 Welcome back everyone to theCUBE's coverage here on location at AWS re-invent, AWS's annual conference, I'm John Farrill, this is theCUBE, part of our SuperCloud 5 special edition where we got on location, cube action, in studio in Palo Alto, enjoy the next three, four days. We have tons of content, special reports on SiliconANGLE.com, research, news, opinions. The battle for AI supremacy is the theme of the industry and we're at Genevieve AI as the hottest term. We've got two great guests here, Narayan Bajwaj and Fred Worden here from VMware and AWS, VPGM of VMware Cloud on AWS. Congratulations and of course, commercial software systems at AWS, you guys are in partnership, welcome to theCUBE. Great to be here, thank you. Thanks for coming on because VMware now part of Broadcom, division of Broadcom, I've got the official title, but it's just closed last week. We've had a lot of conversations of VMware on Cloud. In fact, I remember when Andy Jassy launched with Pat Gelsinger in 2016, the cloud, we kind of were like, finally the clarity around the cloud vision is here, it's been a great run. So congratulations on a successful partnership. Give us the update, what's going on with the partnership? I know the Broadcom thing is not really impacting at all. A little Broadcom's got a customer at AWS, but still, what's the update with the VMware, AWS relationship? Well, it's been a great industry defining partnership when you were there in San Francisco many, many years ago. You've come a long way, but the business has been growing amazingly well. We have recorded some great growth in the past few years, including the last few quarters as well. We are the hybrid cloud for enterprise workloads, for modern workloads, 26 regions around the world in US public sector. We've seen some amazing growth actually in the public sector world as well with 160% growth year over year in terms of the customers and workloads we've been able to adopt. So we're very excited by the partnership which continues forward with a lot of great friendship with AWS and our combined customers and partners in this jointly engineered solution. I mean, we always said the operators running vSphere, every enterprise pretty much has VMware, just adds some more products, Tanzu's got traction, so great trajectory. On the other side, you guys are getting smaller, faster, cheaper on the silicon. Networking's getting better, ultra. You have ultra platform, what was it called? Ultra networking. I mean, all kinds of advancements at the infrastructure layer. Keep coming, we're going to see a lot here. They're not going to be announced tonight. Tonight it'll be announced and then Adam Sileski's keynote tomorrow. You guys are getting better too. So I mean, a lot going on there. What news do you guys have here for re-invent? Well, there's a lot of news. And thanks so much news that it's hard to keep track of. Related to VMware on AWS, we certainly have made some progress and we have an announcement to make on some hard work that we've been tackling to improve the flexibility of customers to move from on-prem to the cloud with better sizing options in terms of a smaller size and better compute performance as well with M7i Nettle 24XL, which is a custom silicon that we've worked with Intel on. It's 15% better price performance and just compute performance than any other cloud provider as well. But what's really neat about this offering is that we've separated storage from being on instance to FSX NetApp on tap. So we can size storage and data is so critical in terms of really understanding how that works with new workloads that we can run on this particular instance. So pretty excited about that coming out and it'll be available in quite a few regions shortly and Narayan can explain a little bit more of what we do and what the other advantages are for this as well. Let's get to the benefits. I wrote my post that ship last night. Chips and models together. So you can now pair the infrastructure with what you're trying to do like wine and dinner, like red wine with some steak, white wine with some fish. You know, this is where we're going. Custom built silicon is a thing because you can match performance and get these use cases. What are the benefits you guys do with the storage? What's the, what does it mean for the customer? I get the performance on the chip side. What does this disaggregated storage mean for the customer? What's in it for them? Yeah, I mean, I think for years we've been running this hyperconverged architecture with VSAN as you very well know. I think what customers told us is, hey, it would be nice to dial up storage and compute independently to get an even better total cost of ownership when you're running in the public cloud with VMware Cloud and AWS. And so we've done exactly that. This is a complete decoupling of storage and compute. You can dial up your CPUs and your storage separately with different storage options, including VMware Cloud Flex Storage and the FSX with NetApp on tap as external storage options. So that decoupling really gives customers a better cost model, enables customers to migrate workloads to the public cloud faster, cheaper. So that's really exciting, that sort of new architecture that we have defined. It's interesting, Fred. I want to get your thoughts on this because we're seeing a lot of this in the hyper-performance computing area. This decoupling, disaggregation, memory is easy. You put a big bunch of memory pools together. You share it amongst a bunch of other different use case or machines, purpose built, the things. This gives an optionality change. As the world starts to go differently where you have diversity in the workloads. Okay, VMware is one use case. You guys are looking at a lot of customers, a lot of diversity. Got some VMware here. I got some new, net new gen AI going on. How important is this to the whole next gen workloads? From a complexity standpoint, from a subtracting way of the heavy lifting to the value of the customer? Well, in this particular case, it's so true that you can have acceleration, but that's only one portion of the equation. You have to have the system performance, whether it's to EBS or your network performance, completely understood and modeled quickly and efficiently for your particular workload. This is one example of being able to do that because we have better performance from a clock rate and we have the rest of the system as we look at the entire performance when you look at particular workloads. Explain that again. I want to just double click on that. System performance versus what performance, acceleration? Well, if you just look at one particular portion of the system, like accelerated computer GPU, okay? How fast you can get the data to that GPU and how fast that GPU can work across the entire system or a cluster is super important. It usually comes back to a couple of things, both your network performance, but also your storage. And then tail latency is super important too. What is tail latency? Well, how consistent are you in terms of operations? Anytime you have, it's kind of like the old book that was, you know, you can only hike as fast as a sluice person in your party. And we know who you put that person in the front of the line. That's right, you put them in the front of the line so that everybody starts carrying their load, right? Well, it takes a lot of engineering work in the partnership between us to get that right because, you know, the networking and the options that we have in terms of making sure it works well with where that data might be on-prem. As people are transitioning faster and faster to migrate, that allows us to work wherever the data happens to be, but we can, we can do this. What I really like about this M7i architecture that we have is this four generation Intel AMX, Advanced Matrix Extensions, right? Coming together with Intel, Custom Silicon, AWS VM where technology is coming together. It really creates what I call this pragmatic AIML in the market, where customers can run their existing enterprise workloads and their AIML workloads side by side on the same cluster. And this is powerful because now with the new Intel capabilities, we have virtualized the instruction set for AMX. And that is now provided up through the virtual machine and the container stack. So all of the machine learning workloads, whether you're doing any kind of data pipelining or, you know, inferencing or testing or optimizations, you can really take advantage of that in this new architecture. So you have the best of both worlds with this CPU based AIML use cases. It's interesting, it's almost as if the dream scenario came through the industry with the whole AI craze because it's the perfect storm of timing. More compute systems thinking around end to end, not just, you know, one point performance piece and then the data being freely available. So the whole January of AI kind of speaks to what we've been trying to do for a decade in big data. Okay, the networking's coming together. You got fast decoupled storage and memory. You got architect the system for the best possible scenario or workload. And make it easier for the developer, regardless of where the data is or the virtualization layer is and then leverage new services that we're releasing, you know, monthly. One of the great things about VMware Cloud and AWS is the ease of use for the operator, right? The same model on-prem, same model in the public cloud with AWS. And so that continues forward where you can have the workload side by side and it's kind of like this multi-tenant environment and you can get the best of both worlds. Well, I'll tell you, Narayan, you guys have done an amazing job and we really appreciate you guys being part of our super cloud event series because when we looked at your community, pre-broadcom, there was an element of, I won't say not coolness, but like you have coolness with AWS and that's helped with cloud. But there was an era of, what's next for me as the VMware operator? What's my career look like? Am I going to be in Amazon or Azure or is multi-cloud going to be a thing? I really don't know that anyone had their hands on that. So you have this renaissance of the VMware ecosystem where there's Vmug or these XP experts out there where they see the super cloud paradigm coming where I get it, VMware is a big piece of the puzzle. It is, I don't have to be just VMware, I can be cloud. And I think to me that is a major success personally for VMware because the psychology is less acquiescing to more engaged in your customer base. And certainly with Amazon, the goodness that you guys bring to the table from a cloud native perspective and with Tanzu around the corner, how does this all tie together? Someone's looking at this equation of VMware is cool with cloud, which I think is the case in super cloud. What's next? How does it come together from here? Yeah, I think we've been big believers in super cloud and we've worked with you for a long, long time now. And I think our multi-cloud mission carries forward as a company, investing in areas like vSphere and VMware Cloud Foundation and Tanzu and many other things. So we continue forward that evolution of our stack towards this multi-cloud, super cloud model. I think we are also looking to harness all of the unique capabilities of a Nitro platform, for example in AWS, right? So you've got to sort of be best of breed in terms of where you are and utilize those capabilities. So it's that really good marriage and that balance that we have in our portfolio that really excites customers. Nitro, it's a good trigger word there. I remember when that came out, like Dave Vellante and I were like, oh man, this Nitro, this is a direction. That was the beginning of the infrastructure scale. Fred, I want to get your thoughts because when a super cloud, most people think we talk about multi-cloud, that's actually not the case. Two reinvents ago, three reinvents ago, we observed a change in your ecosystem. We saw the rise of Snowflake, Databricks, Mongo. They have ecosystems on top of AWS. So we saw the rise of the transition from ISV to platform, ISV independent software vendor, a kind of old term, even Adam last year. Oh, they're ISVs, we're the cloud. Well, next gen cloud's about this next step function change over where you start to see this super cloud capabilities where it's not just, it's Amazon plus a platform. And I think you guys highlighted that and we point to that saying, no, that's super clouds, not just multiple clouds, it's multiple environments. That's right. Ecosystems, I think those partners have really gone all in because of three main reasons. One, security. Nitro really is a premise of security first across the infrastructure, really resiliency and availability. If you can't run and you have an ecosystem of that size, you're not going to do well. And then third, the way that we treat all of our customers and partners from a deep engineering standpoint of being cooperative just like what we've done with VMware, that's what allows them to flourish. And I think super cloud is Amazon web services. You got super chips, you got super apps. I mean, AI is essentially just, I mean, what inference is going to do for AI is going to be game changing. I talked about this, Adam and my exclusive and I posted yesterday. Inference is the new killer app. It's going to be a feature. That's right. Always is going to be game changing. Training, you train once like a sandbox. This is going to be some cost to that, but as companies get their own models that become their intellectual property, the inference from that, whether it's an edge device or core data model, this is going to be the operator's dream because then you have to just set the guard rails up for the platform engineering. And once you get the platform engineering up, then the AI's can do its work. But if there's no data, and that's why I think the storage thing's compelling because you now set the table. This is why the disaggregated storage is interesting because now you set the table for Gen AI. I think it's amazing where we have decoupled storage from compute and so these can follow their own streams of innovation independently. I think with this, we were the earliest adopters of Nitro back when many, many years ago. We were the first workload run on Nitro, the first real workload. It's come a long way, it's super secure. We have jointly engineered this together. We're now entering a new generation of innovation, I feel, with this four-gen and then really unlocking the vastness of the X86 and many other platforms that AWS has with Nitro and really delivering VMware everywhere with those platforms. It's a good said, you have a great relationship. I've always been a big fan. Fred, what's the vibe internally at AWS right now? I think this reinvent, you're going to see a lot of flexing. And not in a too bragging way, but I think AWS is kind of like taking a lot of public criticism for not having AI first when they actually had AI, but there's going to be a lot of flexing going on, I think, here so far it's been great. I wouldn't call it flexing. I'm incredibly excited about the ground route, the routes engineering work that we've done and we take a long-term view. We make sure that we put the customer first. We're very customer obsessed and I think you'll see that in all the announcements that we have. How to enable our partners and how to enable our customers to build on us just like we have always done before. Very excited about that. The thing that I would say the vibe is, everything is accelerating incredibly fast. Like I said, when we met a few years ago, you think it was fast then. It's changing exponentially in terms of year over year, in terms of what developers and customers can go do with ease. Awesome, so. Guys, thanks for coming on. I think it's going to be a whole new level of change. AI's a gift for the industry, in my opinion. I think Genevieve has going to change the stack. It already has. And again, when all the hype is gone and who's got what public perception, it's going to come down to price performance. We're back to the old days of speeds and feeds. It matters. It matters. Energy costs alone. The system performance matters. Yeah, system performance. Great call out there. And I think that's the system's mindset. Guys, thanks for coming on theCUBE. Congratulations on that continued success. Cube's coverage here on Location at Reinvent. I'm John Furrier, part of SuperCloud 5, special edition, we're staying it up for this week. A ton going on. This is Genevieve AI Week. The battle for AI supremacy is on and they're in the arena. Amazon's got a great story with the chips and models and the new stack emerging is here. So we've got all the coverage for you. Siliconangle.com, theCUBE.net. We'll be back on Location after this short break.