Following Canonical and SUSE's announcement of plans to enhance support for NVIDIA CUDA, Red Hat officially confirmed today that it will directly distribute the NVIDIA CUDA toolkit in Red Hat Enterprise Linux (RHEL), Red Hat AI and OpenShift products.
This initiative aims to simplify the developer experience, provide enterprise customers with a more consistent and efficient operation and maintenance environment, and make it easier for users to leverage the Red Hat platform combined with NVIDIA's latest hardware and software innovations.
Red Hat said that in the future, it will directly integrate the CUDA toolkit into its major platforms to enhance the availability of AI and data center infrastructure and achieve comprehensive support from the data center to the edge. Ryan King, head of technology at Red Hat, noted in the company announcement: "Red Hat is committed to providing flexible, open, and scalable platforms for enterprise technology. As AI becomes a core business driver, this mission has become more important than ever. This new agreement with NVIDIA is not just a technical collaboration, but also allows enterprise users to easily innovate AI applications, no matter what stage they are currently in AI deployment."
Regarding some users' concerns about the "closed ecosystem" of CUDA's proprietary software, King emphasized that Red Hat has always adhered to the open source concept. This cooperation is a bridge connecting the open hybrid cloud and the world's leading AI platform, ensuring that users have a more stable and reliable basic platform and can freely choose the most suitable development tools. He added: "The future of AI technology does not rely solely on a single model, accelerator or cloud platform, but on the collaborative innovation of diverse technologies. By integrating the CUDA toolkit directly into Red Hat products, we enable customers to more easily create future-oriented AI solutions."
Details of relevant announcements can be found on the Red Hat official website:
https://www.redhat.com/en/blog/red-hat-distribute-nvidia-cuda-across-red-hat-ai-rhel-and-openshift