This discussion examines modern private cloud as a secure foundation for production artificial intelligence, AI, with a focus on the VMware Cloud Foundation VCF platform, private cloud strategy, virtualization and AI infrastructure, memory tiering and security. The conversation helps technology leaders evaluate enterprise deployment models for AI workloads and assess factors to consider such as cost optimization, operational control and compliance.
Paul Turner of Broadcom VMware, chief product officer, VCF division, and Prashanth Shenoy of Broadcom VMware, vice president of product marketing, VCF division, present VMware Cloud Foundation VCF capabilities and strategy from theCUBE New York studio. Turner highlights memory tiering and combined virtual machine and container infrastructure as cost-saving innovations, and they quantify potential reductions in total cost of ownership up to 46% and significant decreases in server memory cost. Shenoy reports that 56% of organizations prefer private cloud for production AI while public cloud preference falls to 41% and they emphasize security and governance as primary drivers.
theCUBE Research frames the discussion by connecting enterprise trends and vendor insights to practical considerations for AI infrastructure. Key topics include private cloud adoption, virtualization for production AI, memory tiering, workload coexistence across containers and virtual machines, and security posture.
Watch the full conversation to learn how VMware Cloud Foundation VCF and private cloud architectures support scalable, secure production AI deployments and reduce infrastructure cost and risk.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
Modern Private Cloud: A Secure Foundation for Production AI. If you don’t think you received an email check your
spam folder.
Sign in to Modern Private Cloud: A Secure Foundation for Production AI.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for Modern Private Cloud: A Secure Foundation for Production AI
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for Modern Private Cloud: A Secure Foundation for Production AI.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
Modern Private Cloud: A Secure Foundation for Production AI. If you don’t think you received an email check your
spam folder.
Sign in to Modern Private Cloud: A Secure Foundation for Production AI.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to Modern Private Cloud: A Secure Foundation for Production AI
Please sign in with LinkedIn to continue to Modern Private Cloud: A Secure Foundation for Production AI. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Private Cloud Momentum, Insights & Innovation
This discussion examines modern private cloud as a secure foundation for production artificial intelligence, AI, with a focus on the VMware Cloud Foundation VCF platform, private cloud strategy, virtualization and AI infrastructure, memory tiering and security. The conversation helps technology leaders evaluate enterprise deployment models for AI workloads and assess factors to consider such as cost optimization, operational control and compliance.
Paul Turner of Broadcom VMware, chief product officer, VCF division, and Prashanth Shenoy of Broadcom VMware, vice president of product marketing, VCF division, present VMware Cloud Foundation VCF capabilities and strategy from theCUBE New York studio. Turner highlights memory tiering and combined virtual machine and container infrastructure as cost-saving innovations, and they quantify potential reductions in total cost of ownership up to 46% and significant decreases in server memory cost. Shenoy reports that 56% of organizations prefer private cloud for production AI while public cloud preference falls to 41% and they emphasize security and governance as primary drivers.
theCUBE Research frames the discussion by connecting enterprise trends and vendor insights to practical considerations for AI infrastructure. Key topics include private cloud adoption, virtualization for production AI, memory tiering, workload coexistence across containers and virtual machines, and security posture.
Watch the full conversation to learn how VMware Cloud Foundation VCF and private cloud architectures support scalable, secure production AI deployments and reduce infrastructure cost and risk.
>> Hello. I'm John Furrier with theCUBE, here in theCUBE's New York City NYSE studio, of course. We have our Palo Alto studio connecting Wall Street and Silicon Valley. We have a special presentation here on the modern private cloud, a secure foundation for production AI. We have two returning CUBE alumnis from Broadcom and VMware, where they have continuing momentum from a year ago when we talked here. We've got Paul Turner, Chief Product Officer, the VCF division at Broadcom VMware, Prashanth Shenoy, VP of Product Marketing at VCF Broadcom VMware. Gentlemen, welcome back for this featured series. Again, as you're kicking off more private cloud, the modern era is upon us. Good to see you.
Paul Turner
>> Yes. Very good to see you.
Prashanth Shenoy
>> Yeah. Good to be here. Exciting times.
John Furrier
>> A year ago, you guys had that big milestone with VCF bringing everything together. Now that's one year in, a lot's going on. What's driving the momentum? Let's start there. What is the key success now that you're seeing, and what's changed since last year when we kicked this off?
Paul Turner
>> Well, I think most interestingly, it's two letters. It's AI. What we're really seeing is this huge opportunity, I think, for AI is driving a couple of factors. It's driving one, great opportunity, given that I can get... If you look at a PWC report, they say you can get 27% better revenue opportunity if you're an AI-enabled industry. That is pretty powerful. So there's compelling reasons of why people are adopting AI. But one of the things that they need is that the platform that runs AI must be better, and that's what's driving VCF adoption today. If you think about it, it's because of the risk and the cost. AI is also a risk and cost multiplier. So, just think about a few stats. 73% of enterprises see AI-related attacks. I mean, just think about that. That is almost every industry out there, is actually seeing these new attacks that are driven by AI-enabled software. So that is a very scary thing. It's like a Pandora's box. The second thing that we're really seeing happen is, AI is actually a cost multiplier, because it's increasing the cost of infrastructure. It's not just the cost of the infrastructure to run GPU-enabled infrastructure. Let's do what we did with CPUs. Let's virtualize them. Of course, that's the most obvious answer. Bring down the cost of the infrastructure. But it's also the cost. It's driving lack of fab capacity. You're seeing increased costs in just server infrastructure overall. And so, we've got to look differently at how we virtualize, not just GPUs and CPUs, but also how we virtualize things like memory. So, AI is the driver for what's happening on VCF and the adoption that we're getting. It's a driver because of opportunity for companies to actually deliver more productive AI that works in conjunction with their existing application footprint, and builds off their application footprint. It'll build towards an agentic AI future that all companies, I think, will deliver. But the second is the security and the cost aspect of running AI. You've got to deal with the risks that AI can expose as well.
John Furrier
>> That's consistent with the trends we're seeing. Obviously you mentioned revenue. A lot of focus on AI driving revenue, not just cost takeout. Good call-out there. But the cost piece, it's not just the cost of the software, or the cost. It's the cost of what you get wrong. If it's secure, or it gets hacked, that's a cost. If you have the bad data, it's a cost. So, I have to ask, how are those trends translating into the momentum of VCF? Because you guys address a lot of this. In this one year, the execution of this has been key. We're seeing a lot of private cloud, on-premises. How do these trends translate into momentum?
Paul Turner
>> Well, I mean, everything is based on that. So, is it translating? It's actually the fastest release ever for VCF adoption. We have more than 2000 customers already deployed, very accelerated deployments. I think part of that is driven by this AI impact. And the AI impact driving, "I've got to get more efficient on the infrastructure. I've got to get more secure for the infrastructure, and I've got to be able to run AI applications in a very highly efficient way."
Honestly, on AI, the cloud-based AI that we've been used to for so long? Tokenomics. I mean, it is getting very, very expensive to start running services. Companies are suddenly seeing the cloud answer to AI is not the right answer to run production AI at scale, and inferencing AI, what I call inferencing, which is really the runtime of your day-to-day operations. You will not run that on the cloud, because of the cost of operations. You can run it more efficiently, just like you can the rest of your infrastructure, on a private cloud environment. So, fastest momentum ever, more than 2,000 customers. That's a stat from a month ago. We're seeing continued acceleration. I think part of that is driven also by Anthropic and what's happening, in terms of frontier model threats. So, big momentum.
John Furrier
>> Now that you've recently announced the news, and there's a lot going on, what is the key highlights around the innovation of the new VCF?
Paul Turner
>> So, I'll touch a couple of things that I'm really excited about. It makes a huge difference to our customers when we save the money. At the end of the day, we do a lot to make a platform... You're going to hear more about how we make a platform powerful for AI. But it's also very important that we make it cost-effective, that we virtualize, that we provide the best platform possible. So I'll just give you something that's kind of cool. Nobody else can do it. The ability for us to tier memory, to take NVMe drives and use them as an extension of your DDR5 memory. We can actually take that, if you do a cost economic study of that, and you look at a two-terabyte server, let's say, that two-terabyte server with DDR5 is going to cost you $118,000. That's pretty damn expensive, and most of it is memory cost. By just tiering your memory and doing NVMe, which is unique to us, because of the way we work as a virtualization plane, we can actually go and safely tier your memory and bring the cost of that server almost down in half, to $69,000. Massive savings. But it's not just a savings in infrastructure. That's what we do in virtualization. It's also a cost savings when you look at, I should run one platform for virtual machines and containers. You probably think, "Well, why is that so important?" Well, I've got my existing applications that all, for the most part, run on VMs and my database, my data-centric applications will run on VMs, and that's where my data lives. But I want to combine that with AI-enabled applications. I want to even drive agentic AI applications over time. The natural place it's going to be is, you're going to do that as an extension of your application footprint, and to do that, you run it on the same platform. By running your VMs and container platform on exactly the same platform, you get a 46% cost savings for running that infrastructure, on a total cost of ownership basis. So, there's lots of cool things in the platform, but I love going back to what VMware has always been known for. We're about virtualization, we're about saving cost, and of course, doing it on a secure private cloud, and truly secure private cloud. But virtualization, cost savings, who can complain about that?
John Furrier
>> Yeah, it's good. And by the way, what's changed a lot in the past year is the price of memory, so the memory tiering becomes super important when you want to run workloads with certain AI stacks that could leverage that. So, good call-out there.
Paul Turner
>> It's not just the cost of memory, being quite honest to you. It's the cost of the overall server. The memory was the first driver factor. The server overall has become just much more important to virtualize at scale. So what it's driving is what I like to call, virtualization is cool again, because it truly is back into the world of what we do well, which is great to see.
John Furrier
>> What's cool is not only virtualization, it's the AI infrastructure and the things you mentioned, more compute and the more configurations. I think this is, to me, the rise of the private cloud again. It never really went away, but now the importance of private cloud is out there. You guys just released your second edition of the private cloud outlook survey. It looks like the things are continuing to drive momentum. We're seeing all the data on-premise. Premises, that's the gold. A lot of focus on more tokens, more AI, a lot more engagement. Prashanth, talk about that piece, because this is the private cloud's moment to scale even further.
Prashanth Shenoy
>> Yeah. I think that's a great segue, John. Last year when we did the private cloud outlook study, there was a definitive cloud reset happening in the market, where private cloud and the operating model of private cloud to run your mission-critical workload on-premise or in a hybrid environment was on par with public cloud. That was mainly because of the three Cs that we mentioned, the cost, the complexity, and the control aspect of it. Fast-forward to this year, when we did the same survey with 1,800 IT leaders and decision-makers around the globe. Two points coming back again is AI tipping point. Why? Because a lot of organizations are now moving their AI applications from a pilot phase of trying, to production, doing it at scale. When you do run inferencing at scale, the cost, the security, the control aspect just explodes by an order of magnitude. That's why we are seeing a resurgence of private cloud, where a lot of organizations are choosing private cloud to be the defacto preferred platform for running their production AI workloads. Whether it's inferencing, whether it's rack kind of use cases, and making their existing applications AI-smart and AI-ready.
John Furrier
>> Production AI workloads at scale, that's the table stakes. It's interesting enough, these models, people are using them, and they're having success. But when you say, "I've got to run it through my whole company," that's an infrastructure. That's global IT. So, it's interesting to see that private cloud is where the AI goes, because you need to enable that through an infrastructure. This is where AI is heading, is private cloud.
Prashanth Shenoy
>> Yeah.
John Furrier
>> Because all the action's on-prem, and models have crawled the internet. You've got the cloud for cloud workloads, but it's not just some small repatriation. There's actually growth.
Prashanth Shenoy
>> Yeah, absolutely. In the survey we got around 56% of the organizations were picking private cloud to be the model for running their production AI. In fact, in the same vein, when we look at the public cloud trend, it dropped by 15 percentage points, to 41%. So there's a definitive shift in the way customers and organizations are looking. This is very relevant to the data privacy, data compliance, data control, security and compliance aspect of it, because they want to bring the models to their data, and not the other way around. That's one of the key reasons why we are seeing that resurgence, and accelerating the momentum of private cloud for production AI workloads.
John Furrier
>> Yeah. Private cloud seems to be the preferred method for the on-premises, but also you've got to build in security. A lot of them, those crown jewels, is in the data. Is this about security? Take us through the security piece of this, why that's important. You said secure workloads, secure AI.
Prashanth Shenoy
>> Yeah. Absolutely. I think security is pretty much one of the most important concerns in the world of frontier AI models that we're going to talk a lot more around. Two big data points emerge, that pretty much the two top challenges for production AI are data privacy control, and security and compliance. One-third of the customers in the organizations that we surveyed cited that as pretty much the key concern. That's very, very critical, because the IP that organizations have is their data, right?
John Furrier
>> Yeah.
Prashanth Shenoy
>> We want to make sure that you can still gain that opportunity and the revenue that Paul was talking about, but with that cost benefit and the security benefit. So, that's where I think the shift is happening.
John Furrier
>> It's interesting. We've been doing theCUBE, this is our 17th year, and even our first year, the word private cloud. I think one of the slogans was, "Journey to the Private Cloud." If you think about it, that was 17 years ago. Now it's prime time. And also in the past, I'd say a year, I've probably said the word governance, or heard the word governance on theCUBE more in one year than all 16 other years combined. That is, it really points to the data piece of this. What does this tell us about the enterprise market right now? I mean, obviously, when you hear governance kicked around, that means there's a linkage with the infrastructure. How does this tell us the relationship between the public and private cloud? Because now you have that distributed computing paradigm, full throttle, mainstream, do you just say cloud? What's the distinction? What does the data tell us about from the enterprise perspective, between public and private cloud?
Prashanth Shenoy
>> Yeah. I think that's a great point. I mean, there is the security governance aspect from a data perspective that we are talking about. But we see, for the first time, cost taking over as a major concern compared to security, when people pick private cloud or a public cloud operating model. The reason is very clear, it's tokenomics. It's what Paul mentioned. So the cost predictability of fine-tuning the model and running inferencing when you don't have clear observability, predictability, and usage of your AI infrastructure, as well as the tokens that you use for running inference, is being a primary factor. In fact, these organizations that we surveyed, 97% of these organizations said that they're having public cloud spend wastage. So these organizations are fundamentally saying, 97% of these organizations are saying that they're seeing a majority of their public cloud spend being wasted, with half of them saying over 25% of their spend is wasted. So that points to a very clear angle that they are wanting more and more cost predictability and cost transparency as we move towards inference and production workloads. That's where we are very, very focused on making VCF the most secure and cost-effective platform for organizations to run their production AI.
John Furrier
>> We're seeing a lot of tokenomics conversations around those costs. People recently are meeting their budgets on tokens, and then just getting started. So that points to AI workloads and teams building and operating workloads. They want unlimited tokens. Why do I want to pay for tokens when I can get them unlimited? Buy a big machine, put it on-premise, run that private? This is a huge thing. This is not saying, "Hey, I'm going to just move from cloud to on-prem," because the models are involved, right? Why would you want to pay for tokens, when you can give someone, "Here's a zillion tokens. Here's your context window." That's happening on-prem, because the data's there, the people are using it. This speaks to the adoption of generative AI infrastructure. What's your reaction to that? Paul's nodding his head, too.
Prashanth Shenoy
>> Yeah, absolutely. This is where we are seeing some major momentum for the VCF 9 platform. The first major release that we made last year, we've seen tremendous momentum. Over 2000 customers deploying the VCF 9 major release. This has been the fastest ramp of any major release of the platform that we've seen, since the inception of VCF. So this has been great, in terms of the reception, the excitement, and the interest that organizations are showing as they move their workloads, their mission-critical, and their container workloads, to run on private cloud.
Paul Turner
>> Yeah. I just want to add in a little here, because it's something you mentioned, John, that I think is interesting. Imagine a customer had unlimited tokens. That's essentially what you get. You get this incredible power that I'm GPU-dependent, but a GPU, in today's GPUs, I can actually run a fairly big 70B parameter model on actually a single GPU. Generally people run it with a four-GPU server, an eight-GPU server, run a small little cluster of them. My infrastructure is not super costly to actually put in place. I've unlimited tokens on that. I've got this ability to just ramp and run the infrastructure at scale. The way that the charging model has gone for public cloud AI, I think, has really got to a point, and you're going to see it over the next, I think you're already seeing it in many companies, but certainly over the next six months, you're going to see a real revisionist history here happening. Which happened before on the public cloud, where we said, "Everything's going to go to the public cloud." And we said, "Oh, gosh. I can actually do it more efficient on my private cloud." The same thing is happening on AI, and that's happening over the next six months.
John Furrier
>> And that's awesome. They become builders, operators, and investors by default, because they're going to want to save that money and put it towards projects.
Paul Turner
>> That's exactly right.
John Furrier
>> We have the CFO and the CIO and the builder all in one.
Paul Turner
>> Yeah.
John Furrier
>> It's been great to have you on. Congratulations. There's a lot of hard work involved. We covered that last year, now the new release. VMware has come a long way under Broadcom, guys. So talk about that change that's been overdue. You've got the consolidation of the platform. What is the future of VMware under Broadcom? It's a question everybody wants to know about. Give us an update on the innovation and the action.
Prashanth Shenoy
>> Yeah. I think the pace of innovation that we have done as part of Broadcom in the last three to four years has been amazing. I think I've had a few gray hairs. Paul has had a few more, I am guessing. But it's been awesome to see that momentum and the customer adoption of the VCF 9 platform. As we continue, I think the pace of innovation is just going to increase even more. There are three big reasons, and you're going to see us talk a lot more about this at Explore in August. Number one, how can we continue to help our customers manage the hardware supply cost crisis that Paul talked about? All the innovations around that at the infrastructure level, as well as the infrastructure services level. Number two, frontier AI models has completely changed the security landscape. The vulnerability, the volume, the velocity of vulnerabilities that we are finding in our software and open source packages, is by far never seen before. So, how do we help our customers harden their infrastructure, and make them secure and cost-effective? It's the second area of innovation, and continuous innovation you're going to see. And three, every organization is making their application either AI-smart, or AI-native application. Most of these are containerized application. So, how do we help our customers run their existing business critical and their containerized AI-smart application on the same platform at scale for production, in a highly cost-effective manner? So, all the innovations around that. So these are the three buckets where you'll see us really focused on, and you'll see talking about that at Explore.
John Furrier
>> Paul, you're the Chief Product Officer. I say Chief Operating Officer, Chief Product Officer of VMware. I'll give you the last word. What is the future of VMware under Broadcom? You get the keys to the kingdom. Give us a taste.
Paul Turner
>> So, first off, I'd be crappy at operations. I'm very good at innovation. That's me. I'm an innovator. I think that we talked about the core platform work that we've done on security, on virtualization is cool again, and how much we're changing the core platform and bringing TCO value for the customers. But this is about AI infrastructure. This is about the AI platform of the future. We're not building for a data center of today or tomorrow, but for what's happening in four or five years' time, and that is an AI-enabled data center. So, I think in that space is where we're driving all of the innovation. Is how can we help customers use AI to actually run the platform more efficiently, but more importantly, how can they drive productive AI applications? How can they deliver an agentic AI platform with the ability to extend their application footprint, to be able to run agents at scale, to be able to run them securely, to be able to do lockdown mode, to be able to sandbox agents? All of those services are exactly where we're building innovation. As part of that, one of the things... I love hearing from customers. I think one of the most interesting is when you hear their experience, because it tells us also where we can innovate on the platform next. So, I'd love to actually introduce up Chris Wolf, who's got one of our partners, ThinkOn, who's been implementing a whole AI infrastructure on VCF, and hear it from them. Because you hear from the customer, you actually hear where innovation is going. And next steps, you're going to see us build on this agentic AI platform of the future. More to come on that. We've got a big event coming up, and you'll hear a lot more at that.
John Furrier
>> Awesome. Well, great to have you guys on the modern private cloud, a secure Foundation for production AI. More data centers are coming, plural, and of course, you get the hyperscalers expanding with neocloud. So customers are going to be putting together their architecture for this modern era and the modern private cloud to keep things secure, certainly on the top of the list. Gentlemen, thank you so much for coming back on theCUBE, and sharing the updates, and we appreciate it.
Paul Turner
>> Sure, John. Thanks a million.
Prashanth Shenoy
>> Thank you.
John Furrier
>> I'm John Furrier, host of theCUBE. We'll be back with more coverage of this important area of modernizing the current business model, which is AI, AI, AI, and you need the foundation for production AI. Thanks for watching.
>> Hello. I'm John Furrier with theCUBE, here in theCUBE's New York City NYSE studio, of course. We have our Palo Alto studio connecting Wall Street and Silicon Valley. We have a special presentation here on the modern private cloud, a secure foundation for production AI. We have two returning CUBE alumnis from Broadcom and VMware, where they have continuing momentum from a year ago when we talked here. We've got Paul Turner, Chief Product Officer, the VCF division at Broadcom VMware, Prashanth Shenoy, VP of Product Marketing at VCF Broadcom VMware. Gentlemen, welcome back for this featured series. Again, as you're kicking off more private cloud, the modern era is upon us. Good to see you.
Paul Turner
>> Yes. Very good to see you.
Prashanth Shenoy
>> Yeah. Good to be here. Exciting times.
John Furrier
>> A year ago, you guys had that big milestone with VCF bringing everything together. Now that's one year in, a lot's going on. What's driving the momentum? Let's start there. What is the key success now that you're seeing, and what's changed since last year when we kicked this off?
Paul Turner
>> Well, I think most interestingly, it's two letters. It's AI. What we're really seeing is this huge opportunity, I think, for AI is driving a couple of factors. It's driving one, great opportunity, given that I can get... If you look at a PWC report, they say you can get 27% better revenue opportunity if you're an AI-enabled industry. That is pretty powerful. So there's compelling reasons of why people are adopting AI. But one of the things that they need is that the platform that runs AI must be better, and that's what's driving VCF adoption today. If you think about it, it's because of the risk and the cost. AI is also a risk and cost multiplier. So, just think about a few stats. 73% of enterprises see AI-related attacks. I mean, just think about that. That is almost every industry out there, is actually seeing these new attacks that are driven by AI-enabled software. So that is a very scary thing. It's like a Pandora's box. The second thing that we're really seeing happen is, AI is actually a cost multiplier, because it's increasing the cost of infrastructure. It's not just the cost of the infrastructure to run GPU-enabled infrastructure. Let's do what we did with CPUs. Let's virtualize them. Of course, that's the most obvious answer. Bring down the cost of the infrastructure. But it's also the cost. It's driving lack of fab capacity. You're seeing increased costs in just server infrastructure overall. And so, we've got to look differently at how we virtualize, not just GPUs and CPUs, but also how we virtualize things like memory. So, AI is the driver for what's happening on VCF and the adoption that we're getting. It's a driver because of opportunity for companies to actually deliver more productive AI that works in conjunction with their existing application footprint, and builds off their application footprint. It'll build towards an agentic AI future that all companies, I think, will deliver. But the second is the security and the cost aspect of running AI. You've got to deal with the risks that AI can expose as well.
John Furrier
>> That's consistent with the trends we're seeing. Obviously you mentioned revenue. A lot of focus on AI driving revenue, not just cost takeout. Good call-out there. But the cost piece, it's not just the cost of the software, or the cost. It's the cost of what you get wrong. If it's secure, or it gets hacked, that's a cost. If you have the bad data, it's a cost. So, I have to ask, how are those trends translating into the momentum of VCF? Because you guys address a lot of this. In this one year, the execution of this has been key. We're seeing a lot of private cloud, on-premises. How do these trends translate into momentum?
Paul Turner
>> Well, I mean, everything is based on that. So, is it translating? It's actually the fastest release ever for VCF adoption. We have more than 2000 customers already deployed, very accelerated deployments. I think part of that is driven by this AI impact. And the AI impact driving, "I've got to get more efficient on the infrastructure. I've got to get more secure for the infrastructure, and I've got to be able to run AI applications in a very highly efficient way."
Honestly, on AI, the cloud-based AI that we've been used to for so long? Tokenomics. I mean, it is getting very, very expensive to start running services. Companies are suddenly seeing the cloud answer to AI is not the right answer to run production AI at scale, and inferencing AI, what I call inferencing, which is really the runtime of your day-to-day operations. You will not run that on the cloud, because of the cost of operations. You can run it more efficiently, just like you can the rest of your infrastructure, on a private cloud environment. So, fastest momentum ever, more than 2,000 customers. That's a stat from a month ago. We're seeing continued acceleration. I think part of that is driven also by Anthropic and what's happening, in terms of frontier model threats. So, big momentum.
John Furrier
>> Now that you've recently announced the news, and there's a lot going on, what is the key highlights around the innovation of the new VCF?
Paul Turner
>> So, I'll touch a couple of things that I'm really excited about. It makes a huge difference to our customers when we save the money. At the end of the day, we do a lot to make a platform... You're going to hear more about how we make a platform powerful for AI. But it's also very important that we make it cost-effective, that we virtualize, that we provide the best platform possible. So I'll just give you something that's kind of cool. Nobody else can do it. The ability for us to tier memory, to take NVMe drives and use them as an extension of your DDR5 memory. We can actually take that, if you do a cost economic study of that, and you look at a two-terabyte server, let's say, that two-terabyte server with DDR5 is going to cost you $118,000. That's pretty damn expensive, and most of it is memory cost. By just tiering your memory and doing NVMe, which is unique to us, because of the way we work as a virtualization plane, we can actually go and safely tier your memory and bring the cost of that server almost down in half, to $69,000. Massive savings. But it's not just a savings in infrastructure. That's what we do in virtualization. It's also a cost savings when you look at, I should run one platform for virtual machines and containers. You probably think, "Well, why is that so important?" Well, I've got my existing applications that all, for the most part, run on VMs and my database, my data-centric applications will run on VMs, and that's where my data lives. But I want to combine that with AI-enabled applications. I want to even drive agentic AI applications over time. The natural place it's going to be is, you're going to do that as an extension of your application footprint, and to do that, you run it on the same platform. By running your VMs and container platform on exactly the same platform, you get a 46% cost savings for running that infrastructure, on a total cost of ownership basis. So, there's lots of cool things in the platform, but I love going back to what VMware has always been known for. We're about virtualization, we're about saving cost, and of course, doing it on a secure private cloud, and truly secure private cloud. But virtualization, cost savings, who can complain about that?
John Furrier
>> Yeah, it's good. And by the way, what's changed a lot in the past year is the price of memory, so the memory tiering becomes super important when you want to run workloads with certain AI stacks that could leverage that. So, good call-out there.
Paul Turner
>> It's not just the cost of memory, being quite honest to you. It's the cost of the overall server. The memory was the first driver factor. The server overall has become just much more important to virtualize at scale. So what it's driving is what I like to call, virtualization is cool again, because it truly is back into the world of what we do well, which is great to see.
John Furrier
>> What's cool is not only virtualization, it's the AI infrastructure and the things you mentioned, more compute and the more configurations. I think this is, to me, the rise of the private cloud again. It never really went away, but now the importance of private cloud is out there. You guys just released your second edition of the private cloud outlook survey. It looks like the things are continuing to drive momentum. We're seeing all the data on-premise. Premises, that's the gold. A lot of focus on more tokens, more AI, a lot more engagement. Prashanth, talk about that piece, because this is the private cloud's moment to scale even further.
Prashanth Shenoy
>> Yeah. I think that's a great segue, John. Last year when we did the private cloud outlook study, there was a definitive cloud reset happening in the market, where private cloud and the operating model of private cloud to run your mission-critical workload on-premise or in a hybrid environment was on par with public cloud. That was mainly because of the three Cs that we mentioned, the cost, the complexity, and the control aspect of it. Fast-forward to this year, when we did the same survey with 1,800 IT leaders and decision-makers around the globe. Two points coming back again is AI tipping point. Why? Because a lot of organizations are now moving their AI applications from a pilot phase of trying, to production, doing it at scale. When you do run inferencing at scale, the cost, the security, the control aspect just explodes by an order of magnitude. That's why we are seeing a resurgence of private cloud, where a lot of organizations are choosing private cloud to be the defacto preferred platform for running their production AI workloads. Whether it's inferencing, whether it's rack kind of use cases, and making their existing applications AI-smart and AI-ready.
John Furrier
>> Production AI workloads at scale, that's the table stakes. It's interesting enough, these models, people are using them, and they're having success. But when you say, "I've got to run it through my whole company," that's an infrastructure. That's global IT. So, it's interesting to see that private cloud is where the AI goes, because you need to enable that through an infrastructure. This is where AI is heading, is private cloud.
Prashanth Shenoy
>> Yeah.
John Furrier
>> Because all the action's on-prem, and models have crawled the internet. You've got the cloud for cloud workloads, but it's not just some small repatriation. There's actually growth.
Prashanth Shenoy
>> Yeah, absolutely. In the survey we got around 56% of the organizations were picking private cloud to be the model for running their production AI. In fact, in the same vein, when we look at the public cloud trend, it dropped by 15 percentage points, to 41%. So there's a definitive shift in the way customers and organizations are looking. This is very relevant to the data privacy, data compliance, data control, security and compliance aspect of it, because they want to bring the models to their data, and not the other way around. That's one of the key reasons why we are seeing that resurgence, and accelerating the momentum of private cloud for production AI workloads.
John Furrier
>> Yeah. Private cloud seems to be the preferred method for the on-premises, but also you've got to build in security. A lot of them, those crown jewels, is in the data. Is this about security? Take us through the security piece of this, why that's important. You said secure workloads, secure AI.
Prashanth Shenoy
>> Yeah. Absolutely. I think security is pretty much one of the most important concerns in the world of frontier AI models that we're going to talk a lot more around. Two big data points emerge, that pretty much the two top challenges for production AI are data privacy control, and security and compliance. One-third of the customers in the organizations that we surveyed cited that as pretty much the key concern. That's very, very critical, because the IP that organizations have is their data, right?
John Furrier
>> Yeah.
Prashanth Shenoy
>> We want to make sure that you can still gain that opportunity and the revenue that Paul was talking about, but with that cost benefit and the security benefit. So, that's where I think the shift is happening.
John Furrier
>> It's interesting. We've been doing theCUBE, this is our 17th year, and even our first year, the word private cloud. I think one of the slogans was, "Journey to the Private Cloud." If you think about it, that was 17 years ago. Now it's prime time. And also in the past, I'd say a year, I've probably said the word governance, or heard the word governance on theCUBE more in one year than all 16 other years combined. That is, it really points to the data piece of this. What does this tell us about the enterprise market right now? I mean, obviously, when you hear governance kicked around, that means there's a linkage with the infrastructure. How does this tell us the relationship between the public and private cloud? Because now you have that distributed computing paradigm, full throttle, mainstream, do you just say cloud? What's the distinction? What does the data tell us about from the enterprise perspective, between public and private cloud?
Prashanth Shenoy
>> Yeah. I think that's a great point. I mean, there is the security governance aspect from a data perspective that we are talking about. But we see, for the first time, cost taking over as a major concern compared to security, when people pick private cloud or a public cloud operating model. The reason is very clear, it's tokenomics. It's what Paul mentioned. So the cost predictability of fine-tuning the model and running inferencing when you don't have clear observability, predictability, and usage of your AI infrastructure, as well as the tokens that you use for running inference, is being a primary factor. In fact, these organizations that we surveyed, 97% of these organizations said that they're having public cloud spend wastage. So these organizations are fundamentally saying, 97% of these organizations are saying that they're seeing a majority of their public cloud spend being wasted, with half of them saying over 25% of their spend is wasted. So that points to a very clear angle that they are wanting more and more cost predictability and cost transparency as we move towards inference and production workloads. That's where we are very, very focused on making VCF the most secure and cost-effective platform for organizations to run their production AI.
John Furrier
>> We're seeing a lot of tokenomics conversations around those costs. People recently are meeting their budgets on tokens, and then just getting started. So that points to AI workloads and teams building and operating workloads. They want unlimited tokens. Why do I want to pay for tokens when I can get them unlimited? Buy a big machine, put it on-premise, run that private? This is a huge thing. This is not saying, "Hey, I'm going to just move from cloud to on-prem," because the models are involved, right? Why would you want to pay for tokens, when you can give someone, "Here's a zillion tokens. Here's your context window." That's happening on-prem, because the data's there, the people are using it. This speaks to the adoption of generative AI infrastructure. What's your reaction to that? Paul's nodding his head, too.
Prashanth Shenoy
>> Yeah, absolutely. This is where we are seeing some major momentum for the VCF 9 platform. The first major release that we made last year, we've seen tremendous momentum. Over 2000 customers deploying the VCF 9 major release. This has been the fastest ramp of any major release of the platform that we've seen, since the inception of VCF. So this has been great, in terms of the reception, the excitement, and the interest that organizations are showing as they move their workloads, their mission-critical, and their container workloads, to run on private cloud.
Paul Turner
>> Yeah. I just want to add in a little here, because it's something you mentioned, John, that I think is interesting. Imagine a customer had unlimited tokens. That's essentially what you get. You get this incredible power that I'm GPU-dependent, but a GPU, in today's GPUs, I can actually run a fairly big 70B parameter model on actually a single GPU. Generally people run it with a four-GPU server, an eight-GPU server, run a small little cluster of them. My infrastructure is not super costly to actually put in place. I've unlimited tokens on that. I've got this ability to just ramp and run the infrastructure at scale. The way that the charging model has gone for public cloud AI, I think, has really got to a point, and you're going to see it over the next, I think you're already seeing it in many companies, but certainly over the next six months, you're going to see a real revisionist history here happening. Which happened before on the public cloud, where we said, "Everything's going to go to the public cloud." And we said, "Oh, gosh. I can actually do it more efficient on my private cloud." The same thing is happening on AI, and that's happening over the next six months.
John Furrier
>> And that's awesome. They become builders, operators, and investors by default, because they're going to want to save that money and put it towards projects.
Paul Turner
>> That's exactly right.
John Furrier
>> We have the CFO and the CIO and the builder all in one.
Paul Turner
>> Yeah.
John Furrier
>> It's been great to have you on. Congratulations. There's a lot of hard work involved. We covered that last year, now the new release. VMware has come a long way under Broadcom, guys. So talk about that change that's been overdue. You've got the consolidation of the platform. What is the future of VMware under Broadcom? It's a question everybody wants to know about. Give us an update on the innovation and the action.
Prashanth Shenoy
>> Yeah. I think the pace of innovation that we have done as part of Broadcom in the last three to four years has been amazing. I think I've had a few gray hairs. Paul has had a few more, I am guessing. But it's been awesome to see that momentum and the customer adoption of the VCF 9 platform. As we continue, I think the pace of innovation is just going to increase even more. There are three big reasons, and you're going to see us talk a lot more about this at Explore in August. Number one, how can we continue to help our customers manage the hardware supply cost crisis that Paul talked about? All the innovations around that at the infrastructure level, as well as the infrastructure services level. Number two, frontier AI models has completely changed the security landscape. The vulnerability, the volume, the velocity of vulnerabilities that we are finding in our software and open source packages, is by far never seen before. So, how do we help our customers harden their infrastructure, and make them secure and cost-effective? It's the second area of innovation, and continuous innovation you're going to see. And three, every organization is making their application either AI-smart, or AI-native application. Most of these are containerized application. So, how do we help our customers run their existing business critical and their containerized AI-smart application on the same platform at scale for production, in a highly cost-effective manner? So, all the innovations around that. So these are the three buckets where you'll see us really focused on, and you'll see talking about that at Explore.
John Furrier
>> Paul, you're the Chief Product Officer. I say Chief Operating Officer, Chief Product Officer of VMware. I'll give you the last word. What is the future of VMware under Broadcom? You get the keys to the kingdom. Give us a taste.
Paul Turner
>> So, first off, I'd be crappy at operations. I'm very good at innovation. That's me. I'm an innovator. I think that we talked about the core platform work that we've done on security, on virtualization is cool again, and how much we're changing the core platform and bringing TCO value for the customers. But this is about AI infrastructure. This is about the AI platform of the future. We're not building for a data center of today or tomorrow, but for what's happening in four or five years' time, and that is an AI-enabled data center. So, I think in that space is where we're driving all of the innovation. Is how can we help customers use AI to actually run the platform more efficiently, but more importantly, how can they drive productive AI applications? How can they deliver an agentic AI platform with the ability to extend their application footprint, to be able to run agents at scale, to be able to run them securely, to be able to do lockdown mode, to be able to sandbox agents? All of those services are exactly where we're building innovation. As part of that, one of the things... I love hearing from customers. I think one of the most interesting is when you hear their experience, because it tells us also where we can innovate on the platform next. So, I'd love to actually introduce up Chris Wolf, who's got one of our partners, ThinkOn, who's been implementing a whole AI infrastructure on VCF, and hear it from them. Because you hear from the customer, you actually hear where innovation is going. And next steps, you're going to see us build on this agentic AI platform of the future. More to come on that. We've got a big event coming up, and you'll hear a lot more at that.
John Furrier
>> Awesome. Well, great to have you guys on the modern private cloud, a secure Foundation for production AI. More data centers are coming, plural, and of course, you get the hyperscalers expanding with neocloud. So customers are going to be putting together their architecture for this modern era and the modern private cloud to keep things secure, certainly on the top of the list. Gentlemen, thank you so much for coming back on theCUBE, and sharing the updates, and we appreciate it.
Paul Turner
>> Sure, John. Thanks a million.
Prashanth Shenoy
>> Thank you.
John Furrier
>> I'm John Furrier, host of theCUBE. We'll be back with more coverage of this important area of modernizing the current business model, which is AI, AI, AI, and you need the foundation for production AI. Thanks for watching.
John Furrier
