Kubernetes, AI and Hybrid Multi-Cloud Strategies with Himanshu Singh – KubeCon Europe 2025

This is Textron tv. Hey everyone. Alan Shiva here at Cube Cod.

My next guest is Hemanchu Singh. Hemanchu is with Broadcom. He's with the, uh, the, uh, VMware unit within Broadcom.

Haman, you welcome to techron tv. How are you Doing? Fantastic.

Thank you for having Me. Are you having fun here? Absolutely.

Yeah. I mean, KU coupon's got like massive energy. Uh, so we are really excited to be here, represent, you know, what Bmware by Broadcom's doing, and just talk to everybody and like the response from everyone has been just fantastic.

Good. Glad to hear it. Soman, what, what do you do at Broad?

I, I mentioned you're a Broadcom VMware, but what do you do there? So, my role is product marketing. I'm director of product marketing for our compute and AI and, uh, Kubernetes, uh, you know, portfolio essentially.

And so as part of that, you know, my team looks at for, uh, you know, vSphere as our kind of the, the main product that's been around for a long time. Also, our, uh, VMware private AI foundation within Vidia offering. Uh, that's, you know, on the gen AI kinda space.

Yeah. Then all the work we are doing now, and it's probably gonna be most about this, this conversation is Kubernetes as part of the VMware Cloud Foundation offering that we have. Yeah.

And that is, you know, some people used to think that was an oxymoron, right. It didn't, but that's changed over the last years. Right.

I, I, you know, as a matter of fact, so I remember when containers first came out and the discussion was, why would you run containers when you have a hypervisor? Well, you could get a lot more density with containers. That was the, that was the, you know, the throwaway line.

Yep. Yeah. But then I, I think, I don't know, four or five years ago, COVID times, maybe people realized that running containers on top of the hypervisor was not a ridiculous thing.

It was actually a good thing. You got, in many ways, you got the best of both worlds. Yeah.

Um, VMware kind of ran with that notion, and, and you guys have been steadily working on it since then. As we sit here today, what's the state of running VMware on in a container or running a containerized environment on top of VMware? So, As you said, right, I mean, um, it was, we, we kind of got started into this, like pre covid in fact, and, um, made some announcements.

I remember like 2019, if I'm not wrong, uh, around basically replatforming core vSphere and really integrating Kubernetes into it. That's, that was our kind of big change. We, before that we had like V two integrate containers, et cetera.

Then we kind of switched over to like really going after it and replatforming vSphere, and we brought out, uh, V with Kubernetes and, uh, we cloud foundation with Kubernetes as well. And so from that point on, we have been investing in this space very strongly. Lots of partnerships, et cetera as well.

And, uh, I mean, today where we are is, you know, um, especially with the whole kind of Broadcom acquisition that has happened, we've really brought together all these technologies and components together wherein we are able to provide customers one platform to run both VMs and containers as first class citizens. And, uh, you know, provide like a really fantastic experience of what they, you know, what they used to, for example, in the public cloud, be able to get that kind of experience in a, in a private cloud environment essentially, and build their modern apps no matter what components they want to use. I love it.

Absolutely. Now, look, I'm not going to sugarcoat it. There's a lot of people who say, well, you know, since VMware became part of Broadcom, licensing is changing.

We don't wanna be locked in and no one wants, and this is not just a VMware thing. No one wants to be locked into anything. Everybody wants to keep their options open.

It's one of the big attractions of open source, but yet I've seen studies where the licensing changes aren't really the, the real driver, the real driver are people are looking at options around cloud today. Right? This isn't 2010 people aren't shifting, lifting and shifting and moving everything into a public cloud.

Not anymore. We live in a hybrid multi-cloud world. Cloud native world.

Yeah. There people have, people want options and have options and, you know, and, and part of that cloud native world is do I re-architect? Do I digitally transform for my apps, modernize my apps?

Is the word, the code word? Right? Sure.

How do you, at VMware at Broadcom, what's your message to those people? Right. We know you got options.

Why, why go here? Yeah. So there's a few few things in there, right?

Um, and, and yes, we had a lot of like initial kind of knee jerk reactions from customers who did see that as a shock to the system, and right. The changes initially got announced with the acquisition and everything. I think people have kind of moved past that, uh, you know, majority of customers because what they're looking at now is, you know, VCF as kind of the, the one offering we have, you know, as the primary offering we have, at least at this point, and the value it is delivering versus whatever price point, et cetera, we are gonna talk about, right?

So, and as they look at what VCF is today compared to, you know, options that are available in the marketplace, there is, uh, our, you know, there's a massive amount of value that we're delivering with everything coming together. Now, the, the key piece is we had these components before, but they were individual products. Yes.

We never did the job of like bringing them together. Yeah. So VCF used to be a bundle.

Mm-hmm. Now it's a product, right? And that's basically that big jump that we've made as a company.

And, you know, oktan, our CEO has made sure that we are making that jump as well. And I think, um, as we go forward, customers will be, you know, the better for it because now they have a really unified platform to be able to, you know, run all their workloads, et cetera. Now you talk about kind of options and you know, where you deploy things and where do you wanna modernize.

I think based on the conversations we're having with customers, most customers are looking at, you know, being able to, to have that flexibility of figuring out for specific types of like applications or workload, what's the right destination to deploy. At the same time, you've got this, this kind of pressure of, I don't want to create individual silos that are gonna add more complexity. They're gonna add unique tool sets and, you know, unique skill sets, requirements, et cetera.

Um, so the more I can standardize, uh, in terms of a core platform, be able to reuse the skill sets that my people have, be able to reuse all the tools I have while getting the benefits of like, Hey, I want this to be deployed at the edge, this thing in a hyperscaler, this thing in my on-premises private cloud. Um, and, and that is what VCF is able to deliver now through all the partnerships we have, you know, with all the kind of the big, uh, hyperscalers with our edge offering, with our on-prem kind of environment. So I think we are in a really good position to offer customers an extremely viable option for their entire environment.

And, and like the, the kind of response we're seeing that adoption we're seeing from customers is actually a testament to that. I love it. That's good stuff.

Um, you mentioned you were involved a little bit in ai. Mm-hmm. There's a lot of AI going on or being talked about.

I don't know if it's actually going on. Yeah, yeah. How is that playing into this whole, what should I do?

I don't want to be locked in. I've, you know, where, where should I go discussion? So one thing I'll say is, and it's kind also touches to your, the, the previous one is, as VMware, we always believed in building the broadest ecosystem you can, right?

Yes. On the core vSphere platform, we absolutely had the broadest ecosystem. VCF does the same thing.

Our philosophy is the same when it comes to AI as well, right? So one thing we did, and we kind of put the stake in the ground of, uh, the most paramount thing when it comes to gen AI specifically is privacy of the customer's data, right? Uh, the access to that data, all that IP and all that needs to be protected.

We've all kind of heard all these stories where, you know, things being new customers made mistakes and internal IP kind of went into the public domain and all sorts of stuff, right? And that kind of highlights and underscores why we need to make sure that, uh, you know, privacy is addressed at the same time, that same point I made earlier about standardization, right? We want the every customer to be able to treat their gen AI workloads as just another enterprise workload.

It should not become another silo with all sorts of idiosyncrasies and whatnot, right? You want to be able to go back to my same point. Doesn't matter what workload you're looking at, you want to be able to find the right platform for it, get the right tool sets, broadest ecosystem.

So you have the flexibility. So that goes against that whole lockin point, point, right? You standardize on the right infrastructure, be able to get all the options you want, the ML ops providers you want, the, the, you know, the language models you want, et cetera.

So like the offering we have, which is called, it's a mouthful, VMware private AI foundation within Nvidia, and our partnership with Nvidia goes probably more than a decade at this point. Mm-hmm. So this offering essentially kind of brings together the full stack end-to-end offering of, you've got kind of the NVIDIA certified hardware with Nvidia GPOs underneath, you've got the VMware Cloud Foundation layer on top of that.

Uh, and then above that you've got the Nvidia AI Enterprise suite running with all their NIMS and all their blueprints and everything, right? So the, all the engineering work that VMware or Broadcom and, and Nvidia teams are doing together, at the end of the day, you know, customers again, benefit with that. And the idea of making sure that VCF as a platform is providing that security, that compliance and addressing the privacy needs is, is, you know, again, we, we treat that as paramount when it comes to it.

And again, and also making sure that we are not adding, you know, uh, complexity to the customers environment already. GNAI is new, the data scientists are already busy doing, you know, complex work. We don't want core IT platform teams to find new complexities.

And this, we want them to be able to just operate it smoothly, just becomes part of the enterprise fabric without having to, uh, you know, worry about, Hey, for specific workloads, I need to have like these extra things to do. So that's kinda the philosophy we have, uh, across from an ecosystem perspective, simplifying as much as possible, you know, God knows it, and, and platform teams have enough complexity to deal with. We don't wanna add to that.

We wanna simplify things for our Customers. Amen. Amen, my friend.

Where can people at home find out more about this? com, right? com ai ml page there as well.

So those are kind of the two key areas I would say we, we learned more about. We've got tons of content in there, blogs, we've got white papers, we got videos, depending on whoever, you know, who you are, how you would like to consume information, we've got something for you. I love it.

Thank you very much for coming on today. Fantastic. You heard it here.

We're on text drunk tv. We've got more coverage from CubeCon coming your way. Just give us a minute.