Rajiv Ramaswami, Nutanix | Nutanix .NEXT 2026

Yeah, first of all, thank you, Alison. Thank you, John. It's great to be here as always. So our business has shifted very decisively over the last several years, and deliberately so. So maybe I'll take a step back and say where we were, where we are now, and then we can talk about the future, which a lot of this NEXT year was about the future. I mean, I know John, you've known us for many years, tracked us for many years. So if you look at the past, Nutanix was an HCI company. HCI company, and when I got here, maybe six years ago, now almost, when I talk to customers, they say, 'Yeah, Nutanix, great platform for running HCI and my end user computing workloads.' From there, we went to building out a full platform that could run everything, and then we expanded on our multi-cloud approach to be able to work with all the big cloud providers to take our offerings on the public cloud. And then more recently, we've expanded from being a company that's just focused on HCI and hybrid multi-cloud to becoming a platform company in multiple ways. We are not just wedded to our own storage anymore, we support a whole range of ecosystem storage partners. And you saw many of those being announced at the show, new ones being Gleno and NetApp in particular, with some of the other ones coming to market with Dell and even Pure. But then we didn't stop there, right? Now, we are paving the way for the future because our customers are going through a massive architectural shift, right? They are moving from... Of course you got to run today's applications. Of course, they're mission-critical. You've got to keep your businesses going, but you got to think about tomorrow. And what is tomorrow? Tomorrow very quickly is every new modern application being built, is containerized, it is orchestrated with Kubernetes, but more than everything else, tomorrow is about the world of AI and how it's going to impact how businesses operate. And if you don't make use of it, you're going to be left behind from a competitive perspective. So it's all about AI influencing agentic AI where you now move from simple inferencing to being able to really have agents delegate stuff to agents and have more autonomy and enable more autonomy in the enterprise, make things happen so much faster. So that world is our future as well, because we are now moving to this new world of providing a complete platform to enable organizations, companies to be able to go build and run these agentic AI applications in a very easy turnkey way, and also focused on building an ecosystem to support all of that. So that's kind of really what this conference was anchored around. I mean, we spent a lot of time talking about that vision and where it's going.

Oh, 100%, right? I mean, if you think about a CIO today, that CEO has to be thinking and they are about how they can enable AI in the enterprise and do so with the appropriate governance in place, provide the flexibility to all their developer teams to go transform the business. And that becomes more and more central. In fact, after my sessions, we spent a fair bit of time with our CXO or customers who are here and you know what? All the things they heard about, this was top of mind for them. How can you enable me to get AI deployed in my enterprise? Yes, they have to worry about their hybrid and multi-cloud and their VMware migrations. All of that stuff is going on for sure, but they're very focused on what's the future for them, and that is around AI.

I mean, I tell you, one of the things that I'm most proud of as a company is how this ecosystem has grown around us over the last few years. Look at it this year, over a hundred partners sponsored the event and you can look at the expo floor here behind us and it's buzzing. And we've got all the major names in the industry here, the major cloud providers, the major server providers, the major storage providers, the major chip providers. It's just everybody's here, right?

Let me tell you why. It is a sign that what we are doing with... As you said, you use the word platform and we take that very carefully. To me, a platform is around... The value of a platform is directly tied to the ecosystem around it too. And all these partners are seeing the value of the platform that we bring today to the market and realizing the value of being integrated together with us, because it benefits them. It benefits the customer. And that's really why you see this. We've been very deliberate about building that ecosystem over the last couple of years and to where it is today. And now, by the way, it's not done by any means. Now, we've got a whole new ecosystem that we're just starting to build around AI. And we announced, for example, a set of AI pass services today that help developers actually build their AI applications. And that's just the beginning, and that catalog is available this summer, but we've got a lot more to do. But this application layer that we provide continue to provide choice. You talked about AI being now running on everywhere. The inferencing in agentic AI is going to run in the public clouds. It's going to run in neoclouds. It's going to run on-premises. And we want to be the platform that can help our customers run with all of this. By the way, cost per token. Token is now the unit of economics right here. And I talked to a whole bunch of customers whose main focus was, 'Can you help me get my token costs under control?' And a lot of that, providing visibility into what's being consumed, providing guardrails around what's being consumed, trying to reduce the cost by optimizing what you do. All of those become table stakes, and that's what we're doing.

Yeah. Look, I think the interesting thing is our core platform is evolving to become an agentic AI platform. And a lot of the work is to be done in the core itself. I mean, we focus from an AI perspective on four layers where we are innovating. So we talked at the very top about this AI services. And on the AI services, I think it's two things. It's providing this ecosystem again of curated open source components that people can then build applications with. It is providing cost, governance, access control to all the models that people need to run. That's the second big piece as well. And then underneath that is the optimization of the infrastructure to run all of this so that you can drive the cost per token down, optimize it, get better performance. That's actually very core to what we do for as a technology company. We've been doing this for compute centric workloads forever, with the hypervisor capabilities. And what we've done now is to optimize that hypervisor for GPU centric workloads. Being aware of topological constraints and optimizing how workloads get placed onto GPUs and CPUs and the memories so that we can optimize and get best performance. That's actually hard work. And we've done that work, as an example.

Listen, you want to make the maximum use of GPUs that you buy, that you have. Yes, GPUs setting idle are bad because think about it. On the one hand, you're spending more and more tokens, and if you're going to need to buy more and more GPUs to go use that, it's not efficient. Same thing as we saw in compute centric workloads. Back before virtualization came on, utilization was very low; with virtualization became much higher. Same thing is happening now with GPUs. It's the next big resource to be optimized.

Yeah. So there's a different set of constraints that you have to optimize on these things. You want to make sure, for example, you can actually make good use of the memory that you have on the GPUs. And for example, offload some of the stuff that gets used frequently, like key value cache. So that's something you've got to offload. You've got to make sure that the workloads are placed so that whenever GPU talks to memory, you want to be making sure that's as local as possible. You don't want to be able to have to go across through the CPU to the other CPU in the socket and more minimize and try to have locality of everything so that you can get the best utilization and performance. So it's kind of the same problem, but with a different set of things to optimize and manage.

Yeah. So that was the second piece. I just want to finish up on this one. And so then there were a lot of innovations we talked about on the data side, right? Taking raw data that's coming in with our data hub and being able to really pass it, put it, filter it, get rid of private information, for example, out of it, and then vectorize it and put it in a form that can be easily consumed by the AI applications. That's a very important thing. You need to be able to simplify it. AI is only as good as the data that you feed into it. So making sure that you feed the right data into it is very important. So these are the technology innovations that are the pillar of this agentic AI stack.

Yeah. Look, I mean, I will perhaps in this case use the word system and platform a little interchangeably, because the system effectively is everything coming together, and that's what the platform is doing. And the fact is, this is the thing. I mean, if people, the more ecosystem that you have around this platform, the more tentacles you get. And what happens is that the customers get a much more complete solution that they can deploy in the market, and it helps them address very specific problems. Okay? So in this world, for example, supply chain constraint, they want to use and sweat the hardware as much as they can. So the fact that we are broadening our ecosystem of your storage vendors that we support is very critical to them. That allows them to modernize, get to running these new container workloads or everything-

Takes that off the table. It helps ameliorate that quite a bit. That's just one use case for them. Now, as they go to this cloud native world, they're now trying to figure out how they can manage their cloud native environments and their current environments rather than having silos bring them all together. And while doing that together, you also want to be able to stand up these container environments very easily. So container environments need networking, they need storage, they need observability, they need disaster recovery, all those things you want to provide. And part of that is provided by the ecosystem that we build around us. So this is, again, where it all starts coming in to solve real customer problems.

Oh, yeah, it's a new thing.

By the way, so I think this whole move towards sovereignty is here to stay, first of all. I don't think it's going away. It's going to become even more important. And there's a few other things that are driving sovereignty also, which is more self-reliance, more the ability, not just the data piece, which is always part of it, but you want to have your own infrastructure. You want to be in control of it yourself. You want to have US citizens run it and manage it and not be dependent on outside parties. And to the extent they can, you also want to make sure your technology supply chain is robust to support all of that. So that's happening everywhere outside the US and even inside the US, there's concerns about sovereignty. So you go to Europe, you go to Asia, it's everywhere. And that represents a huge opportunity for Nutanix because, again, this is one of the fundamental things. We enable sovereign clouds to be built that meet these needs, provide the self-reliance capabilities, provide the ability for these folks to operate their own infrastructure in a way that matters to their countries and allows them to build. I mean, for example, with these AI sovereign clouds being built, countries are trying to stimulate the economy there. In fact, there's many government incentives out there to say, 'Let's go build this because we believe that this is going to help local businesses use AI easier.'

Oh, 100%.

Oh, the first thing is that it's a government initiative. In many cases, make sure they have multiple sovereign cloud providers in their countries. The second is they do provide incentives for them. They might finance some of this build out. They might try to bring their own government organizations as initial customers into these. So they want to use this as a stimulus for their economy, to your point, say, 'Look, we need our capabilities so that our companies, our industry in our country can actually go out there, get AI trade, AI enabled, and hopefully create more economic value as a result.'

We are the enabler for intelligence is the way I would think about us. Because by having this platform approach, we allow people to go build all these different AI applications, which last year was simple inferencing; today, it's agentic AI; tomorrow, it's physical AI and robotics. I mean, sitting here, I wouldn't really know what's going to happen with that intelligence. There's all kinds of things that could happen. And I don't know, but the possibilities are endless. And just by enabling this, I think it's really going to help transform how every company, every business operates.

The first thing is make it simple for people to consume AI. I mean, that's important. I mean, it's not easy. And just like everything else, the simpler it is, the more adoption you're going to get. The more cost-effective it is, the more adoption you're going to get. And then once you've got the basic stuff, then you've got to make sure that this is properly governed, make sure you have the security, privacy around the data, make sure you figure out who's accessing what, what permissions you're giving it, you're protecting your data, you're protecting all different tenants or users who are using that. All that becomes parcel of the enablement of enabling customers to consume AI, and that's what this is all about.

Yeah, we truly want to be the platform company where all applications run. Today's applications, tomorrow's applications, in this new AI world, we want to become this platform of choice for our customers all around the world.

Great having us. Thank you all.