This Month in KubeVirt News: February 2026

Welcome to our monthly roundup highlighting project news, community writings, and resources devoted to the KubeVirt project!  This series is intended to be an open-source resource for developers, platform engineers, storage experts, and anyone interested in learning more about the ins-and-outs of Day 1 and Day 2 operations of virtual machines in containers with KubeVirt.

In this issue, we highlight the release of KubeVirt v1.7.1, which delivers critical fixes for CPU compatibility during live migrations and streamlines VM image exports. As the community gears up for the v1.8 code freeze, February’s technical discussions centered on advancing vGPU live migration, expanding Arm64 feature parity, and optimizing SR-IOV networking. We also feature a wealth of new community content, including Victor Toso’s insights on VDI at FOSDEM to Zespre Chang’s KubeVirtBMC project for bare-metal provisioning, and ending with a preview of selected KubeVirt sessions at the upcoming KubeCon conference in Amsterdam.

Project and Community News

Version 1.7.1 was released on February 24 with several improvements and bugfixes:

• Simplifies moving VM images or backups out of the cluster. Users no longer have to manually “trick” the system into keeping a Pod active just to ensure the PVC remains accessible for export (PR #16170).
• Improves the stability of core guest-to-host communication channels when dynamic storage operations are performed (PR #16649)
• Fixes instability from migrations to incompatible CPUs by preventing the virt-controller from selecting a target node for Live Migration if the target node CPU vendor does not match the source node (PR #16293).
  

4-Feb: Community Meeting

• v1.8 Release Finalization: The community discussed the final stabilization and “code freeze” enforcement for the KubeVirt v1.8 release, including a retrospective on the release cycle and the scheduling of an upcoming “unconference” to address remaining technical debt.
• Infrastructure & CI Updates: Technical updates were provided regarding the resolution of flaking end-to-end (e2e) tests on the s390x architecture and the ongoing migration of build processes to rules_oci within the Bazel build system.
• Enhanced Reviewer Recruitment: In response to a high volume of new contributors and PRs, the project leads issued a call for reviewer self-nominations to expand the pool of maintainers capable of approving enhancements and SIG-specific code changes.

9-Feb: SIG-Network Meeting

The group discussed debugging performance regressions in SR-IOV workloads and reviewed a proposed fix for how Multus handles resource allocation to prevent Pod Unschedulable errors during high-churn VM deployments.

11-Feb: SIG-Compute Meeting

Primarily about ongoing work for vGPU live migration support. The focus was on ensuring state consistency and managing resource re-allocation on the destination node during the handoff.  The group also reviewed enhancements to CPU and memory pinning strategies to improve performance for high-load workloads

17-Feb: DRA Support Meeting

Featured a progress update on bringing Arm64 support to feature parity with x86. This includes ongoing work to ensure consistent behavior for VM live migration on ARM-based clusters.

18-Feb: SIG-Compute Meeting

A discussion about the rapid growth of VirtualMachineInstancetype objects and how to better organize these cluster-wide resources to avoid cluttering the API.

18-Feb:  Community Meeting

The community finalized the v1.8 release schedule, setting the feature freeze for early March. Key features, such as enhanced VM export capabilities and improved storage hotplugging, were reviewed for readiness.

KubeVirt Blogs and Videos

1-Feb: VDI on KubeVirt (Video) by Victor Toso 

In this talk from FOSDEM 2026, Victor explores leveraging KubeVirt to modernize Virtual Desktop Infrastructure (VDI) by orchestrating Windows and Linux desktops as declarative, versioned Kubernetes resources. He demonstrates how integrating open-source VDI brokers provides a scalable, cost-effective alternative to proprietary stacks while utilizing native features like high availability and live migration.

1-Feb:  EVPN Overlays for Multi-Cluster KubeVirt Deployments (Video) by Miguel Duarte and Federico Paolinelli

Miguel and Federico demonstrated using OpenPERouter with EVPN-VXLAN to provide seamless Layer 2 connectivity for KubeVirt VMs across multiple Kubernetes clusters. This architecture maintains MAC and IP address consistency to support cross-cluster live migrations and legacy workloads requiring broadcast or multicast support.

20-Feb: KubeVirtBMC: Enabling Bare-metal Provisioning for KubeVirt Virtual Machines by Zespre Chang

Zespre introduces KubeVirtBMC, an open-source Kubernetes operator that creates virtual Baseboard Management Controllers (BMCs) for KubeVirt VMs. By providing Redfish and IPMI services, it allows VMs to be treated like physical bare-metal servers, enabling them to be provisioned and managed by external tools like Cluster API or Metal3.  You can also get the code from Zespre’s Github repo.

24-Feb: Dynamic GPU Passthrough in KubeVirt by Peter Orosz

In this blog post, Peter explains how to implement dynamic GPU passthrough in KubeVirt to allow virtual machines to access hardware acceleration without being permanently tied to specific physical nodes. He details the technical configuration required to automate the attachment and detachment of GPU resources, so as to improve resource utilization and flexibility within Kubernetes clusters.

Upcoming KubeVirt Events and Talks

Ongoing:  KubeVirt Community Calendar 

23-26 Mar: KubeCon, Amsterdam, The Netherlands

The Cloud Native Computing Foundation’s flagship conference brings together adopters and technologists from leading open source and cloud native communities for all things Kubernetes, including best practices for platform engineering, application development, AI, data management, and much more!

23-Mar: VMs on Kubernetes Day, nhow, Amsterdam RAI

For platform engineers, Kubernetes operators, and infrastructure teams actively exploring or implementing VMs on Kubernetes. Learn the practical requirements to make VMs on Kubernetes successful in production, with real examples and an opportunity for hands-on experience. 

25-Mar: KubeCon 2026 – KubeVirt Benchmarking at Scale:  A Vendor-Neutral End-to-End Testing Framework for Cloud-Native Virtualization  by Bhumitra Nagar and Dhruv Bhatnagar, Portworx by EverPure

While the KubeVirt community has built a variety of tools for testing and observability of workloads using container-specific metrics, Bhumitra and Dhruv will introduce the first vendor-neutral, end-to-end benchmarking framework designed specifically for virtual machines. 

The KubeVirt Performance Benchmarking Suite is an open-source toolkit that enables operators, SREs, and platform architects to measure the performance characteristics of KubeVirt workloads across any Kubernetes distribution and any CSI storage backend.

26-Mar: KubeCon 2026 – KubeVirt on GB200: Virtualizing a Rack-Scale Supercomputer by Fan Zhang, Kevin Klues, and Alay Patel, NVIDIA

This talk explores how the NVIDIA GB200’s shift to a unified, cache-coherent memory architecture breaks traditional PCIe-based virtualization models, requiring a complete overhaul of the KubeVirt stack. It provides a practical roadmap for managing rack-scale orchestration, guest topology passthrough, and kernel-level integration to treat the entire rack as a single logical system.

Call for Contributors

Do you love working with KubeVirt? Have tutorials or best practices to share? Get in touch with us! We’d love to hear about your work and promote it.

Send us a short abstract of your topic. Be sure to be clear about your intended audience.
Who is this interesting for? Why? What will they get out of it?

Contact us at contributors@kubevirt.blog to get started!

Please note that we do not accept contributions that are promotional or commercial in nature (this roundup is focused on open-source) or AI-generated content.