Nebius Partners With NVIDIA on Physical AI Cloud Platform

Nebius, global industry experts in AI cloud infrastructure and scalable computing platforms, has announced a collaboration with NVIDIA to develop an end-to-end platform designed to support the full robotics lifecycle, from simulation and training to real-world deployment.

The platform combines Nebius’s global AI cloud infrastructure with the NVIDIA Physical AI Data Factory Blueprint, an open reference architecture for large-scale data generation and evaluation. It is designed to provide robotics developers with an integrated environment that addresses key challenges in physical AI development, including fragmented infrastructure and limited access to high-quality training data.

“Physical AI is going to be one of the defining technology shifts of this decade, and the teams building it today are being held back by infrastructure and tooling that was never designed for those workloads, ” said Evan Helda, Head of Physical AI at Nebius. “Working with NVIDIA, we are building the execution layer for the entire physical AI ecosystem — so that any team, anywhere, can go from idea to deployed robot at the speed the market demands.”

“Physical AI is the next phase of computing — where intelligence is trained, tested and validated in simulation before it operates in the real world,” said Rev Lebaredian, VP of Omniverse and simulation technologies at NVIDIA. “That demands tightly integrated systems connecting large-scale AI training with physically accurate simulation to create a continuous data flywheel. By integrating the NVIDIA Physical AI Data Factory Blueprint, Nebius is enabling developers to generate physics-grounded synthetic data and build safe, robust autonomous machines at scale.”

A central challenge in physical AI development is what the companies describe as the “three-computer problem,” where robotics systems must operate across large-scale GPU training environments, simulation platforms, and edge deployment systems. Each environment introduces its own infrastructure and tooling requirements, often leading to significant integration overhead.

In addition, real-world data collection remains costly, inconsistent, and insufficient to cover the edge cases that determine system performance in operational environments. These limitations can slow development cycles and reduce overall system effectiveness.

The Nebius cloud platform addresses these challenges through a unified architecture. NVIDIA OSMO is delivered as a managed service to provide orchestration across the entire robotics pipeline, while NVIDIA Cosmos foundation models generate physics-consistent synthetic data to supplement real-world datasets.

The full stack runs on Nebius AI Cloud, which integrates NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, high-throughput object storage, data management and labeling tools, serverless capabilities, and managed inference services. This enables developers to access scalable infrastructure without managing underlying systems.

Beyond simulation and training, the platform supports deployment through services such as Nebius Token Factory, enabling low-latency inference from cloud to edge environments. This creates a continuous development loop in which operational data improves models, and improved models expand robotic capabilities.

Several companies are already applying the platform across real-world use cases. According to Nebius, RoboForce is using the system to develop robots for unstructured outdoor environments such as construction and agriculture, reducing pipeline setup time and accelerating model deployment. The company also states that Voxel51 is enabling large-scale dataset curation and analysis to support faster model development cycles. Milestone Systems is described as using the platform to train advanced vision-language models, while Porsche is applying synthetic data workflows to accelerate autonomous driving development.

About Nebius

Nebius is an AI cloud platform provider offering infrastructure and tools for data processing, model training, and deployment. Headquartered in Amsterdam and listed on NASDAQ (NBIS), the company supports organizations developing AI-driven applications across multiple industries, with a focus on scalable, high-performance computing environments. For more information, please click here.

Source: Nebius

---

(Editor’s Note: All trademarks mentioned in this article, including company names, product names, and logos, are the property of their respective owners. Use of these trademarks is for informational purposes only and does not imply any endorsement.)