Welcome, and thank you for standing by. I would like to inform all participants that this conference call, as well as any Q&A, may be recorded and made available to clients of JPMorgan. Where a company is presenting, any recording may also be posted on their website. The views and opinions expressed by any external speakers on this call are those of the speakers and not of JPMorgan. Parts of this conference call may also be reproduced in JPMorgan research. If you have any objections, you may disconnect at this time. This call is intended for JPMorgan clients only. Press participants are not permitted on this call and should disconnect now. Unless otherwise permitted by internal JPMorgan policy, members of JPMorgan Investment and Corporate Banking are not permitted on this call and should disconnect now. To ask a question today, please submit any written questions using the Q&A button found on your Zoom toolbar. With that, I would like to turn the call over to Samik Chatterjee to begin. Please go ahead. Yeah, thank you, Harry, and thank you, everyone, for joining. This webinar is part of our CEO Spotlight series that we've been doing this year, and we have the pleasure of hosting George Kurian from NetApp for this session. I do want to thank George as well as the NetApp team for making this possible. I was just going through George's sort of background relative to NetApp, and George, as much as I knew that you've been the CEO of the company since 2015, I definitely didn't know that you were part of Cisco before that, which is another company we covered. For the people in the audience who are not less versed with it, George has been the CEO since 2015 after joining NetApp in 2011. Prior to that, he was at Cisco, and prior to that, at McKinsey & Company, I believe. George, long time at the role here, and we'll get into some of those questions about sort of the longer-term view about the industry. Before I do, we also have Kris from Investor Relations here. Kris, thank you for taking the time, and I'll hand it over to you to go through the safe harbor before I kick it off with questions. Thank you. Thanks, and thanks for having us. Today's discussion may include forward-looking statements regarding NetApp's future performance, which are subject to risk and uncertainty. Actual results may differ materially from the statements made today for a variety of reasons described in our most recent 10-K and 10-Q filed with the SEC and available on our website at netapp.com. We disclaim any obligation to update information in any forward-looking statement for any reason. Back to you, Samik. Thank you. OK, George, maybe just starting with that longer-term picture, now you've been in the CEO role for 10 years plus, I guess. What are the primary changes you've seen in the industry, as well as how have the customer preferences changed that NetApp has had to navigate over this last decade? Yeah, first of all, Samik, thank you for having me. And to the listeners, thanks for joining. Over the last 10 years, there have been sort of three or four things that I would say continue to be consistent. The first is ongoing changes in the market. I think the most pronounced change over the last 10 years has been the growth of public cloud and cloud as an infrastructure pattern. We have been able to be fortunate to uniquely in the market capitalize on the transition to public cloud. Public cloud has also driven the importance of OpEx buying as a growing pattern within the enterprise data center, where people want to harmonize their operating models between public cloud and on-prem. The second has been sort of ongoing technology trends, which is people want to use the latest technology, like Flash or higher-performance networking. There's a lot of new networking things, like RDMA over Ethernet and other things. That's an ongoing set of capabilities that we see that we continue to take advantage of. We got to number one in all-flash as a result of our strength. I think the third, operationally within the client base, is that more and more clients are being challenged to do more with less, especially around infrastructure. It has driven, in addition to the push, emerging push on AI, it has really driven the idea of unifying your data, unifying your infrastructure model into a more consistent pattern. Two other things that we've seen is the large multiline sort of integrated system vendors. These could be Dell or HP or IBM, has lost share in the market. This continues the pattern for the 50-year history of the storage industry. First, it was IBM that was in the 1990s dominant and then lost share. Then Dell and HP have each done large-scale acquisitions and have lost share. That continues. Interestingly, hyperconverged was a trend that was supposed to be a big deal and would take over the data center. It has its place in the smaller data centers. It has really not taken over the large enterprise data centers. That's really the big themes. AI, if I look forward over the next 10 years, the ability to extract knowledge from your data assets is probably the most profoundly important theme that we see in our clients. OK, great. That's on the industry front. If I take that same question and sort of lay it in terms of operationally, what were the primary, maybe over the last decade, what are the primary aspects you focused on in the company in terms of operational transformation? Have those delivered the results that you expected as well? Yeah, maybe I'll talk about it in three ways. One is portfolio. Second is go-to-market. Third is kind of financial and operating performance. I think on the portfolio side, listen, we continue to pivot to the growth areas of the market and do that by shifting resources from mature parts to the growth parts. I think we talk about the fact that there are four drivers that we're focused on: Flash, cloud, block, and AI. From a portfolio perspective, we're pivoting there. Flash was a much, much, much smaller percentage of our revenues when I started. It's now 2/3 of our hybrid cloud revenue. The Keystone program, which is to really deliver a consumption model, has grown strongly across revenue, TCV, unbilled RPO, pretty much all the metrics. Cloud is 10% of our revenue, public cloud, and is growing at a fast clip. It grew 33% last quarter and is very margin accretive to the overall business. That's really the short summary of the portfolio, right? Pivot to where you see growth areas of the market and do so within a disciplined operating model. On go-to-market, we decided to sharpen our focus on the biggest markets. Fairly early in my tenure, we pulled out of about 19 countries and used distributors to go-to-market. Perhaps most importantly, we struck a joint venture with Lenovo that provides us access to the China market without the operating cost structure that's required to participate in a large market like that. From an operating expense perspective, listen, we have been disciplined operators of the business. We compare and contrast. The first quarter as CEO, our operating margin was around 7%. I think you see the most recent quarter was north of 25%. Those would be the three buckets. From a capital return perspective, we had just started a buyback and dividend program in 2013. I became CEO in 2015. We have sustained that and expanded that over the last few years. That's sort of the four buckets on transformation. There are always things that you want to do better, right? You got to number one in Flash. We feel like we should expand our lead from that position. I think on cloud, we feel good about the fact that we've got the big three hyperscalers. There are others we want to do more with. You always have the aspiration to keep pushing. No, great. I was going to bring up the margin aspect and the transformation on the margin front in a bit. We'll talk about that as well. On the portfolio side, NetApp was named again as a leader in Gartner's Magic Quadrant for enterprise storage platforms. As we look at that leadership that you have today, the question is again sort of going back to the operating discipline where you've been growing expense probably half the rate of your revenue, which basically means you've been growing at that level of low single-digit base. How do you envision maintaining this leadership when the experience profile is more mirroring sort of that low single-digit profile at this point? We are seeing more and more companies having to spend more to keep up with the rapid technology transformations. How do you envision maintaining your leadership in that framework? Yeah, I think the foundations are, one, you have a disciplined culture. I think that we have a real focus on, are we moving our resources to the parts of the market that are growing? Then strategically, we have done one thing that is technically very hard to do, but has extraordinary leverage, right? Which is we have one operating system and one platform that allows us to have enormous leverage in our business. There is literally a single code line, one software code line that supports our systems, that supports our Keystone, and that supports our penetration into all the public clouds. Not only does that give us operating leverage, it also gives us enormous innovation capacity because you innovate once, it's available everywhere. That is our advantage. We've worked really hard to sustain that advantage with a single code line and all of that discipline. If I move to sort of a different question, but more on the top line growth over the years, I mean, the way I think we perceive it and investors perceive it is that the growth in data for customers has been consistent over the years. That's a secular driver for your business. However, that's been a very consistent increase in data, has been pretty consistent. The expectation typically has been that that would lead to a more consistent top line performance from the likes of NetApp. Relative to that, when we look at some of the top line performance that you've had, like yours in 2016, you had a 10% decline. In 2022, you had a 10% increase. What drives the volatility in sort of your performance on a year-to-date basis, even though the secular driver seems to be much more consistent and predictable from what the enterprises are seeing in terms of data growth or their storage need growth? It's a good question. While data growth is consistent, customers often buy periodically, especially when there are moments to take advantage of opportunities, right? The first Trump administration's policy on benefiting investment and some tax advantages drove a lot of buying in the early 2018, 2019 time frame. 2022 was a year where clients were catching up from deferred upgrades from COVID, as well as some of the constraints that the supply chain had. Conversely, there have been periods like 2016 and 2020 where buying has been more constrained. I would point out that we focus on two things. One is we have a large part of our business that's more recurring revenues that supports the cost structure of the business. The second is we continue to expand the range of customers we serve and the different set of use cases in the customer so that even if there is kind of volatility in some parts of the customer, we can get access to others. OK, got it. Then another sort of question on the way NetApp operates. I mean, the value of the storage platforms is in the software, or a large part of the value is there. You mentioned there's a sizable amount of recurring revenue from the way you look at things based on that software capabilities. At the same time, you don't appear to be using or relying as much on subscriptions with your enterprise customers in terms of your preferred go-to-market motion, particularly for the hybrid cloud portfolio. What drives that decision to sort of look at it, look at the ongoing business with the enterprises as recurring revenue, which you have visibility into, but not really push them towards a subscription model? Our approach is to meet the customers where they are and to have the broadest range of options available to them. We have several elements of our business that are recurring. Support, which is about 40% of our business, is actually a recurring revenue stream. When a customer buys equipment, even in a capital expenditure model, about 40% of that still goes into a recurring model. The second is cloud is all recurring, and our Keystone model is more of a recurring subscription and consumption model. We offer and push people, hey, we have these models. Do you want to buy them? At the end of the day, our view is if you force customers into a particular model, they may not go with you. They may choose an alternate vendor. We continue to see subscription as a percentage or recurring revenue as a percentage of our total business go up. We also want to meet the client where they are. Many clients have a strong preference to buy a certain way. OK. Block storage market, you entered the market in terms of preparing very specific products for that use case pretty recently. What should we think about as the measure of success for NetApp in that market? Are customers choosing to stay with your unified storage, or are you finding some of them moving over to adopting specific block storage products in your portfolio? We have had block storage for a very long time at NetApp. We have north of 20,000 customers who use block storage with NetApp. They, as you noted, buy it as part of a unified configuration, meaning they want to have one system architecture in a cluster, some of that to support file-based storage, some of that to support object-based storage, and some of that to deploy block storage. We are by far the leader in unified storage. There are, however, customers where they just want a block-only offering. These could be either smaller customers that don't have such a sophisticated set of data needs. It could be departmental environments in larger customers, which look similar to a smaller customer, or in the very, very large customers where they have file teams and block teams in their infrastructure organizations that are separate. We introduced a set of block-only or block-optimized products. The success of those products comes from growth in our All Flash business, and because those are all-flash configurations, the block-only products. Maybe another way to ask it is, are you seeing your deal sizes with existing customers go higher because they're adopting these block-specific products, which they didn't do before? Is it just a migration for these customers from instead of adopting the unified product to moving to block? We are seeing new wallet in existing customers and also new clients, like mid-market clients that never bought NetApp before, starting to choose us. Got it. Great. Going back to your number one position in Flash, you've gained considerable share overall of the market as well in the last few years. Some of that has been C- Series. You also have the ASA Series. How should we think about runway in terms of market share gains? If you take these two specific products separately, how should we think about how much more runway do you have in terms of market share gains? Listen, I think that the market share data that IDC published has us in the 25% range. I tell our teams that's 75% to go, right, in the all-flash market because your aspiration is to have dominant share. There's plenty of market share gains to be had. Our aspiration when we build our financial plans is to grow at or above the market so that we maintain or grow share, right? That's sort of the benchmark that we hold ourselves to. We really want to be gaining share. We believe we can. The C- Series and the ASA products complete the set of things that we need to address the Flash market. It allows us to have price points that we couldn't have before when we only had the A- Series in the unified product portfolio. As we just discussed, a block-optimized offering. I think the portfolio is in a good place. We have some exciting announcements to come on the AI front at our NetApp Insight Conference. I think that we feel confident about driving share gains in the Flash market. George, just to look at it in another way, you brought up an interesting point, which is 75% more to go. How many of your customers adopt only one storage vendor versus how many of your customers do you find typically working with two or more storage companies? I think what we have seen with the growing pressure on operating expenses in most of our clients, the need to have speed, agility, and security, the idea of having so many different vendors for storage is actually shrinking. They really feel like, hey, if I want leverage on my storage vendor, I can go to cloud. Cloud is my negotiation point with the on-premises infrastructure vendors. The pattern of I need multiple vendors for addressing my control and vendor control is starting to dissipate. We see most clients having two vendors and some who want just the simplification going to one vendor. We feel like there's an opportunity for us and maybe one other to sort of grow share in the market. Staying with this market share thread, when I think about you mentioned this early on in terms of the integrated players giving up share to companies that are more focused like yours. When you think about these share gains in the coming years, are there specific customer verticals that we should think of that's more likely, where maybe it's more mid-market driven, that we should think about the share gains? When it comes to competitive companies, we should be thinking more of that coming from like a Dell or HP. How would you encourage investors to think about where some of these share gains come from in the coming years? I think we are focused on the biggest geographies in the world. I think that someone like a Dell or HP has distribution reach advantages over us in the smaller markets of the world, like in Africa or some parts of Latin America. Our view was the strategy that Napoleon pursued, right? When you fight a much bigger army, you kill the middle. When the middle collapses, the sides have to fold into the middle, and you can kill the whole army. We concentrated our coverage in the biggest markets in the world. We have seen that opportunity, right? I think we have taken share. We are number one in eight or nine of the top 12 countries in Europe. We've taken over number one in India. We took over number one in Australia. We're really focused on our go-to-market. I think with regard to where should we gain more share, the U.S. obviously is the biggest market. We really want to gain more share here. The second is with regard to price bands. Our portfolio has historically been strong in price bands seven, eight, and nine. With the C- Series and ASA [Series], we brought more capability into price bands five and six. We feel like there's more room to grow in those two price bands. I think those would be sort of the quick summary of where we continue to take share. With regard to the players in the market, listen, there's lots of kind of legacy players for whom storage is not a focus, innovation is not a focus, ability to attract talent is not a focus, right? These could include the big share donor, which has been Dell, but also IBM, Hitachi, HP. There's lots of opportunities to compete. Maybe moving to the all-flash growth that you've seen, you've been reporting double-digit growth over the last year, except the last quarter when we did see some moderation there. You explained that to be driven by the public cloud vertical. Should we assume that excluding public sector, sorry, public sector, I meant public sector, excluding the public sector, that the double-digit momentum in all-flash is unchanged for NetApp and that you can sustain a double-digit rate if public sector spending comes back to a normal space? I would just say, without giving you all the numbers, if you were to do the math, if public sector had been flat, our Flash number would have been significantly higher than where it was. Without giving you all the specifics, yep, we feel very, very confident that we would have been able to have a much higher number in Flash. OK. I'll take this one question that's come in, which is relative to, again, the all-flash market. The question from the investor is that NAND prices have increased, and some of that has helped the suppliers, that NAND supplier itself in the industry. How should we think about how protected NetApp is from spot pricing of NAND? Will the increase in NAND pricing drive some level of moderation in all-flash market growth? Yeah, I think probably three or four things in that area. One is we offer a broad range of products, both flash and disk-based storage. Our view is that for certain classes of workloads, disk, hard drives, make a lot more sense. Secondary storage, archival, even for media streaming, you really don't need all-flash. The price of flash has to be very close to the price of disk, which it isn't, right? That's the first thing. There are workloads for which disk is much better, right? The second is with regard to the flash pricing itself. Vendors typically have passed through the pricing to customers when prices go up and passed through the benefits when prices go down. As a reminder, only 30%, roughly 30% of our total cost in a system that a customer buys is really NAND, right? There's a lot of software and support and other things that go into the system cost. The last thing is we work with our NAND suppliers to insulate ourselves at least for a period of time from spot prices. There are long-term agreements we structure. We get multiple vendors to compete for our business. One of the advantages when you are number one in the flash market is more vendors want to be part of your bandwagon, right? More NAND vendors. There are lots of ways that we try to manage through that. OK. Moving to public cloud, maybe outline for us what is your vision for NetApp's position in the public cloud storage market where you're a leader, but you've also sort of tried to be in FinOps before you decided to exit using the Spot, exit the Spot business. I mean, how do you envision sort of NetApp to look like in the public cloud in a few years from today and beyond first-party storage? What are the other drivers there that we need to think of? I think that our primary focus is to scale our first-party and marketplace cloud storage services and to build a competitive moat around that. We build that moat in two or three ways. One way is to, now that our infrastructure is in all the public clouds, our storage infrastructure is there, it is to add value in terms of additional software capabilities that allow us to differentiate our technology. For example, autonomous ransomware protection, compliance capabilities, multi-cloud, multi-AZ capabilities. There is an ongoing set of software capabilities that we bring to that platform. The second is to integrate that platform into the software applications that the cloud providers have. This makes procurement a lot more easy for customers, right? Customers may not buy storage. They may buy a SageMaker environment, or they may buy a Vertex environment for AI. In this way, we can get consumed when they provision a Vertex or a SageMaker where they may not know what the storage is. That part of the journey has started. We needed to instantiate our infrastructure in all of these cloud providers. Now we're working with them to, hey, I can make your application so much better using our storage alongside your stuff. You will see us make announcements of that over the course of the next 12 months, right? We're super excited about that. The third is technology is great. Go-to-market is equally important to build a competitive moat. We have a lot of engagements with the hyperscaler sales teams. We are starting to work with their professional services teams around client migrations, customer success, client expansion, and industry solutions. There is lots of stuff to do. We're excited, and we continue to do that work. This is a good opportunity for me to ask you one of the questions that came in from investors on this front, just in terms of how to think about how competitive the market is for talent related to cloud services itself. When we look at some of the newer competition that's coming into the industry, like the VAST and [Veeam] of the world, and their willing intent to sort of try to encroach on your leadership in the cloud services side, what are you finding in terms of your ability to sort of hire and retain talent related to cloud services specialization? Yeah, I think that we have a super strong talent pool. I think that we have continued to evolve our talent pool as the business has evolved. We had a group of people who started our journey in cloud. They were not the right people to scale our journey in cloud. We have really good talent to do that. I think that NetApp's culture has been a strong point for us over 30 years, that we allow people to innovate, to experiment without fear of failure. We have people from diverse backgrounds. We have a lot of talent from the hyperscalers at NetApp because many of them started working on the other side of the fence and then said, hey, I have more opportunity to make impact coming here. You saw the announcements of our Chief Product Officer, who was the person that built the Salesforce Genie platform, was the Chief Technology Officer at Zscaler. We had a new CFO who came to us from Western Digital. We feel good about our ability to get the talent we need for the phase of the journey that we're in. Yeah, no, fair. Before I move to AI questions, AI-specific questions, one more that's come in from an investor. The question is, can George comment on whether the big, beautiful bill and the ability to depreciate CapEx fully in one year for tax purposes will drive any additional customer IT spend? We are waiting to see what the impact of the big, beautiful bill will be in our customers. I think there's always sort of three things that drive customer behavior. The first is their view of macro, which drives their overall posture on spending. IT is a part of that, right? That's the first. The second is, can they get competitive advantage, or do they fear having the opposite happen to them if they don't make an investment? I think that AI is the new thing. Cloud was the thing a few years ago and continues to be an important part of the consideration. Cyber is important. We are positioned alongside those growth drivers. We're waiting to see what all the impacts are as people understand it, what that posture is. OK, good. Let's maybe move to AI, maybe starting with what you're seeing from your enterprise customers and more specifically in terms of what their requirements are as they're thinking about their infrastructure, how their AI infrastructure will look. What are you hearing from them in terms of what their storage needs might be? What drives your confidence that you have the right portfolio to address that? Yeah, I think broadly speaking, in the AI journey, the first phase of the AI journey, which is still ongoing, was primarily about I want to get core technology built, right? This was in the AI, in the generative AI world, really the large language models. That work is ongoing. That's where the bulk of the AI investments have been made. The second phase of the journey was I want to integrate AI into the consumer applications, right? This is, hey, I'm going to make AI smart search like Google. I'm going to integrate that with AI capabilities and maps and so on. It's similar to how iPhones happened, right? iPhones happened in the consumer market and then came to the enterprise. What we see going on in the enterprise is really nobody is trying to build a foundation model. That's way too expensive and frankly no advantage. I think now what we see is the application of pre-trained AI models with companies' data. There's a variety of things they need to do to make that happen. That whole idea is called inferencing. There's underlying terms like RAG and this and graph RAG and all this stuff. Just think about that as inferencing. In inferencing, having high-quality, well-organized, unified data is super important. If you look at predictive AI, it really worked on structured data like databases or tables and things like that. Generative AI works on unstructured data, documents, videos, audio. Agentic AI works on unified, structured, and unstructured working together so that you could have a complete view of a business problem. We feel really good about our position in the AI market for two reasons. One is we hold a huge amount of the data that customers need to apply to their AI models, right? That's generated through their operational applications. For example, you go to a hospital, we hold all the medical images. We are the platform for Epic, for example. When a customer says, hey, I want to use GenAI or some advanced multimodal application to look at which patient has cancer or which patient has COVID, we are the incumbent. The second, as you will see, we have a whole range of super exciting announcements at our Insight Conference, which will give our clients the ability to extract a lot of knowledge from their data that sits on us. I would just tell you, come to Insight. We're going to shock the world. Maybe just another one. I mean, we are early in this cycle, but what we've seen till date is related to AI, storage has seen less of a demand or uplift relative to any particular compared to compute, right, which has seen much more of an uplift. Part of that would be that some of your enterprises might be starting their journey on AI in the public cloud. One, like how do you think about the magnitude of benefit and the timing of the benefit from AI to NetApp as well? Like do you expect that you're benefiting right now or more early on the benefit is on the public cloud storage piece of your business versus related to maybe hybrid cloud and the on-prem infrastructure benefits later? How do you think about the magnitude on each of them? Like where will enterprises eventually land more in terms of their AI infrastructure focus? Yeah, I think first of all, the predominant build-outs of AI have really been for training, right? I think if you look at the vast amount of compute, it's really for training foundation models, for improving the ability of foundation models to do multimodal and a whole range of other things. I think in aggregate, however, there will be more growth in compute than storage. I don't know what the long-term, meaning 5- 10 years out, view is, but if you look at a five-year out view for a mature business, like our FlexPod converged infrastructure business, which is more for enterprise applications, it's typically a 50/50 split between compute, network, and storage, where storage is 50% of the spend, right? That's in the mature enterprise application space. I don't know what the sort of the long tail will look like. I think with regard to storage for enterprise AI, we want to capture it in both places. We want to capture it in cloud. We want to capture it on-prem. You're correct that some of the data science experiments are happening in public cloud where the data science team doesn't know the IT team. They want to start on the public cloud. We have seen some of our public cloud solutions being used for AI. We have customers that spoke at our conference last year. There'll be more at our conference this year that talk about that. We also have clients in our enterprise business. We talked about about 125 wins this year for Q1 compared to 50 wins last year, where it's split between data lakes, which is really how do you unify your data before you do inferencing. We had some sovereign wins for training or fine-tuning foundation models. We had inferencing. It was, I would just say, roughly a third, a third, a third. OK, got it. One of your competitors, I guess we can sort of say Dell, has been talking a lot about parallel file systems. They've talked about it in the context of Project Lightning that they're working on. Can you just sort of make sure, I just want to make sure we understand sort of where NetApp stands on that front, why are parallel file systems important, and what is NetApp's offering on that front? Yeah, I think that the idea that AI has is to speed up access to data so that you can feed it to the GPUs. You parallelize access to the data. In the disk drive, hard drive world, you had a construct called a parallel file system that essentially had a series of independent storage nodes connected into each of them paired to a compute node. This file system striped across all of them. It is a very complex technology. You have a lot of work to be done to load balance it, failover. When the system gets into trouble and you have to troubleshoot, it is supremely complex. As a result, parallel file systems have stayed a very small part of the total market. It's typically been in the sort of the legacy high-performance computing world, the labs in higher education, a few supercomputing environments. As Flash has grown, what you can see is a much more advantageous sort of next-gen architecture is what's called disaggregated storage, where you have the same idea of multiple compute nodes that can access in parallel storage. The storage is connected across a really high-performance network fabric to those compute nodes. What it does is it gives you very, very close to the performance of a parallel file system, but vastly simpler and vastly more efficient. I would just tell you, come to Insight. We'll talk about our approach to providing very high-scale distributed systems to clients. Yeah, got it. I think definitely looking forward to Insight and the announcements there. Maybe just delving a bit deeper into what the inferencing infrastructure for some of these enterprise companies looks like. There's always the concern that as they ramp, as enterprise companies ramp their infrastructure, when they start planning around the inferencing workloads, they typically do budget in dollars. That starts to cannibalize some of the traditional storage spend. One, maybe if you can share any other thoughts. You talked about the data lake, is the stage that most enterprises are in terms of working on data lakes. Where do you see the real inflection in terms of storage needs coming from? Secondarily, as we see that inflection, do you see a pullback in some of the traditional storage use cases just because of the priority towards AI? Yeah, I think the trade-offs that clients make are always driven by their business priorities. I think that, in general, what our approach has been is, hey, it doesn't need to be an either/or. Just like in cloud, people had on-premises and cloud, and they used to have two different silos. We unified them into one architecture so that you could leverage the investments around people and technology and make your operating environment more efficient, meaning less spend, and more agile, meaning more flexibility, right? We are doing the same thing for AI, where we are saying, hey, to use AI with the data that is continuing to grow in your operational environment, we will help you bridge that and make it much, much more efficient. That's our pitch to clients. Hey, you don't need to have two different spends. You can take one and make it much more efficient. Data growth comes from two or three areas. One is if a customer has truly distributed data and organizations that are just not able to work together, they usually create a data science or data analytics and AI team. Those people bring data together. It's a copy of the data, but they bring it to what's called a data lake. A data lake essentially is the idea that you can bring multiple types of data into a single environment. You can do it with NetApp storage without having to build a data lake. You can just do it on the storage itself by marking which data you need. If your data is not organized, you want to create a second copy. That's one. Second is, at the time of inferencing, you have to take your core documents or videos or audio, and you have to create what's called a vector and a graph, which is basically a way that a large language model can understand that data. In some cases, you're not going to do that for all of your data. You're going to do that for a subset. Let's say it's 20% of your data. When you vectorize the data, the volume is typically, our best estimate of doing it most efficiently is 5x growth. That 20%, you will grow it. You may not keep it forever, but you will keep it for as long as you're running the inferencing. The last phase, which is still early, is, hey, when people generate data, how much data will they generate? It's too early for me to speculate. In general, if you look at human behavior in the past, if there's a free resource, they tend to fill it up. Marketing will create 20 copies of slideware that they need, just so that they can compare it and present it. Yeah. OK, maybe the last question on the business side, just in terms of revenue drivers, can you talk about the advantage in AI of already having a large install base? I think on the unstructured data side, you and Dell have a large install base already, but you're trying to do both, which is one, sort of your install base is an advantage, but you're also trying to win share by displacing some of the customers by winning with customers that are part of Dell's install base. How much of an advantage is it having a large install base? How sticky are customers on that front? Are customers in AI trying to use their primary storage supplier across their AI use cases? Do you see opportunities to displace some of your competition on that front? I think the displacing competition is really around their operating environment, right? Meaning the applications that are driving the creation of the data. I think there we have several advantages to go after Dell or anybody else, right? That's the primary displacement mechanism. In terms of the AI side, it's a new wallet that we or Dell might compete in our install base, right, or their install base. I think there, it's so much easier for a customer to be able to bring AI capabilities, meaning all the software capabilities we're working on, to their existing data state rather than lift and shift and copy it over. I would say that when you have a database or what's called structured data, it's easier to do that because it's much smaller volumes. You know, a giant database is probably 10 TB. If you go and look at a giant file, it could be multiple times that. The volume of data you have to move around is much, much, much larger in unstructured data. Therefore, it's much easier if you can do it like we are going to show to bring AI to your data rather than copy your data to your AI. Got it. OK. Maybe now moving to [margins], actually, before I do that, let me take this investor question. The question is pretty short, but I'll let you sort of talk about the partnership here. The question reads, does NetApp work with Nutanix? Maybe just sort of outline the partnership you have there. How material do you think it will be? We have common customers with Nutanix. I would just say stay tuned for upcoming announcements. Nothing to say at this point. Our approach is to work with all the hypervisor vendors north of us, right? We already work with Proxmox, VMware, Hyper-V, Red Hat, OpenShift. There's a long list. Our approach is to work with everybody. If you don't have something today, stay tuned. OK, moving to margins. I was looking back at the sort of last decade, and gross margins have gone from, I think, 61%- 71%, largely over this time period that you've been CEO. Maybe talk about, as you sort of look forward, what do you think are the long-term margins that the company should be hitting, maybe both on the gross margin side and more specifically in the operating margin as well, where you consistently continue to sort of make progress pretty rapidly on that front? I think the broad shifts, maybe I can just go through the elements of gross margin. I think our cloud business, which is growing significantly faster than our on-premises business, it's 10%, but it's growing at 33% CAGR in Q1, right? It's a fast growing number. It grew at 33% year on year in Q1. It is the first party, right? The cloud business overall reached north of 80% margin. We said that our target range would be 80%- 85%. The mix of that business continues to push upward in terms of software. The depreciation of the hardware that we put into Microsoft data centers continues to come off the P&L. We feel good about the trajectory of that business to get to the top end of the target range. I think the most important thing for us to do there is to grow it, right? The second element, professional services, it's a small number, but the growth in professional services is through Keystone. Keystone clients go through a trajectory that starts with lower margins, but as they go through their tenure, that goes up and to the right. You should see that business, as Keystone grows as a percentage of the professional services business, the margin should move up over time. Support is a big number. It's growing in the low single digits. Its margins are north of 90%. 92.3%, I think, was the latest number. It should continue to stay there. In terms of the product, gross margins in the hybrid cloud business, our target range was mid to upper 50s. Our view of that is that it continues to shift towards Flash, which is a higher margin than disk. We are continuing to add more software value into that portfolio. That's kind of the puts and takes on gross margin. Operating expense is a piece of the equation that we managed in a disciplined fashion. We have typically managed it to grow at half the rate of revenue growth. I think that gives you the puts and takes on gross margin and operating margin. What we operate the business as is to really drive gross profit dollars because we have, as you can see, been super disciplined on OpEx. If you can drive gross profit dollars, the incremental dollars convert to income and cash flow at a very high clip. OK, last one from my side. Just M&A and rank order, maybe sort of give us the rank order in terms of where M&A sits in your priorities. When I particularly look at the number of new companies trying to enter the market here, how do you think about the likelihood that there's a wave of consolidation in the near future to return this industry to its usual three or four large participants? Yeah, I think overall, when we look at capital allocation, we said dividends are the first call on capital returns. We typically run that at about 40% of the total number, and then we split 30% of the total cash generation. The remainder we split between returns to shareholders through buybacks or targeted M&A. Over the last few years, we have not done M&A. We've really focused on returning all of free cash flow to shareholders. Sometimes we have done more than 100%. Sometimes it's around 100%. Generally, we've returned all of free cash flow to shareholders. I think that when we look at the market, the on-premises storage market, it's a mature market with a lower growth rate and customers wanting to consolidate how many vendors because they just want simplification. They need to unify their data. That customer behavior generally tends to market consolidation over time. OK, got it. Great. I'll wrap it up there. This was great to get your insights in terms of the industry as well as NetApp. Thank you for the time, Kris. Thank you as well. Thank you to the audience for tuning in as well. Harry, you should be OK to wrap up the call here. Thank you, Samik. Thank you for having me. No, thank you. Thank you, everyone. This concludes today's webinar. You may now disconnect from the call.