Red Hat, NVIDIA expand enterprise AI partnership to support rack-scale AI systems

Red Hat has expanded its collaboration with NVIDIA to support the next phase of enterprise AI infrastructure, focusing on rack-scale systems and production deployments. The effort aligns Red Hat’s enterprise open-source software with NVIDIA’s latest AI platforms as organizations transition from isolated AI experiments to large, centralized deployments.

At the center of the announcement is Red Hat Enterprise Linux for NVIDIA, a new edition of Red Hat Enterprise Linux designed to support NVIDIA platforms from the day they become available. The first target is the NVIDIA Rubin platform, which combines the NVIDIA Vera CPU, Rubin GPUs, BlueField-4 data processors, and rack-scale NVL72 systems. The goal is to give enterprises a stable operating system that works with new AI hardware at launch and integrates with Red Hat OpenShift and Red Hat AI.

Many enterprises are planning to operationalize AI workloads by 2026. This shift requires infrastructure that can support large-scale training, inference, and AI agents while meeting security and operational requirements. Red Hat plans to provide day 0 support for the Rubin platform across its hybrid cloud and AI products, allowing organizations to deploy AI workloads using a consistent software stack across on-premises data centers, public clouds, and edge environments.

Red Hat Enterprise Linux will serve as the base layer connecting NVIDIA hardware with enterprise software. It will support NVIDIA Confidential Computing across the AI lifecycle, helping protect GPUs, memory, and model data. Red Hat OpenShift will add support for NVIDIA infrastructure software, CUDA-X libraries, and BlueField networking to manage accelerated workloads using Kubernetes. Red Hat AI will expand integration with NVIDIA’s open models and support distributed inference across its AI platforms.

Red Hat Enterprise Linux for NVIDIA will remain aligned with the main Red Hat Enterprise Linux release, allowing customers to transition between editions without changing applications. The companies aim to reduce deployment complexity by validating hardware compatibility, simplifying driver access through Red Hat repositories, and maintaining a consistent operational model for enterprise AI systems.

"NVIDIA's architectural breakthroughs have made AI an imperative, proving that the computing stack will define the industry’s future,” said Matt Hicks, president and CEO, Red Hat. “To meet these tectonic shifts at launch, Red Hat and NVIDIA aim to provide Day 0 support for the latest NVIDIA architectures across Red Hat’s hybrid cloud and AI portfolios.”

Jensen Huang, founder and CEO, NVIDIA, said, “In the age of AI, the entire computing stack—from chips and systems to middleware, models, and the AI lifecycle—is being reinvented from the ground up. Together, NVIDIA and Red Hat are industrializing open source to bring AI to the enterprise, starting with the Vera Rubin platform."