From Cloud Reset to Private AI: Key Highlights from VMware Explore 2025

At VMware Explore 2025, a critical question was buzzing in the air: What happens when the hype around generative AI meets the reality of enterprise infrastructure? It’s a question that has been echoing in boardrooms, IT departments, and on the show floor in Las Vegas, and the answers could shape the future of business tech.

Right now, businesses are transitioning from dabbling in AI to figuring out how to make it work at scale. The challenge is clear—deploying AI across entire organizations without sacrificing privacy, breaking the bank, or losing control. But here’s the catch: these big discussions aren’t happening in the shiny, remote data centers of hyperscalers. They’re happening much closer to home—on-premises, or at the edge, where businesses are already working with real data and real-world constraints.

That’s where the conversation around “Private AI” is really taking off. At VMware Explore, it wasn’t just a buzzword; it was the centerpiece of the event.

Key Highlights from the Show:

• Private AI Takes the Spotlight: The buzz around Private AI was hard to miss, as it emerged as the key focus at VMware Explore. As organizations seek control over their AI models and data, the need for privacy-first AI solutions has never been more urgent.

• VCF 9: AI Becomes a Native Service: VMware Cloud Foundation (VCF) 9 is making AI a natural part of its infrastructure. With built-in privacy features, GPU virtualization, and a developer-ready setup, VCF 9 is positioning itself as the ideal platform for enterprises looking to integrate AI at scale—without compromising their core values.

• VCF Intelligent Assist: Generative AI for Cloud Operations: One of the most exciting innovations was VCF Intelligent Assist. This tool demonstrates how generative AI can streamline private cloud operations. It offers contextual, task-aware guidance, making day-to-day management more efficient and reducing the complexity of maintaining cloud infrastructure.

• Agentic AI and Model Context Protocols: These new advancements in AI are transforming how systems integrate across different platforms. By making workflows smoother and more intuitive, Agentic AI and model context protocols finally allow for seamless cross-platform integration, which has been a major hurdle in the past.

• A Cloud Reset: Enterprises Repatriate AI Workloads: As more companies bring AI workloads back in-house, a “cloud reset” is underway. Security, data sovereignty, and cost efficiency are at the heart of this shift, with organizations becoming more cautious about outsourcing their most sensitive operations. The pressure to maintain control over data is rising, and the consequences of getting it wrong could be significant.

What is Private AI & Why Is It More Important Than Ever?

Private AI is more than just a trend or a technical preference—it’s a reaction to the tough choices businesses have been making for years. In the rush to explore AI, many organizations turned to public cloud services to experiment, build prototypes, and get things off the ground.

At first, cloud GPUs were a lifesaver—easy to access, with all the tools you needed right there. But as companies moved beyond the initial testing phase, they started to run into serious problems.

That’s where Private AI comes in. It’s a game-changer for businesses who want to build and run AI workloads on their own infrastructure—giving them full control over everything. It’s not just about keeping things in-house for the sake of security. It’s about protecting intellectual property, securing sensitive data, and keeping everything close to the source. By running AI models on private servers, companies can shield themselves from public exposure while still delivering AI-as-a-service capabilities that feel just like the cloud.

Tasha Drew, Director of Engineering for AI at VMware Cloud Foundation (VCF) at Broadcom, shared a clear message at VMware Explore in Las Vegas: the real pillars of Private AI are built on practicality and protection. She said:

“Protect your IP. Safeguard your data. And if your edge environments are producing massive volumes of data, it doesn’t make financial sense to send all that to the public cloud for AI processing.”

What’s fascinating is that these ideas are no longer abstract or theoretical. They’re now becoming the very foundation of how AI infrastructure is being designed and built.

The whole vibe at VMware Explore underscored the shift. When Broadcom’s CEO, Hock Tan, took the stage, the spotlight was firmly on the move to private cloud solutions. It’s clear—Private AI isn’t just a nice-to-have. It’s becoming essential for businesses that want to scale responsibly, with privacy, cost-efficiency, and control in mind.

From Sandbox to Service: Why Internal Delivery of Private AI is Essential

In today’s fast-paced tech world, enterprise teams are doing more than just safeguarding data—they’re actively looking to operationalize AI. But here’s the thing: simply launching a large language model (LLM) and running a few queries doesn’t cut it anymore. The true power of AI comes from embedding it deeply into everyday workflows, all within a shared, internal platform. That’s where the idea of Private AI-as-a-Service really starts to shine.

But the challenge is real: where do you even start?

Enter VMware’s latest innovation, which aims to solve that problem.

Drew summed it up like this:

"We’ve introduced VMware Private AI Services to VCF. This allows enterprises to build an AI-powered private cloud platform that’s native to their infrastructure. It’s more than just launching a cluster; it’s about embedding AI into the heart of your operations."

VCF Intelligent Assist: A New Era for Cloud Operations

At VMware Explore, one of the most exciting updates was the unveiling of VCF Intelligent Assist, an AI-driven tool now embedded in VMware Cloud Foundation. It’s a game changer because we’re no longer talking about just another chatbot. This is a functional assistant that leverages generative AI to make private cloud operations more intuitive and efficient.

Tasha Drew explained it this way:

"When we talk about ‘intelligent assist,’ we mean it knows exactly where you are in your workflow, understands the context, and delivers the right information—be it documentation, next steps, or actions—to help you move forward. It’s built on the same Private AI services we offer to our customers, so it’s an open, flexible platform."

What makes this stand out is its openness. VCF Intelligent Assist doesn’t just stop with the built-in features; it can be extended to fit your unique enterprise needs. If your team has been maintaining PDFs with detailed setup instructions, compliance checklists, or port policies—now those documents can be indexed and surfaced directly within the platform. Imagine having all that critical information available in-context while you work, seamlessly integrated into your daily operations.

As Drew put it:

"This is about making private cloud easier to manage by using the very same generative AI capabilities that we offer our customers."

In a world where AI is no longer just a tool but a core element of business strategy, being able to deliver it as a native part of your infrastructure—internally, securely, and with ease—could be the competitive edge that many businesses need to stay ahead.

The Rise of Agentic AI and Real-World Workflows: A Game Changer for IT

The future isn’t just about searching for information faster. It’s about automating the way we work. VMware is making a bold shift toward agentic workflows, where AI agents can take over complex, multi-step tasks across various systems. This is where the Model Context Protocol (MCP) comes into play.

Tasha Drew, VMware's visionary, shared her excitement about what’s to come:

“For me, generative AI is a game changer. It finally gives us a practical way to solve the challenges of systems integration. Think about it – Salesforce doesn’t have much incentive to integrate deeply with ServiceNow. But now, both systems can publish MCP servers with agent actions, and suddenly, we’re able to create meaningful workflows across platforms.”

What does this mean in real terms? Cross-platform approvals, automated issue resolutions, or even intelligent task routing— all automated, secure, and with complete control. And here's the best part: MCP isn’t just some idea; VMware is already embedding this into its VMware Cloud Foundation (VCF), making it available to both customers and internal teams.

It’s not just about adding more tools; it’s about creating a seamless, interconnected ecosystem for enterprise IT.

Building Stronger Ties with Nvidia & AMD: The Hardware Backbone of AI

At the heart of any AI strategy is the hardware that powers it. VMware recognizes that reality, which is why it has heavily invested in partnerships with Nvidia and AMD.

Tasha makes the stakes clear: “Nvidia has been a crucial partner. We’re co-developing solutions to ensure that Private AI runs seamlessly, supports lifecycle management, and allows customers to deploy everything from NIMs to NeMo Retriever with just a few clicks.”

With VCF, AI tools run on GPU-virtualized infrastructure that’s designed for high-demand, enterprise-scale workloads. And this isn’t a "figure-it-out-yourself" setup; VMware has optimized everything for effortless deployment at a massive scale.

VMware’s relationship with AMD is also growing, particularly with the MI300 processor series. Tasha elaborates:

“We’re extending Private AI Services to AMD hardware, so customers aren’t locked into a single ecosystem. This aligns perfectly with Broadcom’s vision of openness, flexibility, and integration.”

Why More Enterprises Are Turning to Private AI with VCF

Privacy concerns have always been a driving factor for adopting private AI, but VMware’s customers are increasingly coming for more fundamental reasons: cost control and efficiency.

Tasha explains: “Sure, privacy matters, but some teams are discovering the true value when they compare their cloud AI costs. We’ve had clients who look at their bills, run the numbers, and say, ‘From a financial standpoint, we can’t afford not to switch.’”

VMware makes it easy to compare costs between private and public AI infrastructure with their online cost calculators. But the value isn’t just in the savings; it’s about efficiency. When each team runs its own AI models independently, GPU usage drops, and redundancy increases.

By consolidating everything on a single, internal platform powered by VCF, businesses improve GPU utilization, eliminate unnecessary duplication, and gain better governance—all while reducing costs.

This is exactly the kind of practical, results-driven approach that IT teams, who are constantly under pressure to deliver, need. And with VCF’s support, they don’t have to compromise on security, control, or budget.

The Cloud Reset: A Transformation in Progress

While walking the show floor, we had the opportunity to catch up with Prashanth Shenoy, CMO and Vice President of Marketing at VMware’s Cloud Foundation Division at Broadcom. He shared some insightful thoughts on how AI is evolving within the enterprise space.

He put it simply:

“Last year, AI was still in the pilot phase. People were asking, ‘Show me the value, let’s see the use cases.’ But now, it’s no longer just an experiment. AI is running the business, driving real outcomes, and scaling.”

This shift represents something much bigger, which Prashanth called a “cloud reset.” For years, moving to the public cloud was the go-to solution for many organizations, especially as they started their digital transformation journeys. It made sense back then. But now, as AI workloads grow more complex, the cloud is being reevaluated.

Prashanth shared with us, "Nearly 70% of the companies we talk to are either bringing workloads back on-prem or planning to. About half of them are even building new workloads, like AI, directly on private clouds."

In other words, as businesses push AI to the forefront, they're demanding more control, predictable costs, and better data sovereignty. The growing power of AI is forcing a major rethink of where and how it operates.

But there's more to the story than just technology. A key point Prashanth was passionate about is often overlooked: the people who manage these systems.

“We’ve seen it over and over—if operations teams are stuck in outdated silos, like compute, network, and storage, it's almost impossible to modernize the private cloud effectively,” he pointed out.

That’s why VMware is focused on skills transformation. “We’re not just preparing businesses for the next project,” he said. “We’re preparing them for the next decade. It’s about helping them build, deploy, and operate a private cloud platform that’s future-ready.”

The Bottom Line

What started as a strategy to cut costs and meet compliance demands has evolved into a core element of enterprise transformation. Today, companies want to develop and deploy AI models within their own environments. They need smart, contextual assistants. And they want interoperability between vendors, without adding unnecessary complexity.

At VMware Explore 2025, the buzz on the show floor wasn’t about flashy tech features—it was all about outcomes.

Tasha Drew nailed it when she said:

“We’re not just building cool tech. We want to know: what do you want the technology to do for you?”

That’s the right approach. As AI continues to redefine business operations, its success will be less about where it’s deployed, and more about how seamlessly it fits into the daily work of people across organizations.

FAQs:

1. What is Private AI, and why is it important?

Private AI gives businesses the ability to run AI workloads on their own infrastructure. This approach helps protect sensitive data, saves costs, and gives companies full control over their AI services. In a world where data security and cost efficiency are more critical than ever, private AI is a game-changer.

2. How does VMware Cloud Foundation support Private AI?

VMware Cloud Foundation seamlessly integrates Private AI, making it easier for businesses to handle AI tasks securely. With built-in privacy features, GPU virtualization, and smart assistants for cloud operations, VMware’s platform streamlines AI workloads, ensuring both efficiency and safety.

3. Why are enterprises moving from public cloud to Private AI?

Many businesses are shifting away from public cloud services due to escalating costs, growing concerns about data sovereignty, and the desire for more control. Private AI platforms are becoming the go-to solution, giving enterprises the freedom to manage their AI workloads on their terms while reducing risks and overheads.

References:—

1. Exciting News for Private AI at VMware Explore 2025 (Broadcom)