demand for high-performance computing accelerates worldwide.

The collaboration focuses on developing a full-stack AI cloud, spanning infrastructure design, accelerated computing systems, and production-grade AI software environments. Nebius will expand its hyperscale architecture with NVIDIA technologies, enabling rapid deployment of advanced computing platforms designed for AI workloads.

Importantly, the partnership signals how cloud infrastructure is evolving beyond traditional compute services. Instead, providers increasingly design AI factories, large-scale compute environments optimized specifically for training and inference workloads.

Gigawatt-Scale AI Infrastructure Targets 2030 Capacity

Nebius already deploys NVIDIA hardware across its global platform, including several large-scale AI infrastructure projects in the United States. Through the expanded partnership, the company plans to deploy more than five gigawatts of NVIDIA-powered computing capacity by 2030. This capacity expansion reflects a structural shift underway across the AI ecosystem. Large language models, agentic systems, and autonomous AI agents require dramatically higher compute densities than traditional cloud applications.

Therefore, NVIDIA will support Nebius with early access to the company’s next-generation accelerated computing technologies. These include new system architectures designed to support extreme workloads while maintaining performance efficiency at hyperscale.

The collaboration will focus on multiple areas of engineering integration. The companies will jointly develop AI factory architectures using shared design frameworks, system review processes, early hardware samples, and specialized system software support. They will also coordinate ongoing technical and business reviews to ensure infrastructure deployments scale smoothly across global facilities.

In addition, the partnership will advance a developer-focused AI inference environment built on NVIDIA software platforms, optimized AI models, and performance libraries. This stack will target both enterprise workloads and emerging agentic AI systems that require continuous inference at massive scale.

Nebius will also deploy several generations of NVIDIA infrastructure technologies across its platform. These include the upcoming NVIDIA Rubin computing platform, alongside NVIDIA Vera CPUs and NVIDIA BlueField storage systems, which together form a tightly integrated accelerated computing architecture.

Finally, the companies will collaborate on operational reliability. NVIDIA will deploy advanced GPU health monitoring tools and software optimization frameworks to help Nebius manage fleet-wide system performance across hyperscale infrastructure environments.

Industry Push Toward Agentic AI Infrastructure

The partnership arrives at a pivotal moment for the AI industry, where demand for compute infrastructure continues to surge as models grow more complex and AI services expand globally.

“AI is at another inflection point of agentic AI, driving incredible compute demand and accelerating infrastructure buildout,” said Jensen Huang, founder and CEO of NVIDIA. “Nebius is building an AI cloud designed for the agentic era, fully integrated from silicon to software and powered by NVIDIA’s next-generation accelerated compute. Together, we are scaling the cloud to meet the surging global demand for intelligence.”

Meanwhile, Nebius leadership views the partnership as validation of its purpose-built AI infrastructure strategy.

“Nebius has been built for AI since day one, not adapted from a general-purpose cloud, but designed for what developers actually need,” said Arkady Volozh, CEO of Nebius. “Now with NVIDIA, we are extending that throughout the stack from gigawatt-scale AI factories to inference and software, as we build one of the first and largest clouds for all AI builders everywhere.”

However, the broader significance extends beyond a single cloud deployment. The partnership illustrates how hyperscale infrastructure providers increasingly rely on deep engineering collaboration across silicon, systems architecture, and AI software. As the AI economy scales globally, companies that integrate these layers tightly will likely define the next generation of cloud platforms.

For Nebius and NVIDIA, this partnership signals a clear ambition: building the infrastructure backbone required for the agentic AI era.