Red Hat announces a specially tailored version of RHEL for the Nvidia Rubin platform. The company promises support for the new AI architecture, which will be available later this year, from day one. The collaboration is intended to help companies bring AI projects into production faster.
Red Hat is launching Red Hat Enterprise Linux for NVIDIA, an edition that is fully optimized for the upcoming Rubin platform. The Linux distribution will support all platform features of the latest Nvidia architectures from launch.
The Nvidia Rubin platform consists of six integrated components: the Vera CPU, Rubin GPU, NVLink 6 Switch, ConnectX-9 SuperNIC, BlueField-4 DPU, and Spectrum-6 Ethernet Switch. The Vera CPU has 88 custom cores and delivers up to 1.2 TB/s of memory bandwidth. Rubin GPUs offer 22 TB/s of combined bandwidth via eight HBM4 memory stacks.
Nvidia is presenting the Rubin platform at CES. According to the manufacturer, the architecture delivers up to 10 times lower costs per token for inference with mixture-of-experts models than the current Blackwell platform.
Day 0 support for complete AI portfolio
Support extends across the entire Red Hat AI portfolio. Red Hat Enterprise Linux acts as a bridge between the new hardware and the software ecosystems for modern AI. The OS introduces support for Nvidia Confidential Computing, which enhances security for GPUs, memory, and model data throughout the AI lifecycle.
Red Hat OpenShift adds support for Nvidia infrastructure software and the CUDA X libraries. This gives users automated deployment, configuration, and lifecycle management of accelerated computing. Red Hat has been working closely with Nvidia for some time to optimize AI workloads on OpenShift AI.
Red Hat AI further expands its integrations with Nvidia. The platform adds support for distributed inference with Nvidia open source models on Red Hat AI Inference Server and OpenShift AI. Jensen Huang, CEO of Nvidia, states: “Together, NVIDIA and Red Hat are industrializing open source to bring AI to the enterprise, starting with the Vera Rubin platform.”
Red Hat Enterprise Linux support for the Nvidia Vera Rubin platform will be available when the hardware becomes available in the second half of 2026. Users can obtain the latest drivers and integration tools through the Red Hat Customer Portal. Red Hat Enterprise Linux for NVIDIA remains fully aligned with the main version of the operating system, allowing users to transition to traditional RHEL when their production requirements dictate easily.