OpenShift Virtualization 4.22 certified as NVIDIA-Certified Hypervisor for HGX

Red Hat OpenShift Virtualization 4.22 has achieved NVIDIA-Certified Hypervisor status for NVIDIA HGX systems. The certification validates near bare-metal GPU performance for AI and accelerated computing workloads. As a Kubernetes-native hypervisor built on KubeVirt and KVM, it allows GPU-accelerated VMs and containerized AI workloads to run on the same cluster under a unified API.

What the certification covers

• Certification applies specifically to NVIDIA HGX systems
• Testing covers compute, memory, data-path efficiency, and LLM inference workloads
• Validates accurate hardware topology exposure and key performance optimizations
• Built on KubeVirt (CNCF project) running on KVM
• GPU-accelerated VMs enabled via NVIDIA GPU Operator for automated driver lifecycle management

Why it matters

This certification signals that Kubernetes-native virtualization is viable for production AI infrastructure, potentially accelerating enterprise adoption of KubeVirt-based platforms over traditional hypervisors for GPU workloads.