NVIDIA Donates GPU Driver to Open Source Community

Artificial intelligence workloads have become central to modern computing, with most enterprises relying on Kubernetes to manage these tasks. NVIDIA has announced a significant step to enhance this ecosystem by donating its Dynamic Resource Allocation (DRA) Driver for GPUs to the Cloud Native Computing Foundation (CNCF). This donation, revealed at KubeCon Europe in Amsterdam, shifts the driver from vendor control to community ownership under the Kubernetes project, fostering greater transparency and efficiency for developers worldwide.

The NVIDIA DRA Driver for GPUs is designed to streamline the orchestration of high-performance GPU resources within Kubernetes environments. Historically, managing GPUs in data centers required substantial effort, but this contribution aims to simplify that process. By aligning hardware innovations with upstream Kubernetes efforts, NVIDIA seeks to make GPU orchestration seamless and accessible, as highlighted by Chris Aniszczyk, chief technology officer of CNCF, who called it a major milestone for open source AI infrastructure.

In addition to the driver donation, NVIDIA has collaborated with the CNCF's Confidential Containers community to introduce GPU support for Kata Containers. These lightweight virtual machines act like containers but provide stronger isolation for workloads. This extension of hardware acceleration into a more secure environment allows AI tasks to run with enhanced protection, enabling organizations to implement confidential computing to safeguard sensitive data effectively.

NVIDIA is working with industry leaders such as Amazon Web Services, Broadcom, Canonical, Google Cloud, Microsoft, Nutanix, Red Hat, and SUSE to advance these features for the cloud-native ecosystem. Chris Wright, chief technology officer at Red Hat, emphasized that open source is crucial for enterprise AI strategies, bringing standardization to high-performance infrastructure. Ricardo Rocha from CERN noted that community-driven innovation accelerates scientific , with the DRA Driver strengthening the ecosystem for processing data in both traditional computing and machine learning workloads.

This donation is part of NVIDIA's broader open source initiatives, which include recent announcements like NVSentinel for GPU fault remediation and AI Cluster Runtime at GTC. NVIDIA also unveiled new open source projects such as the NVIDIA NemoClaw reference stack and NVIDIA OpenShell runtime, which offers fine-grained security controls and integrates with Linux, eBPF, and Kubernetes. Furthermore, NVIDIA's high-performance AI workload scheduler, the KAI Scheduler, has been onboarded as a CNCF Sandbox project to encourage wider collaboration.

Following the release of NVIDIA Dynamo 1.0, the company is expanding the Dynamo ecosystem with Grove, an open source Kubernetes API for orchestrating AI workloads on GPU clusters. Grove enables developers to express complex inference systems in a single declarative resource and is being integrated with the llm-d inference stack for broader adoption in the Kubernetes community. Developers and organizations can start using and contributing to the NVIDIA DRA Driver today, with live demos available at the NVIDIA booth at KubeCon.

NVIDIA remains committed to actively maintaining and contributing to Kubernetes and CNCF projects to meet the demands of enterprise AI customers. This ongoing support underscores the company's focus on fostering an open environment where experts can contribute ideas and accelerate innovation. The initiatives aim to ensure that technology evolves in alignment with the modern cloud landscape, benefiting a wide range of users from developers to large research institutions.

The broader context of these efforts highlights the growing importance of open source in AI infrastructure, as noted by industry collaborators. While the donation and related projects offer significant benefits, they rely on community engagement to drive future developments. NVIDIA encourages participation through contributions and usage, with opportunities to see the technology in action at events like KubeCon and GTC Live 2026, where discussions on AI and accelerated computing will feature industry leaders.