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

Okay. Let's get started. I'm Jason Ader with William Blair. Welcome to everybody this morning to our tech talk about NetApp's Public Cloud business. Very pleased to be joined by Anthony Lye, who is the EVP and GM of Public Cloud for NetApp as well as Kris Newton, who's VP of Investor Relations. Welcome, Anthony.

Thanks very much, Jason. Great to be here.

Good. And before we get going, I need to read NetApp's safe harbor statement. Today's discussion may include forward-looking statements regarding NetApp's future performance, which are subject to risk and uncertainty. Actual results may differ materially from the statements made today for a variety of reasons described in the most recent 10-K and 10-Q filed with the SEC and available on the website at netapp.com NetApp disclaims any obligation to update information on any forward-looking statements for any reason. With that out of the way, I also wanted to let everybody know that if you have a question, there's a Q&A box at the bottom. You know the drill. You can enter your questions at any time during our talk. We're going to -- Anthony is going to start out with some slides, and then I'm going to do some -- ask some questions directly to him. So please enter your questions, and we'll try to get to as many of them as we can. So I guess that's all the housekeeping. Anthony, let me turn it over to you.

Thanks very much, Jason. If we could just put the first slide up. I joined NetApp -- it will be 5 years next month. And my background, really, has been exclusively in the application development space, building substantial applications in CRM, ERP, human capital management across a number of different industries. And I was really excited to join NetApp to really think about what NetApp might be in a public cloud. And I think the most important thing that we've been able to do is we didn't just think about storage in the cloud as we had through about storage on-premise. And what we've seen really in the public cloud is a fundamental, I think, restructuring and reorganization and reimagining of sort of technology, really. And what we're increasingly seeing, and it was probably best said by Marc Andreessen that, that sort of software is everywhere. And that every company, even if they were born physical, is now digital. And we're seeing this huge, I think, increase in companies that are born digital. Or businesses within companies that were physical that are now exclusively digital. And so the strategy at NetApp was really to sort of, to think about the problem from the lens of the application, not from the lens of the infrastructure, that organizations would increasingly become application-centric and would focus on data and the use of data in the application to increasingly differentiate itself over its competition. And so over the last 5 years, we have organically built and inorganically acquired, a number of technologies that position us, I think, very much at the forefront of the cloud opportunity. And it's that cloud opportunity that really, really excites us, and we see it as an enormous growth opportunity and one that increasingly broadens our appeal to our existing customers, to customers on-premise that hadn't purchased NetApp and the sizable opportunity we see with organizations that were sort of born in the cloud or these sort of cloud natives. And if you go to the next slide, just a brief sort of history of what we've accomplished at NetApp over my 5 years. I think the things to understand are -- before we go into the sort of the product side of things, is we see opportunities for NetApp with the sort of with the traditional IT operations organizations. IT, prior to the cloud, was structured in sort of horizontal layers. You would have a networking team, you would have a storage team, you would have a security team, you'd have a database team. What we see in the cloud is IT operations are bridging and extending and migrating their existing infrastructure models over to the public cloud. But uniquely to NetApp, we have also, I think, recognized the position and value proposition of our portfolio to these application-centric teams. So sort of a 90-degree pivot from IT infrastructure into application-centric infrastructures. And here, we tend to see a customer that identifies themselves more as CloudOps than traditionally IT ops. And within CloudOps, we tend to see these sort of application-centric teams dividing themselves up into DevOps, into FinOps, into SecOps, where they're sort of constructing a holistic group to manage the SLA, the SLO and the competitive advantage they get through the application. I want to make sure that everybody understands that this has been a huge opportunity for us not just to provide safe harbor and to provide direction and openness to our existing customers, but we are acquiring a sizable number of our competitors' customers who, frankly, have missed the cloud opportunity, who now deny the cloud, who now consider the cloud to be the #1 enemy. An enemy, frankly, I would not want to have. The partnerships that we have with Microsoft, with Google, and as of last September now with Amazon, are very unique to us, unique in the industry, and we are taking sizable share as our customers consider the cloud architectures for extension migration or new application development. But equally, we see development the opportunity in our competitors' customers. We've thought long and hard about the importance of the channel while we had built, I think, for many years, a business that was very successful, directly selling to the end customers. We have, of course, uniquely captured Microsoft, Google and Amazon, our strategic partners and OEMs or first-party partners. But over the last sort of 12 to 24 months, we've really put a lot of time and effort into the needs and wants of the managed service providers, a new category of partner, global systems integrators as well as the traditional resellers and distributors. I'm incredibly proud of our numbers. If you look at what we're showing every single quarter, I hope that you're seeing customer acquisition rates, retention rates and expansion rates that are as good, if not better, than almost any other cloud company. We have really chosen to embrace the cloud fully in our business models. And so at NetApp today, a sizable amount for our business is what we call pay-as-you go. We allow our customers to consume on a monthly basis, our services, as much or as little as they need. We allow and fully support the elastic properties of the cloud. People can use our technology to scale up and to tear down or scale back. As well as providing consumption services, we do provide our customers an annual subscription service. And we've embraced all of the various different channels to sell, of course, the big 3 clouds themselves sell, serve and support our technologies and our services. But we sell not only directly through the NetApp brand and our existing channel, but we've made conscious decisions to embrace the public cloud marketplaces. And we are one of the most significant contributors to the marketplace road maps, and we use a sizable amount of the public cloud marketplace functionality to serve ourselves. Next slide. Just a brief history, 5 years ago, we had investigated our ONTAP technology, and we made a port. We took ONTAP in its entirety and reported it to Amazon. And that formed the basis of a very successful business we now call Cloud Volumes ONTAP. It's deployed by a customer, managed by a customer through an Amazon machine image, and it sells incredibly well to our installed base. But what we did, and I think what we did incredibly well and uniquely was, we decided that we wanted to put ONTAP in the cloud to be as good in a public cloud as it was on-premise. What we're talking about is incredibly high throughputs, incredibly low latencies that will allow customers and give customers the confidence to know that ONTAP in the public cloud was as good as it was on-premise, that we would give them that absolute symmetry. We didn't just stop at the primary storage. We felt that in this sort of application-centric world, customers would want more than just primary storage and that they would want us to take care of things like backup and restores, cross-region replications, compliance, data classification, that we would enhance the primary storage with capabilities like cashing. And we do all of those things through our Cloud Manager product. We made a bold decision to take a very, very successful enterprise software business OnCommand Insight to the cloud, and we reimagined and reengineered the intellectual property of Cloud Insights of -- OnCommand Insight to now Cloud Insights, a very fast-growing cloud business. And it's done very, very well in that it's helped the OnCommand Insight extend their infrastructures to cloud with Cloud Insights. But it's given us a whole new area of opportunity to sell Cloud Insights to companies that couldn't afford to or couldn't deploy OnCommand Insight, whether that's a division or a department of a large company or a small or medium company that previously was sort of unreachable. And then, of course, we made a very sizable decision to acquire Spot, and Spot now forms the basis of our CloudOps platform. You may not know, but we've made 4, now 5 acquisitions with CloudCheckr to bolster the Spot platform, and we believe that we are a leader, a significant player now in CloudOps. And we have a very, very strong bench of technologies that serve our customers in automation as well as optimization, spanning DevOps, FinOps and SecOps. And the important thing to understand here is that we chase, on a daily basis, all of the cloud applications and we chase all of the cloud application teams. We chase the enterprise applications, we chase high-performance compute workloads, we chase the SaaS vendors as they migrate from SaaS 1.0 colo architectures over to the public cloud, we chase the virtual desktop infrastructures, and we chase applications being built on cloud-native technologies like Kubernetes. And most recently, with our introduction of Spot's Ocean for Spark, we're now providing automation optimization services for big data lakes. Next slide, and then we'll go into Q&A. So there are 4 motions really, Jason, that we target. Our field wakes up every day, our channel and our partners wake up every day and we try and serve customers across one or more of these particular motions. Our customers are existing on-premise customers, see the public cloud as a natural extension of their existing infrastructure. They want to use the public cloud for simple things like disaster recovery, backup and restore. They like the elastic properties of the public cloud for things like dev test, they -- for bursting, where workloads require incremental CPU and storage that they can't serve within their own on-premise environments. We help people migrate holistically applications like SAP, relational database workloads, high-performance workloads in oil and gas, pharmaceuticals, genomics, financial services, all now see NetApp and NetApp capability as that sort of symmetrical service that they relied on so heavily on-premise. And we are now running some of the very largest enterprise applications ever created on our platform, on Microsoft, on Google and now on Amazon. And this is where our sort of first-party services have really given us this wonderful opportunity for growth. We help people automate, optimize and secure applications that were built or run in the public cloud. And that's really where we focus our Spot offering. Companies that really see the cloud as competitive advantage, companies that want to embrace CI/CD. They want to embrace speed as the biggest differentiator in their businesses. These are organizations that want to move very, very quickly but are frequently hit by cost overruns, by inefficient and ineffective infrastructure decisions and risks due to missed cloud configurations. Spot plays incredibly well there, and it's been a hugely successful acquisition for us. And as I said on the prior slides, we've made a number of other acquisitions to bolster and enhance and increase the growth rates that we see. And then of course, we give our customers the ability to normalize their infrastructures across the hybrid multi-cloud to monitor and manage their environments, to make decisions as to sort of what stays, what goes to the cloud using Cloud Insights. So technology that was sort of based on a very successful enterprise platform that was used by probably 75% of the Fortune 1000. We've now replatformed and reimagined that as a fully enabled multi-tenant cloud service giving us reach to more and more customers. So these are our plays. We feel that we do address the entire spectrum of opportunity on the public cloud. I think honestly, Jason, we -- what we did was we wanted to make the cloud a first-class citizen. And I think unlike all of our existing or traditional competitors, we made that bold move, and we made it early, and we were able to demonstrate to Microsoft then to Google, and now most recently to Amazon, that we could deliver a service and we could deliver capability at cloud speed, and we would understand and embrace the sort of the fundamental dynamics and attributes of a public cloud. So let me pause there, Jason, and see if there are any questions or how you want to proceed with the conversation.

Great. Yes. I've got a list of questions. And then again, for the folks on the webinar, please enter your questions in the Q&A box if you have any. Maybe to start out, Anthony, that was a great level set on kind of what you guys have built and how you've kind of fashioned this business over the last 5 years under your leadership. My first question is really just, what are the critical factors that drive customers to adopt public cloud from NetApp? Is it a workload-level decision? Is it a business model kind of directive? Or maybe like top-down CIO saying we got to go to the cloud first and therefore -- but we want to maybe work with folks that we know?

Yes. No, it's a great question. So I think we -- the honest answer is we see both. We see companies that are making sort of top-down cloud-first or cloud-only decisions. And these honestly, Jason, are largely made in partnership with 1 or 2 or 3 of the big public clouds, and we are brought in to help the public clouds with application migrations. And here, we chase the sort of the traditional IT infrastructure teams who are tasked with the cloud-first or cloud-only mandate. Now even with that top-down mandate, companies tend to sort of move on an application or workload basis. And let me just point out the difference. An application is, is it's in itself, it's an entirety, like an SAP, and all of the surrounding services or workload is a sort of a subset or a subcomponent of an application. Like disaster recovery or backup, where they may leave the primary application on-premise, but they'll use the public cloud as a secondary. So they'll migrate part of the application environment, and we help people do both. Now second, we've, probably like everybody else, increasingly see that public clouds, more than anything else, have sort of democratized infrastructure. And that now more and more companies are able to take advantage of technologies that were once only available to the largest companies in the world, and we're starting to see more devolved decision-making within companies. Increasingly, the line of business that was either born physical is now building applications to complement physical with digital or we're serving companies or businesses that were born digital. And the line of business creates these application-centric teams, and we chase these application-centric teams around our CloudOps Spot story. What's interesting is, between the 2, think of Cloud Insights as a way that allows those customers to normalize these complex, hybrid multi-cloud environments and manage their cloud opportunities and their on-premise opportunities with data.

Okay. Great. And I know like there's a bunch of other vendors out there in storage that offer kind of the data center extension type of approach. But the first-party services do seem pretty unique to NetApp. Can you give us some examples of kind of NetApp primary file storage in the cloud? Like what types of applications or customers using NetApp primary storage in the cloud through 1 of the 3 public cloud partners that you have?

Yes. So just to sort of ground everybody with the product names, we started off building a service with Microsoft, that's called Azure NetApp Files. It is available now in probably more than 35, maybe 36 Azure regions. It's one of the most successful and fastest-growing Azure services. We are often coined by Microsoft, the platform that migrates the unmigratable. And typically, people are bringing us in to migrate very sizable enterprise applications that had file-based dependencies. Probably the most significant and well-known of these is SAP. We are used extensively by SAP and SAP customers to migrate SAP applications from on-premises to the cloud, whether they are Oracle-based SAP applications or increasingly the more modern HANA-based applications. But we're used also extensively to help people with Hybris, SAP's e-commerce application, SAP's BW, the data warehouse application. And so we do a sizable business in helping companies migrate SAP workloads from on-premise to the cloud. But it doesn't stop there. We help people with Oracle, with VMware-based applications. I said we help people in oil and gas, pharmaceuticals, genomics, migrate or extend their sort of high-performance computer applications. Repsol energy, Chevron, organizations like this use us extensively, GlaxoSmithKline, all of these organizations rely on our enterprise-grade infrastructure in the public cloud. Virtual desktop infrastructure is a very sizable market that was sort of dominated by a Citrix or a VMware on-premise, is now, with this pandemic, aggressively moving over to public cloud, where they see the elasticity and the distribution of regions far more advantageous to virtual desktop infrastructures. So we're platforming Citrix, we're platforming Horizon, and we're seeing a huge growth opportunity with Microsoft's entry into the virtual desktop infrastructure market with Azure Virtual Desktop and now with Windows 365. So it's the enterprise applications holistically or in piece, it's applications in particular industries like Epic in health care or The SaaS Institute, Teradata. It's relational database workloads that were so heavily dependent on our technology on-premise, that can now use the same technology in the public cloud and the virtual desktop infrastructures. Those are sort of the very sort of common examples that we see with our primary storage platform that sold, supported and build as a first-party service by Microsoft, Amazon and Google.

Great. Very helpful. And now what's the value proposition? If I'm a customer and I'm looking at doing that migration, like you talked about, obviously, I have options in AWS. I assume that Azure also has a file services or file storage option for primary storage. What's the value proposition for using the -- either Azure NetApp files or the other FSx? Or the VCS -- or CBS, sorry?

Yes, it's a good question. So I would say, I think that we recognize that customers have a choice, and I think that benefits the customers. The public clouds will offer block storage. They'll offer object storage, and in some cases, they offer fairly remedial file storage. NetApp is the gold standard in file storage, and we have uniquely sort of built a set of capabilities that differentiate our file-based protocols, so much so that the application vendors themselves mandate our infrastructure on the public cloud, and they haven't certified other infrastructures. And they do that because I think we have not just a very scalable file system, but we have a rich set of data services that allow us to do things in the context of the application that, frankly, others cannot. We have built, in partnership with SAP, a lot of intellectual property that allows us to do backups and snapshots and clones in the context of the SAP application itself, not just the storage. We were and continue to be the gold standard for relational database workloads, and we have a rich set of capabilities that differentiate ourselves to run in those particular workloads. And we oftentimes compete with disc and we're able to show at-scale sub-millisecond latencies. We're able to prove capabilities. An example of a customer was they were running about 40 different database workloads on premium discs and they switched over to Azure NetApp Files and found that the performance was significantly better. Of the throughput, they found the latencies on premium disc was somewhere between 3 and 7 milliseconds. With Azure NetApp Files, consistently below 1 millisecond. But to give you an idea of the data services. Backups -- database backups and snapshots on the premium disc were sort of holistic in that they couldn't decipher what was on the disc and were taking several hours. The move to Azure NetApp Files and the use of our ONTAP technology brought those backups and snapshots down from several hours to several minutes. We also provide this rich shaping capability so customers can scale up and down the throughput and the capacity on our file system, a unique capability that isn't offered by any other file system on the public cloud, whether it's the rudimentary services offered by the public clouds themselves or third parties in the marketplace. So it's not given to us Jason, we earn it. But we have, as I think many will testify for 30 years, proven ourselves to be very, very unique and very, very successful in serving the file-based protocols to the applications.

And you have like, I guess, a very tight linkage also with the kind of object-based storage from the various partners as well. So you can just do backups to S3 [indiscernible] ...

Correct. Yes. We aren't limited. That's a good point. So we are -- my team, we build customer experiences. And those customer experiences aren't just limited to our infrastructure. We will take advantage of, as you said, object storage where it makes sense. So customers can use our primary storage in the cloud, backup to an object store using lowest-cost storage that doesn't have the same sort of latency or response time requirements. We do intelligent tiering. So we, for customers, will tear off call data automatically from the primary volume, again, constantly optimizing for price and performance. It's that envelope that I think we drive within our sort of product strategy that our customers benefit from. So yes, think of us as sort of part of what we would call a fabric of infrastructure. And we will, of course, take advantage of any public cloud infrastructure we can to better our own experience for our customers.

Great. Let's drill down a little bit into the various -- the first-party services that we talked about. So with Azure NetApp Files, how are you splitting the responsibilities with Azure? What are you doing? What are they doing? And how long did it take to stand up the service?

Yes. So very simply, Azure NetApp Files is a Microsoft product, and Microsoft and only Microsoft sell Azure NetApp Files. NetApp does not to actually sell Azure NetApp Files. Microsoft sells our product. Microsoft supports our product and Microsoft bills for our product. There is no marketplace transaction. If you are a customer of Azure, then we are a service inside the Azure console. We sit behind the Azure APIs and the Azure CLIs. And so customers that have an Azure account are -- can, in a few clicks, activate Azure NetApp Files. Azure NetApp Files has, I think, redefined enterprise-grade storage. And Microsoft has been incredibly successful with that business. Azure NetApp Files has unique performance characteristics at very low latencies. I talked about this cloud-shaping constructs, the capabilities we've given Microsoft for cross-region replication, backup and restore, the sort of the certifications that we've garnered for Microsoft with companies like SAP, VMware and other databases. What we do is we build and we run the service and we, of course, provide second and third line support to Microsoft. So we are the sort of the engine. We code directly into the Azure console. We provide the solution, and we manage the releases, the updates, the patches. Of course, these are all transparent to customers as it's deployed as a fully managed service. So our file system operates much like dial tone. It's always there and customers can consume as much or as little of it as they want. But Microsoft is really the front end. Microsoft handles selling, supporting and billing, and we manage, we build, engineer, run and deploy the services into Azure on a continuous basis.

My understanding, at least as we talk about AWS and what you've launched there in September, my understanding is that it's somewhat different in terms of how it's structured. Though, I think with Azure NetApp Files, at least I think my -- I think it's correct that you actually have built the hardware there? [indiscernible]...

That's correct. Microsoft wanted a very -- Microsoft and NetApp wants to build what we call a scale-up architecture. Public clouds were built in what the industry calls scale out, and we wanted to capture the enterprise applications that had a dependency on scale-up infrastructure. So in Azure, yes, there is NetApp physical infrastructure that forms the basis of Azure NetApp Files that gives us these wonderful performance characteristics at these incredibly low latencies that we engineered together. And I'll tell you, all credit to Microsoft. We challenged Microsoft around networking infrastructure and networking architectures. And Microsoft made conscious decisions to engineer top-of-rack switches and its networking architectures to really give us this unique and compelling experience that we pass on to customers. Now with Amazon, Amazon has a platform called FSx, and it's used today to support a very simple window server and a sort of a scale-out luster product. And ONTAP is now a first-party service within the sort of FSx framework. So while we build and engineer the service, Amazon sells it, Amazon supports it. Amazon builds it. It is like Microsoft in the console, the Amazon console. There isn't a marketplace transaction for FSx for NetApp ONTAP. It's behind the Amazon API and the Amazon CLI, which I'll tell you, much like Microsoft, Jason, look -- being on the inside looking out is quite different to being on the outside looking in. What I mean by that is Amazon and Microsoft embrace us like any of their own services. So with Amazon, you're seeing just recently our integration with Lambda, our integration with SageMaker, our integrations with CloudWatch. We start to get access to all of the other Amazon services bettering our experiences for our customers. Now the unique thing is, because Amazon has the FSx control plane, Amazon actually runs the service. But we deliver and engineer and provide back line support. So some small, subtle differences. Honestly, for the customer experience, they wouldn't really know nor understand these subtle differences. Think of ONTAP as a fully managed service like this sort of dial tone analogy that we build and we engineer in partnership with Microsoft, with Amazon and with Google.

Okay. And that's clear kind of some of the similarities and differences. When you talked about Azure NetApp files, you talked about, a lot of SAP, a lot of kind of, let's call it, I don't want to say legacy workloads, but relational databases, some sort of existing apps. With FSx, would you say it's more targeted towards more cloud-native type stuff? Or similar?

No. I would say both of them, I would say, look, if you think about the value proposition of File, it probably has more historical value, which would probably push it more into the legacy. And I think it's very clear to us that we can capture a sizable part of that legacy as customers extend on-prem or migrate on-prem, whether it's NetApp or non-NetApp. And that value is very much consistent across our offerings on Microsoft, on Amazon and on Google. We're going for that business. What some people call -- I think McKinsey called it the rejuvenation of the enterprise. And we wake up every single day and we go talk to customers that have a file dependency on-premise that expect or demand that same dependency in the public cloud. Now on the other side of that, we now see these sort of cloud-native, born-in--the-cloud applications that are using container-based technologies, using CI/CD as their release process. They're using Kubernetes. File has a very, very significant role to play in Kubernetes. And we have engineered a capability that we call Astra that sits on top of our storage that purposes our storage to the applications that run within a Kubernetes framework that require application-level data protection, that require things like backups, that require capabilities like disaster recovery. And so we've engineered our file system to be highly attractive to these cloud-native applications. And so today, Azure NetApp Files is integrated with the Azure Kubernetes service, FSx for NetApp ONTAP is integrated with the elastic Kubernetes service on top of Amazon. So think of our file system not divided by vendor. We provide the same set of services on Microsoft and on Amazon that attract the legacy customers, and increasingly, the born-in-the-cloud customers.

Got you. And then just real quick on the P&L impact. Is there kind of a noticeable difference in terms of profitability for you just because in the case of Azure, you have the hardware that you sell into that?

No. I mean first and foremost, we don't really detail this. So yes, with ANF, we supply the software and in-line storage, and we're compensated for both on a monthly basis based on the sold capacity. We do deploy systems into Azure clouds, and their depreciation will go against our COGS. But it's a premium service. There's a very, very nice price point, and we, frankly, benefit from a very, very profitable royalty stream. For FSx, it is, and the rest of our portfolio is increasingly software-only. It runs -- in FSx, it runs natively on the Amazon infrastructure, and we're compensated for the software. So there's no depreciation to go for COGS. But as we've stated very clearly, we think the overall blended margin is the most important thing for us to look at and analyze as we provide these sort of rich services that cater for the various different customer types and the various different applications. And we've stated very clearly, I think that our margins are in the cloud business ahead of the blended margins of our traditional on-premise business. And we're very comfortable and confident that our gross margins on cloud and cloud services will continue to increase as we reach scale within Microsoft, Google and Amazon through the sort of the blended hardware and software solutions that we build.

Got you. And I know you don't want to maybe pick between your children here, but do you feel like AWS, just given that it is the bigger, let's call it, the bigger cloud right now. Do you feel like that has more long-term potential? The fact that it is -- Amazon does so much in storage right now, do you feel like that has more long-term potential? Or is it sort of a tough call?

Honestly, Jason, I wrote down my children. I can't wait to tell [ Sather ] and Adam that you refer to them as my children. I sometimes feel like a mouse dancing with elephants. These are formidable companies that have growth trajectories and market impacts, the like of which we've never seen in the technology industry. I mean I was at Oracle for many, many years, and now Amazon is rapidly approaching double the size of Oracle. Amazon started in 2006 and Oracle was founded in 1997. So we work incredibly hard to provide a service to these organizations. We see huge opportunities, honestly, across all 3. I wouldn't highlight any one versus the other. I think Amazon, obviously, really was the first, and Amazon has defined a lot of the sort of the capabilities now that you'll find in other clouds. But each of the clouds, I think, while having a good base continue to differentiate and innovate in certain areas that give customers, increasingly, I think the competence to make multi-cloud an intentional decision going forward as opposed to an unintentional one that it was for so many years. So I would tell you, we sit here and oftentimes, we have to pinch ourselves. When I joined NetApp, Jason, I never imagined and I never really thought that we would be in this position. I never imagined that we would get to really deliver something into these public clouds. Nobody else has done this. I mean honestly, really nobody else has done the level of integration, the first-party services that we have done -- are not on any one cloud, let alone across all 3. It's really, I think, a testament to the capabilities of NetApp, making ONTAP itself software independent of hardware. But the company has leaned in where many others did not. And I think we've proven to these guys that we are a solid partner and that we bring significant value. So all 3 of them, on a daily basis, Jason, brings us tremendous opportunity for customer acquisition and customer growth.

Got you. And is it fair to say that the -- that part of the business, kind of the first party, the Azure NetApp Files, FSx and now, with the Google partnership, those are going -- you think that's going to be the majority of your public cloud business over time?

No, no, actually not. I think we have a strong, I think, concept and construct that NetApp itself and Spot are very strong brands, and these are services that we will sell directly into the customer base. You can kind of think about it -- I don't know whether it's the right analogy or not, but it's like the razor and razor blades, right? I mean what we really wanted to do was make ONTAP a cloud product. And we wanted to get the confidence of the entire world that ONTAP was a platform that was cloud-ready and would be widely adopted by public clouds. And that has given us not just an enormous opportunity to acquire customers through these massive organizations but it's given customers the confidence that we can provide incremental services, that we can build technologies and solutions that complement or enhance our capabilities. So no, I would make it very, very clear to everybody here that while we've established a very strong channel, and we see enormous growth through the big 3 public clouds, channels that are very unique to us, we have an equal opportunity to extend our own brand and our own capabilities directly into the customer base through our marketplace products, through our direct sales engagements and through the existing and growing channel outside of the big 3 public cloud. So no, we don't have plans to be a provider of infrastructure to 3 companies. We want to be a provider of services to tens of thousands of companies.

Okay. Great. And it sounds like it's helped your traditional business as well as just having these capabilities in the cloud.

Well, I'll tell you, Jason, I get brought into sales deals to present NetApp in the cloud, and I present the absolute commitment to our customers that we are in the cloud as good as, if not better sometimes, and they could do on-premise. And customers will look at me and say, "It's so great you're so committed to the cloud. I now have the confidence to make more on-premise purchases." And they do that because they know that at any time their investment in ONTAP can move, that they're not constrained or limited to a fixed decision. And if that symmetry, I think that's given our customers and, frankly, non-NetApp customers the confidence to move forward with us as a company, whether that's in the cloud holistically, whether that's in the hybrid cloud or whether that's on-premise. We strive for that symmetry and that openness that I think we've done in many ways, better than anybody else has done.

Great. Okay. We have about 10 minutes left. We've got 2 really good questions from investors here, so I want to throw those out at you. First is, can you help us size the impact of all sales members having a separate public cloud quota? How significant a driver is this for NetApp going forward versus your cloud partners selling or your stand-alone cloud sales force?

Yes. So we made a conscious decision to bring in Cesar Cernuda, a Microsoft executive, to run our go-to-market organizations. And we felt that we had proven to the company, to the shareholders, to our customers, that we could succeed in a public cloud. And we have done so largely by having 2 independent sales organizations. We wanted to make sure that we presented specialization to the customer, and we give the customer confidence that we were as committed to the cloud as we were to on-premise. And that really helped us establish the beachhead. So now we've been running this combined model for quite a few quarters, and what we really started to see was, we started to feel that there was a position for us higher in the organization and then we would provide specialist organizations like in the sort of the control plane that we serve through the specialists. So we've taken now a sort of a parallel sales organization of infrastructure, on-premise salespeople and cloud people, and we've essentially put the infrastructure teams on top of the specialist teams. Selfishly, what that gives me, Jason, is thousands of people who now wake up with cloud on their quota sheets, with cloud driving their compensation. And so outside of a very small number of companies that probably don't buy cloud, every single one of our sellers, every single of our solution engineers, every single one of our channel partners is now highly incented to sell cloud alongside their existing on-premise infrastructure. And frankly, in a few accounts, in a few market segments, we only sell cloud. when we sell to these unicorns, these organizations that were born in the cloud, we don't lead with an on-premise construct. We don't mention on the on-premise construct that we have very strong cloud tools. So think of us as selling hybrid infrastructures to the enterprises, the global accounts with specialists that can come in and talk competently about any one of our particular motions and the technologies that support them. And then reaching new opportunities in our competitors and in these companies that didn't have on-premise infrastructure, where we take a strong position and lead increasingly with cloud. It's -- if you think about a Dell EMC customer, I have to believe that they, like a NetApp customer, are facing cloud economics, are being challenged to prove innovation and disruption using a public cloud. And they come to us for that because their existing vendors have not done the same thing that we've done. So every single one of our sellers now has a public cloud quota, and we think that's really starting to show up now in our quarterly earnings as we continue to see the public cloud revenues grow. We've put a lot of emphasis on customer success and customer satisfaction, and I hope you're all seeing that in the revenue retention and expansion numbers that we also publish on a quarterly basis.

Great. And actually, that leads into the next question. Another financial question or a financial question, which is, you've had a strong record of net retention, but it seems like existing customer/expansion has driven all of the growth in the October quarter versus the July and April quarters, whereas new customer ARR seems to have slowed down. Do you expect new customer ARR to reaccelerate as Amazon FSx ramps up?

I can't comment on the details of that particular quarter or month because that's not the impression that -- or what the data would suggest internally. We see a very healthy mix of new customer acquisition and existing customers doing more. I'll tell you, in the cloud, selling something to somebody is where the relationship starts. Cloud is very much a pay-as-you-go elastic capability. And so Jason, we're not really selling petabytes upfront. People aren't capitalizing their infrastructure. Customers can come to us and they can onboard a few megabytes or terabytes, and they can test the service and run the service through a variety of different harnesses to get confident with the service. And so we tend to see a sort of a customer onboarding experience that sort of starts with a small amount of data. They run that data through a whole series of processes. And then oftentimes, they back it off. And they're backing it off the environment now because they've done all the necessary benchmarks, and then it flows into their deployment infrastructures and their deployment decisions. People don't just wake up on a Friday and decide to migrate SAP to Azure over the weekend. That's something that takes many, many months. That's almost always done in partnership with a large global systems integrator. And so think of our customer acquisition as sort of being the first in a lengthy relationship that we try and establish with them. But the first transaction that we do with our customer is small. And we try very, very hard to make whatever the customer wants to do or whatever size and scale successful. And we do it, as I said, on an application-by-application basis. And then we naturally come and expand. So maybe what you're seeing is, because we've got so many customers now who have been through the sort of the landing process and are now starting to ramp their expansions, that maybe what you're seeing is the expansions now starting to sort of -- to grow more and more significantly than the lands. But our customer acquisition rates continue to grow as we continue to either develop or acquire technologies that we think are more and more apparent for non-NetApp customers. So I don't think you're seeing new customer ARR slow. You may just be seeing the overshadowing of sizable expansions that are now taking place with very, very significant rollouts of enterprise applications that may overshadow, but certainly don't distract us in any way from new customer acquisition.

Okay. Understood.

And of course, Amazon then -- there was the comment, do you expect new customer ARR to accelerate as Amazon? Of course. Like literally, it's Christmas when Amazon says, "Oh, just -- we'll sell it for you and we'll make it an Amazon product." That gives you a level of the customer -- a level of confidence and trust in us that Amazon, I think Amazon's endorsement is very special, much like Microsoft's endorsement. They have an ability to affect markets that we don't. And so do we see Amazon as an accelerant? Absolutely. I mean Amazon themselves said very publicly at re:Invent that we were one of the fastest-growing and most successful introductions of infrastructure they've ever seen. And while we're not publishing specific numbers, they're Amazon customers, we're incredibly ecstatic by the Amazon relationship. It's much like Microsoft. These are very unique things that we've worked incredibly hard for. The Amazon thing didn't happen overnight. It's something we've been engineering with them for more than 2 years, and I have had to bite my lip on many occasions not to say anything. But September 2 was a remarkable achievement of, I think, a combination of many years of customer success, of proven customer demand, and I think demonstration that we can engineer a service that stands up alongside any other Amazon service.

Okay. Great. Well, we have like 55 seconds left. I wanted to see if you had any closing comment or anecdote on how the perception of NetApp has changed in the market since you joined any kind of customer accords or anything like that, that you want to share with us as we close here, just to kind of present the group here with sort of how the perceptions might be changing of NetApp.

I think in closing, I would tell you the things that we've done well -- I think probably the most significant thing that we did well was George Kurian, our CEO, afforded me the opportunity to challenge the status quo, to reconsider and to reimagine what NetApp might be on a public cloud. And our uniqueness was, as I said at the start, we were able to take a very specific application-centric lens to the problem, and we engineered around the application, which is very different than the way storage was engineered on-premise. And that, I think, is something that we've done incredibly well. We fully embraced the elastic properties of the cloud. We didn't limit customers to large purchases. we didn't challenge them on subscriptions versus consumptions. We made it as easy to consume our product as any of their Amazon service. I think we made a conscious decision to be bigger than storage. We felt that while we had a strong value proposition for the legacy, there was as big, if not a slightly bigger market opportunity in the sort of innovation opportunity the public clouds brought. And we felt that there was a bigger opportunity for NetApp to platform its services more holistically in infrastructure and increasingly in platform-as-a-service. And that's given, I think, us and our customers the confidence that we are not just thinking about how legacy transforms to the public cloud, but how innovation itself manifests itself in a public cloud. And so I would tell you, as the person that was given this opportunity, we find ourselves in a very unique position. I think our opportunity is significant. I'm often asked what do I think of HP and Dell EMC and Pure and Nutanix. And I honestly can't answer, Jason, because I never see them. I never ever, ever see them in the public cloud. My competition, Jason, honestly, is block storage and object storage. And those are sort of the considerations that customers are making. And much like we did on-premise, we go out every day and prove the benefits of ONTAP and our file services and a rich set of capabilities that surround them. And we're finding, if you look at IDC numbers, we're seeing file storage now really grow very, very quickly in the public cloud. It's about 12% to 13% of the total storage share but growing 2x to 3x faster than block and object. And we think that NetApp can arguably take a very sizable share, much bigger share in the cloud than frankly, we may have on-premise. So think of us as sort of kind of all in. We see the cloud as a primary opportunity, not as a secondary opportunity.

Thank you very much, Anthony. This was really helpful and enlightening, and thanks, everybody, for joining us. We'll have to leave it there, and have a great day.

Thanks very much, Jason.

Bye-bye.