At the Red Hat Summit, the two technology giants are strengthening their common platform to facilitate the transition from AI to production.
As part of the Red Hat Summit currently taking place in Atlanta, Red Hat, the world’s leading provider of open source solutions, announced this Tuesday a major expansion of its strategic collaboration with NVIDIA. This new step aims to strengthen their common software platform, Red Hat AI Factory with NVIDIA (https://www.redhat.com/fr/products/ai/factory-with-nvidia), in order to accelerate enterprise adoption of a new generation of artificial intelligence : autonomous agents operating continuously. The objective is clear: enable organizations to move from the experimentation phase to large-scale production, relying on a secure, compliant and reliable infrastructure.
A secure and sovereign infrastructure for AI
Trust is the key word of this strengthened collaboration. To guarantee controlled deployment of AI agents, the two companies are integrating the open source NVIDIA OpenShell project (https://docs.nvidia.com/openshell/about/overview). The latter provides a secure “sandbox” execution environment, which strengthens policy control, secures the use of tools and guarantees auditable operations. Red Hat, as a major contributor to the project, is actively participating in the standardization of autonomous agent management in the hybrid cloud.
Security is also reinforced at the hardware level thanks to Confidential Computing. Using Confidential Containers, available as a technology preview (https://www.redhat.com/en/blog/red-hat-openshift-sandboxed-containers-112-and-red-hat-build-trustee-11-bring-confidential-computing-bare-metal-and-ai-workloads) and running with NVIDIA Confidential Computing (https://www.nvidia.com/fr-fr/data-center/solutions/confidential-computing/), the platform isolates running agents to prevent any chain compromise. This approach is part of an overall Zero Trust architecture, integrating technologies like SELinux and real-time protection based on NVIDIA DOCA (https://developer.nvidia.com/networking/doca). These controls, applied from the data center to the edge, help companies comply with strict regulatory frameworks such as the EU AI Act.
De l’expérimentation à la production à grande échelle
To ease the transition to production, Red Hat AI Factory with NVIDIA incorporates the latest innovations from Red Hat AI 3.4. These include a new “Model as a Service” (MaaS) offering that simplifies access for developers to foundation models, such as NVIDIA Nemotron (https://www.nvidia.com/en-us/ai-data-science/foundation-models/nemotron/).
The platform also features MLflow-based lifecycle management, providing comprehensive tracking of Large Language Model (LLM) calls and reasoning steps. This unified framework makes it possible to audit precisely how an agent achieves a result, bridging the gap between rapid development and operational stability. To further accelerate value creation, Red Hat and NVIDIA are providing NVIDIA AI Blueprints (https://build.nvidia.com/blueprints) and AI Starter Kits (https://docs.redhat.com/en/learn/ai-quickstarts), validated models for critical use cases like research semantics on proprietary data or the anchoring of agents via RAG and RAFT technologies.
Cutting-edge hardware support and a shared vision
This software advancement is accompanied by leading hardware support. Red Hat Enterprise Linux for NVIDIA 26.01 (https://docs.redhat.com/en/documentation/red_hat_enterprise_linux_for_nvidia/26.01) is now available and supports the new NVIDIA Blackwell architecture. The two partners are already working on supporting the future NVIDIA Vera Rubin platform, thus ensuring the sustainability of investments.
This collaboration illustrates a shared vision for the future of AI in business. HAS” Agentic AI is reinventing business operations. Every business will need an AI factory to design, deploy and manage digital agents at scaleand », said Justin Boitano, vice president of enterprise AI platforms at NVIDIA.
A view complemented by Chris Wright, chief technology officer and senior vice president of global engineering at Red Hat: « Moving AI from the status of an experimental tool to that of an industrial lever requires a sovereign and coherent infrastructure on the scale of the hybrid cloud. Through its strategic co-engineering approach with NVIDIA, Red Hat provides enterprises with the architectural control and open source innovation they need to deploy agentic AI at scale and with confidence ».
Updates to Red Hat AI Factory with NVIDIA and Red Hat AI 3.4 are expected to be available sometime in May 2026.
