NetApp, Inc. (NTAP) Earnings Call Transcript & Summary

All right. Why don't I start? So good morning, everyone. Welcome to day 3 of the Flagship TMT Conference. My name is Erik Woodring. I lead the U.S. IT hardware research coverage here at Morgan Stanley. I am delighted to be joined this morning by NetApp's CEO, George Kurian. George has obviously been at NetApp for over 15 years, been the CEO since 2015. He played a major role in kind of transforming the business into what it's become today. So George, thank you. Thank you very much. Thank you for coming to the conference. Good to see you.

Thank you, Erik. Good morning.

So before we start, just from Morgan Stanley side, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures for important disclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. I don't have the safe harbor agreement yet. But Jeriel, do you want to come up and read it?

We have a safe harbor statement that we would like to read ahead of the presentation.

It's in the NetApp here. Perfect. Thank you, Jeriel. So just from the NetApp side, today's discussion may include forward-looking statements regarding NetApp's future performance, share are subject to risk and uncertainty. Actual results may differ materially from the statements made today for a variety of reasons described in NetApp's most recent 10-K and 10-Q filed with the SEC, and available on their website at netapp.com. NetApp disclaims any obligation to update information in any forward-looking statement for any reason. Thank you, Jeriel.

Perfect, sir. Thank you.

George. Welcome, obviously, a lot to talk about. I think the most natural kind of starting point is just a review of earnings reported last week. And maybe to that, the 2 or 3 things, most important kind of key highlights from the quarter, speak to the momentum in the business, and then we can get into specifics from there.

Yes. Thank you for having me. We just concluded and reported Q3 of our fiscal year. We had a strong performance, as we indicated we saw acceleration in our business in the second half of the year. Revenue growth ex the divestiture of spot was 6%. All of our focused growth priorities did very well. We had close to 300 AI wins. We had 11% growth in our all-flash arrays, again, well ahead of the market. We had growth in our high-margin public cloud business. Again, ex spot total cloud grew 17% with our first parties and marketplace services growing 27% year-on-year, and our Keystone Storage as a Service business grew approximately 65%, 68% year-on-year. So overall, really good momentum across the business. Operating income and EPS were also record highs. So I feel good heading into the final quarter of this year and the momentum heading into next year.

So I want to maybe start just on the AI side. And to your point, you highlighted about 300 AI wins in the quarter. That was up from 200 in the quarter prior. Can you just help us understand those wins, those workloads, how does that kind of differ in either intensity or product as it relates to kind of the traditional backward-looking kind of storage customers?

Yes. I think in the AI landscape, we have 3 different areas of focus. One is what you call data preparation and data lakes, where customers are bringing data of multiple types together to analyze, the second is model training and fine-tuning, and the third is RAG and inference, where you take a model and actually use it for either classic inference or increasingly agentic use cases. We have -- we reported in the quarter that 60% were the first bucket, 20% middle and 20% the last bucket. They are essentially use our all-flash arrays for high-performance use cases, and then they use our object storage, for example, for cold data. A data lake, for example, could be a hybrid of those where they use all-flash for high-performance parts of the data lake and our object store for the big archive of data.

And maybe building on that kind of more of a philosophical or strategic question is how -- as we think about the world shifting in that direction, how do you differentiate from competitors that are trying to do the same thing, position themselves within kind of the AI opportunity?

Yes. I think we have worked on AI for many, many years. I think that we started out with sort of predictive AI. We have hundreds and thousands of customers using predictive AI models with us. In an AI landscape, we differentiate on sort of multiple dimensions. The first is, you've got to have competitive cost performance, and we do. We have introduced a new family of products, which takes performance and scale to yet another level, the AFX family. The second is a suite of tools that allow you to manage your data efficiently because these AI environments, while they're about speed, they also need to have multi-tenancy built in cyber resilience so that somebody doesn't go and hack in and take this data that you are using for your AI models or poison it in certain ways. You've got a variety of data management tools that make it easy to version models and data sets. And we have built those over many, many, many years. I think the third area of differentiation is on hybrid and multi-cloud. We were the pioneer in building integrated solutions in the hyperscale environment. And now we are integrating data pipeline into those hybrid -- into the hyperscaler applications. For example, we announced a capability called S3 access points with Amazon that we had worked on for a long time. This allows an enterprise to take their data and directly pipe it into an Amazon AI tool like SageMaker or Bedrock or variety, a broad suite of tools and makes it super easy for an enterprise to use their data. And then the last is, listen, we are one of the largest holders of unstructured data in the world, and it gives us an incumbent advantage.

Perfect. Maybe, again, staying on AI, I think as a follow-up, you mentioned, I think in the quarter, 40% of AI deals, you won were for workloads in production the quarter prior was 25% to 30%. So if we take that 200 and 300 customers, that means the number of customers that are doing workloads entering production is basically doubling sequentially, 60 to 120 quarter-over-quarter. What does this tell you about, it doesn't have to be an inflection point for kind of AI-driven enterprise data demand. But what does that tell you in terms of the momentum or the inertia? Where are we in that ball game, so to speak?

I think we are encouraged by the momentum. We follow the mix of our business every quarter. I think it's still early days for enterprise AI. We certainly see, for example, regulated industries, who have their data better organized like certain parts of public sector for national security use cases, certain parts of health care and life sciences where their data has been well organized either for advanced patient care, better diagnostics or faster drug discovery. You look in certain parts of financial services for advanced quantitative modeling, risk management, fraud detection. So it's still specific use cases, but we're starting to see better momentum across the broad mix of our business, particularly geographical. Now you're starting to see other parts of the world begin to adopt it.

And to that point, I think what I hear from you is you kind of the U.S. is leading that momentum, but you're starting to see Europe and Asia kind of start to follow.

Correct.

Okay. Cool. You also, in the quarter, I thought something that was important, you highlighted growing interest from neoclouds. You highlighted the neocloud win in the quarter. You're working with them on what you call the differentiated value proposition that brings the AFX conversation into this broader conversation. How significant is the opportunity if we, not next quarter, but just take a big step back and say, neocloud's AFX more broadly relative to your traditional legacy markets? Maybe the first question, is just how big can that actually become?

It's still early to tell what mix of the overall AI market will happen between hyperscale, neocloud, sovereign cloud, colocation and enterprise data center, right? I think there's lots of speculation, but it's still work in progress. We want to participate as broadly as possible in all the sustainable parts of the AI landscape. We have strong solutions that we have developed over many years and strong presence in sovereign environments across the globe, national flag carriers in Asia, many of the leading European cloud service providers. And as they build their AI stacks, we are a natural partner. And so we have several wins there. With the neoclouds, we are pushing 2 sets of capabilities that allow us to stand apart in addition to super high-performance flash-based storage for training and model fine-tuning. We also are working with them on more sustainable, longer-term business models for them by helping them bridge enterprise data into their neoclouds, just like we have done with the hyperscalers. And in the hyperscaler world, listen, we are built into many of the hyperscaler sovereign environments. For example, as they build out sovereign regions or sovereign environments for national security or for European providers, we're part of those. So our goal is to compete as broadly as possible and to help those service providers build sustainable value by bringing enterprise clients in better solutions to them.

Does the go-to-market change significantly for you? Do you have to make significant changes to kind of capture that opportunity?

There's some optimizations like we do all the time, but we know how to co-sell with large cloud providers across the globe, and this is just another version of that.

Okay. And then I want to kind of tie that into then kind of the product set and really how that's changing, talking a lot about disaggregated architecture, private cloud, it kind of brings us back to your event last October. Just how does the rise of kind of Agentic AI and generative AI more broadly impact storage and data requirements, but then also impact the way that you think about innovation and what you need to roll out to make sure that you can stay ahead of your competitors and capture this opportunity?

Yes. Very simply, Agentic AI will hallucinate, deliver bad actions, not just bad data, if the quality of data and the guardrails put around the data are weak and frankly, compromised, right? And so for Agentic AI to work, you need a good data foundation. That data foundation needs to be fast and resilient because these agents constantly go back to the storage to ask for new sets of data to refine their reasoning. And then the second is you need a series of data preparation and organization tools to help the models find the right data. And so we introduced the AFX platform for the first, which is super high performance, scalable storage, and then we introduced the AI data engine, which is a suite of software that allows the organizations to discover, organize and put guardrails around the data for the second challenge.

And if we now kind of take what you've said in past Analyst Days or your most recent Analyst Day, and then kind of layer on the opportunity associated with AI more broadly and how that could expand the TAM. I realize, it's early days, but just how do we think about from a growth trajectory perspective, whether that's how you think about the market or what it could pertain to NetApp. Does that -- is it too early to understand how that could reshape the growth algorithm? Or just your thoughts around how it actually could reshape that growth algorithm?

AI will definitely grow storage because you need more data to analyze your business better, and in the process of analyzing your business better, you will probably generate more data as well so that you can automate various elements of your business process. And so we expect that to accelerate storage growth over time, and we intend to be an unfair beneficiary of that by taking market share in that growing market.

Okay. I want to maybe shift from the market opportunity to the customer set. And U.S. public, we saw a nice recovery in the business last quarter, after some of the challenges that you saw earlier in the year. I guess, the simple question is has U.S. public turned the corner, but maybe the broader question is just help us understand, maybe what's shifting with that customer kind of customer cohort, how that's contributing, how you're thinking about the world as it's emerging.

Yes. I think a year ago at this time, there was a lot of disruption in U.S. public sector. Usually, the year after a new administration takes office, there is some shifting in both budget outlook, as well as which agencies get budget. And so we are accustomed to that. But if you look back a year ago, Doji was sort of running across the U.S. government looking at various elements of contracts and there was a really hard time figuring out, even if you won a deal when the deal would actually flow through the procurement process, we then had in the first half of this fiscal year, a softer U.S. government result because of the shutdown and the duration of when appropriations actually hit agencies and funding vehicles. Q3 was a bit better as we had expected, still soft, but you saw the performance. Our performance met our expectations, although it was a bit subdued. And we are looking to see strength growing through the rest of this fiscal year for the government.

Okay. Very helpful. It took us 8 questions to get into the memory question, but I just want to make sure that we do touch on it. Interestingly, historically, in times of inflationary cost pressures, you've been able to pass through higher input costs to customers you've adjusted list prices. It has been a tailwind to both growth and margins. This period is a little bit more unprecedented to say the least, right? I'd love to just maybe at a high-level start, how does NetApp approach this situation differently, given the differences versus past memory cycles or past input cost inflation cycles?

This feels quite like the year after COVID where there were significant supply chain constraints. We work in a 360-degree manner to -- across all of our stakeholders. I think the first is the industry has over its long history, passed through commodity price increases to customers when they go up and pass through commodity price decreases to customers when they go down. So I don't think that customers see the industry as doing something unusual when we pass through price increases. We have raised prices, and we will continue to if we have to do so again, we will continue to do so. That is in line with our prior practice. We are working with our customers and channel partners, given the dynamic nature of pricing to be more agile, meaning have shorter durations of price protections in our contracts. And conversely, when prices go down to be able to pass it through to them faster. The timing of those adjustments are not always going to be perfectly aligned with the timing of the cost increases or decreases, right? And so we are working super hard to get those balanced. We are working -- we have done prebuys to assure ourselves of supply. Sometimes, the mix of components in the prebuys may be a little different than what we had planned than what we see because some customers will come and say, "I want this component as opposed to another one, and so in Q3, we had to go out and do some open market purchases." And in Q4, we are running ahead of our outlook. We are substantially ahead in terms of revenue relative to what we forecast. So sometimes you have to go out into the open market to procure supply. In addition, we have multiple suppliers. We are working with pretty much every vendor in the industry to qualify alternate sources of silicon because our priority is to have available supply, on the one hand. And the second is to drive to a gross profit dollar target that we set to give us the earnings outlook that we have given the Street.

Yes. So I was going to touch on that. And the question was, I'd love, if you could maybe elaborate on that a little bit, which just as we think about how you can optimize given the input cost pressure, it sounds like a focus on optimizing for gross profit dollar growth versus maximizing gross margins. Can you maybe just double-click on that a little bit?

Yes. We have a broad portfolio of solutions, right? And so we start with, hey, position the right product for the right use case. So for example, we have a suite of hybrid flash products, which are much more economical than all-flash products for capacity and price-sensitive use cases that allows us a competitive advantage over flash-only players. The second is to position tools like Keystone or cloud, where customers don't need to or don't want to buy a large CapEx purchase. I think with those in mind, you're always trying to balance gross margin dollars -- with gross margin percent with gross profit dollars. Ultimately, we run the company for gross profit dollars, right? Because that's the earnings engine and profitability engine of the company. But you're always trying to balance lots of different factors.

Yes. And maybe last related question is, we're still in the early days of pricing increases. And congrats because you nailed the prepurchases last summer. How have your customers responded thus far, not only to the pricing increases, but when you talk about agility in some of those shorter windows, what's the feedback that you're getting in kind of real time from your customer base?

Yes. I think, first of all, we are not the only ones who have raised prices. I think everybody in every part of the tech industry has raised prices including the PC vendors, right? So I don't think we are operating in an environment where customers don't know what's going on. I don't think anyone is happy that prices go up, but they also appreciate that we have a broad range of solutions to offer them, and they are evaluating those alternatives. And the second is, I think that we have been transparent with them, and I think they appreciate that transparency.

Okay. Good. Again, shifting to the maybe customer response, and I wrote one can envision shifts in buying patterns, you brought up Keystone. So shifting from kind of a CapEx to an OpEx model. We talked about a renewed interest in hybrid flash. Can you maybe just elaborate a little bit more on Keystone? I know it's smaller, but maybe the question is when you talk to customers, the intentions of shifting to an OpEx model, does that become some kind of permanency there? Does that become elongated? Is this specific to the environment? Or is Keystone something that in this market, it causes them to take maybe longer term changes to how they buy from you, so to speak?

Our Keystone business has been incredibly sticky. We originally brought Keystone into the market, because we thought clients had temporal use cases where they said, I'm not really sure about my plans. I want to use your Keystone as a bridge to an alternative plan, maybe to go to your cloud offerings or retire an application. What we have found is, it is incredibly sticky. And many customers who started out saying it would be temporary now have run those environments for many years, well beyond the original contract term. I think with regard to these types of environments, we have seen in prior use cases, prior such situations, people moving the mix of how they buy to maybe cheaper alternatives to traditional flash-based solutions. We have seen them look at, hey, should I finance the environment using leasing or some other alternative? And so we expect them to continue to consider those. And perhaps Keystone is one of the really good ways that I would envision a customer looking at choices.

And then just one point of clarity. I believe Keystone is margin accretive. Can you maybe just talk to, again, the financial benefit that you get from Keystone as well there?

Yes. Keystone is, you deploy equipment at the customer premise initially, so it would be a smaller revenue and lower margin initially. But over the course of the term, it becomes very, very profitable. And so we think that as Keystone becomes a broader mix of customers with the balance of mature customers, and new adopter customers, the margin trends up. And for a mature customer, it is accretive to the gross margin profile of the company.

I want to maybe go back to a comment that you made earlier when we talked about international markets or at least geographic trends. You did mention Europe showing improvement last week. What's the strategy that you guys have for expanding in some of these maybe smaller, but because of that faster growth market. So are there specific markets where you see an outsized opportunity to either capture share or where there is faster growth that you can go after, to get to kind of change that growth algorithm?

Yes. I think the first is we have been able to outpace much larger competitors by being focused in where we deploy our resources. When you're competing with companies that are 5x your size, if you try to cover the waterfront in as many places as they do, you won't be able to beat them. And so we have concentrated our attack plan in the biggest markets, and I can tell you that, for example, in Europe, if you lose #1 position in Germany, France and the U.K., the smaller countries in Europe don't matter. They can't make it up. And we have successfully executed that strategy. We have -- early on, many years ago, we decided that China was not going to be a durable long-term bet for American tech companies. And so we partnered with Lenovo to use their scale and local knowledge in China and built a JV that has been a good win for both companies, and we continued to do that. Within the sort of the mature markets, North America, Europe, Western Europe are really the focus areas, and we are -- we have gotten to #1 in many of those markets. And our aspirations are to build a significant moat between #1 and #2 in those markets. We have also invested in select growth markets that we think are longer-term bets. So India, which we think is going to be a big market over time; Middle East, which we see as an AI hub; we're expanding our investments in Korea; we see the Japanese incumbent vendors and storage are mortally wounded, so we're investing in Japan. And then in the rest of the world, we are leveraging strong partnerships with locally knowledgeable local scale players to penetrate the market.

Okay. You -- I don't think you mentioned leveraging AI for software development. And obviously, that's kind of been a big theme over the last few weeks, not necessarily in the world of hardware, but I'd love to maybe just, I don't think we've asked this question, but is there any AI disruption risk when we think about the fact that you do have differentiated software, and it's such a focus point for you guys, is there any AI disruption risk as we think about that?

Yes. We use AI tools in software development. I think that those tools require man in the middle or human in the loop to make sure that they are accurate, that they don't compromise the resilience, the performance of the system. And so you need to have a lot of knowledge of how distributed systems work to actually use AI tools. It's not like AI tools can just generate a file system, right? We are decades away from that despite what other people might say. I think the second is we are using AI tools to build velocity and productivity into our development organization, velocity measured by how much output we are able to deliver in terms of payload to bring competitive capability to the market. And productivity measured in terms of the number of get commits per developer, for example, and we're seeing encouraging trends there. I think the third area is, we are using deep learning and machine learning models to automate the behavior of our systems to optimize the cost performance of data across its life cycle and most importantly, to enhance built-in cyber resilience capability so that we can detect anomalies or malicious behavior instantly and block it.

And I want to ask you maybe building on that in terms of what you're doing internally as we think about companies leveraging AI to be more efficient, to be productive. You mentioned some of those initiatives. What's the impact that you -- that we should maybe expect on your operating model, headcount needs, your ability to drive leverage in that OpEx base?

Yes. As you know, we have been a steady grower of operating profits over many, many years. I think our operating margins this past quarter were north of 30%. And so we feel good about the trajectory of continued discipline. We are using AI tools broadly across the company to make our go-to-market more efficient to deliver more value in products to streamline the back office. And our expectation is to continue to do that. We will balance sort of margin accretion to invest more in the business as well. We see an opportunity in the market where we have some tailwinds from AI and cloud and some of these growth opportunities and also several weakened competitors that we can attack.

Yes. Okay. So just to make sure that -- maybe to summarize that is we'll find ways to drive efficiencies and gain efficiencies, but there is also an opportunity to reinvest, and we don't want to necessarily not do that, so to speak.

Correct.

Right. Great. Last 2 questions for me. Just kind of capital allocation and capital structure beyond share repurchases and dividends. If you could just -- I don't think anything is changing there, but just kind of maybe reinforce that. And then the second part would be just the strategic importance of M&A to either the financial model or expanding, complementing your current technology set?

Yes. I think we have been disciplined stewards of capital. We have returned in the range of 100% of free cash flow to shareholders over the last many years through a combination of dividends and buybacks. I think when we look at the go-forward opportunities, I think we will be, again, on M&A with large-scale M&A, never say never, but our strong preference is for tuck-ins, especially software tuck-ins that enhance our differentiation in capabilities like cyber or AI data prep and data pipelines.

Okay. Great. I want to give you the opportunity as we're wrapping up here. Just to maybe share the final word, and that is either what you're most excited about as we look forward, that could be what you think Wall Street might not fully appreciate about the NetApp story. But just what's the kind of final message you want to leave everyone with here?

Yes. We are excited about the opportunities that data has to transform businesses. AI is really dependent on high-quality data and unifying it is a requirement for all enterprises so that they can be successful. And we have been stewards of unifying data from 2004. We talked about unifying different types of data, and then we talked about unifying all of the locations you have your data. And then we are adding a suite of services and capabilities to make it easier to secure your data to protect it and use it for competitive advantage. So our multiyear view of what the enterprise will want and the bets that we have made are paying off, giving us the ability to gain share in all-flash storage, to have a unique and fast-growing high-margin cloud business that continues to create a competitive wedge against our traditional competitors and to innovate in new areas like AI, data prep, cyber resilience and new business models like Keystone. So overall, from a top line perspective, feel really good. We have an experienced management team that has been through many of these cycles before. And so we have kind of capabilities to manage through kind of memory costs. We have a broad portfolio and different ways to serve customers, and we are disciplined operators. So I feel like we have strong long-term value creation potential in a market where traditional competitors are much, much weaker than they ever have been, and we intend to take advantage of that.

Perfect. We are just up on time. Thank you very much.

Thank you. Thank you very much for having me.

Thank you.