NTAP - NASDAQ

Good day, and welcome to the NetApp Third Quarter of Fiscal Year 2026 Earnings Call. [Operator Instructions] Please note, this event is being recorded. I would now like to turn the conference over to Kris Newton, Vice President, Investor Relations. Please go ahead, ma'am.

Hi, everyone. Thanks for joining us. With me today are our CEO, George Kurian; and CFO, Wissam Jabre. This call is being webcast live and will be available for replay on our website at netapp.com. During today's call, we will make forward-looking statements and projections with respect to our financial outlook and future prospects, including, without limitation, our guidance for the fourth quarter and fiscal year 2026, our expectations regarding future revenue, profitability and shareholder returns and other growth initiatives and strategies. These statements are subject to various risks and uncertainties, which may cause our actual results to differ materially. For more information, please refer to the documents we file from time to time with the SEC and on our website, including our most recent Form 10-K and Form 10-Q. We disclaim any obligation to update our forward-looking statements and projections. During the call, all financial measures presented will be non-GAAP unless otherwise indicated. Reconciliations of GAAP to non-GAAP measures are available on our website. I'll now turn the call over to George.

Thank you, Kris. Good afternoon, everyone. Thanks for joining us today. We delivered another strong quarter with Q3 revenue of $1.71 billion, an increase of 4% year-over-year. Excluding the divested Spot business, total revenue was up 6%. Our accelerating growth, coupled with continued operational discipline, has enabled us to drive profitability metrics higher. Operating income and EPS achieved record highs. We are in a strong position to deliver sustained growth and are on track to deliver our strongest year yet. I'm proud to share a marquee moment that underscores our pivotal role as the intelligent backbone for modern data-driven innovation. The Super Bowl is more than the biggest sports event of the year. It is a global showcase of innovation and partnership. During Super Bowl LX, our technology transformed Levi's Stadium into an interactive data center. We managed billions of data points, powering everything from video boards to real-time inventory and security operations. In this most demanding no-fail environment, we demonstrated our ability to deliver flawlessly. In the AI era, organizations face security threats, fragmented architectures, shortage of expertise and operational complexity, which make it difficult to unify and harness data for its full potential. We help enterprises solve these pressing data challenges by delivering a data platform that is optimized, secured and AI ready. Customers rely on NetApp technologies to be the data foundation to support AI innovation, modernize data infrastructure, strengthen cyber resilience and transform cloud strategies. In Q3, approximately 300 customers selected NetApp to help prepare their data for AI and to be the storage foundation for their AI innovations. Last October, we announced major enhancements to our enterprise-grade data platform for AI workloads. These new solutions, AFX and AI data engines are generating significant customer interest and engagement. AFX is our disaggregated storage system purpose-built for AI that gives customers the benefit of enterprise-grade security and capabilities coupled with extreme performance and scale. We are excited to report strong early momentum with AFX, in its first quarter of shipment. We have secured significant AFX wins across key industries, including neocloud, financial services, and semiconductor. An example of an early AFX win is for model training and fine-tuning at a neocloud. AFX stood out among competitive disaggregated architectures for its multi-tenant management, container integration, cyber resilience and replication capabilities. One of the biggest challenges in AI is data. The AI data engine helps to improve time to value in AI projects by simplifying workflows with integrated data discovery, curation, policy-driven guardrails and real-time vectorization for Gen AI. By understanding where their data is, customers can ramp AI projects faster, boost result accuracy and slash time to insight. Our early access program has been highly successful, engaging customers from key industries such as semiconductor, media and entertainment, financial services and IT services. AIDE will be generally available in Q4. NetApp helps customers modernize their environments with a unified, adaptive data foundation that extends across on-premises and cloud, delivering high performance, availability and integrated security. As data center demands grow, customers are leveraging all-flash arrays to achieve density and power requirements. Strong customer engagement and interest in our unified and block-optimized all-flash storage portfolio delivered another record all-flash array revenue quarter, growing 11% year-over-year to $1 billion in Q3 for an annualized run rate of $4.2 billion. The importance of our robust cyber resilience capabilities cannot be overstated as customers look to safeguard their most valuable asset, data. The NetApp data platform delivers comprehensive ransomware protection, backup, disaster recovery and data governance in a single, secure foundation that reduces the time to detect anomalies, recover data and get our customers back to business. Embedded protections and our Ransomware Recovery Guarantee help customers confidently withstand today's sophisticated threats and prepare for tomorrow's challenges. These capabilities foster trust in NetApp in an increasingly volatile digital landscape, enabling us to win new customers and displace competitors. An example of this is a European financial services company needing to refresh its entire data center to enable scalability and compliance with current regulations. In Q3, this customer selected NetApp all-flash systems to replace multiple competitors. Critical to the win were our anti-ransomware services, data classification, and Write Once, Read Many snapshots, delivering business continuity, robust data protection and regulatory compliance, while positioning the customer to manage future growth. Keystone, our storage as a service offering, continues to perform well, as customers navigate infrastructure transitions, cloud migrations and rising memory costs. Keystone revenue grew approximately 65% from Q3 a year ago. In Q3, an insurance technology company planning a multiyear migration to the cloud selected NetApp Keystone as the storage solution to enable this transition. This new-to-NetApp customer selected Keystone for a fast and efficient way to eliminate a competitor's storage as a service footprint that did not offer a true path to the cloud. As that example demonstrates, our first-party relationships with the hyperscale cloud providers is a real differentiator for us. Adjusted for the Spot divestiture, our Public Cloud Services revenue grew 17% year-over-year, driven by first-party and marketplace services, which grew 27%. These services are a powerful driver for new customer acquisition. About half of the revenue driven by new first-party and marketplace customers in Q3 came from new to NetApp customers, highlighting the role cloud plays in expanding our customer base. Let me share a couple of examples of how our cloud services are displacing competitors. A multinational insurance company, aiming to overcome the complexity of its legacy infrastructure and improve agility, selected Azure NetApp Files for its proven performance, ease of use and enterprise-grade reliability. ANF is now the cornerstone of their cloud transformation. Similarly, a retailer, after experiencing a ransomware attack, moved off a competitor's infrastructure to the cloud. They chose AWS FSx for NetApp ONTAP for its support of immutable volume copies, providing data protection against cyberattacks. FSxN is now their default storage service in AWS. As a leading enterprise storage solution provider and the only one with first-party data storage services native to the public cloud, NetApp is uniquely equipped to help customers easily connect their data with the leading cloud-based AI applications and accelerate modern workloads like AI in the cloud. In Q3, we introduced a new capability, enabling Amazon S3 Access Points for Amazon FSx for NetApp ONTAP. This allows enterprises to make their workflow simpler and more efficient by connecting the many AWS AI and Analytics Services directly with their NetApp data, both in the cloud and on-premises. Also in Q3, we announced the public preview of Object REST API on Azure NetApp Files, enabling seamless, real-time integration between an organization's data and Azure's advanced analytics and AI services. With direct and secure access to enterprise data, companies can extract actionable insights and make data-driven decisions faster, giving them a competitive edge in an increasingly data-driven world. Already, these connections are being used by customers. In Q3, a multinational manufacturing company selected FSxN as the high-performance data layer for its AI workloads on AWS. The customer is leveraging our recently introduced S3 support to bring AI to its large, existing file-based datasets without having to duplicate or re-platform its data. Before I wrap up, I'd like to address how we are managing through the unprecedented inflation in memory prices currently affecting the global market. First, we have raised our pricing and will do so again, as needed. Second, we are working with our customers and channel partners to be more agile in this dynamic environment. Third, we are working with our multiple suppliers to address availability and manage costs, as we have successfully done in the past. And finally, unlike our all-flash only competitors, we have a broad portfolio that includes hybrid flash arrays, giving us the opportunity to better service price-sensitive workloads. In summary, solid execution and operational discipline delivered another strong quarter. Customers are choosing NetApp for our unified data platform that delivers exceptional value and operational efficiencies, solidifying our position as the intelligent data backbone for the AI era. As I look to the future, I am confident in the opportunity ahead and in our ability to successfully execute on our strategic plan. We will continue to invest in key areas that drive growth and provide long-term value for our shareholders. I'll now hand it over to Wissam.

Thanks, George, and good afternoon, everyone. As George mentioned, in the fiscal third quarter, we delivered strong results, exceeding both the midpoint of the revenue guidance range and the high-end of the EPS guidance range. Total revenue for the quarter was $1.71 billion, up 4% year-over-year. Non-GAAP earnings per share was $2.12, up 11% year-over-year. Excluding the divested Spot business, which generated $25 million of revenue in the fiscal third quarter of the prior year, total revenue was up 6% year-over-year. The effect of foreign currency exchange rates was favorable to revenue growth by approximately 2 percentage points year-over-year, while it was immaterial relative to guidance. Looking at revenue by segment, Hybrid Cloud revenue of $1.54 billion was up 5% year-over-year, driven by Product, Support and Keystone. Keystone continues to build momentum, with revenue growth of approximately 65% year-over-year. Public Cloud revenue of $174 million was in line with last year's third quarter revenue. Excluding Spot, Public Cloud revenue was up 17% year-over-year, driven by strong demand for first-party and marketplace storage services. At the end of the quarter, our deferred revenue balance was $4.63 billion, up 12% year-over-year and 9% year-over-year in constant currency. Remaining Performance Obligations were $5.11 billion, growing 14% year-over-year. Unbilled RPO, a key indicator of future Keystone revenue, was $482 million, up 38% year-over-year. Moving to the rest of the income statement. Please note my comments will be related to non-GAAP results unless stated otherwise. Gross margin for the fiscal third quarter was 71.2%, up 50 basis points year-over-year, driven by Public Cloud gross margin expansion. Gross profit was $1.22 billion, up 5% compared to Q3 2025. Hybrid Cloud gross margin was 69.6%, down 1.8 percentage points sequentially as product gross margin declined by 4.2 percentage points to 55.3%. This was primarily due to an unfavorable revenue mix and, to a lesser extent, the need to make market purchases to be unexpectedly higher demand for certain products. Our Support business continues to be highly profitable at 92.5%. Professional Services gross margin was 31.3%, improving 100 basis points sequentially driven by higher Keystone revenue mix. Public Cloud gross margin was 85.1%, up approximately 2 percentage points sequentially and approximately 9 percentage points year-over-year. Operating expenses of $686 million were down 3% sequentially. Operating expenses were up 3% year-over-year in part due to the unfavorable effect of foreign currency exchange rates. Operating income was $533 million, up 8% compared to Q3 2025. Operating margin was 31.1%, up 1.1 percentage points year-over-year. Earnings per share was $2.12, growing 11% year-over-year, exceeding the high end of our guidance range. Our results demonstrate strong execution on key revenue growth opportunities in all-flash, Public Cloud and AI, along with a continued focus on operational discipline, resulting in record highs in both quarterly operating income and EPS. Cash flow from operations was $317 million, and free cash flow generation was $271 million. During the quarter, we returned $303 million of capital to our shareholders, with $200 million in share repurchases and $103 million paid in dividends of $0.52 per share. The Q3 diluted share count of 200 million decreased by 8 million shares, or 4% year-over-year. Cash and short-term investments were $3 billion and gross debt outstanding was $2.5 billion, resulting in a net cash position of $522 million. I'll now turn to non-GAAP guidance, starting with Q4. We expect revenue of $1.87 billion, plus or minus $75 million. At the midpoint, this implies a growth of 8% year-over-year. Excluding the divested Spot business from the year ago comparison, our revenue guidance implies a 9% growth. We expect Q4 gross margin to be between 69.5% and 70.5%. Operating margin is anticipated to be in the range of 30.5% to 31.5%. We expect EPS to be between $2.21 and $2.31. Turning to full year 2026. We now expect fiscal year 2026 revenue to be between $6.772 billion and $6.922 billion, which at the $6.847 billion midpoint, reflects 4% growth year-over-year. Excluding Spot, our revenue guidance implies a growth of 5% year-over-year. We expect gross margin to be in the range of 70.7% to 71.7%, and operating margin to be in the range of 29.3% and to 30.3%. Other income and expenses are anticipated to result in approximately a $24 million net expense. For the year, the tax rate is expected to be in the range of 20.2% to 21.2%. EPS is expected to be in the range of $7.92 to $8.02. In closing, as we look ahead to the rest of the fiscal year, we remain committed to our strategic vision and are confident in our ability to navigate this dynamic environment. Our focus on revenue growth and disciplined execution is yielding positive results and record profitability. We are dedicated to delivering exceptional long-term value to our customers and shareholders. I'll now turn the call over to Kris for Q&A.

Thanks, Wissam. Operator, let's begin the Q&A.

[Operator Instructions] We'll go first to Param Singh from Oppenheimer.

Really appreciate if you could give some more color on what's driving this incremental growth on the product side. What are you seeing in terms of adoption? And how persistent is that?

We have, as we guided from the start of the year, been working on large deals that for many quarters, some of those deals closed in Q3 and a number of them are expected to close in Q4. And so we're excited about the momentum in our business and the ability to close these large deals.

Got it. And maybe one for Wissam. Look, the product gross margin, obviously, is going to be a little bit difficult to manage in this pricing environment. Help us think through what can you pass through to customers? What's your supply chain agreements? And can you get enough componentry, not only on the DRAM and NAND side, but also on the HDDs.

Yes, sure. Thanks for the question. So look, as we've said all along, we are operating in a dynamic environment. Commodity prices are increasing at a fast pace. However, we do have agility and the ability also to manage the situation through either adjusting prices, working with our customers and partners to be more agile. We also have our products qualified at several suppliers that we work very closely with to understand what the supply is doing and negotiate prices in advance where we can. And also, we have the ability to offer alternatives for some of our customers for price-sensitive workloads. We have the ability to offer our hybrid flash array or for some customers who have much more interest in the consumption model, we can go for our Keystone, which is storage as a service or our Public Cloud business. The -- with respect to maybe the second part of your question, you mentioned also hard drives. We are seeing some price increases there, but it's nothing compared to the rest of the components we're seeing on the -- for instance, on the NAND. Memory, obviously, has been in the news for quite some time and other types of semiconductor supply chain.

Katherine Murphy from Goldman Sachs has the next question.

You talked about 300 AI deals signed in the quarter, up from 200 in the previous quarter. And I was wondering if there's anything you could share here about the mix across use cases and customer types, if there's any areas you're seeing particular traction. And just as a follow-up, anything you could share to help us think about the potential contributions of AFX and AIDE as we think about 2027 and what the potential margin profile would look like there?

We have seen strong momentum in our AI business across multiple industries, public sector, manufacturing, health care and life sciences and financial services in particular as well as early signs of adoption of AI in semiconductors, for example. In each of those cases, our hybrid architecture, our ability to deliver performance at scale with a rich set of data management capabilities has us standing out. With the introduction of the AFX, we are also seeing momentum in the neoclouds for use cases such as model training and fine-tuning. In terms of the outlook for AFX and AIDE. AFX is a new architecture, so it requires qualification by customers. It comes with a proven software set, but still, customers will test and adopt it. We were ahead of our expectations in the first quarter of availability, but we are also not predicting a ramp like you saw, for example, with the C-Series. AIDE provides differentiation not only for new environments, but also for brownfield environments where customers can expand the value from their existing investments with us. And we'll tell you more on the outlook of these products as we head into the next fiscal year.

The next question is from Samik Chatterjee, JPMorgan.

George, maybe if I can start with your price increases and if you can share what are the magnitude of the price increases you've taken? And how are you seeing sort of customers respond to it? I assume if prices are going up, customer budgets don't flex up to the same extent? Or what are you seeing sort of as typical customer response to that? And I have a quick follow-up.

Yes. We raised prices at the start of this quarter, and those increases were roughly in line with what you saw in the market. They varied, of course, by type of product. With regard to customers, as we have said for many, many, many years, customers budget in dollars, not in systems, and those dollars are tied to their IT spending priorities. We offer them a range of options as Wissam mentioned earlier, hybrid flash, all-flash, Keystone consumption, cloud offerings as well as a huge range of features to optimize the use of flash in their existing systems to give them more value. And so those conversations are ongoing, just like they've been every quarter for many years.

Got it. Got it. And maybe for Wissam, if you can walk us through the gross margin drivers at the company level between sort of 3Q and 4Q? And is it really the moderation just driven by product gross margin? And I think on the last call, you had outlined that you still believe that when it comes to fiscal '27 gross margins can still be sort of basically flat over fiscal '26, even with memory prices going up. Is that still your expectation because memory prices have changed quite a bit over the last sort of 90 days?

So let me start with -- there's a few parts to the question. Let me start with Q3. Look, Q3 to Q4, really, the dynamics moving from quarters is very much driven by the revenue mix, the component of the revenue mix. When we look at Q3 versus Q4 revenue, we're seeing growth across the board. And so just the revenue mix is pretty much what's driving these components and what's driving the ultimate 70% midpoint, let's say, on the guidance. When I look at fiscal '27, one, I think it's too early for us to really guide with '27. And I don't recall necessarily saying that we're expecting to be in line as much as to say that we would be very actively managing our business, like I mentioned earlier, with respect to working with our suppliers to track the plan as much in advance as possible and where there is opportunities for us to lock in prices to do that, but also working through our -- with our customers to maintain that agility and help them find the right solutions for them in an environment that's dynamic; two also adjusting prices if needed to make sure we continue to protect our profitability. Look, as we think through our product margin, I would say really a couple of things. One is our long-term product gross margin target is still that sort of mid-50s to high 50%. And then the second thing I would say is we -- from a company perspective, we really are focused on the total company margin. But more importantly, we're also focused on the gross profit because we believe that gross profit for the company is really what drives our earnings growth.

The next question is from Aaron Rakers, Wells Fargo.

Just to kind of build off that last question a little bit. Wissam, maybe it would be helpful to appreciate what are you seeing from a pricing perspective, particularly on NAND flash? What's kind of embedded in your gross margin assumptions today? And then talk a little bit about the duration or how the duration of your supply commitments have changed. I think in the last couple of quarters, you kind of talked about having good visibility through the fiscal year. But how has that evolved to kind of think about next fiscal year and the inputs into that?

So look, I mean we talked about also when you look at Q3, for instance, sometimes we have mix that is not as predictable. We did have to buy a little bit in the open market or basically replenish some inventories in Q3 because we had some unexpected high demand on certain products. But as we look for fiscal '26, the dynamics haven't changed from what we discussed before. We're still mostly covered from our prebuys that we did earlier in the fiscal year. For Q4, we may have some demand where we're replenishing inventory. I'd rather not talk much about fiscal '27, given that we're still focused on really closing the fiscal '26 in a strong way. And there's a lot of -- I mean, as you know, this is a very dynamic environment. I wouldn't want to sort of make a statement on duration of cycle or any comments on the price, that's not necessarily my area of expertise.

Maybe I can just comment on the price increases. We did raise at the start of the quarter. As we have said many times, it takes a little while for those to be actually actionable at customers because we want to give them some amount of time to plan their purchases. And so we typically give customers a period of time like 90 days to 120 days, to manage their purchasing agreements with us. So we raised prices at the start of this quarter, and we'll give you an update when we report the quarter.

Yes. Very helpful. And as a quick follow-up, in this environment, I'm curious, you kind of alluded to, I think, the fourth point you highlighted in your prepared comments, you do have a hybrid product portfolio. Are you seeing customers actively maybe move away from deciding to go all-flash back to hybrid, maybe Keystone versus CapEx-centric purchases? Any change in behavior that you're seeing across customers in that regard?

It's too early to draw a trend. We are certainly seeing more interest from customers around some of those topics that you mentioned, hybrid flash, Keystone and other alternatives.

Wamsi Mohan from Bank of America has the next question.

Maybe, George, you can just help us think about some of the purchasing behavior trends you're seeing in storage if there's any concern around supply availability, which is causing people to perhaps prebuy, perhaps prebuy ahead of some of the price increases that the industry knows is coming. It'd be helpful to get sort of what you're picking up from your customer conversations. And I have a quick follow-up.

Yes. Thanks for the question, Wamsi. While we can certainly understand the behavior that you were describing, our Q3 results were not a result of pull-ins. And importantly, our Q4 guidance does not rely on pull-ins. As we have said from the start of this fiscal year, we expected increasing momentum through the second half of the year for a few different reasons. First, as we said, Europe started to get better through the second quarter. And as you saw in Q3, we have increased momentum in Europe. We said that U.S. public sector would be less of a headwind in the second half of the year, and while it hasn't fully recovered, it was -- met our softer expectations in Q3. And we said that we were working on large deals that would happen in the second half of the year at the start of our fiscal. And we have been working on those deals. And we have seen some of them come through in Q3, and we expect more to come through in Q4. Finally, with regard to what we see broadly in the market, IT spending has always been tied to customers' business outlook. And what we see today is that business outlook is pretty favorable. It is quite similar to what we saw last quarter, but in certain markets like Europe, things are picking up. And so IT spending, you can see in the public reports is expected to be reasonably durable this coming 12 months. Customers then prioritize business projects and associated with those business projects, infrastructure projects, like we said, cyber resilience, cloud transformation, data center infrastructure upgrades, as well as AI projects, and we are well positioned to capture our share of those markets. And we are seeing that reflected in the mix of our business, high performance flash to support AI workloads, growing number of AI use cases and of course, growth in cloud. We'll tell you more. It's a dynamic environment. We'll tell you more when we report next quarter.

Okay. And a quick follow-up. How do you feel about the supply availability on SSDs as you go into next year? There's been a lot of talk about memory companies coming back to renegotiate pricing at a faster rate, not necessarily honoring all the LTAs that were put in place prior. Obviously, you guys have a lot of scale. So just wondering what you're seeing on that end from a component both availability and negotiation standpoint of if that's -- the window of that is shrinking from maybe a quarterly negotiation to a monthly negotiation? And any color on LTAs, too.

Yes. I think we work very closely every year and every quarter with a broad base of suppliers for virtually every component in our systems. We are not experiencing any supply shortages at this time and are not aware of any that is upcoming. Of course, it's a dynamic environment. And so we are staying super close to our suppliers, qualifying multiple different components so that we have, first, access to supply so that we can meet customer demand; and second, competitive price points for those supply. So these are dynamic. It's not -- we don't see any specific trend to comment, Wamsi, other than, listen, we're engaged like we always have been with our suppliers.

The next question is from Erik Woodring, Morgan Stanley.

George, not to belabor the point here on memory inflation, but if you take some of the public views on future memory price hikes, we could be in store -- the industry could be in store for multiple price hikes this year. I would just love to know how you approach balancing protecting margins and keeping those product gross margins within the mid- to high 50% range, while also limiting the risk of demand destruction or even market share losses? I know historically, storage has an elasticity of less than 1, but we're in pretty unprecedented time. So just curious how you're thinking about balancing that in this environment.

Yes. Listen, I think that we anticipate pricing and the sort of the tight supply environment to continue for a period of time. And so we are not expecting this to be a short-term issue. I think, as we said, the first thing is to make sure we have adequate supply across multiple suppliers to both assure that we can meet customer demand as well as so that we can have competitive positions in the market from a pricing standpoint. With regard to our operating model, listen, it's -- every quarter has puts and takes. There are deals that we want to be competitive on and there are deals where we say, listen, we probably don't want to be a part of that specific transaction. And I think those continue to date. I think with regard to our kind of operating model, as we said, as Wissam mentioned, our core focus is to drive gross profit dollars because that, in turn, drives the earnings per share model of the company. I think, of course, within that, we look at matching supplier cost to us with our pricing to customers while respecting that it may not happen immediately because we want to give customers some time to adjust. And so that's kind of how we are operating. I think that -- with regard to demand, I think as we have said many, many, many times, customers budget in dollars and they budget against business priorities. It is our responsibility to be in the spending priority stream of our customers and to give them the best value offering for the budget dollars they have to spend, and so we bring a lot of different options. We bring competitive storage-efficient flash, we bring hybrid flash solutions. We'll bring cloud solutions. We'll bring Keystone and we'll bring the full portfolio. Just like we do every quarter, certainly now it's an elevated environment, and so we'll do even more of that.

Okay. Understood. And then just a quick follow-up maybe for you, Wissam, and maybe it's a clarification. But I think when we look -- when you were talking about fiscal 3Q product gross margins, you were talking about kind of unfavorable mix as the reason why margins were down sequentially. All-flash was up 11%, that was stronger than your overall product growth. So can you just maybe help us better understand why mix was the headwind to product gross margins when that part of your business, when all-flash was outperforming. We just want to make sure I understand that dynamic.

Yes, of course, Erik. So when we talk about revenue mix, it could be the multitude of things or a combination thereof. Revenue mix is driven by geos, by customer type, by customer, by product. And so there's multitude of things that contributed to that. I wouldn't want to get in more details than this.

Krish Sankar from TD Cowen is up next.

I have two of them. George, first one, I'm just kind of curious, where do you think we are in enterprise AI storage adoption cycle. It looks like last year, they were still going through pilot and data preparedness. Do you think we enter production this year? Or do you think that's still TBD? And then I have a follow-up.

I think, as we said, our AI business has grown in terms of AI customer wins. I think this quarter, a year ago, it was around 100-plus wins, it's now close to 300. So we're seeing acceleration in our AI business. Within the AI business itself, there are industries and use cases that are certainly more advanced and more repeatable in customers, and there are others that are further behind. So for example, in regulated industries where their data is well organized, you are seeing customers put stuff into production or if you want to call it a pilot, it's a very large-scale pilot. Health care, life sciences, some parts of public sector, manufacturing, there are lots of use cases there that are starting to progress well beyond what's the pilot. Within our mix of business this quarter, roughly 60% were still in the data prep, data readiness, data lake models and 40% were in production, training or production inferencing use cases.

Got it. Got it. And then just a quick follow-up. In the past, like your visibility on lead times for storage used to be a few months. Has that changed now with the memory dynamics, i.e., commodity costs going up, you're raising prices, has that really changed the visibility time line?

Listen, it's certainly a dynamic environment. We are staying on top of lead times. I think that it would be inappropriate of me to say that lead times are uniformly extending. We're not experiencing that. There are, of course, specific components that might be on one configuration that are causing a longer lead time. We're not experiencing anything across the board and so far. And we are working hard to bring alternate silicon into those configurations so that we can meet reasonable lead times. As you know, in the prior supply chain situation that we experienced a few years ago, we were able to manage through that without having really long extended lead times because we use merchant silicon that's off the shelf. And we have multiple suppliers who we qualify for a particular piece of silicon. That being said, it's dynamic, and we're staying on top of it.

The next question is from Steven Fox, Fox Advisors.

I guess I was just curious if you can provide more color going forward as NAND prices go higher, your assumptions on sort of hybrid arrays selling more? Are you seeing -- can you maybe talk about how much replacement demand you're seeing from customers that were looking at maybe lower-end AFAs now are going into hybrids replacement into other consigned business, et cetera. How does that mix look going forward?

I think it's a great question. It is a topic of ongoing discussion at our customers. As we said earlier, we raised prices at the start of this quarter. And so that has triggered some of the discussions with customers about what's the optimal architecture, and we are seeing more discussions of HFA. It's too early for me to say it's a trend, but there are certainly a lot more discussions once we raise prices, and those price increases were more on AFAs rather than HFAs.

And I know you said that, obviously, HDD prices aren't going up nearly as much as NAND prices. But can you just give us a sense for ability to procure HDDs at this point going forward?

Yes. Listen, I think that we are well aware of the constraints in the market. So far, on HDDs, we feel good about our ability to procure what we need even with an elevated demand picture potentially. I think that specifically, we use HDDs that are dissimilar to the ones that are in short supply. We are not buying the highest capacity HDDs, which are the ones that are more constrained. I hope that gives you enough color.

Tim Long from Barclays has the next question.

Let's see if I can get a prize for asking 2 without one of them being on memory. So maybe first one here. Can you just touch on the Public Cloud business, maybe 2-parter on the Public Cloud business. Excluding Spot, I think we've been running in this high teens range for the last several quarters here. Talk to us a little bit about going forward how you think we could potentially break out of that range? Are there any new offerings or new customer sets or anything that you see in the pipeline that could maybe accelerate the growth in that piece of the business? And then related to that, obviously, the gross margin in the Public Cloud was really high, I think, kind of above ranges. So Wissam, if you could talk a little bit about profitability of that business and how much more room is there? Or are we kind of tapped out on the margin side for Public Cloud?

I'll take the first one. Thank you for the question. With regard to continued growth in Public Cloud, it's sort of threefold. One is to take the new customers who we continue to add at a good clip, get them to adopt more and more of our portfolio. We call that cohort management, and it's really expanding within our growing customer set to use more of our technology. That includes offering different price points for storage solutions, adding new capabilities and so on. The second is to connect into the AI growth rates of the hyperscalers. I think you saw us bring innovations to the market like S3 access points for AWS that allows customers to use their AWS' broad suite of tools alongside our storage solutions. This gives us the ability to expand into new wallet opportunities and customers. And then the third is to scale go-to-market. I think we have done a good job in certain countries. We need to leverage more indirect routes to market as well as broaden the number of places we are growing. And you will see us continue to look at ways to invest to grow the business. It is a highly profitable business, and it's extraordinarily sticky. And so I'll turn it over to Wissam for the second part.

Yes. And Tim, for the second part, we did provide the long-term range being 80% to 85% from a profitability perspective on the gross margin line for the Public Cloud business. Obviously, we are at the high end of that range in Q3, and we're comfortable operating at this level for now. I wouldn't say we're tapped out, but it's also too early for us to sort of move from that range. The nice thing about the Public Cloud business, when you sort of put it also in the context of the overall company, if we go back to the comments around the margin for the total company and the gross profit for the total company, it is an accretive business and it's growing at a faster pace than the rest of our revenue streams. So all in all, it helps us on the margin line.

Next question comes from Jason Ader, William Blair.

Just firstly, on the Q4 gross margin guidance. It just seems to imply -- if we keep everything else kind of equal, it seems to imply that the product gross margin will be sort of flattish sequentially, and I wanted to just confirm that, that's the right way to think about it, Wissam.

So Jason, without getting into a lot of the details on line by line, as I said earlier, when you look at the Q4 guidance for gross margin, really, the dynamics quarter-to-quarter are driven by very much the overall components of the revenue, basically, the revenue mix.

Right. But I mean like the other parts of the business, the cost of services has been pretty much 84-ish percent, right? So that was -- I'm just trying to do the math here based on what your guidance was. I just wanted to make sure that there's nothing on the maybe cost of services side that should be called out that would vary from where you've been over the last couple of quarters?

I mean, to your point, typically, there's sometimes minor fluctuations quarter-to-quarter on these lines. And so I really would rather not get into the line-by-line guidance because we typically guide to the total company. But yes, when you look at sort of the various components, there's fluctuations -- minor fluctuations quarter-to-quarter.

Okay. Great. And then for George. George, as you kind of look at the just dramatic improvements, especially on the software engineering side that AI is bringing to bear, how are you thinking about the development organization at NetApp? Are you continuing to hire? Is this an area where you think there could be some pretty substantial efficiencies over the next couple of years at NetApp? Just what's your overall philosophy, strategy at this moment? You may have seen there's some chatter today, the company, Square is like laying off like half its workforce. So I know that's maybe a very extreme example, but -- and they're doing great. So it's just -- it's basically about AI and efficiency. Maybe just speak to that larger topic.

Yes. I think, listen, we have been prudent stewards of expenditure. Our operating margins are north of 30%, right? And so we will continue to look for efficiencies. With regard to using AI to develop software, we already do so. And our first priority is to, at this moment of weakness for many competitors, to accelerate the amount of innovation we put into market. And so while we'll always look to be prudent and optimize spending to the right parts of the business, we also see that it's important for us to continue to bring innovation so that we can capitalize on some of our weaker competitors. And you'll see us provide more instructive direction on that as we go through the next fiscal year.

We'll take the next question from David Vogt, UBS.

So George, I jumped on late, I might have missed this. Have you seen a recovery in sort of the federal category that's been sort of an albatross in the industry for a while. And the reason why I'm asking is you said you saw some orders that you've been working on for a couple of quarters close. And what is kind of your expectation or that particular vertical as we move forward given sort of all the noise, particularly out of D.C. as of late. And then I have a follow-up for Wissam.

Yes. At the end of last quarter, we said that our outlook for the second half of the year was cautiously optimistic. We said that we saw people starting to come back to work in our Q3, but it was too early for us to see broad-based funding into programs translating into orders. We met our cautious expectations in Q3. Our expectations for Q4 is that it improves from Q3. And then we'll tell you more about next year when we get to next year. I think that Q2 was, of course, because of the shutdown, a severe impairment. Our expectation is that Q3 and Q4 were better than that. It's too early to comment whether it's robust yet.

Got it. Okay. And for Wissam, I know I might have missed this again. I didn't catch in either the prepared remarks or any responses to a question regarding your purchase commitments. I know you talked about you're covered based on the inventory and kind of the commitments that you've made. Can you kind of share with us sort of the magnitude of the purchase commitments that you have on balance and -- so we can kind of get a sense of how to think about that playing out over the next couple of quarters? And when we think about that also, what is sort of the working capital commitment or obviously, your program or your strategy to mitigate sort of the higher component costs that are flowing through the market right now?

Yes. So we didn't talk much about the purchase commitments. But I would say for fiscal '26, there isn't much of a change from what we talked about last quarter, David. I mean we did replenish some inventory in Q3, we're probably doing some in Q4 just to make sure that we have what we need from a mix perspective to deliver the revenue. It's too early to talk about fiscal '27. Obviously, our supply chain team works with all of our suppliers on a regular basis, as George mentioned earlier, and we continue to work to -- with -- the top priority for us is to ensure supply availability and then, of course, to be able to negotiate whatever prices -- better prices or best prices that we can. Relative to working capital, look, we have a very strong balance sheet and there is no concern there if we need to take action to secure supply or protect or better pricing or do a combination of the two, then we have plenty of flexibility. And also sometimes we work with third-party logistics companies that could help with that. So there's -- from my perspective, there's no concern on working capital fluctuations.

The next question is Simon Leopold, Raymond James.

I wanted to see if you could talk a little bit about the competitive environment, specifically in light of price adjustments. And what I'm trying to understand is, is every participant, every vendor raising price by the same amount at the same time? How much variation is there? And how should we think about this dynamic?

Thanks for your question. It's always been a competitive industry, and our -- most of the players in the market are rational. With regard to the price at the customer versus the list price, different vendors take different approaches. Some of them raise list prices and have the same discount level. Others may not raise list prices as much, but restrict discounting. The net of it is everybody is roughly in the same boat in terms of the cost structure of their commodity supply. And so it really becomes around the software value you bring, the range of offerings that you can present to customers and the discipline at which you run your business. And so we don't see much deviation from that. I think one of the places that we have are sort of working on is to position the right offering for the right use case, particularly HFAs are starting to see a lot of interest from customers. And it's too early to comment whether that's a trend, but there's certainly a lot more discussion about that going on.

The next question is from Alek Valero, Loop Capital.

Alek on for Ananda. So my question is, how do you see the rise of AI agents impacting, first off, your AI business? And how should we think about the progression from where we are today in early adoption to a point where AI agents become a meaningful part of -- drive a meaningful part of your revenue and demand for your products?

I think it's a long question, but I'll try to give you a couple of short answers, right? I think, one is AI agents depend on high-quality data and good guardrails so that they don't make mistakes. We have introduced a series of technologies that we call the AI Data Engine that makes those data preparation and guardrails and enforcement much easier to do. The second is AI engines typically repeatedly go back and do what's called a recursive kind of request for data so that it can improve its analytical outcomes. And so our higher performance systems like our AFX system, some of the caching technologies that we have built to optimize the data pipeline, those become useful as the world moves to agentic. And there's a lot more innovation we are working on that we will tell you more as we bring that to market.

Our final question today comes from Asiya Merchant, Citi.

George, I don't know if you commented on this, and again, apologies if I joined in late. But just your opportunities to win business with some of the neoclouds, I mean, not necessarily the hyperscaler. And I know you have your first-party services on a lot of these hyperscaler offerings. But how do you look at the sovereign or the neocloud opportunity? And was there any traction there that you're seeing? Just any color you could provide on that sort of customer segment would be great.

Yes. I think that once we introduce the AFX solutions, we are seeing growing interest in the neoclouds. As we said, we had a win in a large Asian neocloud this quarter with the AFX in its first quarter of availability. We are working with other neoclouds to provide a differentiated value proposition that includes the rich suite of cloud-ready data services like multi-tenancy, like integrated security, all of the things that we have proven in the hyperscalers, which now these neoclouds want in their environment. Of course, performance, scale, all of those capabilities. And then hybrid use cases where as they move their business towards addressing enterprise AI, our incumbency and our ability to bridge on-prem to neocloud is quite differentiated. And so we'll tell you more about it over the next few quarters. We recognize that, that's a large segment that we can pursue and we're making a focused effort to go after it. It's early, but we'll tell you more, and we are encouraged by what we see.

Thank you, Asiya. I'll now hand it over to George for some final comments.

Thanks, Kris. Thank you for joining us today. Our strong execution and operational discipline enabled us to deliver another outstanding quarter. Our unified data platform delivers exceptional value and operational efficiencies, solidifying our position as the intelligent data backbone for the AI era. And our broad portfolio positions us well to navigate the current inflationary memory environment. We are on track to deliver our strongest year yet. And looking ahead, I am confident that our visionary approach to a data-driven future will enable us to outpace market growth and capture additional share, driving significant value for our customers, partners and shareholders.

Thanks, everyone.