Great. Good afternoon, everyone. Thank you again for joining the UBS Global Tech Conference. We're excited to have with us this afternoon from NetApp, George Kurian, Chief Executive Officer. Before we jump into a Q&A session here, we're going to have Kris Newton read the obligatory disclaimer.
Hi, everyone. 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.
Great. Thank you, Kris. George, thank you for joining us. So we've been hearing a lot. We'll start at a high level. We've been hearing from companies around the macro. And I figured, you know, given your commentary on your last earnings call about two weeks ago now, not really seeing a tremendous, tremendous impact from macro, but would love to kind of get your firsthand experience of what you've seen out there, vis-à-vis your competitive positioning, and maybe how would you frame sort of the macro environment going into 2025, particularly given sort of the administration change that we're expecting in, you know, a couple of months.
Yeah. So thank you. Thanks for having me. Thank you for joining us today. The macro today is certainly better than it was a year ago. And is it yet a broad-based kind of refresh of the infrastructure? Not yet. A year ago, when you talk to clients, they would be prioritizing only the most important projects, and they would do those in phases with lots of approval levels for each phase. Where we are today, it's still the most important projects, but being done in one shot rather than in multiple phases. And, you know, the next level of improvement will be, you know, if the macro gets more constructive. And how what we monitor for that is, in the U.S., the first step towards the improvement was the Fed lowering rates.
I think they've done that a couple of times, but you really want to see that to be a trend for a while before everyone gets constructive. I think when I talk to clients with regard to the impact of the elections, I think the positive is that now legislature and executive is in one party, so you can get things done. It remains to be seen what the specific policy agendas are, so we'll have to wait for the new administration to be sworn in. We are taking share in pretty much every market that we participate in.
It's a combination of, you know, customers valuing what we bring to address their priorities, you know, simplicity around one architecture to serve all of the landscapes that they want to operate, which improves, you know, both the cost of the operation, but also reliability and security, the integration of on-prem with cloud, the ability to use their unstructured data for AI. So I feel really good about our competitive position heading into 2025. If the macro were to improve further from here, that should be a tailwind.
So maybe just as a quick follow-up, so when, if I turn the clock back to the summer at your investor day, you laid out multi-year targets with the macro conditions at that time. And so they haven't gotten dramatically better. So to your point, if the macro does recover in 2025, whether it's lower taxes or a more pro-growth environment, do you think, you know, not to say you should change your targets, but is that obviously clearly a tailwind to those targets? I guess maybe another way of saying it is, are those targets reflective of a relatively modest macro backdrop continuing into 2025 and maybe a little bit beyond? Or do you need to see a degree of recovery in the out years?
We are delivering to our long-term model this fiscal year based on the guidance we just issued at the end of the last quarter, and we have every reason to believe we can do that next year. I think if I were to look at the long-term model, if the macro were to get more constructive, we would be at the high end of the range in the model on top line, and on margins. And then conversely, if the macro were to deteriorate, we would be at the low end of the range.
Got it. And then just one more on sort of the macro. Obviously, Trump 1.0, we had supply chain during COVID. Anything on the margin that you are thinking about that you may or may not need to do differently with this administration? I mean, the tariffs seem to be somewhat scattershot at this point, so nobody really knows. But from a supply chain perspective, like, how are you thinking about where you're positioned and what you may or may not need to do going forward?
We made a set of changes to our business prior to the last Trump administration where we moved our China business to a JV with Lenovo, and that has been a really good move. I think, that allows us to operate our business with focus on the rest of the world while leveraging, frankly, the best partner you could have for a China market position. And they have grown their share and conversely our and in concert our share of that market. I think with regard to the supply chain, we have moved 98%-99% of our footprint out of China. So even if there were tariffs imposed on China, it would not be material to us. We use contract manufacturers for our, you know, supply chain, and so they have given us the flexibility to move to where it's, you know, tariff advantage.
We'll have to wait and see, but we feel good about having taken the changes.
Some of your CMs have exposure in Mexico. Is that right?
Correct.
So Mexico.
Could be. Could be. But we can move it quickly to other locations. Because we use industry-standard components and systems, we are able to move it very quickly.
Understood. Okay. That's helpful. All right. Let's zoom out a second. So you mentioned you took share. All the industry data looks like you've taken share for multiple quarters. Some of it, I believe, is, you know, obviously your refreshed product portfolio, QLC-driven all-flash. Can you talk about what that opportunity looks like from a tech perspective and why that new C-Series has been so successful over the last, call it, 12 months or so since initiation or launch? And what is the outlook? You know, and we've talked about this before. You've been growing flash AR, high teens, low 20s. How should we think about that competitive position given what maybe some of your peers might be thinking about doing going forward from a new product perspective?
Yeah. Thank you. I think, first of all, if you were to look at the opportunities in the flash market, you know, there've been high-performance drives, which are 15K drives. These are used for highly mission-critical applications like transaction processing or, you know, high-scale electronic commerce or, you know, the most intensive genomic research. Those were about 10% of the storage market pre-flash. They have all moved to flash. It took many years to move, but they have pretty much the installed base has moved, and now it's in a refresh mode. The second tranche of drives, which is about three times the size of 15K, is 10K. And those are in the early innings of a nine-inning ball game to move, right?
So we are much earlier in the life of the 10K to QLC movement, you know, probably second year that it's been sort of mature, maybe two and a half years in. And the scale of that installed base is three times the size of the 15K, so a long runway. With regard to what we are doing that allows us to be differentiated is, you know, I think the most important was that we were able to combine our superior software with the price point of QLC, right? When we didn't have QLC, we had to compete with a much higher price point, still having our software as differentiated. And then when we brought out our software with the price point of QLC, it has allowed us to pick up share from pretty much everybody else in the market.
I think the second is we are now seeing a long-term positioning around, hey, have one architecture, have a complete portfolio, high-performance flash, capacity flash, and block storage, giving customers the view of, hey, I will just consolidate my entire environment on NetApp. This morning, we just won a large transaction in Canada where a service provider that had Pure and Dell and us consolidated everything onto us because they said, hey, I have one operating system with NetApp, I have four with Dell, three with Pure, I just want to go to one. And so we're seeing more of that pattern of wins where one architecture plus cloud integration is allowing us to take share.
And to your point about QLC taking share, you know, mass capacity or mass storage, how much of that has been driven by sort of the macro, right, where maybe your customers are not looking for the highest performance flash and it's cheap enough now that it replaced the drive market that you referenced? Like, how much of that is like sweet spot, meaning right place, right time for this like product transition or market migration? And the market's a little bit tough. And if the economy gets better, does that slow that product transition, or does it accelerate the product migration roadmap from the drive perspective?
You know, I think that there have been long-standing patterns of storage, you know, landscapes where some are super high-performant. In the high-performance category, flash made complete sense because it was just much higher performing than disk, right? So at almost any price, flash would beat disk. For the QLC, meaning the 10K, flash has many advantages over disk, but performance, you know, is only modestly more of an advantage. And so the cost had to get to a place where it made sense. And for the near line, which is the large capacity environments, listen, performance is not a value proposition there. So you really got to get the cost down for that to transition.
You know, I think in terms of the, you know, the environment, the macro effect on, you know, flash versus disk, listen, I think that macro drives the total storage budget and IT budget, frankly, for any customer. I think that how customers choose to spend it depends on how strategically they think. If they really are like, hey, I want to future-proof this environment, they'll go to QLC going forward.
Got it. And then you mentioned this Canadian customer consolidated from three vendors to one. Was that different profile of flash and storage medium? So block, file, object, like you had different aspects of their solution and it all consolidated under the NetApp umbrella?
Correct. So we're selling them, you know, our high-performance flash, our capacity flash, and our block storage. And then object is a tier.
Right. Block would not have been.
Correct. On NetApp.
Right. On NetApp. And is that, you know, so you've talked about being sort of a disruptor in block. What was kind of the selling point for that particular customer to obviously use NetApp block when you're the new entrant in that particular market?
We have introduced a block-optimized version of our product so that the block team can operate the same software, but in a tailored environment for them. And the second is having one operating system gives their IT staff, you know, you can learn once and use it everywhere. You have one certification process, you have one fleet management process, you have one security architecture, you have one failover model. So it's a much more, like, from a cost of operating your environment is way better.
So maybe just sticking on block for a second. So when you look at the block opportunity, is it a penetration of existing logos where you have another portion of a customer storage budget and you can now offer block alongside, let's say, file? Is that kind of the go-to-market in the near term, or are these new logos where previously you had no traction with that particular customer and you have a you know far more attractive solution with the optimized blocks you know offering that you have today?
We are prioritizing the former.
Okay.
We are, of course, winning the latter as well through our channel and other sales motions. But I think the priority within our large enterprise customer base is about 5,000 accounts to go after and expand our wallet share, and there's a large opportunity there. You know, block in aggregate on the hardware side is $20 billion in full stack, meaning hardware, software, and support, is a $40 billion addressable market.
So then when you think about what your competitors historically offered, right, it was somewhat siloed, right? So you want block, you go to one vendor, file, you go to NetApp. Do you think there is an opportunity for your competitors to kind of attack the market or approach the market, I should say, the same way that you're doing? Like, is there a way for them to develop sort of incremental capabilities or technological platforms to be, you know, a best-in-breed, full-suite solution offering as we sit here today over the intermediate term? Because, you know, Dell's portfolio is in a little bit of a messy situation right now. I'm just trying to get a sense for what the competitive landscape looks like over the intermediate term and how much of a greenfield opportunity that you have going forward.
Listen, I think people have tried to develop file systems for years, right? It is a supremely complex technology, which is why the world's biggest IT providers, Amazon, Microsoft, and Google, all have chosen our file system to be their enterprise file system offering, right? So it is very technically complex. People have tried to build unified storage for many years, and no one really has a very credible offering. So it's a high hill to climb. I think with regard to the near term, the biggest disruptions are, frankly, in the large server manufacturers where their storage business is no longer a priority for them while they chase AI or other opportunities, as well as particularly for Dell, the disruption in their portfolio created by the separation of VMware has been extraordinary. So we are taking footprint from them.
We took the largest VMware vSAN environment in the world over to NetApp. We haven't yet, you know, we won the design, and now we're going to get the business. So we're taking some really big footprints away from them.
I want to ask you about hyperscalers and your public cloud business. So let's do just straight-up hyperscalers first. So obviously, you have a long-standing relationship, as you just referenced, with the hyperscalers. I think you've said publicly AI goes to where the data is, correct? And so how are you viewing your portfolio when you're offering today, given this insatiable demand for data ingestion, synthetic data, AI training? Where are you best positioned, and how are you thinking about the opportunity as we go into 2025 and 2026?
Yeah. I think the first thing is that enterprises broadly have started to recognize that pre-train, you know, first of all, training a foundation model is an insanely expensive exercise with very unclear returns, right? Not even the big foundation model vendors are making significant returns. So more and more of those vendors are essentially taking pre-trained models and then grounding it using their data, a process that's called fine-tuning or inferencing or RAG. There's lots of different. So it's essentially taking a pre-trained model and using your data. We think for storage, that is the majority of the market, and other vendors have also said the same thing, right? Training is 10%-15%, but the real growth opportunity is in inferencing. We are attacking that problem in multiple ways.
One is, you know, continuing to deliver technology and reference architectures with NVIDIA and the hyperscalers so that customers can say, "Okay, now I know how to do inferencing." We have a very, very large installed base of the unstructured data. Jensen said, "Hey, we have 50% of the world's unstructured data or the enterprise's unstructured data." So the first priority is to bring inferencing capabilities to that. You have seen some announcements at our customer conference. There's more coming to make it really easy with your in-place data. The second is allowing customers to use hybrid models. So we have built integrations with Bedrock and OpenAI Studio and Azure Data Lake and Vertex and all the hyperscalers' tools so that you can pre-integrate.
You can take your data and use it with those cloud providers' landscapes and our storage in the cloud and our storage on-prem to make it really, really easy. And we are starting to see that funnel crank up. And then the third is people that want to just use it in the cloud. We are bringing out more price points, more scalable options. And so I feel really good, like lots of different choices for clients, to use cloud or on-prem or a hybrid model with AI and inferencing.
And how much of your model is predicated on these incremental opportunities that you just referenced over the next several years, like to hit the low end, the high end? Like how much incremental uplift do you think you could see or do you need to see to hit the model?
AI should be an uplift towards the high end of our model.
Right.
We have said consistently that storage is not, there's not a big, huge driver at the moment from AI for storage. We are starting to see the, you know, the data prep phase getting built out with data lakes and some small, you know, inferencing projects. And we think that that'll come online in the second half of calendar year 2025.
Okay.
So it's about halfway into the three-year model, right? And so it supports the back half of that model.
I think what do the companies have to do to prep the data in terms of clean the data? Like what are the mechanical steps from your seat to get to that point where we can actually have data that is actually usable for inference or RAG at some point, let's say in 2026?
Yeah. We are doing some of these AI projects ourselves for NetApp, right? And so I would say there's three or four things. First, you got to have a really clear business case for what you're trying to solve, with AI. The second is you need to understand what data sets help you with a coherent understanding of the business process changes that you will get to transform that part of your business. And so what we advise clients on is, "Hey, if you want to go solve a customer support use case, where are all your technical documents? Where are all your case management tools? Where are all your, you know, kind of case notes, things like that?" And then you want to unify that data.
And so people who have highly fragmented data sets create a structure called a data lake, which is essentially a way to pool all of the relevant data sets together. It's a copy of all their existing data, so it's incremental buildouts. And then they create what's called a vector of that data set, which is a numerical representation of a word or a sentence or whatever it is. And that vector database is also an incremental growth opportunity for storage. And then they run the models against the vectors, right? And one of the significant advantages that we have is lots of ways to streamline that entire life cycle, make it a lot easier, make it a lot more efficient for the customer. And so that's kind of where we're at.
You know, we see several customers building out these data lakes, some of them doing early pilots of inferencing, and that's why we believe that second half of next calendar year will be when we see more scale.
How about your public cloud business? Some fits and starts over the last couple of years. Obviously, it sounds like it's found some solid footing as of late. Should continue to grow. Kind of maybe can you give us an overview of where the portfolio sits today and some of the key drivers as you see it today versus kind of you've pruned that business over the last 18 or 20 months or whatever the case may be. Kind of where we sit today and how should we think of that public cloud vector being a contributor to the overall growth? It should be growing faster. You've already got relatively strong gross margins as of last quarter in that business. So kind of how do we think about how that fits into the overall strategy?
Yeah. We were affected by two things in the public cloud business. One was customers essentially going through a phase of cloud optimization where they had, you know, overconsumed cloud or felt that they hadn't used it wisely. And we saw two changes in the hyperscalers as a result of that. One was a movement from subscription to consumption so that people felt that they were spending more fairly and paying for only what they used. The second was a realignment of some of their go-to-market models more tailored towards industry rather than horizontal, you know, segments, right? And then the second, which was our own making, was we built the wrong go-to-market model in terms of trying to get our frontline sellers to sell both cloud and systems. That wasn't successful, in fiscal year 2023.
And the second was we were hopeful that a new market called the DevOps market would expand, and we were going to build out a CloudOps suite to support that. Neither of those two worked well. So what we said heading into 2024 was we're going to shift our, you know, portfolio to consumption. It was already underway. Consumption at that time was about 50% of our business, now 80%. So pretty much all of our businesses moved to consumption. The second was we said we're going to align go-to-market together with the hyperscalers. And, that took us a quarter to get going in 2024, and then now we've started to see strong results. And on the portfolio side, we said, "Hey, we are going to focus on our first-party and marketplace cloud storage, which is our most differentiated products." And we've seen strong growth in that business.
You know, we saw 35%, 30%, 40%, 43% year-on-year growth in that. We are in a tornado market that's on fire. We're just going to add more capabilities. We continue to bring online more and more capabilities with the cloud providers, and we are going to continue to invest in go-to-market to scale that out. We don't see any barriers to that business. That has driven our cloud business to have, you know, gross margins of 74% last quarter, clearly on the trajectory to get to our target range of 75%-80%, as well as, you know, we have got the cloud business to 9% growth as the power of that business has overridden the rest of the business. We expect it to be in the teens in the second half of this fiscal year and really well-positioned heading going forward.
We're going to hit the teens as we exit fiscal 2025. Is that kind of the general run rate of the business on an organic basis as you add go-to-market capabilities? And do you need the historically, the old model was you were looking to add, you know, capabilities through acquisition? That sounds like that's off the table, right? So this is an organic go-to-market strategy that should grow in the mid-teens. Is that the proper way to maybe characterize the opportunity?
Yeah. We don't need, you know, additional technology in our cloud storage. It's our core technology.
Core technology, right?
We acquired a few things several years ago to integrate it with the hyperscalers' tech, but we're in great position now.
Great. Got it. And so when you think about one final thing on sort of where the technology is, so you've got, you know, great block portfolio today that's winning share, QLC Flash that's been winning share, the public cloud business has kind of right-sized and now starting to re-accelerate. Where in the portfolio today do you think there is an opportunity or maybe that requires some more care to kind of get it going in the right direction? I mean, the legacy drive business is relatively small, less of an emphasis. Just, you know, put your strategic hat on as the CEO.
Like where, you know, if we have this conversation a year or two from now, you know, what else has kind of turned the corner or do you feel fairly confident that all of these different moving end markets and pieces and technologies are moving in the right direction over the midterm?
We feel really good. On the other hand, when you feel good, you want to do more, right? So we are putting the pressure into the teams to accelerate competitive takeouts to, you know, kind of become the single data platform provider for the enterprise. So it allows us to sell cloud data services, you know, kind of block, file, everything, right? And so I would say that would be the push we're giving our sales teams as well as, you know, talking to clients about. I think we're closely monitoring AI to make sure that we are dialed into the right early adopter customers so that we can build to where the market's headed. And, those would probably be the two.
We didn't talk about consumption storage. I wanted to leave that for the end. So obviously, Keystone's an important part of the story, relatively small from a TCV or SaaS, however you want to view it. How do you think about consumption storage in the context of, you know, the traditional go-to-market model for storage? Are you selling it just where customers want it? You're meeting the customer where they want to buy storage? Like how do we think about how that develops over the next couple of years?
Our belief has always been to let the customer have choice and to meet the customer where they are. That was the thesis around cloud that played out well for us. It is the same thesis on Keystone. We're like, "Hey, if the customer wants to go there, great, we'll make it available to them and no hindrance." Our sales teams are compensated to be neutral to it. They get paid the same whether they sell a CapEx deal or an OpEx deal. I think what we are doing is using it to drive our growth. We're using Keystone as a way to migrate competitors off their existing footprints where they may say, "Hey, I'm not ready to make a large capital purchase mid-year into the life of another, you know, capital equipment provider.
Consumption.
Yeah. Or to bridge them to cloud where they say, "I'm moving off Dell or Pure. I'm not long-term in the data center. I want to go to cloud. You have a good cloud offering b ridge me," right? So there's a few different ways to work on that.
Not to be overly maybe negative then, does that mean that's kind of potentially a triage solution for some customers or it's like a temporary solution?
In some customers, it is. That's not the predominant. There's a lot of clients that have learned from using the cloud that now say, "Hey, I want to use my on-prem environment." What has been surprising to us is, you know, in a good way, right, is that we've had zero churn. There's not a single customer, and we have hundreds now on Keystone that has churned off us. So yeah.
All right. Just maybe in the interest of time, anything we didn't cover that maybe I think the market, from your perspective, the market hasn't fully appreciated about the NetApp story, the narrative, the opportunity as we move forward?
Listen, I think we have four secular drivers of growth, two secular, you know, the disk to flash movement as well as AI, two company-specific, which is cloud and block. We are executing our game plan. You've seen us do that for multiple quarters. We have a leveraged business model so that when we drive top-line growth, a big chunk of that falls to the bottom line and to cash generation. I think our long-term model was mid to high single digits on the top line, double-digit EPS growth. We're on track for doing that this year. We're not guiding next year, but we feel good about our ability to do that going forward. So thank you.
Great. Thank you very much for your time. Thank you, everyone.
Thank you.