The Rise of Private AI
At VMware Explorer, Broadcom made an announcement that could shape the next decade of enterprise AI. Chris Wolf, a leader within the company, called out the growing need for private AI. His message was clear: as excitement for generative AI continues to grow, enterprises must think about governance, regulation, and the secure handling of sensitive data.
Tasha Drew , who leads the engineering team responsible for private AI services at Broadcom, explained why this matters. “We’re talking about models trained on private data,” she said. “We need role-based access control, we need to protect intellectual property, and we need to make sure nothing unsafe is injected into sensitive codebases.” Broadcom
Building AI Into VMware Cloud Foundation
VMware Cloud Foundation (VCF) now integrates a complete AI platform. This includes a model gallery, runtime services, ML API gateway, data indexing and retrieval tools, and the Intelligent Assist features that Broadcom uses internally.
The goal is to give enterprises an AI-ready infrastructure that is both powerful and manageable. Expensive GPU resources can be scaled across teams. Models can be governed. Data can flow seamlessly into vector databases for retrieval-augmented generation (RAG) applications.
As Tasha noted, Broadcom is not just providing these tools; they are using them. “We’re dogfooding our own platform every day. That’s the best way to make sure it works in real-world conditions.” Broadcom
Moving Beyond Bare Metal
When generative AI exploded, many assumed all workloads would remain on bare metal. That is still true for training, but production is different.
“Bare metal is fine for experimentation,” Tasha said. “But once something goes into production, customers want it running on VCF.” Broadcom
The reason is simple. Enterprises need the ability to schedule downtime, migrate workloads, and scale services. VMware has been delivering that for twenty-five years. Now those same strengths apply to AI.
An Expanding Ecosystem
Partnerships are central to Broadcom’s strategy. At VMware Explorer, announcements highlighted support for AMD, Intel Gaudi, Canonical, and Ubuntu.
This reflects an open ecosystem approach. “We want customers to adopt best-of-breed technology in a seamless, manageable way,” Tasha said. broadcom
That flexibility allows enterprises to choose the accelerators and Linux distributions that fit their needs without compromise.
Agentic AI and Autonomous Operations
The industry is buzzing about agentic AI and the Model Context Protocol (MCP). Broadcom is building these directly into VCF.
Upcoming releases will bring OAuth integration, distributed tracing, and curated tool galleries. This will allow enterprises to create secure, controlled workflows that automate common tasks.
As Tasha explained, even a simple workflow can save time. “I want my users to ask in Slack, list VMs, and get that list without checking spreadsheets.” Broadcom
Customer Stories That Inspire
Customer use cases show how powerful this shift is becoming. Walmart is standardizing on VCF for AI. Vehicle manufacturers are running air-gapped workloads, with racks dedicated to each production line.
These examples show private AI is not an experiment. It is a production reality across industries.
The Road Ahead
Broadcom is focused on execution. Private AI is now included with VCF. MCP features are being added to Agent Builder. A tech preview of Intelligent Assist is scheduled for release early next year.
As Tasha said, “We’re heads down delivering. There’s a lot of excitement, and we’re ready to put private AI into customers’ hands.” Broadcom
Broadcom’s vision is simple but powerful: the future of enterprise AI is private, secure, and built on VMware Cloud Foundation.
