Broadcom Bets Private Cloud Can Consolidate AI Infrastructure Without Extra Stack

The shift toward production-grade artificial intelligence is prompting enterprises to seek integrated solutions rather than piecemeal collections of AI tools. The industry's response centers on packaging AI capabilities into complete systems that accelerate deployment timelines.

The central premise is that the same infrastructure handling virtual machines and container workloads can also manage AI models and agents as a unified offering, rather than forcing organizations to build their own integrations. This perspective comes from Prashanth Shenoy, chief marketing officer and vice president of the VMware Cloud Foundation Division at Broadcom Inc.

We have reached a very critical juncture in the world of AI. A lot of our organizations are moving from pilot to production at scale, so there are big concerns around cost, tokenomics, security and privacy concerns of their data. A lot of our organizations are looking towards private cloud as the preferred platform for deploying their production AI workloads.

Prashanth Shenoy

Shenoy shared these remarks during an exclusive conversation with theCUBE's John Furrier at VMware Explore 2026, where discussions covered the VCF AI Factory, private AI services and frontier AI security.

Product marketing moves toward turnkey AI factories

When organizations must independently combine virtualization, storage, Kubernetes services and model repositories, the result is fragmented systems that delay time-to-value. Broadcom has pursued a different approach by certifying servers from multiple vendors, collaborating with Advanced Micro Devices Inc. and Nvidia Corp. at the processor level and working with MetalSoft Cloud Inc. on heterogeneous firmware and hardware initialization. The strategy aims to consolidate these elements into a single management layer.

All of this is integrated into the VCF Ops Console. The operations become easy, and the provisioning of this hardware and management gets reduced from months [or] weeks, to now minutes.

Prashanth Shenoy

The selection of models carries equal importance to hardware validation. Broadcom has tested and optimized roughly 150 models for VCF deployment, encompassing open-source, open-weight and proprietary variants, plus an AI gateway enabling connections between on-premises models and over 40 cloud-based model providers.

Not every use case that enterprises have requires a frontier LLM model. It's all about purpose, fit, governance and cost, which is very, very critical.

Prashanth Shenoy

Security represents the complementary pillar of this strategy, with frontier AI emerging as a focal point at this year's VMware Explore gathering. Broadcom strengthened VCF 9.1 from the foundation upward, shifted to monthly patch cycles and introduced new professional certifications. These efforts support the Frontier AI Security Readiness Program, which encompasses assessment, architecture, implementation and workforce development.

The volume, the velocity and the type of variety of these AI-driven threats have just exploded. Attackers don't take weeks or months. They take hours or minutes to get into the system.

Prashanth Shenoy

Prior decisions to streamline the VMware product lineup into a smaller number of offerings now appear prescient, particularly capabilities like memory tiering that were embedded before hardware availability constraints intensified. The advantage is that AI workloads do not demand a separate infrastructure footprint.

https://www.youtube.com/embed/KR9uXtz3-vY?feature=oembed

You don't need to create another siloed infrastructure. The same infrastructure that you've tried and tested for running your VMs – for running your containers – can be used to run your agents and AI workloads, with the same unified operations management and the security and data privacy.

Prashanth Shenoy