• Announces first general availability of NVIDIA GB200 NVL72 capacity in Europe;
• NVIDIA B200 capacity to be available on-demand through Nebius self-service platform, as well as via NVIDIA DGX Cloud Lepton;
• Strengthens software capabilities of Nebius AI Cloud for enterprise use-cases and sovereign AI projects.

Nebius (NASDAQ: NBIS) today announced the first general availability of NVIDIA GB200 Grace Blackwell Superchip capacity for customers in Europe, as it continues to build out full-stack AI infrastructure globally to accelerate AI innovation at scale.

Arkady Volozh, founder and CEO of Nebius, said:

“Our mission is to provide the infrastructure that accelerates AI innovation in Europe and globally. By integrating NVIDIA accelerated computing and enterprise-grade software more deeply into the Nebius AI Cloud platform, we are giving European innovators – whether startups, enterprises, developers, or researchers – access to the tools they need to build, scale, and deploy the next generation of AI applications.”

Nebius today also announced other integrations with NVIDIA’s accelerated computing platform, including expansion of Nebius AI Cloud with the NVIDIA AI Enterprise software stack; first-to-market access to NVIDIA Blackwell via NVIDIA DGX Cloud Lepton; and first-in-Europe access to NVIDIA Hopper GPUs through the NVIDIA Brev platform. Nebius will also offer instances featuring the RTX PRO 6000 Blackwell Server Edition later this year.

“Nebius is helping bring the power of the AI factory to the forefront of Europe’s AI transformation,” said Dave Salvator, director of accelerated computing products at NVIDIA. “With its strong focus on performance, scalability, and compliance, Nebius enables innovators across sectors to develop and deploy the next generation of reasoning, agentic and physical AI faster and more efficiently.”

As an early adopter of NVIDIA Blackwell, Nebius worked with research group LMArena in collaboration with NVIDIA to bring LMArena’s Prompt-to-Leaderboard (P2L) system into production using GB200 NVL72 infrastructure. Trained in just four days via NVIDIA DGX Cloud on Nebius, P2L routes real-time user queries to the most accurate and cost-efficient models, leveraging live feedback to optimize performance. The project demonstrates the flexibility and scalability of Nebius’s infrastructure for open-source AI innovation. “Working with Nebius and the NVIDIA GB200 platform fundamentally transformed our ability to scale P2L rapidly,” said Wei-Lin Chiang of LMArena. “The system’s performance gave us the flexibility to experiment, iterate quickly, and deliver a real-time routing model that adapts to live user input. We’re seeing improved accuracy and efficiency as a result.”

The addition of the NVIDIA AI Enterprise stack to Nebius AI Cloud expands Nebius’s support for enterprise use cases and national AI projects. NVIDIA AI Enterprise provides access to a robust suite of tools to accelerate AI development and deployment, including NVIDIA NIM microservices, NVIDIA NeMo, and other foundational components.

Nebius customers are already realizing the benefits. Prima Mente, a biotech company focused on neurological diseases like Alzheimer’s, is leveraging Nebius’s infrastructure to power its epigenetics research.

“Nebius’s GPU Cloud and AI-centric infrastructure give Prima Mente the elastic, cost-efficient HPC we need to train and fine-tune our 90M to 7B-parameter epigenetic foundation models at scale,” said Ravi Solanki, CEO of Prima Mente. “Their life-sciences-ready platform lets us spin up secure clusters in minutes, compressing model-development cycles and accelerating delivery of next-generation brain-health diagnostics to researchers and clinicians worldwide.”

The first European-headquartered Reference Platform NVIDIA Cloud Partner (NCP), Nebius aligns with EU regulatory standards and provides the performance and transparency needed for national and sovereign AI projects. That makes Nebius a natural fit for strategic AI initiatives like Milestone Systems’ Project Hafnia, which in collaboration with NVIDIA and Nebius is developing a compliant, anonymized video data platform for training AI vision-language models. These models are then used to help cities develop smart city AI solutions powered by NVIDIA digital twins and AI agents. All data is stored and processed within the EU.

“It is a core commitment of the EU to lead the way when it comes to transparency, fairness and regulatory integrity within AI and technology,” says Thomas Jensen, CEO of Milestone Systems. “Aligning with the legal framework is a key part of this. I am proud that Milestone's Project Hafnia, built on NVIDIA, can now provide the world’s first regulatory compliant platform – and with Nebius serving as European cloud provider, visual AI model training with high-quality, compliant video data can truly be rooted in Europe.”

About Nebius

Nebius is a technology company building full-stack infrastructure to service the explosive growth of the global AI industry, including large-scale GPU clusters, an AI-native cloud platform, and tools and services for developers. Headquartered in Amsterdam and listed on Nasdaq, the Company has a global footprint with R&D hubs across Europe, North America and Israel.

Nebius’s AI Cloud has been built from the ground up for intensive AI workloads. With proprietary software and hardware designed in-house, Nebius gives AI builders the compute, storage, managed services and tools they need to build, tune and run their models.

Nebius is one of only a handful of companies globally to hold Reference Platform NVIDIA Cloud Partner status, underscoring its expertise in designing and deploying a full stack of hardware and software infrastructure to NVIDIA’s Reference Architecture.

To learn more please visit www.nebius.com

View source version on businesswire.com: https://www.businesswire.com/news/home/20250611485160/en/