KubeVirt has graduated from the Cloud Native Computing Foundation (CNCF), a milestone for the project and for efforts to manage virtual machines alongside containers through Kubernetes. The CNCF accepted KubeVirt into its sandbox seven years ago. Its graduation was announced at KubeCon + CloudNativeCon North America on November 12, 2026.
The project reflects a shift away from treating virtual machines and containers as competing approaches. Many enterprises continue to run applications on VMs, while cloud-native development and artificial intelligence workloads are increasing demand for newer infrastructure. KubeVirt’s aim is to let organizations operate both types of workloads within a Kubernetes-based environment rather than maintain separate platforms for them.
One Kubernetes platform for different workloads
Red Hat is applying that approach in OpenShift 5, which it describes as an environment for managing containers, VMs and AI workloads. Fabian Deutsch, Red Hat’s director of engineering for OpenShift Virtualization and Modernization and a KubeVirt maintainer, said the goal was to make it easier to run different compute workloads through one platform.
Deutsch framed the challenge in practical terms: organizations have applications running in VMs, alongside newer container-based applications, and may not want to operate two separate systems. KubeVirt offers a way to carry existing VM workloads forward while adopting Kubernetes for containerized applications. The approach is intended to reduce the divide between established virtualization environments and cloud-native operations, not require enterprises to replace all their current infrastructure.
That distinction matters as organizations take on workloads with different requirements. The source of KubeVirt’s recent development activity increasingly includes AI-related needs, Deutsch said. He argued that organizations should not have to rebuild or “repave” a data center simply to support a new workload. Instead, the ability to run workloads in the form that suits them can give operators flexibility as requirements change.
Virtualization and AI infrastructure
AI is also influencing how the KubeVirt community views virtualization. Deutsch said virtualization’s isolation model can be useful in the AI era, and identified passing GPUs, CPUs and memory into VMs as work on the project’s agenda. Those capabilities are presented as next steps, rather than as a claim that every such function is already available across KubeVirt deployments.
The project’s underlying proposition is to offer a consistent Kubernetes API across different deployment footprints. According to Deutsch, KubeVirt already runs on edge devices and in the cloud. A common API across those environments could help teams work with VMs without abandoning the Kubernetes operating model, although the source does not detail specific deployment configurations or performance results.
Deutsch named Nvidia as an early KubeVirt adopter and said Microsoft is now stepping up its involvement. The comments point to interest from major technology companies, but do not specify the scope of either company’s use or involvement. KubeVirt’s graduation establishes a new project milestone within the CNCF; it does not, by itself, establish how widely the technology is deployed or guarantee a particular operational outcome.
Carrying existing investments forward
For enterprises, KubeVirt’s significance lies in connecting a large installed base of virtualized applications with Kubernetes-based management. Organizations can continue running workloads in VMs while using the same broader platform for containers and, increasingly, AI-oriented infrastructure. That could help avoid a forced choice between preserving long-running applications and adopting newer workload models.
The graduation marks a point in the project’s development, while the work described by Deutsch indicates that integration with evolving hardware and AI requirements remains important. The next phase will depend on continued project development and on how operators put the platform to use. For now, KubeVirt’s central promise is a shared Kubernetes environment for VMs and containers, with the flexibility to carry existing workloads into infrastructure that also supports newer applications.
