KubeVirt graduates CNCF, targets AI and confidential computing
KubeVirt—the open-source project that brings virtual machines into Kubernetes clusters—just graduated from the CNCF. Red Hat engineers are now steering it toward AI workloads, confidential computing, and dynamic GPU allocation with direct contributions from Nvidia.
Editor, Lazyfounder
30 SEC SUMMARY
- KubeVirt, an open-source tool for running virtual machines inside Kubernetes, just graduated from the Cloud Native Computing Foundation.
- Red Hat engineers are now steering KubeVirt toward AI workloads, confidential computing, and GPU allocation with help from Nvidia developers.
- The project lets teams run VMs and containers side-by-side on the same Kubernetes cluster—no duplicate infrastructure required.
- Confidential computing is making a comeback as companies look for secure ways to handle sensitive AI tasks, especially in the public cloud.
- Red Hat is also eyeing KubeVirt as a way to isolate AI agents, using features like live migration that other backends can’t match.
TABLE OF CONTENTS
KubeVirt graduates CNCF, pivots to AI and secure workloads
KubeVirt—the open-source project that lets virtual machines run inside Kubernetes clusters—has graduated from the Cloud Native Computing Foundation. The milestone arrives as Red Hat engineers refocus the project on AI workloads, confidential computing, and secure sandboxes for AI agents.
VMs and containers converge on the same cluster
KubeVirt was built to end the split between VMs and containers. Instead of maintaining two separate stacks, teams can now run both on a single Kubernetes cluster. Nvidia engineers are helping extend that idea, adding dynamic GPU allocation so the same cluster can spin up training jobs and inference services as needed.
Red Hat maintainers Itamar Holder and Vladik Romanovsky call the pairing complementary. VMs handle isolation and specialty hardware setups; containers deliver speed and scale. For AI workloads, that combination matters more than ever.
Holder also wants KubeVirt to become the isolation layer for Red Hat’s AI agent sandbox. Its live-migration feature—something lighter-weight backends lack—lets workloads move between nodes without interruption.
Confidential computing stages a comeback for AI
Confidential computing is staging a comeback. By locking sensitive workloads inside hardware-encrypted enclaves, it gives enterprises a way to run AI tasks on public cloud infrastructure without trusting the provider. Holder believes this could finally unblock regulated industries that have held back because of data-security concerns.
A broader move toward open, secure AI infrastructure
The shift fits a larger push to make AI infrastructure both open and secure. Red Hat and Nvidia recently launched OpenShell, an open-source project under CNCF that locks down AI agents by default. Meanwhile, Broadcom open-sourced its Kubernetes VM API, signaling broader demand for tools that straddle the VM-container divide.
What this means
Lazyfounder analysis — our interpretation, not reported fact.
KubeVirt’s graduation signals that the cloud-native world is embracing virtual machines—not as relics, but as secure, flexible companions to containers when AI workloads demand extra isolation or specialty hardware.
Key takeaways
- Teams can now consolidate VMs and containers on the same Kubernetes cluster, cutting infrastructure sprawl.
- Dynamic GPU allocation—backed by Nvidia—lets the same cluster handle both training and inference workloads.
- Confidential-computing support could open the door for regulated industries to move sensitive AI workloads to the public cloud.
- KubeVirt’s graduation adds weight to the argument that VMs and containers are partners, not rivals.
Related on Lazyfounder
Sources
- SiliconANGLE · 2026-11-12
A decade in, KubeVirt pushes virtual machines into the AI agent era
This story is an original summary drafted with AI by Lazyfounder from the reporting listed above and checked by automated validation. Facts are attributed to their original publishers; sections marked as analysis are Lazyfounder's. Where a source is in another language, facts were machine-translated and quotations are reported, not reproduced. Read the original coverage via the links, and see our AI policy and corrections policy.
About the author
Editor, Lazyfounder
Tarun Mottlia edits LazyFounders, covering Indian startups, funding rounds, AI and product launches. Every story on the site is AI-assisted and checked against its cited sources before publication.
More stories by Tarun MottliaGet the LazyFounder Brief
Startup, funding and AI news in a five-minute read. Join the early-access list.
