Very happy to have with us NetApp from the company, Cesar Cernuda , President, as well as Kris Newton from Investor Relations. I'll do a 30-minute fireside, but Kris is going to do a safe harbor for all of us to wake us up.
That's right. 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 on 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 the show.
Nice work.
Thank you.
I'm new to the story. I think maybe some other investors are as well. The company's been around over 30 years, started with disk storage, obviously has made a big transition to flash and cloud. Maybe you can kind of just, from a very high level, walk us through how you think about the business, where it's come, and where you are planning to push it.
First of all, thank you. Good afternoon to everyone. Great pleasure to be here. As you said, we are a 32-year-old company, an experienced company. We define ourselves as an intelligent data infrastructure company. We help our customers to build their intelligent data infrastructure. We help customers on their data management, data storage capabilities, reliability, security around their data. We've been building the company on top of our ONTAP software, which is our biggest differentiator. The way we've been monetizing this ONTAP software has helped us with data management, data storage. The way we monetize historically has been selling our hardware and software together. As you have rightfully said, we start with a value proposition around disks. We moved into flash arrays, and lately as well on cloud storage. Lately means the last 10 years we've been already working on the cloud storage.
I think that's one of the biggest differentiators, because most of our customers, many of our customers are truly hybrid cloud. We have managed to build a very in-depth partnership with the three hyperscalers. Ten years ago, we started bringing our technology, in this case our software, into the hyperscalers' marketplace. Since five years ago, we were the first ones, and actually the only ones that have managed to bring that software, that technology as a first-party service inside the hyperscalers. In other words, five years ago we launched Microsoft Azure NetApp Files with Microsoft. That's part of their own console. Two and a half years ago, we launched with AWS FSx ONTAP, which is again in their console. They offer this service storage with our technology. Since one and a half years ago approximately, we announced Google Cloud NetApp Volumes with Google Cloud.
All in all, a lot of progress made. The mix of our business has been changing a lot in the last years, flash becoming two-thirds of our overall hybrid cloud business, public cloud, mostly cloud storage being the driver of the business. As you all know, we report the business into hybrid cloud, which we have flash, and our public cloud business, which is what I just talked about.
Maybe can you give us a perspective of the weightings between those two? How big is hybrid versus cloud, and kind of what are the relative growth rates, and kind of how do you think about the go-forward growth rates for the different segments?
I'm going to take the numbers and then I talk about the growth levers.
Yeah, sure. Public cloud makes about 10% of our total revenue. The remainder is the hybrid cloud business. As Cèsar mentioned, that's a combination of our flash business, which has been growing significantly fast. We gained about three points of share last year in All-F lash, outpacing any of the other competitors. Then a hybrid flash hard disk business, which has been declining as customers shift to All- Flash-based technologies. However, we expect that to asymptote out as there's a segment of kind of cold storage that's most cost-effectively hosted on nearline-based disks today. Those are the primary components of the hybrid cloud segment.
I think from a growth levers perspective, as we think about the future, we see four big opportunities for us. The first one is flash alarm . As Kris just said, last year for us was a 14% growth. We actually outpaced the market. We grew almost three points market share in the last year. That is an important growth opportunity. It is a market that is growing, and we are outpacing the market. There is no doubt that for us, that is a big investment area. I will talk later a little bit more about some details there. Second big opportunity for us is block, block storage. We are not new to block. We have been selling block for many years. As you know, we have approached our customers with a unified storage value proposition. We have thousands of block customers, but they were using block mainly through that unified storage approach.
Since one year plus ago, we brought to the market a new product, which we call it ASA, which basically is focused on bringing to customers only block. We recognize our customers that only need that functionality. They do not want to pay for all the unified storage. They do not need all the features that we are bringing with our unified storage value proposition. That block storage has also been a good growth engine for us. We actually captured one percentage point market share in the last 12 months, which is as we were expecting. There is another big growth opportunity. The third one is public cloud. That is one of the areas, and I said it at the beginning about NetApp, but this is one of the areas that it is critical for me that we all understand the differentiation that we bring to that market.
First, we're talking about hundreds of millions of revenue already with last year 43% growth. This is a really important growth engine for NetApp. The biggest differentiation is that, while many other companies are talking about the partnership that we're building with the hyperscalers, we're the only ones that have become a first-party service in those hyperscalers. In other words, today there's thousands, thousands of ISVs, thousands of companies that are part of the hyperscalers' marketplace, and NetApp is one of them. We have different products or services in the marketplace. To be very transparent, we have had that since 10 years ago. Ten years ago, we started bringing ONTAP into the marketplace. We call that a third-party service. We've been doing that for many years, and it's a growth engine.
However, as I said before, the biggest differentiation in that cloud storage strategy and relations with hyperscalers is that they have decided to bring our technology as a first-party service in their own console. They are the ones selling that technology, that product through their sellers, through their channels. You could think about it somehow as an OEM business for us. We do the co-engineering, and they are the ones going to market with that product. It is not truly an OEM business because we have invested, given the strong partnership we have with them, bringing cloud specialists that can help the salesforce because we are experts on that infrastructure and storage piece. We have also invested with some cloud architects on storage. We have also built and created a customer success team.
Once those customers have deployed or have acquired that product from the hyperscalers, we bring our customer success team to help them on the deployment, on the ramp-up, in that customer success, customer value in the public cloud. The fourth growth lever, so I said flash, I have said block, public cloud, and the fourth growth lever, which I'm sure we're going to go deeper, is AI. What we talk about there is enterprise AI, where we believe that we have a huge differentiation as well on value proposition. Very high level. If you think about AI, AI is basically built on unstructured data. That has been, and it is, the sweet spot of NetApp, unstructured data. We have a huge install base, and that huge install base has a lot of unstructured data. It's a very natural conversation for us with our customers.
I think the last years what we have seen is many projects, many customers thinking about building large superpods or large LLM models, learning models, or training models. We have been competing there, and we have been bringing value proposition to our customers. The real big opportunity that we see just got started, which is what we call enterprise AI. Three things. I do not go three things there, but yeah.
Yeah, we'll dig into them. Maybe one more just kind of high-level question for you in terms of the, you talked about gaining three percentage points of share in hybrid cloud. Can you maybe just give us the landscape of who you compete against and why you're winning? And then maybe the same kind of question for cloud. You mentioned the three major hyperscalers. They have their own storage products. So who's your competition in cloud versus hybrid cloud?
Yeah, look, on the hybrid cloud space, we compete with companies that have been there for years on the storage side. You'll know the names. We believe that we have a better value proposition because we have actually the first thing that I'll say is we have refreshed in the last 18 months pretty much all our products. So we have a complete new set of products that brings great performance, reliability, security. We have a 32-year history. That's one thing that is important for our customers is the trust. We have a very loyal customer install base. And the trust is based on the reliability of our technology, especially in the security side, data management, differentiations, et cetera. On the flash side, on the hybrid cloud, certainly flash has been a big growth lever. As I said, block has also been a big growth lever.
We're working on modernizing data centers, modernizing their infrastructure. Certainly, and we see a lot in Europe as well on their sustainability piece, how our technology can help them on their ESG scores and on their sustainability agenda. All in all, that has been our hybrid cloud overall value proposition. It's true that AI is a combination between hybrid and public cloud. Let me kind of move to the public cloud and then come back to AI. On the public cloud, here's what we see. We see some NetApp customers. We have a combination. We have thousands of customers in the public cloud. We see a combination of existing NetApp customers that have moved workloads that were working with NetApp or workloads that were working with our competitors on-prem to the public cloud.
They want to have that with our technology with ONTAP because they need that performance, they need that reliability, and the interoperability between their on-prem world with the public cloud world. At the same time, we have a single control plane. Many of you know about BlueXP. That central control plane can help our customers have all the data management capabilities in that central control plane of the data they have on-prem, but also in the public cloud. This public cloud, once again, is a multi-cloud because they can do that with the three hyperscalers. For us, it has been existing customers moving existing workloads or new workloads into the public cloud, but also a great source of new customer adds. Customers that wanted to go and move some of their workloads to the public cloud.
The public cloud, the hyperscalers decided that the best place to kind of store their data was with our technology. That has been a great source for us of new customers. As you can imagine, those customers working with our customer success team, I'm talking the non-NetApp customers, we talked to them about their on-prem needs as well because they might have also their own data centers. That is a new source for us to go and bring our hybrid cloud business. The last piece is the AI. We approach it there with different motions. Think about AI. Many of our customers or many customers have been building their LLMs and training models with hyperscalers. As they go and think about the inference, the modernization of the data lakes, think about the RAG or fine-tuning, they're thinking in doing that in their on-prem environments.
That is a huge opportunity for us to connect hybrid cloud with public cloud. That is a good growth opportunity for us.
I'd like another kind of overview question because I wanted to talk about product sales versus subscription recurring revenue. So maybe you can walk us through the accounting of if you sell an appliance, it's flash array, that'd be a product sale. But what's the recurring amount and how big is that of the overall mix?
You got it. When we sell a traditional product, it's a combination of hardware, software. More than half of the value is related to the value of the on-top software package. The revenue for that product sale is typically recognized in the period that it was sold. Associated with the product sale are support contracts for the hardware and software. It's a 100% attach rate at point of sale. Those contracts tend to be three to four years in nature. The revenue for them is recognized ratably over the life of the contract. What's not recognized goes into our deferred revenue balance. We receive cash for the entire thing upfront, and that's why we have such great cash flow. We also offer storage as a service in our hybrid cloud segment. That is a baseline revenue agreement. You've got a minimum commit.
There's incremental charges if they go over that commit. The contracts are multi-year in nature, recognized ratably on a monthly basis. The unbilled portion of those contracts goes into our unbilled RPO. Finally, on the cloud side, that's largely a consumption business. That is recognized as the data is used or the capacity is.
I don't know if you have an overall percentage, but I'm assuming since the cloud is growing so fast that the percentage of subscription recurring.
Cloud is consumption. It's a pay-go model. That's how the public clouds tend to sell. The support revenue stream is about half of the hybrid cloud business. And Keystone, while we don't disclose the revenue, is growing pretty rapidly. Those two elements compose the subscription services. Support alone is about half of the hybrid cloud segment.
Trying to figure out as more and more moves to the cloud, what's the impact. I mean, you talked about you have some hybrid customers that might migrate. You might get some share from other people's hybrid. You might even win some cloud customers as well. Maybe you can just talk us through, 43% growth is quite strong. Is there a way to kind of think about what's contributing from those factors to that growth?
Yeah, I mean, you can go through the numbers. I think from a strategy point of view, little numbers. The way I will share is, look, as we think about the footprint of our customers in their on-prem environment, we're gaining share through our big push and new products on the flash side. I think, Kris, as I said before, is our block value proposition, modernizing those data centers. I think Keystone has been a good growth engine for us as well. I mean, given the choice for customers to consume our technology as a service in an on-prem environment through Keystone is a good value proposition for them. I said before a couple of times, I'll go deeper on the enterprise AI. I think this is a good opportunity for me to share that because this is a really good growth engine for us.
I said before, in the last years, you've seen companies really focusing on creating these training models, LLMs, SuperPODs. As I said, we've been participating in all that value proposition and offering. We have several customers have gone with us there. What we think is going to be the biggest opportunity is now what we call enterprise AI. Think about enterprise AI in three vectors or three areas. The first one is modernizing data lakes. The first discussions that we have with our customers is, we have all this unstructured data. We have all this backup data. We have all this legacy data. How can you help me to ensure that it's ready for AI? One of the key messages that we openly share with our customers, hey, don't be confused.
Your job is to make sure you bring AI into your data, not the other way around. In other words, let's go and work on building that intelligent data infrastructure where you kind of modernize those data lakes and take advantage of all the data that you have. We can then really go to the next level, which is what we call AI factory. In other words, how do we do the inference? How do we do the RAG? How do we do the fine-tuning? Some of those models are going to come from the work that you have done with hyperscaler. Customers have realized, hey, if I go and do all the inference, the hyperscaler is going to be extremely expensive for me. What we're seeing customers doing is, hey, I'm going to do it in my own on-prem environment. I have the capabilities.
I need NetApp to help me build that intelligent data infrastructure with your hardware and software. By the way, given your interoperability with hyperscalers, I'm in a really good position to go and take advantage of that in my own data center. We have seen as well customers saying, I'm going to use a third party. I'm going to use some of the NetApp partners that have been building those AI labs to have those AI factories, WWT, CDW, other partners that have been building all those capabilities with us. We also have seen the hyperscalers themselves trying to bring all that functionality into an on-prem environment for them as a stack. We're working with them. For example, Google Distributed Cloud, which, as you know, we did a public announcement where basically we're teaming up and ONTAP is integrated there as well.
We're seeing that as part of the overall AI factory. First one, modernizing data lakes. Second, what we call AI factory. Third lever is what we call AI as a service. Think about AI software companies, companies that are basically integrating AI elements and they need data management capabilities, data storage capabilities to really help them to operate with a hyperscaler, but also in their on-prem environment to serve their own customers.
Or public customers.
Or public.
A lot of tier two clouds are building AI as a service on NetApp.
Very good point.
Yeah, I guess is there a way to parse it out? Because it sounds like AI is the driver of them consuming the same services. Are you getting any value add for if it's an AI application, do you make more money on that? Or is it more about just driving the share to NetApp in terms of storage?
I'll say when you think about that modernization of the data lakes or you're thinking about, well, each of those scenarios basically require NetApp technology, both, most of the cases, both hardware and software. That is a new opportunity for us. We think about it as a new workload inside our AI umbrella. The truth of the matter is we're seeing the peak of the iceberg. I think we're seeing the starting point of what enterprise AI will become. We have some wins. The pipeline has been growing. I think customers are really preparing their data for that inference. They've been preparing their models. They've been doing the training. We are quite optimistic for the next years to come.
I want to ask you kind of a question. I mean, I've known, I cover semis, but I've known your company for a very long time. I do feel like investor sentiment has swung over that period. What investors typically miss on your story? I mean, you're talking about some very good growth in cloud and AI and such. Obviously, the sentiment on your stock hasn't always been high. I think I remember some lows a couple of years ago. What are people getting wrong? Are people still getting stuff wrong?
I don't think you're getting it wrong. Probably we need to tell a bit of our story. And that's what I'm here for. What I'll say is our growth lever, the company has been transforming a lot the last years. That will be the first thing that I'll say. The places where we're growing the business, our growth levers, and big bulk of our business is in areas that are growth engines. Flash is one of them. As I said, 14% growth, winning market share. Certainly, block is another good opportunity for us. But probably, I think the two areas what we probably could do a better job telling the story and going deeper is the hyperscaler differentiation. We're talking about hundreds of millions of dollars. And last year was a 43% growth.
Of first-party and marketplace cloud storage services.
Cloud storage services. AI, the way we're approaching AI with that enterprise AI approach that I share and our market share on unstructured data, the partnership with hyperscalers and that interoperability between the on-prem world with the public cloud for AI is key. Probably those things are key things for me to land. I don't know, Kris, what do you think?
For those of you who followed the company five years ago, I would say we're radically transformed. The composition of our revenue has shifted from low-growth markets to higher-growth markets. The quality of the revenue is significantly higher as well. We are now servicing and growing on All- Flash, which is a growth market, our public cloud services where we're uniquely differentiated and seeing tremendous market growth. Those two business areas also carry higher gross margins than the old business. We're seeing greater gross profit. You can see that in the improvement of the consolidated gross margin of the company. We continue to be very fiscally prudent as we manage the business, returning up to 100% of free cash flow to shareholders in the form of a dividend, which takes about $400 million per annum to service and the remainder going to share buybacks.
We're slowly bringing down share count to help accelerate earnings growth.
Just in terms of your cloud business, these three hyperscalers, can you walk us through the timing of that? To be 10% of revenue, was that all three kind of contributing? Or is that still layering in, I guess? What stage are you in terms of that deployment of cloud?
I mean, the first partnership we went on was with Microsoft. We start earlier with Microsoft. The second one was with AWS. The third one has been with Google Cloud. Each one of them was one year, half approx later. All of them are contributing to our growth. All of them are actually performing quite well. We're winning new logos with them. Yeah, we don't disclose the numbers for each one of them, but we start different time frames.
I'll point out that's for the first-party natively integrated services that they're branding and selling. We actually started down our cloud journey over a decade ago where we posed a question internally that said, we don't make storage. We monetize ONTAP through other people's storage that we build into an integrated system. Why can't we look at the Amazon cloud as a similar monetization vehicle for ONTAP to put more data under the auspice of ONTAP management? Whereas a lot of our competitors, those who have server businesses, looked at cloud as an existential threat because servers are fungible and why not move to the cloud. We were very early to adopt the cloud. I think that enabled us to have these unique relationships that we do today, coupled with the fact that our history is in file services.
That was a gap in the clouds where our legacy, our enterprise tested robustness, the breadth of what we do around file services really did help the cloud providers fill a gap for transferring large enterprise file-based workloads without having to refactor them. We were kind of a sweet spot business, but we also looked at it as a huge opportunity for us. I think it's paid off very well.
Just curious, people are always scared of hyperscalers in the semis as well as in software. It's the same thing. People always are worried that they could do their own thing. For the hyperscalers, do they actually even care? I mean, you're kind of a value add. You have a software layer on top. They're still, I'm assuming, getting economics on this. Are they actually pushing your product? Or is it kind of, and would you.
Each one of them, as they go and push it, they're going to get some enterprise applications, applications running on their cloud. There's only a piece for them. The rest of the piece, and basically, they can move that workload if they have the technology there. We are an enabler for them to go and move that workload.
Exactly. They do get economics from it. We have a revenue share arrangement with each of them. Right in Google and Amazon, it's our software running on their infrastructure. They are getting compensated for their infrastructure. We are just getting the software piece. In Azure, we deploy our system. We are getting compensated for the software as well as the underlying storage. It gives Azure an opportunity to drive more and more enterprise migrations.
I want to ask you maybe the final question just on gross margin. You mentioned it. In terms of these different pieces, can we talk about how that would affect the mix? Generally, people look at flash storage, very cyclical market, limited sector calls. How much of an impact does that have on your gross margins on the product business?
All right. I'll start with the pieces of the least mobile gross margin pieces and we'll end with the most volatile. First, support revenue. I mentioned it's a significant portion of the company's total revenue. That's been a fairly stable gross margin, roughly 92%. We leverage AI and a lot of our support capabilities, which is why we've been able to get our support numbers so high. Next, it's a fairly small portion of our business, but professional and other services. I'm sure most of you know PS is historically a very low-margin business. Our Keystone Storage- as- a-s ervice offering is recognized in that line. It's what's been growing our professional and other services line and also what's been driving up the gross margin there. Cloud, a growing, a fast-growing part of our revenue. It's also an improving gross margin.
We gained 11 points, so 1,100 basis points of gross margin in the past year on that business. More and more of what we're deploying is software only. As we're seeing revenue scale on fixed assets, that margin is improving. We exited Q4 at 79.3% and said we expect continued improvement above 80% as we move through FY2026. Now finally, product gross margin. I mentioned earlier that half of what customers buy from us is software. The impact that flash has is only on the hardware piece. Flash is less than half of our COGS. It is 25% of the total customer purchase. Generally, our industry is one where the commodity price fluctuations are passed through. There is always some timing difference. When you're able to pass through price increases or hikes, it's not a perfectly efficient market.
You may see some near-term tail or headwinds to product margin. Over time, the underlying commodity fluctuation should not really have that much of an impact. We have said our long-term target for product gross margin is in the upper 50s. We have been there. We did see a bit of a dip down in Q4 and expect Q1 to be fairly flat quarter- on- quarter, but then again, continued progression upwards.
Thank you very much.
All right. Awesome. Thank you for having us.
Thank you. Thank you.