Research Finder
Find by Keyword
Broadcom Doubles Down on Open Source
A shift toward open-source stewardship and ecosystem flexibility aims to stabilize VMware Cloud Foundation for the enterprise AI era.
3/27/2026
Key Highlights
Broadcom has contributed the Velero backup project to the CNCF Sandbox to foster vendor-neutral governance.
The vSphere Kubernetes Service includes extended 24-month lifecycle support for minor versions as standard.
A new networking add-on framework allows for the native integration of third-party CNI implementations.
Partnerships with F5, Kong, and Tigera address the complexity of securing and managing microservices at scale.
Support for Red Hat Enterprise Linux 9 expands operating system choices for Kubernetes cluster nodes.
The News
At KubeCon Europe 2026, Broadcom announced several updates to the VMware vSphere Kubernetes Service (VKS) designed to simplify long-term operations and improve ecosystem interoperability. These include the move of the Velero project into the CNCF Sandbox and the release of VKS 3.6 with enhanced performance tuning and lifecycle management. The company also highlighted new validated architectures with partners such as F5, Kong, and Tigera to bridge the gap between traditional infrastructure and modern containerized workloads. Check out the Broadcom blog site for more information.
Analyst Take
The transition of Broadcom from a hardware-centric giant to a steward of the private cloud has been a journey marked by skepticism from many in the open-source community. However, the recent moves at KubeCon suggest a calculated effort to reposition VMware Cloud Foundation (VCF) as a foundational layer that plays well with others rather than a closed garden. We met with the team on the ground earlier today, and the dialogue was about community and ecosystem more than we have seen from Broadcom in a while, if ever.
We see this most clearly in the decision to donate Velero to the CNCF Sandbox. Velero is the standard for Kubernetes-native backup and restore; by moving it toward community-driven governance, Broadcom is signaling that it understands that enterprise trust in the cloud-native space is built on vendor neutrality.
It is a pragmatic move. According to HyperFRAME Lens data, 79% of enterprises anticipate having multiple foundation models concurrently deployed, signaling that multi-model and multi-platform architectures are becoming the standard. If Velero remains seen as a proprietary tool, its adoption might stall in these heterogeneous environments; as a CNCF project, it becomes the industry standard that Broadcom just happens to support best.
What Was Announced
The technical core of the announcement centers on vSphere Kubernetes Service (VKS) 3.6, which aims to deliver operational stability for "day two" and beyond. VKS 3.6 adds support for Kubernetes version 1.35 and includes 24-month extended support for each minor version. This is architected to allow large organizations to maintain their own upgrade timelines without being forced into the rapid-fire release cycles typical of the upstream project.
The update also includes a networking add-on framework designed to provide a supported path for partner-validated Container Network Interface (CNI) plugins, such as Cilium and Calico, to integrate natively with VKS. For data-intensive workloads, the introduction of declarative TuneD profiles aims to enable safe kernel and sysctl tuning, which is designed to optimize nodes for high-throughput networking and database performance without requiring unsupported host customizations.
On the security front, VKS 3.6 simplifies AppArmor profile management by allowing administrators to define profiles as Custom Resources, which are then automatically synchronized across worker nodes. The release also introduces centralized, API-driven management of node-level firewall rules, allowing platform teams to open ports for HostPorts and NodePort Services through standard cluster configurations. Furthermore, the addition of Red Hat Enterprise Linux (RHEL) 9 as a supported operating system for cluster nodes provides enterprises with more choice in their underlying Linux distribution.
We see these updates as a direct response to the "complexity gap" that often plagues platform engineering teams. By integrating with F5 BIG-IP Container Ingress Services, Broadcom aims to deliver automated, production-grade L4-L7 services that bridge traditional network appliances with dynamic container environments. Similarly, the validation of Kong Konnect on VKS is designed to provide a unified API management layer, which aims to govern traffic for both microservices and AI workloads with minimal latency. These integrations are not just about adding features; they are about reducing the "custom glue code" that often makes enterprise Kubernetes environments fragile and difficult to upgrade.
The focus on AI is also notable. VKS is now positioned as a CNCF-certified Kubernetes AI Conformant Platform, intended to ensure portability for AI workloads. By supporting NVIDIA’s Dynamic Resource Allocation (DRA) driver, donated by NVIDIA this week to the CNCF, the platform aims to automate GPU driver management, which is architected to make it easier for teams to deploy VCF Private AI services on top of their existing VMware infrastructure. This is a clear attempt to keep VMware relevant as the focus of enterprise IT shifts from generic virtualization to specialized AI infrastructure.
Looking Ahead
Based on what we are observing, Broadcom is moving toward a more sophisticated "co-opetition" model within the Kubernetes ecosystem. The key trend that we are going to be looking out for is whether the community truly embraces Velero under its new governance or if the move is perceived as too little, too late. Our perspective is that Broadcom is correctly identifying that the next battleground for the enterprise is not the orchestrator itself, but the management of stateful, data-heavy applications like AI and large-scale databases.
This transition is urgent; HyperFRAME research reveals a stark "Execution Gap" where only 23% of AI/ML projects launched in the past year reached production while meeting original ROI objectives. By stabilizing the underlying infrastructure, Broadcom hopes to help customers bridge this gap.
Going forward, we are going to be closely monitoring how the company performs on delivering on 24-month support cycles for VKS. While this is highly attractive to conservative enterprise customers, it places a heavy maintenance burden on Broadcom's engineering teams to backport security fixes across a wide range of versions. The announcement places Broadcom in more direct competition with Red Hat’s OpenShift and Nutanix’s Karbon, both of which have long touted their integrated networking and storage stacks.
However, Broadcom’s advantage remains the sheer ubiquity of the vSphere hypervisor. By making VKS feel like a natural extension of the vSphere environment rather than a bolt-on product, they are lowering the barrier to entry for traditional infrastructure teams. HyperFRAME will be tracking how the company does in converting its massive legacy install base to these modern VCF services in future quarters. Early signs are good, but we will get the first proper read as those initial 3-year contracts penned in the first few weeks of VMware’s new ownership come up for renewal. The wider success of this strategy will depend on whether they can maintain the "openness" they preached at KubeCon while still providing a seamless, vertically integrated experience.
Steven Dickens | CEO HyperFRAME Research
Regarded as a luminary at the intersection of technology and business transformation, Steven Dickens is the CEO and Principal Analyst at HyperFRAME Research.
Ranked consistently among the Top 10 Analysts by AR Insights and a contributor to Forbes, Steven's expert perspectives are sought after by tier one media outlets such as The Wall Street Journal and CNBC, and he is a regular on TV networks including the Schwab Network and Bloomberg.