One way to understand Prashanth Shenoy's job is to picture an enterprise cloud diagram. Over time, the diagram acquired more boxes: public clouds, private clouds, virtual machines, containers, data services, accelerators, security tools, management planes. Every box promised simplicity. Together, they produced a new kind of complexity. Shenoy, the CMO and vice president of marketing for Broadcom's VMware Cloud Foundation division, now has to turn those boxes into one story.
The story is a reversal with an asterisk. For a decade, the default direction of travel was out of the company data center and into a hyperscaler's cloud. Shenoy argues that production AI changes the calculation. Inference is expensive. Sensitive data has gravity. Sovereignty rules matter. Public-cloud experimentation can become a surprisingly large bill when it grows up. In his telling, private cloud means a cloud operating model on infrastructure an enterprise controls.
That is a difficult sentence to sell because it asks buyers to separate a place from a way of working. Shenoy's preparation for the task began far from the marketing department. He arrived at Cisco in 2000 as a senior software engineer. Five years later, he moved into marketing management. He did not leave the technical world so much as change which layer of it he worked on.
The engineer who kept the wiring diagram
Shenoy holds a master's degree in computer science from Texas A&M University and an MBA from UC Berkeley's Haas School of Business, completed while his Cisco career was gathering speed. The combination became his professional grammar: understand the system, then explain why the system matters to a buyer.
Over more than 22 years at Cisco, his remit widened from engineering to product management and marketing across enterprise networking, data centers, cloud, mobility, IoT infrastructure, and developer products. By the end, he was the global marketing head for a cloud infrastructure and software portfolio that CRN described as a $25 billion business. He had led a team of 80 marketers in 25 countries and helped introduce Cisco's Future Cloud automation work and Cisco Plus offerings.
Their expertise is adding value, elevating both their career and their organization to a new level.Prashanth Shenoy, writing at Cisco
Former colleague Greg Dalzell once identified the useful skill beneath those titles: Shenoy could absorb a technology and quickly grasp its value for a customer. Dalzell also pointed to honesty, work ethic, and a collaborative nature. Those are LinkedIn-recommendation words, but here they describe a specific kind of operator. Infrastructure products are too interdependent for the marketer to bluff. A claim about a platform eventually reaches an architect who can see every loose wire.
He joined VMware before the floor moved
In August 2022, Shenoy announced that he was joining VMware as vice president of cloud platform, infrastructure, and solutions marketing. His new brief stretched from on-premises private cloud to hyperscaler and service-provider clouds. Mark Lohmeyer, then a VMware senior vice president and general manager, pointed to Shenoy's technical background and understanding of customer problems. His official debut was set for VMware Explore later that month.
The timing was almost theatrical. Broadcom had already agreed to acquire VMware. The deal closed in November 2023, and the company Shenoy had joined became the center of a much sharper operating model. Product bundles changed. Perpetual licensing gave way to subscriptions. The partner network was narrowed. Broadcom concentrated VMware's sprawling catalog around VMware Cloud Foundation.
Customers and partners did not experience that as a tidy slide transition. They objected to parts of the new pricing and packaging, and they asked what kind of company VMware would become. Shenoy's public language met the discomfort more directly than corporate transformation language usually does. “Change is never easy,” he wrote in February 2024. Speaking with RedMonk's James Governor about the first year and a half of integration, he joked that 18 months felt like five years.
Although, as you said, 18 months, for me, it feels like five years.Prashanth Shenoy, in conversation with RedMonk
The joke works because it concedes the obvious: simplification can be complicated for everyone living through it. Shenoy had to announce the new portfolio and make its choices intelligible. In a 2025 interview, he emphasized VMware's focus on being a product and platform company, with partners still important and the product sitting at the center.
The pitch behind VMware Cloud Foundation is less about a single workload than a shared way to run several kinds of work.
AI gives private cloud a new argument
VCF 9.0, launched in 2025, gave Shenoy a product around which to organize the thesis. Broadcom describes it as a unified platform for traditional applications, cloud-native applications, and AI. Cloud administrators get tools to build and operate infrastructure. Developers get self-service access. Virtual machines and Kubernetes containers can live under a consistent operating model. Private AI work with partners including NVIDIA adds another reason to keep models and data near controlled infrastructure.
The argument became more pointed in Broadcom's Private Cloud Outlook 2026, a blind survey of 1,800 senior IT leaders. Fifty-six percent said private cloud was their preferred environment for production AI inference. Sixty-two percent were very concerned about the cost of agentic and generative AI. Eighty-three percent were considering moving workloads from public to private cloud, while half said they had already repatriated some.
Private Cloud Outlook 2026 · 1,800 senior IT decision-makers
Survey percentages do not settle an architecture debate. They do show why the old vocabulary feels incomplete. “Cloud first” sounds decisive until finance asks which cloud, security asks where the data sits, and an application team asks how to get a GPU. Workload placement becomes a portfolio decision, not an article of faith.
In the 2026 survey, 97% perceived waste in public-cloud spend. Half had moved some workloads back, and 39% cited cost predictability as a repatriation driver.
Shenoy adds an organizational layer. In a June 2026 essay, he located the gap between enterprise AI ambition and results in infrastructure and operating models that have not kept pace. The proposed answer links three chores that companies often separate: modernize existing applications, mature the private-cloud platform, and align infrastructure teams with platform engineers who package capabilities for developers.
This is where his engineer's view and marketer's view meet. In his framing, AI readiness is a chain of dependencies: applications, skills, governance, operations, hardware, and money. The public-facing story has to be simple enough to remember without pretending the work itself is simple.
He keeps returning to the people running the system
In interviews, Shenoy tends to move from platform architecture to the people expected to operate it. The traditional virtualization administrator, he says, has to become a cloud operator. That means learning architecture, implementation, automation, networking, storage, and Kubernetes. Broadcom has attached certification and training programs to that transition. The strategy needs the installed base's knowledge, but it asks that base to stretch.
That is also the human tension in Shenoy's own career. He changed functions without throwing away the old function's way of seeing. Engineering taught him to decompose a system. Product management put choices and tradeoffs in view. Marketing made him compress the result. Executive leadership turned that compression into an operating rhythm shared across countries, products, partners, and stages.
The private-cloud pitch will be judged by customers who can measure costs, deployment times, reliability, and developer experience. It will also be tested by the history attached to VMware and by buyers who remember every abrupt change. Shenoy cannot erase that context. His more credible move is to acknowledge it, then make a narrower promise: fewer disconnected products, one platform, and more deliberate choices about where work runs.
Cloud computing was once sold as freedom from the data center. Shenoy's current work asks whether freedom might instead mean freedom to place a workload where its economics and constraints make sense. For production AI, he believes that answer is increasingly private. The lasting idea is broader. Infrastructure fashion moves in cycles, but the operator still has to make the diagram work.