Okay. Thanks a lot, folks. My name is Simon Leopold, Raymond James Semiconductor and Data Infrastructure Analyst. So we're here at our TMT+C conference here in New York City, Fireside Chat. And we have from NetApp, Jeriel Ong, who is a director of the Investor Relations team. So Fireside Chat format, but folks have questions in the audience. But Jeriel.
Before I start.
Sure.
Do you want to do a Safe Harbor?
Yes, the Safe Harbor statement. The team will kill me if I don't read this. I almost forgot, so today's discussion may include forward-looking statements regarding NetApp's future performance, which is 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 our website at netapp.com. NetApp disclaims any obligation to update information in any forward-looking statement for any reason.
You're now safe.
Yes, I'm now safe.
Okay, good. It's interesting. Only about a third of my companies are reading Safe Harbor. I don't.
The rest of them are taking their risks.
Yes. Do whatever you want, whatever makes you happy. We're good. We're accommodating. So I'd like to sort of start off in order to maybe set the groundwork for maybe you've got some investors who are interested but new to the story. So how do you like to introduce NetApp to a prospective new investor?
Yeah. So NetApp's a data infrastructure storage provider. We make your data valuable. We make the insights valuable. We deliver this variety of mediums as a service through our solution called Keystone in the public cloud on three major public cloud providers, and then in a traditional kind of CapEx delivery way through our HD and hybrid solutions, as well as our all-flash solutions.
Great. And I want to sort of split it up. I definitely want to get to sort of the AI discussion, but I want to get to kind of the foundational aspect of the business, which is selling to enterprises. So I feel like we've gone through the last couple of years post kind of the supply chain dynamics from the pandemic with worry about macro recessions, slowing enterprise. So what's your take on sort of core enterprise demand?
Yeah. I would classify core enterprise demand as it's not amazing. It's not terrible. I think you can kind of see that in our most recent results. Our revenues, excluding the U.S. public sector business, grew mid-single digits in the private Americas business in APAC, as well as EMEA. So that is, if you look historically at NetApp, not the strongest revenue growth that NetApp has done in the last five years. Historically, not the weakest either. So it does feel like it's a tepid environment, but neither too strong nor too weak.
And are you observing any sort of notable deviations either by geography or by vertical other than the federal? And one comes to my mind is historically, and I don't know the reason, you guys have typically had very high exposure to Germany. What sort of ebbs and flows or differences are you seeing, or is it really uniform across the globe?
Yeah. We are a global company. About half of our business is based. We have about a third of our businesses in EMEA, predominantly Europe, and then probably somewhere in the mid-teens of our revenues are in Asia-Pacific. To what Simon mentioned, we do have a pretty strong position in Europe. Germany's number one market share geography as an example. There are other European countries who are number one, and in the Americas, we're solidly two or worse. Dell is by far the largest company in the United States in terms of share. Yeah, but so in Europe, we are a little more macro exposed relative to Americas, where I think some of the company-specific performance can outstrip a broader macro environment. Yeah, there are countries out there that are underperforming and overperforming.
I think on the whole, across these big geographical splits, there's nothing really to call out in terms of something that is overly one way or another. I think the main callout to your point and to your original question is in that U.S. public sector business, which is down year on year in the most recent quarter that we reported, and down year to date for the first two fiscal quarters of the year we had so far as well. And that's been driven by a bunch of cost cutting and cost control efforts that's happening at the U.S. federal government. The U.S. federal government is the majority of the U.S. public sector business. We have some state and local business in there as well, but it's probably about three-fourths of our U.S. public sector revenues, and that's really been what's been pressuring.
The way I look at it is the business really excluding the sector grew mid-single digits, but our total revenues grew closer to 2%-3%, and that's been really weighed down by this public sector dynamic.
So some of the other companies we follow selling into federal vertical have said, "Oh, but our federal business is mostly DOD, and DOD is not getting cut, and so we're going to be fine. Don't worry about it." Others have said, "Well, if you're exposed to agencies like Department of Education, you're going to feel it." What's sort of maybe double-click the mix of your federal exposure?
Yeah. So our U.S. public sector disclosure is, any given quarter, anywhere from 10%-14% of revenues, depending on whether it's a heavier quarter or a lighter quarter. U.S. public sector is about 75% of that. So let's call it high single digits % of our total revenues. As I split down that high single digits, it is pretty split evenly into three categories. So there's a military angle to that. That's probably about a third. There's FBI, CIA, DEA, those kind of agencies. That's probably about another third. And then the remaining thirds are what I would call civilian-based agencies. I'm not sure I have any triple-click to click down on what exactly of the federal government's being impacted between these three segments, but that kind of gives you a rough sense for the mix of our business.
That federal segment as a whole is really what's holding down our U.S. public sector revenues in terms of declines relative to that state and local portion.
So I want to pivot the discussion to supply chain and memory chips. And maybe as a first step, educate maybe somebody who's new to the story. You've got different types of memory. There's DRAM, there's HBM, and there's NAND. NAND presumably is the important one for you guys, but maybe give folks a little bit of education, and then let's talk about why this is important to you guys.
Yeah. So we are an enterprise storage provider. And so what that means is customers are buying our products because they have a certain workload that they want to store those bits on. I think that it's important to highlight that customers are buying our products not for direct access to bits. They're buying it because they value ONTAP as a software system on top of those bits. So I would argue if all you want is access to bits, any enterprise storage provider is actually a very expensive way to get access to those bits because we are a 70% gross margin company. So when customers are buying our products, DRAM is a very small portion of our bill of materials called low single digits historically. And then SSDs are a growing portion of our bill of materials because the mix of our revenues are increasingly becoming all-flash revenues.
We still buy HEDs. We buy SSDs, and SSDs are growing. If I look at our COGS, which is about $2 billion annually in that range, not even anywhere close to the majority of this is related to these memory components. It's less than half related to, let's say, HDDs and SSDs. SSDs growing the percentage of that mix. And so it is, there's a lot of money in COGS and other things like other semiconductors, assembly, freight, EMS costs, warranty costs. There's a lot in there that is not related to commodities.
I guess one of the challenges for the investment community is, I think, we're bombarded with headlines about spot prices on memory. Presumably, you don't buy much on the spot market. How does the company manage the supply chain? I'd say, in your defense, gross margins were actually pretty good this quarter. People thought they were going to be terrible. What is sort of the operational aspect that allows you to sort of not pay these spot prices?
Yeah. So we buy in volume. We buy in much higher volumes. And so that's the first thing. When you have a volume, there is a discount very relatively to spot that you can get from that. But we're also opportunistic. So we don't buy on a steady cadence. So if we need, let's say, one terabyte in month one, we don't just keep buying one terabyte every single month through the year. We buy it in spurts. And so yeah, we work with different providers effectively to achieve our needs. We have a procurement team. They forecast. They form their own view on the prices of these products and potentially which vendors to work with over the course of the year or two. And yeah, we remain kind of flexible to when we purchase commodities for our products.
I guess talking to you guys and your peers, it sounds like basically there is inflation, but there may be somewhat different strategies as to how you as a vendor go about raising price. We're hearing about combinations of, well, we do less discounting. We're raising list price. What's sort of been NetApp's approach and philosophy managing the price hikes with your customers?
Yeah. So what Simon is referring to is a little bit more complicated than perhaps just changing the price of an iPhone on a website. So we have a list price that is officially listed on our website or perhaps given to our channel partners. Typically, there's a discount rate that is applied to that. And so there's an effective price that the customer experiences. And so different companies in the storage industry manage the effective price in different ways. Some of them raise the list price and then have the same discount rate. And some companies reduce their discounting but have the same effective list price. But effectively, you're changing the actual customer experience price depending on your strategy. Our strategy is to try to manage both. I think that we raise our list prices if the commodity costs do increase, and we have in the past.
In fact, we did it recently, and yeah, then we try to manage up that effective price that the customer experiences through both efforts.
Now, in the past, I think maybe the prior cycle might have been 2017, which might have been before you were over there. But the perception we've had is that sort of when the market share leader raises price, everybody just follows. Is that sort of been the dynamic in this cycle, or is it somewhat different nowadays?
Number one, the cycle's yet to play out. So we'll see how this one plays out.
That's fair.
Yeah. I think that there is a little bit of a, let's wait to see if we are in line with everybody else type dynamic with these price changes. I think what happens is, again, prices are typically on a decline. They kind of edge down over time because commodity costs are coming down over time. That's kind of what enables these storage systems to address more and more storage over time. Yeah, on the price raise, they tend to be a little bit more abrupt, I guess, or a little bit more of like a step function up. And so yeah, there is a little bit of a what is the industry doing dynamic. I think the reality is the market leader is smaller today than they were in 2017.
So, there's perhaps a little bit of, like, a "let's wait to see what one company does" dynamic in the industry. Yeah, but we'll see how the cycle plays out.
But just to make sure I understand, your guidance, the forecast you've given the street for this year suggests that your margin should hold in and/or improve through the year. You're not expecting further degradation. Is that accurate?
Yeah. I think if you look at our fiscal 2026 guidance and what we've done this year certainly implies that gross margins kind of maintain in a similar range, and depending on your assumptions, we don't guide specifically to the product gross margin line. It could be in the range or maybe slightly lower, but not materially lower than the 59.5% that we just reported.
So I want to pivot to sort of the AI topic because that's sort of what's exciting these days. And I think it's interesting in that when we look at the IT space as a whole, compute's obviously sort of the center of the storm. And storage has not readily participated. Now, I feel like the narrative that we've been hearing from NetApp and peers sets it up to be a beneficiary. How do you envision sort of the storage angle of kind of the AI theme playing out?
Yeah. I think that our thesis is that as the AI industry shifts its spending from training AI models to inference, that storage has a bigger presence in that inference cycle in terms of A, maybe contributing data that's been created in enterprise over sometimes decades to improving that AI model, but also as that AI model is used again and again to create outcomes, that data that is created needs to be stored. And that is the inference step that will generate more data for an enterprise storage provider than perhaps the training stage has so far. So we still think that growth is still to come. Certainly, in the most recent quarter, we've already seen evidence of that. We talked about 200 design wins, AI design wins in the quarter, up from 100 a year ago. So that evidence is happening.
But that still feels like more so the tip of the iceberg in terms of future growth than it is. We are right in the thick of that AI inference stage really taking off.
So what can you tell us about these 200 design wins? Are there common themes? Are they 200 completely different projects? What's sort of the nature of the customers and the applications?
Yeah. There isn't really a geographical or, let's say, an end customer mix that really heavily skews one way that I can call out. What I would say is that there's still a lot of data lake monetization wins in that number. And what data lake monetization is, it's kind of like an in-between step in AI where you have a broad data state. Perhaps it's extremely siloed across different vendors, and you are aggregating that data into one data lake so you can improve your AI workloads. So it's a copy, effectively. It's a copy of your data. There's a lot of that still happening, which kind of indicates perhaps we're not as much in that inference stage strongly as we'd like to be. But I don't think that's specific to NetApp. I think that's more indicative of where the industry more broadly is at in this.
And so yeah, I think we hope that these AI inferencing workloads continue to take off. And we really think NetApp will be a beneficiary as that kind of plays out.
I think one of the logical and probably as yet unknown debates is when enterprises engage and use AI, where? So what's the mix that you would expect on public cloud, private cloud, on-prem, and what makes a difference to NetApp?
That's a tough one to predict, if I'm being honest. I think that's where just because the inference stage isn't in its maturation stage, I think a lot of industry participants are just in a prediction stage. I would say one hat, throw in the ring why it makes more sense on-premise is that a lot of the data that will improve and augment and improve the quality of your AI outcomes is data that's from on-premise, number one. So that's the first point. I think the second point is high performance, low latency workloads tend to be more on-premise. So if you look at the current spending in storage, if you have a lot of backup, a lot of recovery, a lot of low touch stuff, you don't care if the cloud is down for an hour. You don't really touch that workload very frequently.
But if you have a workload that is essential to running your company on a daily basis, you want that to be a low failure rate, low latency, high performant workload. That's an all-flash workload. That's an on-premise workload. So it feels like if AI workload looks more like that, that there could be a higher chance it's on-premise. But we'll see how the market kind of plays out.
Yeah. I have not made up my mind either, so I don't want to give you the impression that I have some bias on that. I think it's probably the interesting debate. We'll get more clarity over the course of the next year is my hope.
Yeah. Yeah. I think, yeah, we'll see what happens.
Then about a year ago, you guys announced a platform called the AFF, which I think inserted you more strongly into these AI discussions, maybe helped people understand what is it about the platform that was an evolution and what makes it more relevant to AI use cases.
Yeah. So AFF was announced at Insight. This happened in October of this year, so relatively recently. It's in the hands of some customers. So we'll see. We'll get a better sense as we move through fiscal 2026 what the traction on this is like. This product is fundamentally a little bit different than what NetApp has delivered in the past. So NetApp was founded in 1992. It was essentially a one-product company, something called FAS. It was HD-centric product. Eventually, we released something called A-Series. This was a high-performance all-flash product, NAND or SSD only. And most recently, a few years ago, we released C-Series, which is also an all-flash product. The common theme through all these three products is that we are delivering our products in an integrated approach.
What that means is they get access to ONTAP, but the only way they get access to ONTAP is when they buy our storage arrays. It is a product. It's similar to if you buy an iPhone. If you want iOS, the only way you're going to get iOS is if you buy an iPhone. You cannot get iOS without an iPhone. That's how the entire industry delivers storage technology to its customers. But really, AFF is disaggregated. It's AI-focused. And the AI market has kind of increased the significance of disaggregated solutions. And so this new solution delivers our technology in a fundamentally different way for a different market. And so we will see the traction that this has. And I think that it's really where customers are wanting to consume ONTAP technology.
How does this contrast with the AFF, the A900, is it, or A900?
Yeah. So it contrasts in this kind of integrated versus disaggregated approach to the technology. But also, the product that you're specifically describing is a very, very high-performant solution. So in terms of where exactly AFF stand in that stack is still kind of yet to be seen. We have a lot of options that consumers can offer there. But if it goes the way of all-flash solutions so far, I think AFF will probably be a pretty high-end solution as well.
So you mentioned sort of, well, you can't get iOS without the iPhone, but you guys have your PCS, your public cloud services. So isn't that sort of a way to get ONTAP without necessarily?
Yeah. Yes. Yes, Simon, so you make a good point here. So yeah, we had some experience already kind of disaggregating technology. So when we first started selling ONTAP in the public cloud, we actually had to install our hardware on the public cloud itself. But as the business has matured, it's really shifted to more a software-only solution where the public cloud providers sourced and installed their own hardware. So in that sense, yes, it is kind of similar in that sense where we have offered our technology in a more disaggregated fashion. Our public cloud line is increasingly a lot more software-only, so to speak. But yeah, I think in terms of that enterprise customer, they really haven't had a disaggregated kind of solution to consume unless they go for the public cloud. So it's a little bit different in that respect.
The way I've thought about it is, and I won't be embarrassed if you correct me, is that this is the way that an enterprise customer can have the experience with ONTAP but be in the public cloud. So they're using your software, your application, your interface, but it's public cloud economics, right?
Yeah. I think that's fair. Yeah. I think there's thousands of customers on our public cloud solution, big and small. I think on the small end, what's interesting is there are customers that really perhaps are large customers but start off at really small workloads that maybe wouldn't have chosen NetApp or NetApp wouldn't really have had a chance to address this customer in the first place. And then they become very big customers either through the growth in the company or the fact that it was just a testing environment that really worked really well that then they're now growing pretty materially. So we've seen a lot of good use case growth from this public cloud business. I think there's a sense out there that it's kind of a little bit of a left-hand, right-hand thing where they would have otherwise bought it on-prem.
Now they're buying on the cloud. Certainly, that doesn't feel like that's the case. Half these customers are new to NetApp customers, never gotten a core product for our business ever until they were a public cloud customer. And so it does feel like we're growing our customer base with this product. It's growing really well. The first-party marketplace revenue disclosure, which is about three-fourths of our public cloud revenues in the most recent quarter, grew in the low 30% range year on year. So this business is gaining a lot of traction. And yeah, it continues to grow really well. It's a good business for the company.
And not arguing that point, I guess one of the things I've sort of struggled with is, well, why would the public cloud want to introduce competition to its own storage offering? So for them, what's the logic?
Yeah. This is a good question. So ONTAP offers a set of features that is pretty fully featured. And it offers a different offering than what the public cloud has and has had already for coming on two decades now. So that's the main logic is that each public cloud vendor has an incentive to offer as much on the public cloud relative to what the customer could choose on-premise at a high level. And so NetApp offers that opportunity to move a workload that maybe would have been on-prem onto public cloud instead is really the single line. And really, within that category of fully featured products, ONTAP is, frankly, the best solution relative to the other core vendors that we compete with on-premise. And so every time a vendor chooses something in our category, so to speak, ONTAP is really winning in spades.
That's what we feel really good about this business. I kind of acknowledge the risk kind of side of the question you're saying, which is that, well, if you're getting so big that perhaps you're bigger than the public cloud itself on the public cloud, why wouldn't they just build a solution themselves? We're just such a small, insignificant revenue relative to their own solutions right now that it's really not even a question or a topic right now that really I think they're even thinking about.
Now, I want to talk a little bit about the competitive dynamics, partly because one of your CEO from one of your competitors likes to characterize as a knife fight in a phone booth. And that's just such a profound visual to characterize your market. Now, it seems as if it's competitive, but it also seems as if everybody's doing pretty well right now in terms of not getting hurt by memory costs. Maybe it's not quite as brutal as that. But how do you like to sort of think about or characterize the competitive landscape? It seems as if you guys have generally been a share gainer. What's enabled that?
Yeah. We're generally a share gainer, so our share has been increasing according to IDC data. Roughly speaking, we're in the high teens in market share in the all-flash market at this point. At some point, if you kind of rewind it back a few years, we were probably in the low teens, so we've been gaining share, and the market's consolidating. The top seven vendors in all-flash have 90% market. I think what's enabling us is, yeah, it's our operating system. It's our launch of our C-Series product a few years ago that enabled QLC NAND in all-flash enterprise workloads. That was an expansion of the all-flash market, and I think in that expansion phase, I think there were certain vendors that did better and certain vendors that did worse, and we happened to be one that did better in that transition.
But yeah, I think at a high level, yeah, it is competitive. I wouldn't say it's scale one to 10, is it 10 out of 10 competitive? I mean, I wouldn't say it's like crazy, crazy or crazy uncompetitive. It feels like we're still in that kind of middle ground where things are competitive. There's always, especially for a new customer that is considering a new very large workload, yeah, it's going to be competitive to win that. But yeah, I like NetApp's odds in general. I think we've done well.
I think one of the things we've observed, really, I'd say roughly in the last year is when we've done channel checks, it sounds to us that NetApp and Pure Storage are encountering each other more often than they used to. I felt like in the past, each of you had kind of a swim lane, and you did your own thing. It seems like you're competing more. We're also getting questions from investors about, I'll say, noisier private companies, the Wekas, the VAST, Hammerspace. What do you see in terms of the smaller players and the newer players?
Yeah. I think the smaller players and newer players are more disaggregated versus integrated in terms of the solution delivery strategy. They have a little more customer concentration, a little less features. But this is actually what I'm describing is actually very consistent with the history of storage startups. You can never just start a solution from the ground up and have a fully featured set. You have to kind of pick and choose what features and what products you want to deliver. And then as you grow as a company, you expand that. And so when a customer chooses them, this is why some of these companies are winning in the AI training phase is because they don't need these fully fledged out features that the rest of the industry already provides. So that's what we kind of observe there is that they're winning AI training.
We'll see what happens. Who can make this leap to AI inference or perhaps more enterprise use cases? But ultimately, you have to win in enterprise if you want to become a fully fledged storage company because the vast majority of the spending is from an enterprise customer. I think as it relates to Pure, I mean, yeah, they've gotten big. And so are we. I mean, I think our business, our all-flash business is in the $4 billion run rate size. And they're slightly smaller than that. And so as companies get big, you just kind of rub shoulders a lot more. I think this is also true for Dell or HPE. I think the market is more consolidated. The top seven manufacturers in all-flash have 90% market share, like I said. But in the non-all-flash market, the top seven have 55% share.
You just rub shoulders with a lot of different companies when you're talking about a hybrid or HDD estate versus the all-flash market.
Believe it or not, our time's pretty much flown by. I like to close with the following. What do you think is either the least appreciated or most misunderstood aspect of the NetApp story?
I would highlight just how NetApp has changed over the years. Like I said, NetApp was founded in 1992. At that point, all-flash did not exist. It was an HDD-only market. And so we've built up an all-flash business that is now two-thirds of our revenues. The cloud business didn't exist five years ago in a material size. And now that's run rate at $6.7 billion, built largely from the ground up. And so I think I'd just highlight NetApp's ability to kind of continue to drive innovation, continue to change itself as a company. A lot of companies in tech really have a hard time progressing beyond their first successful product. That is really the history of technology, frankly. And NetApp's really done that twice. And we're kind of in this kind of reinvention stage, so to speak. And so yeah, I think that will continue.
I think that's really the most underappreciated aspect of the company.
Great. Jeriel, thanks for joining us. Folks, thanks for joining us. This is our session with NetApp. Thanks a lot.
I think.