This presentation and the statements made on behalf of NetApp during this presentation may contain forward-looking statements within the meaning of the Private Securities Litigation Reform Act regarding our strategy, products and services, shareholder returns and our future results, performance or achievements, financial and otherwise. Such statements or projections reflect management's current expectations, estimates, and assumptions based on the information currently available to us and are not guarantees of future performance. Actual results may differ materially from our statements or projections for a variety of reasons, including without limitation, the effect of general global political, macroeconomic, and market conditions on our business, our ability to expand our total available market, and grow our portfolio of products, customer demand for and acceptance of our products and services, and our ability to successfully execute on our strategy.
These and other equally important factors that may affect our future results are described in reports and documents we file from time to time with the SEC, including the factors described under the section titled Risk Factors in our most recent filings on Form 10-K and Form 10-Q available at www.sec.gov. The forward-looking statements made in these presentations are being made as of the time and dates of the live presentation. If these presentations are reviewed after the time and date of the live presentation, even if subsequently made available by us on our website or otherwise, these presentations may not contain current or accurate information. Except as required by law, we undertake no obligation to update or revise any forward-looking statement based on new information, future events, or otherwise. This presentation also includes non-GAAP financial measures.
Reconciliations of these measures to the comparable GAAP measures are available in the appendix to this presentation. NetApp's investor relations website at investors.netapp.com contains a significant amount of information about NetApp, including financial and other information for investors. NetApp encourages investors to visit that website from time to time as information is updated and new information is posted. The content of NetApp's website is not incorporated by reference into this presentation, and any references to NetApp's website are intended to be inactive textual references only.
Disruption almost feels too small to describe all the changes we've faced. But amazingly, we humans always find a way to rise to the moment and move forward. At NetApp, helping you turn disruption into opportunity is what we do with you by making your data infrastructure intelligent. It's built on unified data storage, reimagined to let you store any data across any environment. Integrated data services, which help you protect, govern, and get more value out of your data. And Cloud Ops solutions, which enable you to work faster and more efficiently with intelligence. Others might try to convince you their same old storage will meet your next era needs, but this moment demands so much more. We make your data infrastructure more seamless, secure, sustainable, and high-performing than ever before. So whatever your opportunity, whatever the moment, we can rise to meet it together. NetApp.
All right. Hello, everyone, and thank you for joining us at our 2024 Investor Day. For those of you who haven't met me yet, I'm Kris Newton. I head up the Investor Relations Program here at NetApp, and I really appreciate you joining us today. We've put together a really great agenda for you. You'll hear from our CEO, heads of product and cloud operations, César Cernuda, our President, who's in charge of all the go-to-market activities, and of course, Mike Berry, our CFO. After the formal presentations, we'll have about an hour for Q&A where you guys can ask all the questions. We'll also be taking questions from those of you on the webcast. And then for everyone who joined here in person, we'll have lunch, and over lunch, we'll break into small groups, and you'll have the opportunity to ask questions of all the executives.
So I encourage you to stay for lunch. It is always a great meeting. All right. And with that, no further ado, I'll invite George Kurian to the stage. George?
Good morning. Thank you, Kris. Welcome to NetApp's Investor Day. I am excited to share with you the momentum that we see in our business because we are solving hard problems for our customers in the age of data and the value that we are delivering to shareholders through our leveraged business model. I want to begin by just sharing with you a few important messages that we are uniquely differentiated, that we have a strong track record addressing a growing market opportunity with a focused strategy that is delivering and will continue to deliver great returns to shareholders. Let me tell you about our agenda today. We will cover four topics. The first, I'll share with you the results of this past quarter and the past year and the underlying reasons for those strong results.
You will hear about our differentiated approach and the hard problems that we solve for our clients in unique ways. You will hear about how we are taking those capabilities and addressing large and growing market opportunities with a focused strategy that delivers profitable growth. Let me start. I'm super excited about the results of this past year. We took action at the start of the year to focus more clearly on our enterprise storage portfolio and optimize our approach to cloud. As you saw in the past quarter, our all-flash storage portfolio, which is now approximately 60% of our hybrid cloud segment, had very strong results, up 17% year-over-year to $3.6 billion run rate.
Keystone, our storage as a service offering, has enabled us to win new workloads and footprints in our customers and has reached a total contract value of about $150 million, growing in triple digits year-over-year. Our cloud storage, hyperscaler native and marketplace offerings grew very strongly to about 38% year-over-year growth and is now approaching about half of our total cloud revenue. Our focused bets are growing strongly, and we are exiting fiscal year 2024 with those bets now being a much more significant part of our overall business. Not only that, we coupled our momentum in the product and service segment with an operationally disciplined business model that delivered record financial results across virtually every element of our P&L, coupled with a shareholder-friendly capital allocation model. So now let me tell you why it is that we are delivering these results.
We call this the NetApp Edge, the unique set of capabilities that allows us to win the trusted advisor position with many of the world's largest institutions around their most important problems. We have a deep understanding of the market and a track record of helping our customers adapt to the changing business needs and leveraging their data for competitive advantage. It is because we have built and architected uniquely in the industry a storage and data management platform that is built for the age of data. We are able to do that because we have the talent. We are the employer of choice for the best talent in the storage and data management industry, and because we have over many years worked as the partner of choice for the leaders in the technology industry.
Across many eras of the technology industry, our ecosystem is the ecosystem of innovation and disruption, serving to bring our customers forward as their business needs evolve. This has enabled us to have a global customer portfolio and a financial track record of disciplined operational management, translating to strong results for our shareholders. So now let's talk about our customers. We serve the world's most important customers and organizations on their most important problems. We are in the age of data. You hear it every single day. Data is not only more important, but it is also much more complex to manage. The rate of growth of data, the diversity of data types, and particularly when you need to unify and secure your data for AI, you find that your data is spread across the hybrid cloud landscape.
And this requires an intelligent data infrastructure to bring about the two important elements: flexibility to deal with the challenges of your business, to unify your data for AI, to integrate the cloud seamlessly into your technology and business architecture, to be able to modernize your application portfolio, and to advance security while simultaneously meeting the needs of simplicity at scale so that you can be efficient, you can be performant, you can have the cost structure that your business demands, you can be secure, reliable, and sustainable. An intelligent data infrastructure is what NetApp enables for our customers through four key building blocks: unified storage and data management so that you can unify all of your data for any application anywhere and seamlessly and consistently manage it as your business needs.
You have integrated with that unified storage and data management platform, data services and operational services, either embedded into the software or closely integrated with it, so that you can simplify how you deploy security. You can recover from malicious attacks. You can instantaneously discover when someone is either maliciously or inadvertently compromising your security and access control mechanisms. And you can deliver massive productivity gains to your IT team, but more importantly, to the scarce data science and data engineering teams that are trying to progress your AI priorities. We have always-on intelligence to enable this architecture to be autonomous and to be self-healing and to be most efficient as the needs of data and business evolve over time. And we have integrated these capabilities into the world's leading ecosystems, the world's biggest hybrid public clouds natively.
We've integrated the biggest software architectures, and we've integrated it with reference designs such as with Cisco and Lenovo and NVIDIA. NetApp uniquely enables customers to build intelligent data infrastructures for the age of data. We do that because we have over 30 years of our history being building on a unified platform, a platform that is based on a unified intelligent storage and data architecture that combines that with software, with AI and ML intelligence, and APIs that allow all the world's applications to connect to it in a very, very efficient manner. At the foundation of that architecture of the platform is an intelligent data fabric consisting of a data plane, a metadata plane, and AIOps capabilities that power consistent data experiences and optimize storage operations for all of your data anywhere you need it.
You will recognize many of those capabilities as part of our ONTAP operating system. We have a very rich, the industry's richest suite of services, both data and operational services integrated with this unified storage and data management platform. We have, over the last couple of years, built a unified control plane called BlueXP that delivers unified control of all of your data across on-premises, the cloud, the edge, and any of the other places that you deploy your data. As I've mentioned, we've integrated this to the ecosystems that matter to make it risk-free, to make it efficient, and to be able to give your customers the ability to bring their IT teams into the new worlds.
As I mentioned at the outset, this delivers uniquely transformative flexibility to unify your data for AI, to modernize applications, to advance security, and to integrate the cloud, as well as the simplicity at scale that your business needs as data grows and your resources are constrained. We have led this development over many eras, and we've always believed in the promise of this unified platform. Now, recently, as many, many of our competitors have recognized the wisdom of our approach, they've started to copy the narrative and try to claim that they have the equivalent, but that isn't true. No one else has a single unified data storage operating system that combines any data anywhere for any application. No one else has the richest suite of data and operational services integrated into this foundation or embedded in it.
No one else has brought the capabilities of these have brought these capabilities and have natively integrated it into the world's leading public cloud. So we feel really strong about our position in the market, and it is not something that you can replicate easily. It has required discipline and excellent execution and architecture over many years to get to this place. It is helping us to serve the world's leading organizations. You see, winners choose winners, and the winners are increasingly choosing NetApp, as we proved through the second half of last year. Let me give you an example. This is one of the world's leading financial institutions. You might be able to actually see them if you turn around. We began our work with them decades ago, where we started to serve their unstructured data environment, unifying Windows and Unix and Linux environments for file storage.
We then brought their backup environment onto our platform as they started to move away from tape to online backup. Then I joined in 2011, and one of my first efforts was to work with them on their mission-critical environment, where we started with a Metro Cluster Continuous Availability Architecture that then allowed us to win their virtualization environment. We then were able to work with them to steadily and successfully displace our competition in every part of their mission-critical environment. Today, we are their leading storage provider, serving about 500 PB of installed data, which is north of 80% of their total wallet. Platform approach, delivering compelling value to the customer together with strong customer-facing teams has allowed us to grow our wallet with this client enormously over many years. We can do that in every large global institution around the world.
So now let's talk about the markets. We said we have a uniquely differentiated architecture and capabilities, and we are applying them to large and growing markets. The foundation of our market drivers is the fact that data is growing. Data is ever more important in the world of AI, and data is more complex to manage. The largest aperture is hybrid cloud and public cloud, which today we serve about a $100 billion market in fiscal year 2027, growing approximately 7% CAGR between now and then, with a larger part on-premises and a faster growing part in the public cloud. Data is fragmented across these landscapes, and we have uniquely positioned ourselves to have the largest footprint across all these landscapes because we not only have a strong presence in the enterprise data center, but in all the world's leading public clouds.
Now let's double-click on four important sub-markets that are growing at an accelerated rate or where we believe we can disrupt. You will hear about me talk about Flash, Block Storage, Cloud, and AI, which are parts of this overall market. Let's start with Flash. We introduced 3 new AFA families, all-flash array families for performance, capacity, and block-optimized storage. We believe we have the industry's best all-flash storage portfolio because we combine flash-optimized system design with the world's best storage and data management. We are growing faster than the all-flash storage market so that we believe we have a compelling opportunity in this $40 billion market in 2027. Second, Block. Let's start with File because we address any data type through our unified approach, combining File, Object, and Block.
We have a strong leadership position in file that we are expanding on by unifying file and object in the ONTAP operating system by integrating the public cloud so that customers can build very, very large file-based storage environments and use the public cloud for applications like AI or advanced analytics. We have deployed over the last year, without question, the best data security capabilities for the file storage market. We are absolutely confident that we can not only defend, but we are expanding our leadership position in the unstructured data market. We are the agent of disruption in the mature block storage market. Let me repeat that. We are the agent of disruption in the mature block storage market. We have block-optimized flash storage, and our software plus cost-effective flash gives us the price-performance sweet spot in the block storage market.
We are winning more and more customers with our block storage offerings. We deliver the most complete and the best integrated storage portfolio for any data, any application, anywhere. We allow our customers the broadest range of options to consume that portfolio any way they like: on-premises with CapEx models or storage-as-a-service models, or with elastic cloud offerings. Third, gaining share with cloud migrations. We uniquely in the storage industry saw the fact that the public clouds were going to become an important part of customers' IT architectures. Data management across the hybrid cloud was an incredibly complex problem to solve. We embedded our tools in all of the world's leading public clouds. The big three cloud providers, Amazon, Microsoft, and Google today, based on public data, control 72% of IaaS and PaaS in the public cloud market and are gaining share inexorably in that market.
As they gain share, we are positioned to win with enterprise workloads being deployed in the public cloud and born-in-the-cloud native workloads. Our team will tell you how we are expanding the innovation in the public cloud market, how we are reaching the distributed cloud and edge environments together with the hyperscalers. César will tell you about our successful go-to-market model with them to bring the best of cloud with the best of storage and data management to solve our customers' problems. Now, let me tell you about AI. It is, of course, the most important topic of conversation in the industry today: data. Unstructured data, meaning documents, files, video, audio, chat, log data, is the foundation for generative AI. We at NetApp are the unquestioned leader in unstructured data management. We hold the most important unstructured data for the world's most important organizations.
We are uniquely positioned with a leadership position that we intend to exploit for competitive advantage. We are also meeting customers' use cases for the entire AI lifecycle, from data lakes and data foundations to data pipelines for RAG and inferencing and tuning and training, as well as for deploying hybrid cloud architectures. We combine our industry-leading all-flash systems performance together with the richest suite of data services that allow customers' data science and data engineering productivity, rapid time-to-market with their new models, and the ability to deploy responsible AI quickly and productively. As you saw last quarter, we continue to see strong growth in our AI business, and we expect the market to grow even more substantially over the next couple of years.
I told you about the fact that we serve large and growing markets and that we have a focused approach to grow our share in specific areas: Flash, Block, Cloud, and AI. I will now start to hand it over to my colleagues by telling you about our strategy for profitable growth along four axes. I shared with you that we have a $100 billion market opportunity and that we are aligned to the fastest-growing segments in that market. Harv and Haiyan, our two product leaders, will tell you about our uniquely differentiated technology and product strategy to enable us to be the leader in those seg ments. César Cernuda, our head of go-to-market, will tell you about our focused go-to-market plan to address the needs of each customer segment in unique ways.
Then Mike Berry, our CFO, will talk about how we are driving growth with high-margin products in these fast-growing large markets, but coupling that with a disciplined operating model to drive earnings growth and a shareholder-friendly capital allocation model. So with that, let me invite Harv Bhela. Harv?
Thank you, George. Hello, everyone. How is everyone doing? Wait, like, you know, like, I'm sure the financial analysts can have more energy, right? Come on. How are you doing? Good. Good. All right. Excellent. Hey, I'm Harv Bhela. I'm the Chief Product Officer of NetApp. And joining me soon will be Haiyan, my wonderful colleague, to walk you through, you know, the presentation we have for you on the product. All right. Well, hopefully this works. Yay. Awesome. This is an amazing time for NetApp.
Haiyan and I are really pleased to be here to discuss our product and our platform, to talk about the growth opportunities we have, and give you a little bit of a glimpse of what lies ahead. Okay? So as George said, we are in the age of data. Customers tell us that for them to really harness the value of data, they have to overcome these six major hurdles that you see up there. Well, number one, you know, that data is exponentially growing. Another example would be that data is increasing in complexity. Their data is everywhere. It's on-premises. It's in multiple public clouds, fragmented across all of those different places. Data fuels AI, but what data should I use for AI? You know, what data is safe for me to use for AI?
These are tough questions or complicated questions for customers to kind of work through. So at NetApp, you know, we realize that these challenges for customers are daunting. They lead to extremely high management complexity. They make it hard for data and AI to come together, for data and workloads to come together. And all of this makes it very hard for customers to benefit from the transformation potential of data. You know, if you don't overcome these challenges, how can data transform and help you achieve your goals? So at NetApp, we started with their biggest problems. We traced back from that to see how we can help customers turn their data into a competitive advantage. And we have built the NetApp platform to do exactly that. It's a complete solution.
It's a complete stack that provides this intelligent data infrastructure that enables customers to solve those gnarly data challenges I mentioned before. We'll walk you through in this presentation how that enables us, you know, to solve those problems directly for customers better than anyone else and better than anyone before us. There are three primary pillars to our platform. The first one is any data, any workload, any place. Oops, this went too far. All right. Our platform, you know, enables data to be delivered to customers where they need it and how they need it. No excuses. You know, starting with the NetApp platform in the middle here, you know, it's the most powerful and robust storage and data platform in the industry. It is everywhere data is. It is everywhere where data is, from edge to core to cloud. It's on-premises, of course.
George mentioned how it is natively integrated into all three major public clouds. So it is in all of the places where customers have data. And then on top of that, we provide, we allow the customers to have complete flexibility in how they want to consume the product. They can buy the product upfront from us. They can consume the product as storage as a service. They can do this with us. They can do this with any of our partners. The customers get complete flexibility in how they want to consume the product. We don't force one way or the other. You know, they use what's best for them. So in all of this, what is then the customer benefit of having this any data, any place, any workload, you know, platform?
So, you know, if you look along the bottom here, you know, I try to call it out where we say, hey, by having one platform across all those different places, data, and workloads, the customers, they get a consistent automation and management of their data everywhere. They get consistent data protection and security of the data everywhere. They get consistent performance for their workloads everywhere. They got a unified control across all those different places, and they can run all of their workloads anywhere they want. This is amazing flexibility and simplicity for customers. Like, if you're using a competing solution, an organization would have to, like, find, like, n different ways of managing the data. They had to figure out how to manage the data in different ways depending upon the workload, the place, the data. That's a nightmare from a manageability point of view.
But we have made it, like, really easy for customers to do this in a consistent way everywhere. At NetApp, we have eliminated all the silos that get in your way. No silos for apps, no silos for where to place the data is, no silos for the type of data or the workloads, no silos at all. That leads to dramatic simplicity and dramatic savings for the customer. And by the way, I should tell you, this is not an easy thing to do. It takes a decade of engineering to be able to build a platform that is able to provide this level of simplicity everywhere. For example, in the cloud, we have worked the last 5 years with every hyperscaler to integrate our systems directly into the platform so that customers can have this simplicity for managing the data everywhere. That simplicity just cannot be easily replicated.
It's extremely hard to be able to do what I just described. All right. The next pillar of our platform is called adaptive data management. It's taking all the goodness of NetApp Storage. It's building content-based intelligence right inside storage, smart workflows right inside storage so that we can help the customers solve those gnarly data challenges I mentioned before. These adaptive data services, that's what makes our storage platform intelligent. You know, that's kind of what leads to the intelligent data infrastructure. So I'm going to give you a few examples. Initially, we are focused in these four categories of data services. For example, number one, security, helping protect customers' data from ransomware. Number two, protection, you know, in ensuring better business continuity in the event of data loss. Number three, seamless mobility, wire scoping, syncing, tiering, caching.
And finally, the last example is, you know, data governance to help customers meet all the regulations that they have to go that they are that are imposed on them, essentially. Now, if you are using our storage, our goal is that if you are using our storage, you shouldn't have to worry about any of these things at all. Our platform allows us to solve these larger problems for the customer. That, in turn, expands the value of our platform to customers. That, in turn, extends the upselling opportunities we have with the customers. So let's take an example. The autonomous ransomware protection is just one example of these intelligent data services. We were the first in the industry to build autonomous ransomware protection right inside storage. We are data's last line of defense.
There are 3 amazing things that I want you to remember about our ransomware solution. Okay? Number one, and you may already know this, but security attacks in the world today, they are over in minutes. They start and finish in a few minutes. If your protection is not real-time, it is sort of useless. Our autonomous ransomware solution, we are in the path of data. We protect customers real-time. You know, that is, you know, there is no delay. You know, it is real-time protection for the customers. Okay. Number two, the security landscape is changing constantly. If you had protection for yesterday, that is generally not very good for the threats that happen today.
In our case, our ML models, our machine learning models, we update them daily in customers' on-premises environment so that you always have up-to-date protection against ransomware, not yesterday's protection, but today's protection. None of our storage vendors do either that real-time protection or this always up-to-date protection. Finally, not only do we do detection, but we also do this automated recovery workflows. These are done at the workload level, at the application level, so that customers can easily recover in a matter of minutes. Once again, our competitors don't actually do this level of recovery in their ransomware solution. So others might say, "Oh, we have a ransomware solution too." But none of them actually have this level of capabilities I just described. Real-time, always up-to-date recovery built right inside storage.
And this is why we back all this up with the industry's most comprehensive ransomware recovery guarantee. We want to make sure that if customers use our storage, they can have complete peace of mind. And that's what you get when you buy, when you use NetApp Storage. You know, without this example of, you know, an intelligent data service and the other intelligent data services we have, your storage would be kind of just dumb storage. You know, look, the world has moved on. The customers, you know, have these crazy, these complex data hurdles that need their, that they need our help to overcome. And you need this fuller stack. You need this full solution that NetApp provides. You need that intelligent data infrastructure that NetApp has built.
With that, I'm going to invite my colleague, Haiyan, to walk us through the rest of the pillars of the platform.
Thank you, Harv. Good morning, everyone. My name is Haiyan. I run the CloudOps business here at NetApp, and we focus on building and delivering operational services on the NetApp platform. NetApp powers the data infrastructure for some of the most demanding workloads in the world. Operating that with the highest reliability, performance, security, and efficiency is mission-critical. I'll dive into how this operational services will help our customers intelligently, automatically operate and adapt their infrastructure to meeting the shifting needs of their modern application. We have a suite of offerings in this area. They're all anchored on the observability and instrumentation for AIOps that were built into this data infrastructure.
This service is intense differentiation for NetApp and also helps to drive the adoption and expansion of the core storage business. For example, the NetApp AIOps services, they deliver intelligent insights to the data infrastructure way beyond just the NetApp storage infrastructure. It brings full-stack visibility for workloads, which are crucial when you have to worry about how to reduce the mean time to resolution, MTTR as we say it, and maintain your SLAs. Many of the world's leading institutions, financial institutions, some of you are sitting here, leverage that technology to manage your heterogeneous environment. Over the years, we have also enriched those capabilities and went beyond just monitoring. So we have capabilities to help you with infrastructure optimization, workload automation, FinOps. And also, we've added capabilities for application services.
As the world adopts more open source, we help you manage that technology that's powering your most demanding open source applications that really require expertise that usually does not come natively. These services help customers to get the best of the intelligent data infrastructure. But in the meantime, they actually help NetApp to get in front of many new personas, users, and decision-makers. And these new users and decision-makers are crucial in architecting and running some of the new initiatives in the company: digital native, cloud native, AI projects. This truly has helped NetApp to become more and more part of the new strategic initiatives that matter into the future. Many of you know NetApp first embarked on the CloudOps business to help customers to operate in the cloud efficiently.
Today, we have expanded that capability to really help customers to manage their data infrastructure in the most intelligent way: anywhere, any place, any data, any workload. This enrichment has played a key role in customers' journey to transform, to modernize, and to migrate to their cloud. I talked about an example of AIOps, and Cloud Insights is one of the services in that area. Because of the in-depth observability capability and the reason that we manage and map all the workloads in their existing, you know, infrastructure, some of the leading organizations, one of the major hyperscalers, actually have been using that technology to conduct the majority of their storage assessment workshops and with that produce a lot of migration projects successfully.
In addition, the NetApp team and our partners are also using the similar technology to do the assessment and planning for data center modernization, and that has yielded replacement projects that have driven our flash and block footprint in the customer. You know, getting the workload to run in the cloud is an important step, but it's actually just the first step. To continue to run, to scale, to build more workloads and scale the infrastructure actually requires continuous optimization. These capabilities will allow them to grow and scale with efficiency, meeting the sustainability goals and meeting the SLAs. And doing so are the reasons behind our integrated solution and the design we put into it to holistically help customers in the entire journey to the cloud. And we want to provide the peace of mind to them as they embark on that journey.
Cloud continues to be the center of innovation, especially in the era of AI. We will continue to add new capabilities to help customers to power those workloads in the cloud and help our ISV partners to build more workloads to leverage that infrastructure. This is actually a good segue for us to talk about our ecosystem. You know, the ecosystem partnership has always been a strategic priority for NetApp, and the platform we have has always enabled us to integrate with many key technologies. César would talk about the go-to-market aspects of the broad ecosystem, and I'll focus on the technology and the solutions we built together here. Many of you know key workloads have always fueled the growth for NetApp, be that Exchange, you know, the VMware era and the SharePoints. That's pre-cloud era. We've always worked with those partners closely to optimize the solution.
On our platform, to make storage choice for them is NetApp. But today, the ecosystem has to expand dramatically. This is because it's no longer sufficient to have one workload, one data type at one location. That's not a data infrastructure. It's just one workload. Today, the growing workload, the new projects actually occupy the intersection of cloud, data, and AI. This is where we come in. As the largest independent storage vendor, we're uniquely positioned to work with many of the partners to build that new data infrastructure. They know we have the best technology, the platform that enables them, and the best practice in partner engineering. Many ISVs, as you know, are navigating this new dynamic as well. This is where they truly value what we bring to them.
As George mentioned, the transformational flexibility and the simplicity are key as they build and expand their services to their customer. Many ISVs know that successful integration in this area brings a virtuous circle for growth. As we build the infrastructure for the ISV solution stack, they have total freedom to run and manage those themselves or enable their customers to run those in the cloud or in the public cloud. AI makes this new dynamic even more powerful because customers who are running on NetApp infrastructure have freedom to serve those data to any of the AI services in the cloud. And the ISVs can add new offerings to their system and workloads wherever the workload is running. Powerful platforms don't equal complexity. Intelligent data infrastructure that's powered by NetApp is simply managed at scale through a unified control plane that George mentioned.
It's enabled because we have a powerful platform. We have the service modules on top of that. Customer experience can be personalized per the roles, by the components they manage, and by the service they choose to use. To summarize the platform section, the NetApp platform provides a strong foundation and key building blocks for intelligent data infrastructure. It provides unified control through an open and modular architecture that supports unique solution building that you can choose your storage. You can choose the data type. You can choose the data and operational services. The platform approach enables customers and partners to use these building blocks to customize and configure specific solutions for their industry as they expand their use cases for flash, for block, or as they embark on new initiatives in the cloud and in AI. Thanks for your time.
Let me just pass the mic back to Harv. He'll tell you how the unique, powerful solutions and platform enable us to grow in the four key markets.
Thank you, Haiyan. All right. Hi again, everyone. So I'm going to, in this section, tell you how we are competitively positioned and leveraging our platform to target large and growing markets. You know, as George mentioned earlier, we are focused on four growth markets: flash, block, cloud, and AI. Let's start with the first one, flash. In flash, we now have the most comprehensive portfolio in the industry. It provides optimized performance and design across every price point, from lowest cost secondary use cases to best price and performance to best performance on tier one workloads.
From file and, you know, from unified and file all the way to block, all you now need is NetApp for all your needs on the enterprise. Now, we are winning in flash for four major reasons. Number one, customers love that they are able to meet all of their flash needs with NetApp, but it has a unified platform underneath. Simplicity, no silos. That's number one. Number two, customers love that our on-premises solutions are seamlessly integrated with all three public clouds. Remember that customers have data in all public clouds. They have data everywhere. That strategy really resonates with customers. That was number two. Number three, customers love the anti-ransomware, the real-time, always up-to-date anti-ransomware solution that's built right inside our storage. That's number three.
Number four, not only do we have high performance, not only do we have the lowest cost over the data lifecycle, we also have the best-of-breed data management features. And that's critical for GenAI and enterprise AI. The sum of all this is that customers can get the most simplicity, the lowest TCO across all of their needs, across all the different price points when they use NetApp. And that's why we are winning in flash. Second market. Actually, before I go there, let's talk a little bit of an example. There are numerous stories that I could share. I'll briefly highlight one. This is an ISP that only used NetApp in the past for NAS and file workloads. When we announced the new block products, you know, earlier this year and last year, they moved over to using NetApp for both.
They love the flexibility of using NetApp for both NAS workloads as well as SAN workloads. They separately manage both, but they benefit from having the same underlying platform: low TCO, simplicity, one unified control for this customer. The next vertical is block. I am really excited about this. George mentioned earlier how we are going to be the agent of disruption for this block market. Look, we already have an awesome product that is modern, resilient, and, you know, simple. We have tons of customers who already use us. We are the leader in VMware deployments. But now, with the new block products that we released last year, we offer everything. We offer all the block capabilities that customers want. We offer it at a price that's better than any other entity in the industry. And it's an integrated approach. It's the same platform for file and block.
That integrated approach across file and block really resonates because customers, again, get simplicity and low TCO. Same block capabilities as others, but for far less. You know, same for less is unbeatable. Why would you use anyone else? Here's a good example. You know, this customer is a leading healthcare provider. They had never used NetApp before for anything. But when we released our block products in the last year, they realized, "Oh, we can use that ASA, our ASA C-Series for their high-capacity block workloads." They can use the ASA A-Series for their high-performance SAN workloads. They can meet all of their SAN needs with us. Now they're even looking to move their NAS, their file workloads to NetApp as well. Same for less. Again, same for less is unbeatable.
I feel confident we are going to soak up the share in the block market like a sponge. You know, who's going to stop us? Well, we have same for less in the block market. Number three, cloud storage. So let me tell you why we are winning in cloud storage. We are the only vendor who is natively integrated with all three major public clouds. No one else is. At the same time, we also provide full enterprise storage capabilities to all customers while being integrated into all the hyperscalers. For example, you know, we provide integrated data protection in all public clouds. We provide multi, you know, zone high availability. We provide cross-region disaster recovery in all major regions, in all major data centers, in all major public clouds, everywhere. You know, you don't have to compromise. Sorry about that. It's working now. Yeah, excellent. All right.
So where was I? It provides you the same enterprise storage capabilities in every single public cloud. Next, we now support a broad spectrum of workloads in every public cloud, from low performance, low end, to high end across every single price point. All you need in cloud storage is NetApp. Another example why we are winning is that we are natively integrated inside the hyperscalers' own consoles and their APIs. At the same time, we provide a single control plane, a single console across customers' on-premises and cloud. So you get the best of both worlds. And finally, we get very wide customer reach because it's the hyperscalers and their huge enterprise sales force that sells our products and is sold against the commitment that customers have already made to the hyperscalers. There is nothing like this. There is nothing like this in the whole storage industry.
So let's take an example. This is a, you know, a leading semiconductor company that uses NetApp for their on-premises EDA or electronic design automation workload. And thanks to our cloud storage, they have now taken their on-premises workload and they've extended it to the AWS cloud. This means that when they need that elasticity, when they need that additional resources, they can burst from the on-premises to the cloud. They're also extending this now to the Google Cloud and the Azure Cloud. It's an awesome example of a workload that's a hybrid workload. Why did they use us? They get better cost performance than they would get from any of the other hyperscaler storage offerings. They get all of the enterprise storage capabilities that they need. And they have to deal with one platform everywhere. They don't have to manage it in different ways in different places.
The simplicity of using just one storage platform everywhere. Now, a question for you. Who could they even do this with? You know of anyone else that this customer could even do this workload with? No. There is no one else with such a platform. We are the only ones with a platform that enables customers to do that. All right. The next pillar, the fourth pillar, you know, is our opportunity in AI. And before we dive into it, I thought it might be good for us to take a step back and walk you through. You probably already know that GenAI is used in many different scenarios. Okay? So one example on the left side here, you know, is when customers use it in these high-performance large language model training, you know, large language model training. We have an amazing solution for that.
But we also know that only a few large enterprises will have or will do the scale of LLM training. And then the right side of this is enterprise AI. You know, this spans from data prep and data lake modernization to fine-tuning to RAG and inferencing. You know, this allows enterprises to use the large language models. They combine them with their proprietary enterprise data. And that's what they use for RAG and inferencing. If you want to use GenAI, this is what you're going to have to, like, do. For most enterprises, this is how you use GenAI to unlock the value of your data. This is going to be the largest TAM for GenAI in the future. And NetApp is very uniquely and best positioned for this world of enterprise AI. For example, you know that GenAI largely operates on unstructured data.
Roughly half of the world's unstructured data in the enterprise is already in NetApp. Another reason we feel we are well positioned for this is that AI is inherently hybrid. Training is done in the cloud. RAG is done on-premises where your data is. NetApp is the only one that has an amazing hybrid workflow for AI. Yet another reason. We have amazing data management capabilities, data mobility, metadata management. This is what made NetApp and ONTAP famous and loved by customers. And now, with all the work we have done to integrate with MLOps toolkits, with, you know, RAG and inferencing pipelines, that combination, you know, data management plus MLOps integration, that's a deadly useful, you know, useful combination. It's required for building AI workflows. It's necessary for responsible AI.
And it is critical for driving up the productivity of a data scientist, which is the most valuable resource that most enterprises have. And finally, that ransomware solution that I mentioned earlier, you know, that becomes even more critical when more and more workloads in the enterprise become AI-driven. Those are the reasons why we feel we are very well positioned to go win enterprise AI. Look, innovation is not innovation unless everybody in the world can do it. Everybody in the world can benefit from it. And at NetApp, we are helping democratize enterprise AI. We are making it easy for every enterprise to do this by using our intelligent platform as their data foundation for AI. And here's a great example of that in action. This is a customer who is a leader in the aerospace industry. They have built an AI center of excellence within the company.
They enable the rest of the company to do a bunch, to do a variety of AI workloads, you know, from training to tuning to RAG. You know, they do this all on NetApp solutions. They consume NetApp using a storage-as-a-service model. They tell us that they love the productivity that they get for data scientists. They love the security capabilities they get. And they get no silos. You don't need a silo for AI. You should be able to use an existing storage infrastructure for AI. No silos needed. Every enterprise is going to be like them very soon. So those were the four markets that George talked about and why I feel, why we feel that we're going to win in those markets. We got amazing, you know, differentiation in those four places. But then we also know the world moves fast.
I hope you have noticed that we have delivered a blistering pace of innovation in the last year to extend our market leadership. I mean, let's just talk about just these four markets and just a few examples just in the past year. Look, in flash, we delivered, we, you know, launched a completely new line, the AFF C-Series with capacity flash. That has become the fastest-growing product in NetApp history. We just launched a completely new hardware design for AFF A-Series. I believe that completely resets the bar for flash in the industry. In block, we never had any block-specific products. In the last year, we released the ASA C-Series and A-Series. And now customers can meet all of their needs of block with us. In cloud, last year, actually this year, we released the first-party cloud storage in Google.
So now we are in every single public cloud, all three major public clouds. We also have released new tiers of performance, both low-performance tiers and low-price point tiers to high-price point tiers. The idea being you can meet all of your needs in cloud with just NetApp. But how about AI? BasePOD certification, SuperPOD certification. We just launched a new AI Pod together with NVIDIA and Lenovo. So customers have a turnkey simple solution that they can go buy. And then common across all of them, you know, in the last year, we released this AI-based autonomous ransomware solution and guarantee. We released BlueXP, the single control plane across on-premises and cloud. New, new, new, new, new. Oh, my fricking lord. You know, this is just in the past year. These are just a few examples.
You know, I don't think there's anyone in the storage industry that has even done 10% of this in the last year. Okay? We're just getting started. You know, we have a lot more planned. I mean, look at this list. We're going to keep doing ongoing updates to our hardware design, optimized for performance, cost savings, efficiency, density, and so on. Number two, we're going to introduce new performance-price tiers in all three clouds: Azure, AWS, Google. So we can go from lower-end, you know, workloads to all the way to the high end. All you should need in the cloud is NetApp. Number three, we're going to release even more robust ransomware protection. That automated recovery workflows release that for more workloads, more, you know, more applications.
Number four, release new premium add-ons for security, protection, governance, privacy so that we can help customers solve those gnarly data challenges we talked about before. Last one, you know, add more, release more integrated data services and MLOps integration so that customers can do enterprise AI more easily and benefit from enterprise AI. You know, last year was not a fluke. As a company, we are on a roll. You know, I love the speed at which we are moving. I love the full vision we have and the completeness of the solution we are building for customers. Now, our competitors are generally building, I would call, dumb storage. You know, like, even if you're building your own SSD, it's still fricking dumb storage. You know, the game has changed. You know, you need to solve those larger problems that customers are facing with data today like we do.
In our world, if you move the fastest, if you move, if you deliver the most value, and if you can do this with the most complete stack and solution, you win. And I'm very confident that we at NetApp, we're going to deliver a pace of innovation our customers simply can't, our competitors simply can't, customers will get it, our competitors simply can't. And the market has not seen because of the underlying platform we have. You know, you know, it is decades of engineering. It is the most robust and powerful platform in the industry. And also because of the culture we have built in the team. We are customer-obsessed, and we move at extreme speeds. Okay, so in closing, here are the four key takeaways. We deliver unmatched customer value, any data, any workload, any place. No excuses, no compromises.
When we do that, customers really appreciate. They love the simplicity, the cost savings, the world-class protection, you know, they get from us, and also the flexibility in consumption. They can use it the way that works best for them. Number three, we are competitively positioned to win in Flash, in cloud, and AI. And as George said, we are going to completely go disrupt the block market. And then finally, number four, our unique platform and culture, it enables us to out-innovate everyone else in the industry. Hey, on behalf of Haiyan and me, you know, I hope this was useful. You know, thank you for taking the time.
Oh, I'm curious as well.
Thank you, Harv. I appreciate it. We are going to take a quick break. We'll reconvene sharply at 9:55 A.M. So go grab a cup of coffee, stretch your legs, corner one of our executives, and we'll be back at 9:55 A.M.
I can feel you're losing hope. You've been hanging on a rope. Something tells me you're alone. Baby, just keep holding on. I can feel you're losing hope. You've been hanging on a rope. Something tells me you're alone. Baby, just keep holding on. Just save me, my love. Save me, my love to keep you from falling down. Just save me, my love. Just save me, my love to keep you from falling to the ground. Call me up, I'll take you home. I've been waiting by the phone. Don't you know that you're not alone? Don't you know you're not on your own? Just save me, my love. Save me, my love to keep you from falling down. Just save me, my love. Just save me, my love to keep you. 'Cause I don't care how far. I'll run to wherever you are.
'Cause I don't care how far. I'll run to wherever you are. Just save me my love. Save me my love to keep you from falling down. Just save me my love. Just save me my love to keep you. Just save me my love. Just save me my love to keep you. Just save me my love. Just save me my love. Just save me my love. Just save me my love. Dizzy, we're calling it a spin. If I was older, it would you still let me win? Cool, so cold, and I won't let you say again that it's life. Don't go, so cold, you're all the same. Seriously, it must have bothered me. Under the moon, I might desire to be. Never taking myself that seriously. Don't think it bothers me. Chasing debts on a set. I heard you love me. Disabled in consequences.
Cool, so cold, and I won't let you say again that it's life. Don't go, so cold, you're all the same. Seriously, it must have bothers me. Under the moon, I might desire to be. Never taking myself that seriously. Don't think it bothers me. Seriously, it must have bothers me. Under the moon, I might desire to be. Never taking myself that seriously. Don't think it bothers me. I think I need a change. Too much of the same things. I'm going insane. Too much of the same things. Rolling all my life. You know I'm a rock star. Fuck a 9-to-5. It's a game, and I'm the nice guy. I think I need a change. Too much of the same things. I'm going insane. Too much of the same things. Tired of the lines. I wonder where the something really's, yeah. Look into my eyes.
I need my feelings to be feeling, yeah. I think I need a change. Too much of the same things. I'm going insane. Too much of the same things. Tired of the lines. I wonder where the something really's, yeah. Look into my eyes. I need my feelings to be feeling, yeah. Can't let you go. When we're moving with the music, I-I-I can't let you go. You're the healing. Keep me breathing. I-I-I can't let you go. Push, pull, and hate. I'll be on my way. Hot summer night. Don't let it be. Lost track of time. Like when we were kids. You got me high. I don't want it to end. My hands in the sky. Feeling the desert. I think that I'm gonna stay here forever. Can't let you go. When we're moving with the music, I-I-I can't let you go. You're the healing. Keep me breathing.
I can't let you go. Can't let you go. You're the crimes I can never get over.
All right, we're back. It's 9:55 A.M. Everyone get settled. All right, give you guys a second. All right, well, welcome back, everyone, in the room and on the webcast. We'll be getting to our last couple of presentations before we start the Q&A. So with that, it's my honor to introduce César Cernuda, our President.
Thank you so much, Kris. Good morning to everyone. Thank you so much for being here with us today. It's my great pleasure to be presenting to all of you today. And my name is César Cernuda, and I run go-to-market here at NetApp. You've seen this morning George talking about the vision and the opportunity that we have ahead of us. Harv and Haiyan talk about our product portfolio, our roadmap, and what makes us unique. My presentation is going to be about how we can take advantage of all of that and make it happen. So in other words, how are we going to execute? And I'm going to do it in two steps. This works. First, I'm going to talk a little bit about the momentum that we have. Where are we? How can we build on top of this momentum?
And the second piece is going to be about sustainable growth. How are we going to execute on our strategy and make things happen for the future? Winning share and making sure that we have more customers with us at NetApp. So let me start with a slide that George already has shared, which is the good momentum that we are in. And I want to be very open. We made changes at the beginning of fiscal 2024. We decided to be way more focused. We said, let's do two things in a big way. Let's go on winning flash and let's win with first-party cloud and marketplace. In other words, cloud storage.
What we did at the beginning of the fiscal year was let's align our field sellers more and more into the flash opportunity that we have in front of us and let's specialize even more our cloud specialists to win on cloud storage. I'm happy to report back that we feel extremely proud of the progress that we have made focusing there. Our customers have appreciated that and our partners as well. At the end of the day, we go back to something that you have heard since the beginning of this morning, which is the power of choice. Letting customers decide how they want to consume, how they use the technology, and serving them in their life cycle, in the journey they have, either in the cloud, on-prem, or hybrid. One of the numbers that you see here is Keystone.
You know, the fact that we have managed to grow that to the size that it is without forcing our sellers or our partners to move customers into a storage as a service, but actually letting them go and decide how they want to use the technology is a great differentiator from our side as well. So all in all, we feel very proud with the changes that we have made. We're building this on top of more than 30 years of history where customers trust NetApp, not just our technology, but also the way we serve them. Our sellers, our partners, our customer success team. I always say that we are truly and really an enterprise-ready company. What does that mean? We serve customers 24x7, 365. For sure, many other companies can say that, but we run mission-critical applications, and we are a truly global company.
This is important. Not being just a U.S. company, but having presence in the street, in EMEA, in APAC, because many of those global customers require that level of attention and care and understanding of all those markets. We serve in all different types of industries, but I want to call out that the last year we have seen a lot of great momentum, specifically on the public sector side. Some of the, you know, product pieces and features that Harv have shared have really served us to help on the public sector side, financial services, think about all the protection, the ransomware approach that we have to our technology. But we have seen as well high-tech, media and entertainment as industries that are growing with us, and we're growing our presence with them.
the last year, we also have seen some progress on what we call the commercial segment, which is kind of the small and medium enterprise. That's been driven by two things. Our capacity flash portfolio has helped us, but also our cloud approach has helped us to win more position in this segment. I want to call out Keystone as well. Keystone has been a great asset for us the last year to win new logos and win more share inside these customers. We do this with partners. Haiyan has shared a slide with a partner ecosystem more from a product perspective. I'm going to talk a little bit more about the go-to-market perspective. We have thousands of partners worldwide. NetApp can probably say that we have been and we are a partner-driven company.
But as you said before, that we want to have presence in those markets because we believe that the power is to go with a partner, go with our specialists, and serve customers all together. Let me talk a little bit about the hyperscalers. You have heard us saying many times, we're uniquely positioned. Well, it's true. Many companies talk about the partnership with hyperscalers. And we also have that type of partnership. So let me kind of separate two things. We are an ISV for these hyperscalers. There's other companies that are ISVs for this, you know, for these hyperscalers. We are part of their marketplace. We've been recognized several times as being a top ISV in their marketplace. Now, what really makes us unique is not that marketplace approach, which we feel very proud of that work as well.
What really makes us unique is that we are part of their console. We are a first-party vendor. I was receiving, you know, a small summary because today there's a big launch that Microsoft is doing actually in Spain. They have their artificial intelligence day where they're launching the new data center. And NetApp is part of that. There's 4,000 customers attending that event. They just sent me some pictures because, you know, we have all these people interested in Azure NetApp Files, which is our technology with Microsoft. So in other words, what I really want to make sure that we take away from a go-to-market perspective is a scale. This partnership, this what we're doing with hyperscalers is helping us to have thousands of customers through their sales force with our cloud specialists, which is very important.
Many of their cloud service providers are starting to work with us as well through the hyperscaler. Second thing are alliances. Some of you, you know, have followed some of them, some of these alliances because they have been for many years. For example, Cisco, we have FlexPod. We're embedding our technology, and we're going to market together with our joint partners or their partners or our partners with joint offerings. Lenovo is OEM in some of our storage. They come with their compute, with our storage, et cetera, et cetera. Here we put global partners. We have all types of partners, local partners, regional partners, and global partners. But I want to call out some of these partners that have been working with us for years, building and offering solutions which are managed service for customers.
In other words, when we talk about Keystone as storage as a service, many of them have been doing this with us for many years. And on top of that, they build their solutions or they bring in an ISV. And this is a huge market for us. You know, we've been working with them globally, and we have actually as well regional and, you know, local service providers that we team up with. And the last piece that I want to talk today about is AI. Everyone is talking about AI. I'll cover it later, but George said it very well, which is, you know, AI is a huge opportunity for NetApp. Most of the GenAI runs on unstructured data, and unstructured data runs, you know, on NetApp technology. So all in all, we have great excitement right now. Our partnership with NVIDIA is getting stronger and stronger.
I can share that in our last sales kickoff, I had the EVP joining me on stage with a joint presentation talking about the partnership, the NVIDIA commitment to NetApp, the NetApp commitment to NVIDIA as well. As a matter of fact, two months ago, you know, WWT, NVIDIA, and NetApp announced, you know, in their labs, we're having this advanced AI workloads where customers can come and work with us. One of the things that we've been working with NVIDIA is to make sure that their top partners are part of our top partners that are having joint solutions and AI offerings. You're going to see more in the next months to come. We have a big partner program called Partner Sphere. All in all, that's one of our biggest differentiators as well. The trust that our partners have on us is huge.
They really appreciate the investments that we have made and the progress that we have done, not just on the cloud side, but with block, you know, as Harv have shared, you know, anti-ransomware, you know, and AI is a big differentiator for them. So we're looking forward to grow the business with them in the next years to come. And we're recognized by the industry. It's not that I'm telling you all this. So we've been recognized by the industry, you know, the industry analysts for the last years in different aspects. And one of the ones that I like, which is not industry analysts, but is the fact that we have been recognized by one of the hyperscalers as, you know, and I want to read it well, the technology partner of the year in infrastructure storage.
This is not the first time that we won one of these awards from the hyperscalers. All in all, my goal is to say, given this momentum I already shared, how can we go and execute in a way that we can even accelerate more, you know, in the market? This is part two of my presentation. The first thing is we are the intelligent data infrastructure company. And in other words, this means we're helping our customers to really build that intelligent infrastructure where they can work with their data. I think George and Haiyan have done a great presentation, so we all understand how that platform is working for them. What makes us unique and why more and more customers want to use this? It's a great seamless experience.
It's not just AI ready, but actually the best way to enter GenAI and work with AI. You know, and this is a huge differentiator from our side. But I'm going to be talking about products. I'm going to be talking about how do we go to market and attend customers. So first of all, we have different roles like other companies. We have sales, we have marketing. One of my goals has been to align sales, marketing, and I'm going to talk later about customer success. And I align them or we work with them all together through different segments. So we think about the market in these four big buckets. Global enterprise. Think about the top 100 global enterprise companies that when we sit down with them, we don't talk about the next 12 months. We do account plans for the next three years. What are your plans?
What are your top of mind? We bring GSIs, and we sit together and we bring the customer success team. We bring our specialists. You see on the right side of the slide, you know, some of the different roles that we have from technical, in-depth technical people, specialists. I'll cover that, you know, a bit more. Then we talk about large enterprise. Think about companies. By the way, these global enterprises are the ones that require us to be present in those markets to support them, to work with them, to understand their local needs as well. Then we have what we call large enterprise. Think about the top 3,000, 4,000 companies across the globe. I'll come back to public sector. Basically these, you know, the customers, we do an annual account plan with their partner as well.
We bring our customer success team, we bring the specialist group, and we talk about the projects that they have for that year where we can cover and support them. Given what you just saw before, we have a huge opportunity to expand our footprint in each of those account plans, but I'm going to talk about that later. Public sector. This has been a huge opportunity for NetApp. It's very interesting because in those markets where we are market leaders, most of the times we have a very strong penetration in the public sector side. This year in fiscal 2024, we have decided we have always had a U.S. public sector team outside of the other groups. In our top markets worldwide, we have created this public sector unit.
So in Germany, we have a public sector leader now, in the U.K., Japan, France, you know, you name it. Because we want to track that business separate from the rest. We have a great value proposition, and we believe this is a big growth opportunity for us. And then what we call internally the commercial segment, which is the small and medium enterprise, this is very much driven by partners and through partners. And we're seeing great momentum with the new offerings that we have. So all in all, those are the segments that we have. And we support the segments with specialists. This year, we have created an AI specialist team. So on top of training our 3,000 field people, we have also created an AI specialist team engaging in special deals that I'll cover later. In-depth technical resources supporting those engagements.
Cloud specialists, as I shared before, working with the hyperscalers and their go-to-markets. Something very important. I shared this probably some of you two years ago. We had a plan to centralize renewals with a team. Well, we have executed on that plan. Basically, we have a team dedicated to renewals. Why this is important? Because we have freed up our seller's time. We have freed up time for our sellers to go and hunt for new projects, for new workloads. At the same time, we've been providing better customer satisfaction, better customer renewal experience with specialists doing renewals, and we're way more productive. So how are we going to grow the business? There's four growth initiatives. We could talk here about many, but one of our learnings is fear is better. Going deeper is better as well.
So we're going to talk about modernizing data centers. We're going to talk about expansion of our existing customer install base, which is big. The third opportunity is winning new customers. And the fourth is how we're going to address new markets. Pretty much AI. I'll talk about it as well. Modernizing data centers. There's a huge opportunity out there. There's a big install base of disk and hybrid flash. Hopefully, Harv have shared all the great things that we have on our product portfolio to help our customers modernizing the data centers. I can tell you that we can talk to customers on how they can reduce 30% of their costs by moving to our high-performance flash, for example, modern, you know, flash portfolio. We talked to customers how they can reduce 42% their energy consumption, which in parts of the world are very important, right?
We have a great value proposition about that, and they can reduce and help on their ESG ratings, you know, so our customers demand that as well from us. Certainly, our ransomware protection and other pieces like our hybrid cloud approach is helping us on these scenarios around modernizing data centers. To give you an example, you know, one of the large governments in the world, we're thinking on centralizing and actually, you know, consolidating older data centers. Some of them were with NetApp, some of them were with other vendors, and they decided to go with us, not just for cost savings, but actually the flexibility, you know, the unified storage experience that George has talked before about. So this has been a great win for us last year. When we think about expansion, block is a huge opportunity. Harv has shared that with us.
Many of our customers have been using NetApp on their file NAS environments, and they love ONTAP. So we have many customers that have decided to test and start using, you know, ASA with us. For me, one of the mandatory things that our sellers need to do is to go to every single one of our existing customers. I'll talk later about new customers and talk to them about their block environments and ask them to try us and do a POC with them. And our partners have been trained to expand our share of wallet with those customers with block. We also believe there's a huge opportunity, and we're executing against that opportunity with VMware. We have thousands of joint customers on VMware, and each one of them has different plans. Some of them want to keep VMware, but they're looking to reduce cost.
We have seen, you know, projects and plans where we can reduce 30% of their cost with our technologies in VMware. Some of them want to move to the cloud with VMware, and some others are thinking of replacing VMware. So all in all, great opportunity to expand our share of wallet inside those customers. You know, we have a customer who has been 15 years a NetApp customer, has been using file storage. Last year, they decided to try block with us. We had a huge and great, you know, implementation. So they decided as well to start using public cloud, you know, with NetApp. So all of those are great opportunities to increase our share of wallets and grow our business with a huge installed base that we have. Winning new customers.
Certainly, you know, our portfolio, our new flash portfolio can help us to win new customers, you know. But when I talk about winning new customers to our team, I said, "Hey, let's go big in two things. Let's go big with cloud and let's go big with ransomware." Customers are very concerned, of course, on the overall cybersecurity situation, and we have a unique value proposition. So let's make sure that we talk to new prospects and customers about their data and what they're doing and how do we differentiate ourselves. And the second piece is let's leverage what I shared at the beginning, which is the scale of the hyperscalers to reach more and more customers through them. Because that's a unique opportunity to real scale. Let me give you a real example.
One of the largest bottlers companies on the planet had one of the biggest SAP DB2 implementations. So Microsoft worked with them and basically pulled NetApp in, and we migrated in five months one of the largest DB2 SAP customers to Azure NetApp Files. The outcome? 30% operating cost savings for the customer, great customer satisfaction, more flexibility in their environment. One customer added for NetApp. We're talking now to that customer, you know, to see if some of their data centers they can use our technology. Now they have tried ONTAP. So when we talk about new markets, so when I talk about winning new customers, as I said before, there's plenty of opportunities.
Capacity Flash has been a great opportunity as well for us, but I'm trying to go and talk about, hey, this piece is in a way that we can understand how we're focusing our field and our sales force. So let me kind of go here and talk about new markets. And when I talk about new markets, basically we're putting all this effort on AI. Our marketing team is helping our teams, our selling team. We have enabled and we have trained the 3,000 people that attend our sales kickoff around AI. What makes us unique? What is the value proposition? Today, our sales team is knocking the doors of our customers and prospects, trying to understand what are you doing with GenAI and AI in the next 12 months, 18 months. And this is what we can do to help you.
Our partners have been trained and enabled as well to have that type of conversation with our customers. We have built this AI specialist team to make sure that we support our field people and our partners in many of these engagements. So all in all, we think the opportunity we have with AI is huge. Let me talk about this biotechnology company. They were building a massive, quite big, you know, DGX SuperPOD environment. NVIDIA, NetApp, and a partner of ours, TeamUp, who invited us to get them to work with us. All these engagements that I'm talking about, some of them start with our sellers, some of them start with our partners, but some of them start with our customer success, which I'm going to talk now about. That's the way we connect the dots.
We team up and we bring that value proposition to our customers. So before I close, let me touch base on a very important slide for me, which is the progress that we have made the last years, moving from support and services organization into an AI customer success organization. And this is very important because one of our goals has been we start talking about how do we digitalize the overall customer success experience. And now we're saying it's not just digitalizing, we need to have more and more AI enabling that relationship. I'm going to stop starting in the middle. I didn't talk about it before. AIOps. Basically through AI, you know, AIQ, our Active IQ, we're able to track, monitor, and have the telemetry of the data, you know, our customers have that telemetry of their data. And we support them through that.
The information that we get helps us to go in the renewal process to address better their customer needs. It helps us well to understand are they reaching max capacity or not and anticipate and talk to them about a potential add-on opportunity or a refresh. What are they going to be doing? So that AIOps is connected to the first line, which is the renewals team. So the seller is not even involved in that process. We have technology supporting our renewals team that is also taking care of some of the add-ons engagements. They involve the partner, they involve the customer, and our seller. Support. 98%, 98% of all the support cases are starting a digital way. Customer satisfaction has been going up. We're able to do preventive, you know, do it before we have the issue. We're reaching out to them.
We're seeing this in your environment. So all that is super important, not just for the customer satisfaction, but also we're making it more productive for us. You know, how do we operate? And that's freeing up time again for our sellers to hunt, which has been our big strategy. How do we free up time to grow our business and have our sellers focusing more and more selling products? I'm going to skip Keystone. We have talked today about it, but I'm going to close with the customer success piece on the cloud side because it's very important. And I got this question a couple of times. One of the great things that we have is having the hyperscalers selling to customers. And I said it before, sometimes our sellers are not even involved in that sales, which is fine.
The hyperscaler is providing the support to the customer, which is also fine. We provide the support to the hyperscaler. But we have built a customer success team that works with the hyperscaler directly with the customer. So when a customer acquires, you know, one of these products or services, we help on the enablement side and we support the customer on that cloud journey. We build intimacy with the customer, and that helps us as well to do some cross-sell and up-sell. So that helps us to build a relation with those cloud customers as well through the customer success team. So let me kind of wrap up and share with you, you know, I will say the five pieces that are more important for the go-to-market. The first one is we have a strong customer relationship. We are a global company, enterprise-ready, global presence across industry.
Our partner ecosystem is unique, starting with the hyperscalers, the alliances, the AI partners that we have, the service providers that we have, our partnership program. We have a unique value proposition. Harv and Haiyan have shared that. But when you think about the different scenarios, modernizing data centers, expanding our share of wallet, you know, being able to win new customers and new markets, we feel very strong with the value proposition that we have. And our customer success is part of the first piece as well. It's helping us, you know, to improve our overall productivity, improve customer satisfaction, serve customers better, learn more about our customers and grow with them. And the last piece, and George opened with that, you know, we are the employer of choice. People want to come and work with us at NetApp.
So with that, I want to say thank you, and I'm going to pass over now to our CFO, Mike Berry. Mike, are you here? Thank you. Show us the numbers.
Yes. Thank you, César. Thank you. Oops. All right, good morning, everyone. Great to see everybody again. And for those of you on the webcast, great to see you as well. So, hey, there's a lot of work goes into these for a lot of good reasons. This is, I've had the honor now to do three Investor Days at NetApp. The first one, we were all in a conference room in COVID with masks on. It was terrible. We had a wonderful time in March 2022 when we came to New York, and now we're here. Hey folks, this is a very different company. As Harv said it, last year was not a fluke.
The difference that we're going to bring you today is we talked a lot about cloud in March of 2022. It's still a great growth vector, but what you're going to hear today is a much more balanced story on both storage and cloud, and I hope you got that from the earlier conversations. My job is to bring it all together. We started with George, talked about all the great things in the market, the growing TAM, all the different sector growth drivers that we have. Harv and Haiyan did a great job saying, "Hey, look at all the product stuff." What was that? Blistering pace of innovation. I think five news. I love that. Five news. And then César talked about all the changes in go-to-market that we've made to really focus the team on what the customers want. Now I'm going to bring it together in numbers.
I'm going to give you a little bit of view of 2024 just to add, I'm going to say credibility to what I'm going to show you at the end. Yes, at the end, I'll talk to you about a long-term model. We're going to talk a lot about the NetApp business model, and then we're going to talk a little bit about our capital structure, and I'll go through that. Again, all of this, folks, leads to the long-term model that we're going to talk about at the end. So everything we've done today, everything I'll talk about leads to that.
If there are four things that I would like you to remember after this, and we might do my favorite country singer, Luke Bryan, I may ask you to sing along with me at the end because I'm going to say it every time, is, "Hey, there are real industry and company secular growth drivers that we have." All the teams walk through those, and we're going to hit that a bunch of times, not only in product, but then the discipline around go-to-market. That's number one. Number two is our revenue mix continues to change towards more high-margin products. That's number two. Number three, and this is, if I say this once, I'm going to say it 50 x. The NetApp business model has operating leverage. You're going to see it in the historicals, and our goal is to keep driving that going forward. That's number three.
Number four is we will continue to be good stewards of your capital. We have a very shareholder-friendly capital return and allocation, and you're going to see that continue going forward. So those are the four things. You're going to hear that throughout my presentation, and then we're going to summarize that at the end. Okay. So let's take a little bit of a step back into 2024. And this is the historicals. It's always fun to say, "Hey, what timeframe are you going to use when you do these?" Hey, I picked the last 8 years. Could have picked six, could have picked 10. The growth rates are largely similar. The great part about this slide that jumps out at me is the model has leverage. Even though revenue has stayed relatively consistent, 2% CAGR over this 8-year period, we've been able to grow earnings per share by 14%.
That's not only more earnings, but it's also now driving down our share count. So again, strong EPS growth, even in a relatively flat macro environment. And yes, we've had one bump. We've talked about that in 2020. We're not going to go back to that. Going forward, we do expect to be able to drive sustainable growth going forward. So George had this in his slide as well, and there's a lot of great numbers on here. I'm just going to hit on a couple of things. The one that I love the most is the operating cash flow. You all know that I love cash. I have my green shirt on today. It's all about cash. It was great to see that finally bump up again over $1.5 billion. We've been sitting at that $1.3 billion for a long time.
Also, multiple records across virtually all of our profitability measures, most of them for the last three years, and then we've certainly guided for continued growth and record earnings in fiscal 2025. And then again, $1.3 billion capital return to shareholders, and I will talk about that at the end of the presentation. Okay. Before we go on, hey, I'm going to show you a new long-term plan at the end. You're going to say, "Yeah, but that's great. Mike, what'd you tell us in March of 2020, and how'd you do?" So this is a comparison against that. The left side is what we showed you at that point in terms of what our three-year model was. The middle is the midpoint of what we've guided for fiscal 2025. Certainly, the economy changed pretty materially around when we did March 2022.
There were a lot of geopolitical conflicts that happened, and then, of course, the resulting economic slowdown. But during that timeframe, even though the dollars were not what we wanted on the top line, we well exceeded all of our margin numbers. And that was really our goal. We've talked about on every earnings call, control what you can control, and that's what we've done. So again, that was our model. We didn't hit the top line, but a lot of what I'm going to talk about today is, and this is what Harv talked about. Folks, it's a very different company in terms of the growth vectors that we have, both from a company perspective and secular trends. And I'll bring that all together when we talk about the long-term model at the end. Okay. This is my Luke Bryan thing. Strong business model.
Let's jump right into the numbers. Okay. Before we do that, though, you heard all about the platform and the benefits that customers get from our platform, and it's material, and Harv and Haiyan did a great job walking you through that. That's on the left side. But there's great advantages to me as a CFO and us as a business model as well. It brings three clear advantages to NetApp. Number one is greater revenue opportunity because of the ability to cut across all the different ways they want to consume. Number two is it helps us lower our cost of goods sold. Plus, hey folks, we learned from the supply chain issues we all had before. Well, you can't completely eliminate any risk. We've done a great job in terms of our supplier concentration to make sure that we're better prepared if that happens again.
And then near and dear to my heart is all the great R&D leverage that we get. It's one platform across all products that enables us to share that R&D. And you can test it in the cloud. You can deploy it on-prem. It's a great, from us, financial advantage to be able to do that. So real financial advantages to the platform. Okay. Let's jump into some of the numbers. So I'm going to start with revenue, and we'll work our way down. Again, all of this, folks, is to show you when we get to the long-term model how these influence what we think is the next three years. And this is the revenue growth. What you see on the bottom two is that's our hybrid cloud segment, and then you have public cloud on top.
Revenues staying relatively consistent, but the mix has changed relatively dramatically, and I'm going to walk you through that in a separate slide. Support and services revenue, the large portion of that is support, but services continues to grow. That's where Keystone shows up. So you're going to start to see that grow as well. And that is the highest margin business that we do right there in that middle chart. Public cloud, even though it didn't hit the $2 billion we like to talk about, we've really started to see, and you saw in Q4, hey, we're working through some of that subscription headwind, and you saw the first party and marketplace growth start to drive year-over-year growth. Importantly, talked about being good stewards of capital. Well, we continue to look at acquisitions. Those are mostly focused around tech and talent, not adjacencies.
So most of the growth you should expect to see there will be organic. Okay. Services revenue. This is all services revenue, and some of our other on-prem friends talk about all of their recurring revenue. Ours is a mix of recurring, start and end date, and then also consumption. Even though it may not have a start and end date, it is very sticky revenue. So again, in recurring, this includes mostly support, also our cloud subscriptions and Keystone, which is a subscription. So that continues to grow nicely. And then the consumption part of it, the majority of that is our cloud consumption. That is now 80% of the cloud revenue, and it's very sticky. While customers will bump up and down, burst up and down, Harv talked about that, we rarely lose a customer in first party and marketplace. So that's a very sticky piece of business.
Bless you. I believe this is about 53% of our revenue guide in fiscal 2025. I don't want it to be 80% because I want product to grow as well, but it continues to be a big piece of our revenue stream and very high margins as well. Okay. This is one of my funnest slides. I love this slide. Hey, at the top is product revenue by quarter. We know this. It has linearity, and it moves around. The best part about support is it stayed very consistent throughout that. While product revenue during this timeframe, the last eight years again, has declined on a compound annual growth rate of about 1%, support has actually grown by 2%.
What you see is in the last four years, if you take out the extra week we got in fiscal 2021, support revenue has grown by six, five, three and three . Very consistent. You typically see mid-single-digit growth when product revenue is growing, and then if it's lower growth, you'll see a little bit lower growth in support, but still growing. The key there is the mix helps, but also renewals is becoming a bigger piece of what we do every day in support, and it's a huge part of our business. That helps mitigate any volatility from product revenue growth, and that base will just continue to grow. So again, very stable, and again, very high margin. Almost half of the hybrid cloud revenue is now support, and it's something that obviously we work a lot on.
Harv's team and Haiyan's team to deliver that value to our customers. All right. I'm going to give you a second to look at this chart. There are two new pieces of information on here. Can't do an Investor Day without giving you something new. So the two middle ones are new disclosures. Excuse me. I had to get a drink of water. The top we've talked a lot about. That number for all flash as a percent of hybrid cloud revenue finished the year at 60%. These are annual numbers, so 58% for the full year. And then the bottom number we've disclosed before, keep in mind cloud storage, higher retention rates, CloudOps a little bit lower with the subscription business. So the cloud storage continues to do better.
We expect that to stabilize and then grow, and we expect all flash as a percent of hybrid cloud to continue to grow. Support dollar-based net dollar-based retention is very similar to the support revenue growth because this is our existing customers only. We took net out, net new, and this measures how well we renew our existing customers. Super important number. It will typically move along with support revenue, but this is really what we look at, how good of a job are we at keeping our customers happy. And then the third one, a lot of questions about you talked about first party and marketplace. How big is it? It ended the year, folks, about almost half of the total cloud revenue number. Keep in mind we've told you cloud storage is about two-thirds. CloudOps is about 1/3.
This is 48% of the total number, and we expect that to continue to grow. And then that's going to feed right into the margin story we'll talk about. New stuff. Hey, folks, we're not going to give you this every quarter. We're going to give it to you every year, but it's important that we wanted to set the baseline as we go into fiscal 2025. Okay. This is a big one, and we get a lot of questions about margin for all the right reasons. If you look at our total company margin now, we guided 71%-72%. If you go back four years, and you can largely get to these numbers because we give this enough to you. Before, the top part, which is our hard drive business, this is flash product revenue. This is the rest of product revenue.
This is all support, and this is all public cloud. Okay. As the mix shift continues to move to more all flash, that has higher margins. As more goes to support, that has higher margins. As we grow public cloud, that has higher margins. As the HDD and hybrid business shrink, that's the lowest margin business that we do. So this is a big driver for us being able to sustain greater than 70% total gross margins, and it helps mitigate the fun that we all have with commodity costs that are going to bump around every year. So again, remember I said, "Hey, we got four things." This is number two. The revenue mix continues to change, and we'll help drive higher margins. Okay. Gross profit. Leading to that, we've talked a ton about hybrid cloud, so I want to spend a bunch of time on that.
We know where we are in fiscal 2025. We talked a good bit about this in the Q4 call. Support, bless its heart, continues to be at greater than 90%, which is great. Let's talk about public cloud. This is the year we're going to see the inflection in margins really for three reasons. Yes, we expect revenue growth, but the majority is going to be us actually driving down our costs in that business. We've invested a lot in the last several years, and now we're going to get the benefit. Number one is finally that asset depreciation for the hardware that we've deployed, mostly in Azure, is going to start to decline. We get the benefit of using those assets for greater than the three-year depreciation life. That's number one.
Number two is the team's done a great job, especially in the products that haven't grown as much as we would like, driving down our cloud costs. And that's going to take dollars out, and you saw most of the benefit you saw in Q3 to Q4 was that. And then smaller but important is we did end of life some products in fiscal 2024, and we took costs out there. So you should expect to see that cost continue to go down, which is why we feel really good about the 75%-80% margin, and we fully expect to make good progress towards that goal as we go through fiscal 2025. Okay. Net income and cash flow. You don't see cash flow on this slide because we didn't guide for it, so I'm not going to put it on the slide. But here's the important part.
Net income has continued to grow. As you look at our cash flow generation, it will typically move in line with net income. During the year, you see some working capital bumps, and there's four areas for you to think about as you model our cash flow. Number one, linearity matters. It matters a lot. If you see AR bump or you see DSOs move, that typically means we had different linearity. Number two is incentive compensation does move around, and I talked about this in fiscal 2025. You're going to see a big cash outflow, thank goodness, in Q1 for the payments of incentive comp. Number three is supply chain payments, and it depends there on how much we want to do in pre-buys to make sure and mitigate any gross margin issues going forward.
Then number four is we all have to pay them our tax payments, and those move around as well. You're going to typically see those move during the year, but it's going to even out over time. Okay. Working capital here doesn't have a huge impact on cash flow. If you look at the annual numbers, it's relatively consistent with net income, which is great. Okay. Before I move on to the capital structure, hey, earnings per share. It's what we get paid to do, right? Drive EPS growth. 14% compound annual growth rate for this period as well from 2021 to the midpoint of our guide. Super important that we continue to drive earnings, but then we're also going to continue to drive down share count. I didn't hit it at the beginning, so let me back up.
I use fiscal 2025 guide, not 2024 actuals, because I don't want people to think we're going to stand up and take credit for profitability when we had low component costs. Hey, 2025, costs are actually higher. So what we wanted to show you was the best apples to apples in terms of our growth rate from a profitability perspective. And again, 14%, this measures our guide in fiscal 2025. Okay. We will continue to be aggressive on our buybacks, and again, we've talked about 100% of free cash flow going to dividends and buybacks. Okay. We will be good stewards of your capital. We realize the cost of equity is very high. Okay. Let's talk a couple of minutes about capital structure and maintaining our financial strength.
Before we get to there, you should expect us to, again, spend about 100% of our free cash flow in dividends and in share buybacks. Dividends this year, because of our confidence in our cash flow growth, we've actually bumped the dividend from $0.50 a share to $0.52 a share per quarter, and we do expect to see steady increases in those dividends. And then we will continue to make sure not only do we offset dilution, but drive the share count down by at least 1%-2% a year. Okay. If I was better at forecasting cash flow, I would have spent more than the $1.3 billion this year. We only got 86%, but that's fine. We'll make sure next year, again, we're targeting 100%.
As I said at the beginning, acquisitions are focused on smaller ones, which is why we feel comfortable allocating 100% of free cash flow, and that goes to the next slide. Okay. We ended the year with about $3.3 billion of cash. Awesome. About $900 million of net cash. Cash less our outstanding debt. We're very comfortable running the business at about $1 billion of total cash supplemented by the revolver. Okay. So when those two debt maturities come up, one's in June, $400 million. I'm sorry, September, $400 million, and then next June, which is in our fiscal 2026, $750 million. The current expectation is we will pay those down with available cash. Okay. What that leaves us is a lot of flexibility to when we want to hit the debt markets. Folks, we will always have debt in our capital structure.
The cost of debt is so much lower than the cost of equity, but we also want to make sure that we maintain our investment-grade rating. So we'll always have debt in the structure, but it will be a prudent level. Okay. And this gives us the optionality around when we go back and hit the debt markets. Even though yields have come down, I'm sorry, spreads have come down, the rates are still a little bit too high. At some point, they will lower rates, hopefully at least a little bit, and then we'll be opportunistic about when we hit the debt markets. Okay. Hey, Mike, that was all fun. Take us to the end, please. Okay. This is a big slide. If I would have stood up in March 2022 and said, "What are the revenue levers?" there would have really been just three: flash, cloud, and support.
Thank you to Harv and Haiyan and all the 12,000 employees at NetApp. Holy cow, folks, what a list, right? Five new are all on this. So let's start at the top. All flash is everything that we've talked about: the ongoing 10,000 replacement cycle, product refreshes, block, and AI. Some of these overlap, but it all goes to flash. Block, we've talked about. We will be a disruptor in that market, and any growth there is additive. We've talked about cloud, the focus on first party and marketplace, and all the great, and Harv said it. We have all three. Hey, folks, nobody even has a native offering in one. Never mind two, never mind three. So we feel great about that cloud market. George talks about AI every earnings call. We talked about it today.
We feel very favorable because of all the unstructured data we have, and we have a great customer base to cross-sell into. I also talked about it. Hey, hard drives are going to continue to decline. We will cannibalize that business on purpose. Okay. There's still some of that that will stay around nearline drives, but we do expect that to continue to decline. Keystone, we talked a little bit about two years ago, but not much. Hey, almost $150 million of total contract value will continue to grow, and we feel great about that. Remember, that shows up in services. That's where that is, and you'll start to see that growth. And then, hey, folks, the thing that pays for everything we do is support, and the stability of that revenue line is super important. So again, high margin, stable.
These are all the growth drivers that we have, very different from two years ago, which is why I feel a lot more comfortable showing you these numbers. So our full expectation over the next three years is we can drive mid- to upper-single-digit growth in revenue. There are company and sector-specific growth drivers that we have. The product lineup is unparalleled, and all the work that César's team has done around the go-to-market has been awesome and bringing discipline. The economy will be what the economy is, but we feel much better about our ability to withstand some of those bumps because of our product offering. Now it's just not one; we've got the whole gamut around storage. So we feel very good about, on average, being able to drive that growth, and our goal is to grow above market and to drive share gains.
That's what we have: above market growth, drive share gains. That will then equate to operating margins above 30%. Yes, we hit it in Q3 of 2024. We were all helped by the low component costs. We guided 27%-28% in fiscal 2025, but we expect that to grow and stay above 30% as we grow through the three-year plan and certainly by 2027. Double-digit EPS growth is something that we really focus on. Again, over time, 2025 we guided down a little bit. Hopefully, we can continue to drive that up as we go, but that's really what we want to do: not only driving by growth, but then also driving the share count down, and then free cash flow margin sustainably above 20% of revenue. It was at in 2024, slightly over 20%, and we expect to keep that there.
Let me go back to the top line. Folks, the North Star that we have is the Rule of 40. We're not there yet, but we fully expect to get there over time. And it won't just be 40% margins and no growth or 40% growth and no margin. This will be a balance, and we feel very good about being able to get there. You may not see it in the three-year period, but hey, everybody needs a North Star. Everybody needs a plan, and that's ours. So that's the long-term model. Feel very good about all the things we've talked to you today about being able to get to these numbers. Okay. So this is my Luke Bryan moment. Remember the beginning? You can almost sing with me on this. The platform has advantages to customers and NetApp.
If there's one thing I would ask you to take away from this discussion, it is the model has leverage. It has, and it will continue to have leverage. You're talking to a management team that's very disciplined in terms of what we spend. Do we invest in new stuff? Absolutely. I give the whole team a lot of credit. First thing we ask is, what can you reallocate to drive growth before we need to spend incremental dollars? And you see that in the results. And then, of course, we will be good stewards of your capital. We'll drive a shareholder-friendly allocation, and our goal is to drive double-digit EPS growth going forward. So those are the takeaways. Yes, since I have so many numbers, I have a ton of GAAP to non-GAAP slides, so we will go through them. These are all, of course, on the website.
If you have questions on these or you want to go through them. And then, of course, Kris and her team are here whenever you want to talk about any of the numbers. Okay. So with that, hey, now the fun starts. We'll open it up for Q&A, and I'll bring Kris back to the stage. Thank you for your time. Thank you for your attention. Great to see everybody again. Thank you.
All right. Thank you, Mike. We're going to get some guys up here to help us with chairs, and I'll invite all of our presenters up, and we're going to start the Q&A session, which I know you guys have been waiting for. For those of you on the webcast, there's a question function. You can put it in. We'll have it read here in the room for you. For those in the room, we have two mic runners, and so raise your hand. Someone will come to you with a mic. Do not speak without the mic. Mike, you take the far end. We'll need so you can be heard on the webcast. George, you take this chair. Everyone else in the middle. So we can hear you on the webcast.
You want me here?
I want yours.
Okay.
Okay. Sorry about that. All right. I see you, but we're going to start over here.
Thank you.
Hello. There we go. Thanks. Nehal Chokshi from Northland Capital Markets. Mike got a question for you. Why are you projecting support revenue to be stable rather than growing, given that four or five product categories that you point out are expected to grow?
Yeah. So here's what I'd say: stable growth. It's grown again. Six, five, three, three Stable does not mean it doesn't grow. Stable means it stays somewhere in that low single digit. When we drive higher product revenue growth, hopefully that then goes up to mid-single digits. Stable growth does not mean zero.
Okay. Great. For the broader executive group here, what are your thoughts on what's going to be storage revenue to server revenue for AI-related revenues? Right now, it seems like it's a really low percentage. I hear people expecting it to go up as inferencing becomes a bigger portion, but we'd love to hear some perspective on that.
Yeah. Maybe I can answer that question. Listen, we are in the build-out of the new computing architecture for AI. It is built on a new computing model with the GPU model, and it is running on existing data. To be honest, ChatGPT was trained on the internet's public data. We didn't create a second internet to train ChatGPT, but it used a lot of GPUs because CPUs were not as effective for the training elements of AI. When the model—we are reaching the stage where the training part of the AI lifecycle is getting more mature, and enterprise AI customers are starting to now move from using foundation models to actually moving them into inferencing. And that's where the data growth will really start to happen, both in terms of building unified data foundations to accelerate RAG and inferencing, but also to deal with new data that's being generated.
As we've said before, we expect the growth in storage to be about a year or two out.
All right. We got questions. I know Asiya had her hand in the pink. She caught my eye first.
Thank you. If you can just elaborate on the block, I mean, it seems like it's an exciting opportunity for you guys, and the TAM certainly seems interesting. Just for fiscal 2025, what are your expectations for that block to grow? And will it cannibalize some of your file offerings that you have? If you can just elaborate on that. Thank you.
Yeah. I'll start the answer and then bring Harv in. So first of all, we won't be disclosing our block numbers. The majority of our block focus is in the flash part of our business, and so you should see the growth rates be part of our overall flash growth rates. They will not cannibalize file. These are workloads that are typically running on block environments that we are going to displace competitors, either as part of infrastructure upgrades of the competitor footprint or as new workload deployments on block footprints. Maybe I can hand it over to Harv to talk about why we are differentiated in that market.
Yeah. No, I can maybe talk more from the product side, George, which most customers, the cannibalization between block and file doesn't happen because they run SAN workloads, which are different than the workloads which are the NAS workloads. So traditionally, NetApp has been well-penetrated in the NAS or file market, and today we have a smaller share in the block market. Now, by releasing products which are really optimized for block, we never had that before. This is the first year. Last year was the first year we actually introduced block-specific products. Essentially, we are expanding our value to customers. We are saying, "Don't just use us for NAS and file like you've always done. Use us for block as well." We had some customers using us for block in the past, so block isn't new, but these are products which are optimized for block.
The software that runs on them, how it is configured, how the performance is done, the price we can charge for it is very different for these block products than the previous unified products that we have. So just from a capabilities point of view, it has all the capabilities of others, and then it's optimized for the SAN workloads. I don't see any cannibalization opportunity at all.
Yeah. And actually, just to complement this piece, think about the following. There's real time. When we close the fiscal year, we spend the month of May normally doing the account plans with our customers. Those account plans in the past, we're talking about our file, overall NAS penetration in those accounts. What happened during the month of May is in those account plans, we're talking not just NAS and file, but also their block environments. And we have done the same with our partners. Partners have been working with us on the file space or NAS space. We're now discussing with them how much business we're going to do as well in the block space. So that's why we feel so bullish for the opportunity.
Yeah. All right. I'll get a question down here in the front.
Hi. Samik from JP Morgan. Harv, if I can just ask you, going back to one of the slides that you had where you outlined the AI opportunity between data prep, model training, and fine-tuning, and then RAG and inferencing. I know in one of the initial slides, you had the $14 billion AI opportunity. If you can just share how big each of these individual TAMs do you think are within those three areas, and how do you think about enterprises starting in terms of timing on each of those? Is it more all three kick in at the same time, or are you thinking about a different timing? And just for Mike, second question, Mike, any more thoughts in terms of public cloud growth trajectory in your long-term model? What's embedded in there for the public cloud growth? Thank you.
Yeah. I don't think I know the TAM of those individual categories within enterprise AI. Maybe one of you or people in here might actually know that. I think of that space as generally the broad enterprise AI because to do enterprise AI, you need to have a workflow that spans from training to data prep to tuning to RAG and inferencing. And most enterprises I know, they don't have the scale to do large language model training. So the rest of the steps they still need to do. They still need to do data prep and RAG and tuning and inferencing and so on. So I think about that space all up as one space. I think in the context of AI, I don't think our customers look at them as separate product categories.
They're essentially trying to build an AI workflow, and it happens to span across those stages of AI workflow. So that's why I can't separate out into individual things. I don't know if anybody else here wants to add.
Maybe I can add to that. I think what we see in clients is the data lake, data prep, data foundation happens regardless of whether it's generative or predictive AI. You need to get a good analytic foundation, and you need to have the right coherent data sets to help you to build the right models, regardless of whether it's generative or whether it's predictive AI. So that's happening. It's been going on for many years. Remember, we've been in the AI market for now almost seven years. And so we've seen that.
I think there is a broad-based replatforming from the Hadoop model to more of an open-source data lake, and we see that going on. The foundation model training environments are going to be, if the way the world works is bigger and bigger models get better and better returns, and the cost of dealing with content and copyright issues, there are going to be very few very large-scale entities, and it'll be a concentrated environment for model training. I think the majority of the market will be in inferencing and RAG and fine-tuning, taking a foundation model, using enterprise data to train that model even better rather than starting from ground up. So that's really the sequence.
I'm going to riff on that. I'm going to say that, look, the exciting thing for me is that, look, people's data is everywhere, and you have to bring AI to where the data is. You can't bring all the data because that's everywhere to where AI is. And so one of the exciting things for us is how do we actually provide this in-place RAG, this in-place inferencing. The data stays where it already is in NetApp, and you can actually kind of apply AI to it. I think that's really exciting ways in which we can help the customer be able to use GenAI more easily for the data they already have.
We have a huge advantage because we hold a huge amount of the world's unstructured data, and we're going to enable our clients to do in-place inferencing.
That's right.
Mike?
Sure. Thanks, George. So, Samik, I'd ask you folks to think about cloud revenue in kind of two pieces. One of the reasons why we gave you the breakout on first-party marketplace is, hey, ease of math, it's about half the business. That grew in the last three quarters over 30%. The law of large numbers will start to play a little bit. If you go back to one of the slides, which is we fully expect to gain market share in all the products. So that will come down, but that's an industry, and you can look at your favorite outside group in terms of what that is. It's probably more like mid-teens growth as an industry, so you should expect to see that come down.
The rest of the business probably grows at a little bit lower rate there as we still get through some of the subscription headwinds. So as you model it, I'd encourage you to model it, which is, hey, first-party marketplace, and then everything else. So hopefully that helps in your model.
All right. We'll take a question from the side of the room. Anthony's choice.
Hi. It's Krish Sankar from TD Cowen. I had two questions. One for Mike. Just on your mid to high single-digit revenue growth target, can you segment it? How much is cyclical? How much is structural like AI? How much is share gains driving you to get to that? And then I have a quick follow-up.
Okay. Thanks, Krish. Hey, so we gave you the mid single digits, the upper single digits. We feel a lot more in control of our ability to grow than we did a couple of years ago, largely due to all the great work here and then what César is doing in high-end as well. That's probably dependent a little bit more on what the economy does. Specifically on AI, as George talked about, we expect that really to start to pick up in 2026 and 2027 as that moves to storage. On Samik's question, cloud revenue will start to become a bigger driver in the future, so that's under our control as well. So really, the variable within that is how does the economy do? How does AI or other stuff pick up as we go through? And then how does cloud revenue grow?
So much of that, thankfully, is relatively stable. Again, growing but stable support revenue. So that's, Krish, when you do the model, I'd kind of break it up that way. Hopefully that helps.
Then a quick follow-up for George or the management team. Obviously, Microsoft has been driving a lot of this GenAI through OpenAI, and you have the ANF service for Microsoft. Can you talk a little bit about how that has evolved pre-GenAI to now, and how do you see that evolving over the next few years? Thank you.
So our view has always been that our unique value in the public clouds is a proven architecture for high-performance unstructured data with the hybrid or multi-cloud capability. No one else has that, not even a hyperscaler. So what we've been working on with many of the hyperscalers is to enable an enterprise to bring their data to the hyperscaler tools so that they can do AI work using the hyperscalers' AI platforms, but with all of the controls, compliance, protection, security, and governance. And so stay tuned. We got lots of innovation coming. We'll tell you more about that. Maybe Harv, you can hit that.
Yeah. I think that when I talk to customers, almost every time they talk about their AI workflow as being hybrid. This is true for any hyperscaler. I know you asked about Azure, but it's true for AWS as well as GCP as well. The hyperscalers have some really good tools that people like to use for training purposes. But then most of the data is kind of on the enterprise side. And so it might be that perhaps AI is the first true hybrid app in the world. Maybe every time else we talk about hybrid, or maybe this is the first true born hybrid workflow in the world. And so for many customers, it kind of spans across the cloud and spans across enterprise because that's how the AI workflow works across those places.
When you do that, what really becomes important then is how do you do data management and copying of data and protection of data? Because that egress, ingress of data, who has access to what, how long does it take to make a copy of this data in the cloud and that thing on the enterprise? That's going to kill you. When we talk to customers, they say that 80% of their time in AI is spent in just fricking copying data. Only 20% of the time, the data scientists are actually doing useful work.
They're like, "Oh my God, that is crazy." And so from my perspective, if you have an AI workflow that's a hybrid AI workflow, which we can do because we are on the enterprise, we're in every single public cloud, we can save you that 80% of the time your data scientists are wasting. The data management becomes really important in that scenario.
Maybe I'll just add one quick thing. You asked about what has taken place since the initial work. So the Instaclustr services we provide, the Postgres solution we've integrated with ANF, and that performance that we were able to demonstrate is 3x more in terms of throughput and other things. So as we speak, we're working with some of the large healthcare providers to help them modernize their proprietary database workloads onto ANF. And I think that's just to George's point. It's part of their modernization, part of reimagining their data lakes and things. I think that's one of the driving forces. We're super excited about how some of the newer solutions we bring in would help drive the cloud storage.
All right. More questions over here. All right. We'll get Aaron.
Yeah. Thank you, Aaron Rakers at Wells Fargo. I'm going to apologize. I'm going to build on a couple of prior questions if I can. The first question is, it sounds like everything you're saying, you've got a pretty good line of sight into some of these AI opportunities. I guess, first of all, do you see that line of sight? Can you actually do you envision a period, Mike, in front of you where you could actually quantify how much of your business is exactly being driven by AI? Is that something that you'd think about? And then the second question is going back to the block storage. On one of the slides, you referenced 20,000 customers for the block storage products. I'm curious if you can give us any reference point of that. What was that a year ago?
How much has that penetrated into your existing installed base? Just any kind of curiosity because that does seem like a key vector of growth opportunity for the company. Thank you.
You want me to start? So yeah, we look at all opportunities. We understand. And it's really for us, Aaron, it's the workloads. What is the workload? We also know if it's connected to a GPU at some point. And César talked about what we have for most of our customers in terms of being able to see that. In terms of quantifying it, there's a lot of hybrid stuff. We've thought about it. It is a big part of the pipeline. Obviously, it's a big part of what we look at when we project the future. And we typically will track it by workload. We know that more on-prem. We're starting to find out a little bit more in the cloud because of the work with the hyperscalers. So yes, it's something we view. I know some folks have tried to talk about how big their business is.
Hey, we've been in this a long time. We've done a lot of AI for a while. We expect that only to grow.
Can I?
Yeah, go ahead.
So certainly, we have our internal targets, and we have people compensated on AI. I think what Mike is saying, which is we're fully aligned. There's some workloads around AI, which is straightforward. So this especially is working with the field people. You have a SuperPOD or BasePOD . You name it, it's super simple. We're talking about AI analytics is a bit more difficult. So we track that separately, but there's some specifics that we're tracking. We have a pipeline that we review every week around AI. So this certainly, we have put the discipline to track that.
We have, as Mike mentioned, we have good telemetry into what applications and workloads people are running on our environments. The data we are sharing with you and the milestones that we have shared, they are true AI workloads. They run on GPUs. They are built for either predictive AI or for generative AI. We have a good line of sight into what is actually being consumed on our infrastructure, whether it's on-prem or in the cloud. We will not disclose all of that. It will align with the disclosures that Mike and I would want to do. I think with regard to the opportunity ahead, we feel really good about it.
To your SAN question, the 20,000 customers are customers who run block workloads either on unified storage environments where they can consolidate file and block onto a single architecture, or where they run block workloads on our dedicated flash storage devices, whether it is a unified storage device but configured only to run block, or whether it's the new ASA block footprint. So it's a large part of our customer base. We have a proven track record, which is why we are really confident about the opportunity to expand wallet share.
All right. We'll get one more here.
Thank you very much. Simon Leopold with Raymond James. I guess one of the things I'm struggling to envision is when enterprises are engaging in their AI projects, what's the mix of data being pushed up to the cloud for those AI projects versus an accelerated refresh of the on-prem storage and the large language models coming to the enterprise user? How do you envision which way are things going? Is the data moving up to the cloud or the large language model coming into the enterprise? Thank you.
It depends a lot on the industry and the use cases. I think in regulated industries, for example, there's a lot more of the on-premises training environment or really inferencing environment that's being built out where they will take a large language model, either an open-source model or a previously trained version of one of the well-known public cloud models, bring it into a lab or a sort of a center of excellence environment. We have several examples of wins we talked to you guys about during the earnings call. And they'll build a kind of an inferencing lab in their own data centers. In customers where there is less of an on-prem environment, like a SaaS company, we have an example of one of the really large SaaS software as a service companies that's using AI work that's building AI into its environment. They're using a hybrid model.
They use us in the public cloud alongside some of the big hyperscalers, but they also have one of their own dedicated environments. It really depends on the industry and the kinds of concerns around privacy and protection and so on.
I'll add more to that, George. Hey, I feel like when we talk to customers, they seem to have gone through four phases or two phases, I should say, for AI. Initially, there was a lot of, I would say, experimentation and piloting. Everybody wanted to learn how to kind of go train their own language model. And so if you were in that experiment stage, if you were in the pilot stage, people would take the relevant data from on-premises and move it to the cloud because the cloud has the best tools for training. Now, as AI becomes more mainstream, most customers are realizing that, hey, they don't actually need to go train the large models. Those are already done. How can I take the large LLM model trained by others, and I'm going to go combine it with the proprietary data I already have in the enterprise?
And then this is where I think RAG has changed that. Inferencing has changed that. And that's how they're deriving the most value for GenAI. And so customers are now mainstreaming or beginning to, I would say, mainstream this Phase II of AI. And that model is most of the customers we talk to is largely on-premises data.
Inferencing.
You can't really I mean, data, you can't just move all your data to the cloud. There's no magic wand. This is impossible. That data is kind of really distributed. It's on the edge. It's inside SaaS apps. It's in SharePoint. It's in all those places. It's not physically possible. It's a physics problem to try to move all the data into one place. It's better to bring AI to where your data is, which is on the enterprise.
All right. We've got a question from the webcast. OK. This question is from Jason Park of UBS. Why would your customers actually prefer your OS, hybrid disks, and NAND SSDs over lower-cost direct flash offerings from your closest competitor?
Let me answer that question. There is no magic bullet on direct flash modules. There are trade-offs that a direct flash vendor has to deal with, just like there are trade-offs we deal with. If you look at the usable cost structure of direct flash modules, meaning when you buy a raw module and then convert it to being usable for data, it is a lot less efficient, meaning for 1 terabyte of raw disk, you get a lot less usable capacity than you would get if you used one of our devices. And so the idea of a cost structure advantage on raw is basically not true. Second, there's a lot of claims that DRAM as a part of direct flash architecture is better than the DRAM being used in an SSD environment. That is not true.
If you look at the percentage of DRAM in a large capacity device, which is the de facto model going forward, it is essentially a fixed percentage of the SSDs and the direct flash module. It is a tiny and irrelevant portion of the cost of goods sold. Third, if you want to go and build your own direct flash module, you will end up accruing an enormous amount of R&D expense to do for yourself what the industry does for solid-state devices. That means scale testing. That means reliability testing. That means environmental readiness. So there are trade-offs on operating expense that becomes a part of the cost of a device as opposed to cost of goods sold.
Fourth, if you look at the cost structure of the whole system, the fact that you have a raw NAND flash module makes you have to now spend more on the compute horsepower in the system because you are actually doing a lot of the tasks that the processors in the SSDs had, which is why if you compare the data sheets environmental impact of one of the direct flash module vendors against ours, we are more efficient in terms of environmental footprint. I would just tell you that every architecture has pros and cons. This magic bullet bullshit that our competitors keep talking about is just not true. I'm sick of it. I don't want to hear it anymore. Get the facts. We'll tell you about the technology.
I also think this is like solving yesterday's problem for customers, maybe. Because yeah, I mean, it's like if the customers look, the problem that customers are running into are how do I use which data for AI? How do I protect the data? How do I do governance for the data? How do I do security for the data? That's where in some way the IP is. That's where that's what a solution has to kind of go solve for the customer. I don't know how this direct flash thing solves any of the problems that customers say are the ones that are preventing them or preventing them from harnessing the value of data. You can invest in that, or you can go invest in helping customers harness the value of data by solving the security, protection, governance, privacy, AI problems.
Our approach and vision has been, wow, we've got to have an expanded solution, not just dump storage, but an intelligent storage platform that helps customers solve these broader problems. I think that's solving today's problem as opposed to solving yesterday's problem.
All right. I know there's some more questions. Let's go over here to Meta .
Hey, thanks. Meta Marshall from Morgan Stanley. I think three years ago, you guys had laid out that about 20% of customers of the cloud products were existing customers. And obviously, you're talking a lot about that being a growth vector. Is there a way to think about what percentage you think will be kind of these combined hybrid customers going forward, or just ways in which you're enabling kind of the third-party sellers to have more of a list of who these customers are to kind of go after to ensure more of that hybrid business? And then maybe just for Mike, as you think about using AI internally, just contextualizing that, I know you laid out the AI-enabled customer service, but just ways you're thinking about kind of savings within the business from using AI internally. Thanks.
Maybe I can start the answer to your first question, Meta, and then have César address it, and then give the second to Mike. First of all, I think we are super excited by the fact that public cloud has been an awesome source of new customer growth. I think when we first introduced the story of public cloud, many investors thought that it would be purely a substitution of our on-premises business. That has not proven true. We really are excited about the fact that the hyperscalers enable us to get into new customers, new buying centers in existing customers, new workloads, and are able to drive a lot more growth with us. That being said, we have a disciplined model with them to target the largest opportunities, many of which are in the largest accounts in the world, which are our existing customers.
César has built a focused go-to-market model to help address that.
Yeah. So great question. I will say the beauty about the model is we're getting both. So you start think about when we started, the reality is many of our existing customers took advantage of that. But the reality is the hyperscalers are really scaling. They've been reaching new customers. And as I said before, we have our cloud specialists supporting them. So let me talk about the two pieces. One is the hyperscaler going to market with their own specialist group, their own sellers, their own cloud service providers. And actually, each one of them having position in these solutions in different ways. So they're basically reaching out to the markets, and they come to us for help when they need our support. So our specialists are mapped to their own guards, in other words, to their own segmentation, supporting them.
Then we have our own sellers that talk to existing customers of potentially new opportunities. And our own sellers will bring the specialists to come and discuss that opportunity. And the specialists will bring the hyperscaler as well. So our seller can go and focus on an opportunity, and the specialist will go and do that. So answering your question, we're seeing customers joining us more and more outside of our install base. And we have existing customers, as George said, taking advantage of it. Now, the last piece that I want to share, which is very important for us, all those customers that were not NetApp customers, where our sellers have not even engaged with them, we have our specialists engaging with them. The hyperscaler, in a way, is owning, and I'm putting owning in other brackets, that customer.
We build this customer and create this customer success team that is working with the customer success team of the hyperscaler, and it's engaging directly with the customer as well. That's very important for us for many reasons, enabling the customer, ramping up the customer, ensuring that we really have a customer experience with them and the hyperscaler. But at the same time, they can take advantage of some of the marketplace offerings that we have, not just the first-party service. We're starting to see opportunities through the customer success team. Like scenarios, the one that we have talked, we're saying we're doing this on AI, but they want to bring this on-prem, and they want to do it with ONTAP.
Our goal is that investment on the customer success to help us not just to retain the customer on the cloud side, but also on the on-prem side, if that makes sense.
Meta, thanks on the AI question. Here's what I tell you. We have so many use cases that we're actually having to shrink the list. When we look at NetApp, and we talk about our North Star of getting to the Rule of 40, and again, all sing with me, the model has leverage. Part of that is we need to bake AI into everything that we do. There's a lot of great work, not only in our customer support area, in the R&D team, in marketing, in the way we serve our customers. From my perspective, I would love to be able to take all the manual stuff out of what we do as we roll new systems. There are multiple use cases. We are using fiscal 2025 to really prioritize those. Where can we drive efficiencies first and then also serve our customers better?
You'll start to see us talk about the efficiencies as we go into fiscal 2026. That's when you'll start to see it roll through the P&L. But we realize, hey, this is a competitive advantage that we need to have.
I think the other thing I'll add just a minute is that rather than thinking about our cloud substituting our on-premises, maybe you should think about our cloud substituting a competitor's on-premises deployment. Because if you are on a competitor's system on-premises, and you have a VMware workload, you want to do disaster recovery of that to the cloud, or you want to move the entire VMware workload to the cloud, the best option for you is to go to NetApp. And so every time there's a big movement like that of on-premises to the cloud, it's a great way for us to go disrupt the existing market. So we should think about the existing market as the entire on-premises market, and how can our cloud offering disrupt the existing on-premises market and substitute competitors' on-premises deployment with our own cloud storage?
This is why we see net new in cloud rather than just substituting our own on-premises.
OK. I know, Lou, you've been trying to get a question.
Hey, thank you. Lou Miscioscia, Daiwa Capital Markets. So my question is a go-to-market question. So over the last five to 10 years, there have been times when you've had too much of a concentration in certain areas, maybe the global customers. I'm not sure. So can you give us a breakdown of the different ones you just talked about, as much color as you can, for global, large, public sector, and so on? Actually, you disclosed public sector. And then secondarily, at various times, you've had to shift around the sales force, either hire more or retrain. Just, is there anything going on that doesn't sound like it? But just clarifying that would be helpful too.
Sure. So we didn't disclose the segment piece, but certainly, I will say we probably can say that we serve the largest companies on the planet, and we want to keep doing that. Going back to the discussion we had at the beginning, I think it was the first question, we believe our share of wallet can even increase more with those accounts. Now we have block, AI, et cetera, et cetera. But during the last years, we've been building the sales force that is able to go and penetrate even deeper. And that's what I share about the public sector side. I have recognized that it's a big opportunity for us to go deeper, not just in the U.S., which we have done the last several years, but in many other countries.
So that's why we're investing more to go and say, let's go and penetrate even more that public sector side. Probably the biggest area of opportunity has been for us what we call the commercial segment, which is the SMB space. It's a bit tricky because many of those customers have been served through our service providers. So we look into a service provider as our customer, but the end user is one of those SMBs. With the new product features, I mean, the product launches that we have done with capacity flash and our cloud, that's helping us to come and address in a better way that segment. So all in all, our goal is how do we balance that mix? But that mix for me is not about, hey, should I go lower in the large enterprise? But I want to grow in each one of them.
You're going to see in the next years, hopefully, each of those segments will be growing, their top line, and the number of customers there as well. Probably the large enterprise will not grow the number of customers. We always will have 100 approx, but the rest, that's our goal.
If I were to just summarize two things, Lou, that I would leave you with, one is our opportunity in the larger accounts is much larger today than it was the last time we were here because of what Mike and Harv and Haiyan talked about, as well as César's focused approach and go-to-market. I think the second is we have leveraged ways to address the mid-sized enterprise segment, both through our partner network, the hyperscalers, and managed service providers.
All right. I'll get to David in the vest. And I see you, Wamsi, in the back.
Great. Thanks again, guys, for doing this. David Vogt, UBS, different part of the bank. Maybe a question for Mike to start. Mike, when you put up that really detailed slide about the revenue trajectory for 2025 through 2027, you talked about the growth vectors. But I'm curious, when you think about support being such a large part of the overall mix, where do you see the growth coming from in terms of above-market growth? And can you help us understand what underpins that growth, whether it's in all flash and block and some of the other categories? And then you mentioned separately GDP. When AI compute becomes a bigger part of the overall portfolio, how are you thinking about maybe IT prioritization away from traditional workload storage to storage related to AI?
Does that have an impact on that part of your business going forward over the next couple of years? Thanks.
You want me to take the first one? OK. So yeah, when we look at the 2025 to 2027 long-term plan, so that's just the underpinnings. Support we expect to be a stable grower, super important. That's not only new product revenue, but also renewals, it's more important. What we talked about in terms of market share are all the things that I listed. In terms of, hey, we will be a disruptor in block, all of that is additive, David. Flash, you add the AI workloads, and block is a part of that, and AI as well. And then you have Keystone, which is now going to start to be actually a material part of the growth as we go forward. And then importantly, cloud revenue, where we've kind of flattened out, and we expect that to grow as well. So support is a result of all of that.
It underpins everything we do economically. But I would look more at, hey, the product revenue growth, which will, again, drive support growth in the future as well. Does that answer your question? OK.
I think with regard to your question about spending priorities, listen, I think first of all, we are well positioned against the spending priority as the enterprise. We see data security becoming a higher part of people's concerns. We see cloud acceleration, VDI reappearing. We see that, hey, people get a lot of economic and sustainability benefits by moving from a legacy disk-based architecture to a flash architecture. Those are all trends that happen over multiple years. So nothing that we shared with you shows a hockey stick in any of the models. It's like we are in very large markets. We are well positioned under a range of scenarios. I think with regard to AI, what I would tell you is in the long arc of computing, unstructured data has actually about 70%-80% of an enterprise's data.
And for the longest time, there has been no mechanism to actually extract insight from unstructured data. It was always on the structured data side, databases, data warehouses, BI tools, and on and on. With generative AI, you actually now have a scaled application that can derive insight from your unstructured data. And for us, that is a very, very positive thing because it suddenly moves your unstructured data landscape from being a source of cost and, hey, I've got to keep it around for compliance reasons, to now actually can derive insight from that. And so we feel good about the long-term trajectory on that.
All right. We'll get to Wamsi in the way back.
Yes, thank you so much for doing this. Wamsi Mohan, Bank of America. I don't think I caught what your TAM growth assumption was overall, I think, to get to the $100 billion. What is the aggregate CAGR as we look at the TAM growth to start, please?
The overall TAM is growing from $84 billion to $103 billion at a 7% CAGR, roughly speaking. We can get you the details on the sub-TAMs.
OK, thank you for that. Then maybe just to follow up on the last question, when we think about the support being very stable and contributing, I don't know, maybe low- to mid-single-digit kind of growth, and aggregate total growth that you're talking about is mid- to high-, that would really mean that your product revenue is going to be fairly strong over the next few years, going maybe high- to even double-digit growth. Is that the right way to think about it? Product historically has been kind of volatile. So if you look at it over the historical period, maybe you can just bridge why you expect this kind of more re-accelerated growth. I know you spoke about seven different levers, so obviously there's a lot there. But just to maybe characterize that in context of history would be helpful. Thank you.
You want me to take it? OK, so thanks, Wamsi. Hey, the last two quarters, product revenue has grown 10% and 8%. George just talked about the growth in TAM. The reason why we feel very good about being able to drive product revenue growth is that slide with all the drivers in it. Connect it to the history. We never had all of those selling opportunities. We couldn't address all the workloads. César has done a wonderful job, and he talked a lot about going to get net new. Again, this is very different than the past in terms of all of the products that we have and the focus around go-to-market. Yes, in a mid- to upper single-digit growth because of the stability, again, thankfully, of support, that means product revenue typically grows higher.
The variables for you to think about in your model is, hey, cloud revenue at some point, we talked about it, we'll start to re-accelerate. We feel good about that as we go through 2025. Keystone will start to add as well. So there are multiple growth drivers in that. There's no, I would say, sacrosanct number in terms of product revenue, but we feel so much better about the products that we have, the ability to address the TAM, and the go-to-market to support that. So again, coming off Q3 and Q4, we feel good about that. We haven't gone crazy in terms of product revenue. Those other pieces will also start to contribute to growth as well, which again, Wamsi, and we both know you haven't seen that in the past.
All right. Question from Sean. And then Ananda, we'll get to you.
Quick one for Mike. If I look back fiscal 2022 to 2025, a lot of the leverage came from gross margins. There was some at the OpEx line, but most was gross margins. Looking forward, can you give some color on what is embedded in operating or the key drivers of the operating margin going forward? Thanks.
Sure. So here's the way I, most of the gross margin leverage you saw was in 2024, right? And that was product gross margin. As we guided 2025, and we talked about this on the call, Sean, the majority of the gross margin dollar incremental dollars is coming from support in cloud. Because again, we have to get through the component cost thing, and we'll do that as we get to 2025. And you have all the numbers. You can look back. Most of the gross margin growth in 2024 was product. Great. It's now support in cloud going forward. I'm going to push a little bit, which is a lot of the bottom line leverage has also come from OpEx. That really only moves around as it relates to incentive comp. And most of the growth you saw this year was, thankfully, we had a good year.
Last year, it sucked, right? So that was a big difference. There's also been a little bit of leverage as well in OIE because we now actually get interest on our cash. So both of those have contributed. I would push a little bit. Not everything came from gross margin. Some of it was also what we did in OpEx and then also with our interest expense. So going forward, what I'd say is once you get through 2025 and we get through the higher component costs, we think that that normalizes. Here's a gosh darn who knows what happens there, right? And really the game there will be pricing probably more than cost. Completely confident that cloud will continue to drive gross margin growth because the structural costs have changed there. And the growth in first-party and marketplace are the higher margin growth as well.
Now we'll be able to utilize that asset base that we've deployed. Support, no concerns about that staying above 90%. Again, that will drive a lot of it. The variable there will be what happens with component costs. So going forward, once you get through 2025, it's probably more of an equal mix. My other comment, Sean, is I don't know where component costs go. As I like to say, it's like high school English. The drama is crazy, right? It's going to go where it goes. We will react to it. Support in cloud, under our control.
All right. We'll take our last question from Ananda.
Yeah, thanks, Kris. Thanks, guys. So Ananda Baruah at Loop Capital. Two if I could. The first is George. Is there any useful distinction to be made in the share gain conversation, both within existing accounts and for new accounts, share gains occurring from sort of on the point basis, the various incremental opportunities you have versus customers taking a fresh or maybe an increasingly holistic view of their storage strategy given new market realities and saying, hey, who has the best holistic solution set? NetApp is that person or is that entity. So is there any meaningful distinction to be made there? And then the follow-up is probably for Mike. If you guys were to achieve the high single digit versus the mid-single digit, what would be the things you would think would occur for that to happen? So those are mine. Thanks, guys.
I think with your first question, it is increasingly evident to more and more clients that they don't have a broad range of IT skills and resources available for infrastructure, right? So what you see in every part of the global IT landscape is that the number of people that are available to an IT department is less, significantly less than it was a few years ago. If you look at the jobs data, the rate of acceleration of compensation is about double digits. It's 9%-10% CAGR for the last 10 years. So the idea of having five different architectures with different operating systems, different security models, different APIs, different monitoring is beginning to become a bigger and bigger challenge.
You know, I would just tell you we are winning with the idea of a consistent architecture, even some of the largest accounts, which have the resources and can really have multiple different environments. But I think that in general, what I see is the lower end of the customer base will even find greater value from having a consistent architecture because they have fewer resources. I think with regard to competitive technology, we feel really, really, really good about our portfolio. We feel really good about our portfolio. We've got optimized flash. We've got a cloud story that nobody has. And we are well positioned for AI.
Did you ask about the compensation of our people between a refresh on a new project on block or AI or something like that? Or did I misunderstand that piece of your question?
Yeah, thanks. Now, the follow-up was if you guys were to do the high single digit of the guide versus the mid-single digit, what do you think would be the things that would have occurred to allow that?
So I'm going to lean forward. When I expect that will happen, here's what the drivers will be. And hey, Sean, one more thing on that margin thing. That mix is super important. Don't lose sight of the hard drive down to flash. That is the secular trend as well. So I just want to leave you with that. So Ananda, on your question, feel really good about flash. I think the thing that will drive even higher growth than we expect is where we do better in block. That is a little bit of a newer market. First-party and marketplace, we feel really good about that growth. Keystone, you saw tremendous growth. That's because it's a subscription, we'll start to add as well. So we feel really good there. I think the variable there is the timing of AI.
We've talked a lot about it in terms of, hey, storage will come at some point. I think that depends. And a lot of that is, call it out of our control. Companies have to get ready. They have to get their data ready. We've talked about that. I think that's the variable in terms of what's under our control.
Just to comment, and I know we're running out of time, but from a compensation point of view for our sellers, AI has accelerators. Block has accelerators as well. And new customers have accelerators. So basically, those three elements are key things for them to go and pursue.
All right. Well, thank you all. I'll ask everyone except for George to leave the stage.
Thank you.
Thanks to everyone who asked questions. Apologies if we couldn't get to yours. In the room, you'll have opportunity for more Q&A over lunch. From the webcast, just shoot the IR team an email, and we'll get back to you. With that, I'll hand it over to George for some closing remarks.
Thank you, Kris. Thank you to everyone who joined us here today, both in person and on the webcast. I want to just tell you that we appreciate you spending your time with us. I want to close with a couple of comments that reiterate what were the important takeaways that I wanted you to all hear from us. The first is that we are uniquely differentiated with a software-developed data platform for the age of data. That is building on 30 years of innovation, disciplined innovation, where we have continuously expanded the range of capabilities of that data platform. And now we are addressing a much broader range of opportunities. The inherent growth rates of our markets, as we said, is about an $80 billion market growing to a $100 billion market at a 7% CAGR.
But we are exposed to high-growth segments of that opportunity: flash, block, cloud, and AI. And we have a focused strategy. César talked about it. Harv and Haiyan talked about it in the technology side. And I hope that over the last many years, Mike and I and our leadership team have demonstrated the capability to have a leveraged business model that delivers value to both customers and to our shareholders. You know, I'll close by saying we are addressing a $100 billion opportunity with a uniquely differentiated platform that combines software, data intelligence, AI and ML capabilities, and data and operational services integrated into the world's biggest ecosystems to deliver unique value to customers. We have a focused go-to-market plan to actually capture that value and those opportunities within our customers.
As Mike said, we have a track record of a leveraged business model, high-growth, high-margin products in growing market opportunities, combining operational discipline and a shareholder-friendly capital allocation structure to deliver earnings growth for you. I want to thank you. I hope you have a wonderful rest of the day. We will disperse to take Q&A. Kris?
Thanks, George. Thanks, everyone, for sticking with us. I appreciate all of your time today. This concludes our webcast. Please reach out to the IR team if you have any questions or you want follow-on conversations. We're happy to get together with you. For everyone here in person, we'll be doing small group meetings over lunch in the hallway where you came in. You can grab lunch. There are three rooms that you can get settled in. Pick any room. It doesn't matter. Every executive will circulate through each of the rooms. You'll have an opportunity to spend about 20 minutes with each exec. We're pairing up Harv and César and Haiyan and Mike together. George will be on his own. You'll kind of have 20 minutes with each group of executives as they circulate through your rooms to ask more questions.
Hopefully, you'll find that valuable. Grab your lunch. Grab a room. We'll see you in there. Thank you very much.