The community behind KubeVirt is releasing version 1.8 of the platform. The release is compatible with Kubernetes 1.35. It demonstrates how KubeVirt continues to evolve within Kubernetes environments.
According to CloudNativeNow, the update was presented during KubeCon + CloudNativeCon Europe 2026. While KubeVirt was originally developed to run kernel-based virtual machines within containers, its focus is now shifting toward supporting multiple hypervisors.
A key innovation is the introduction of a hypervisor abstraction layer, which allows other backend hypervisors to be integrated alongside KVM (Kernel-based Virtual Machine). With this, KubeVirt is evolving from a KVM-driven solution into a broader virtualization layer within Kubernetes. This is happening without directly changing the existing default configuration.
In terms of security, the release adds support for Intel TDX attestation, which allows virtual machines to verify that they are running on hardware that supports confidential processing. This aligns with the growing demand for confidential computing. In this model, data remains protected even during processing.
Additionally, KubeVirt addresses the increasing demand for AI and high-performance workloads within Kubernetes. Support for PCIe NUMA topology enables more efficient use of hardware, bringing performance closer to native levels.
Network management becomes more flexible
Improvements have also been implemented on the networking front. The passt binding (passt stands for Plug A Simple Socket Transport) has now become a core component of the platform. This allows network configurations to be adjusted without restarting a virtual machine. At the same time, dependency on Network Access Device definitions has been reduced. Consequently, this leads to fewer API calls and better scalability of the control plane.
For storage, KubeVirt introduces, among other things, ContainerPath volumes, which allow container storage paths to be directly linked to virtual machines. Additionally, incremental backups with changed block tracking support, storing only modified data. This reduces storage requirements and shortens backup windows.
The platform’s scalability has also been further tested. Internal benchmarks show that the control plane continues to function stably as the number of virtual machine instances grows. This is accompanied by a predictable increase in memory usage.
KubeVirt was originally developed by Red Hat and has been under the Cloud Native Computing Foundation since 2019. It currently retains the status of an incubating project. At the same time, adoption is growing, partly because organizations are using the platform to run traditional, monolithic applications alongside cloud-native workloads within a single Kubernetes environment. This allows them to reduce the need for separate infrastructures.