Broadcom Contributes API for Virtual Machines to CNCF

Broadcom today, at the KubeCon North America 2026 conference, revealed that it has contributed an application programming interface (API) for virtual machines (VMs) to the Cloud Native Computing Foundation (CNCF).

Additionally, Broadcom is adding its weight to a Workload Conformance Working Group launched today that seeks to establish a shared, objective standard for ensuring workloads deployed on Kubernetes clusters are truly portable.

Prashanth Shenoy, chief marketing officer for the VMware Cloud Foundation (VCF) Division at Broadcom, said contributing the VM API as a sandbox-level project is a step toward ensuring that VMs are first-class citizens within cloud-native computing environments. At the core of that effort is Sonobuoy, a diagnostic engine developed by Broadcom that is already being used to certify distributions of Kubernetes adhere to another conformance program administered by the CNCF.

Long used to deploy monolithic applications, many IT teams actually deploy Kubernetes clusters of VMs to ensure workloads are isolated. In fact, a report published today by the Linux Foundation finds VMs account for 53% of production applications, with just over a third of respondents (34%) expecting to expand their VM footprint as they deploy AI and agentic workloads.

While a VM introduces additional overhead, its impact on most applications has been proven to be negligible. Broadcom is now making a case for standardizing the APIs to make it simpler to provision them declaratively, regardless of whether they are proprietary or based on an open-source project.

Broadcom, since acquiring VMware, has been making a more aggressive case for centralizing the deployment of multiple classes of workloads on top of a VMware Cloud Foundation (VCF) platform that combines a hypervisor with virtual networking and storage services that a Kubernetes cluster can be deployed on. The overall goal is to reduce the total cost of IT by making it simpler to centralize the management of monolithic, cloud-native and, more recently artificial intelligence (AI) workloads running on a single platform. Otherwise, IT teams will find themselves managing different platforms for each of these types of workloads. As part of that ongoing effort, Broadcom has now validated 30 offerings from partners to run on VMware vSphere Kubernetes Service (VKS), its specific distribution of Kubernetes that has been optimized for VCF.

It’s not clear how many cloud-native workloads have been deployed on a distribution of Kubernetes, but Broadcom claims more than 2,600 organizations have licensed VCF 9, which has been deployed across more than 24 million processor cores. The expectation is that as more of those organizations deploy cloud-native and AI workloads the number of IT teams that opt to deploy VKS will also continue to increase. Of course, many IT teams in the wake of changes to the terms under which VMware software is made available are also considering all their options.

Regardless of what platform is ultimately deployed, the one thing that is certain is the total number of workloads deployed in the AI era is only going to exponentially increase. The issue then becomes how best to deploy them all without breaking the proverbial bank.