NetApp, Inc (NASDAQ: NTAP) Q1 2026 Earnings Call dated Aug. 27, 2025
Corporate Participants:
Kris Newton — Vice President – Corporate Communications and Investor Relations
George Kurian — Chief Executive Officer and Director
Wissam Georges Jabre — Chief Financial Officer and Executive Vice President
Analysts:
Krish Sankar — Analyst
Mehdi Hosseini — Analyst
Erik W. Woodring — Analyst
Samik Chatterjee — Analyst
Wamsi Mohan — Analyst
Tim Long — Analyst
Simon Leopold — Analyst
Ari Terjanian — Analyst
David Vogt — Analyst
Ananda Baruah — Analyst
Asiya Merchant — Analyst
Presentation:
Operator
Good day and welcome to the NetApp First Quarter of Fiscal Year 2026 Earnings Call. All participants will be in a listen-only mode. [Operator Instructions] After today’s presentation, there will be an opportunity to ask questions. 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.
Kris Newton — Vice President – Corporate Communications and Investor Relations
Hi, everyone. Thanks for joining us. With me today are our CEO, George Kurian, and CFO, Wissam
Jabre. This call is being recorded 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 second 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 estimates are available on our website.
I’ll now turn the call over to George.
George Kurian — Chief Executive Officer and Director
Thanks, Kris. Good afternoon, everyone. Thank you for joining us. We delivered a solid start to fiscal year 2026, achieving revenue of $1.56 billion, above the midpoint of our guidance range. As anticipated, robust performance in the Americas enterprise offset year-over-year declines in US public sector and EMEA. This success was fueled by strong demand for our all-flash offerings, first-party and marketplace cloud storage services, and AI solutions.
By maintaining an unwavering focus on helping customers with their top priorities, we are driving success in a dynamic market. While we continue to observe some macro-related spending caution, the emerging enterprise AI market is driving urgency among customers to modernize data infrastructure, advance cloud transformations, and bolster cyber resiliency. The data infrastructure demands of AI applications are complex and relentless, involving the need to unify, search, and organize massive volumes of data scattered across multiple silos on premises and in the cloud. Legacy architectures are increasingly becoming bottlenecks, limiting the ability to scale and adapt to new requirements.
Organizations are turning to NetApp for data solutions that deliver competitive advantage and operational efficiencies. Our unified data architecture, capable of handling any data type anywhere, empowers customers to break down silos, eliminate complexity, and accelerate their AI journey.
Customers seeking to build future-proof, AI-ready infrastructure choose NetApp. In Q1, we secured a number of wins highlighting the unique benefits of our intelligent data infrastructure platform, powered by ONTAP, as customers modernize their data infrastructure. For example, a leading global bank turned to NetApp to manage massive amounts of highly sensitive structured and unstructured data. The customer needed a centralized hybrid cloud data lake to be the foundational source for its AI initiatives with stringent requirements for cyber resiliency and data protection.
Our ability to deliver a secure, high-performance, scalable platform with unified management, supporting all data types and spanning both cloud and on-premises environments made NetApp the clear choice. In Q1, our all-flash array revenue grew 6% year-over-year to $893 million, an annualized run rate of $3.6 billion. Exiting Q1, 45% of systems in our installed base under active support contracts are all-flash. Healthy customer engagement and strong interest in our unified and block-optimized all-flash storage portfolio have enabled us to displace competitive all-flash and hybrid-flash footprints. This strength propelled us to the number one position in the all-flash array market for calendar Q1 2025 as reported by IDC.
Let me share a couple of examples that highlight why customers are moving to NetApp. In Q1, an aerospace company selected us to replace a competitor’s footprint. With ONTAP-based all-flash arrays, the customer gained cutting-edge security and performance for its defense initiatives with consistent data management across multiple data types and domains, streamlining operations, boosting productivity, and enhancing security, data protection, and resilience against cyber threats.
Another example from Q1 is a UK-based financial services firm that replaced a competitor’s storage infrastructure with our block-optimized all-flash arrays. The customer selected NetApp for our enterprise data management capabilities, which helped them to secure, protect, and efficiently store sensitive client data. Our modern data infrastructure also helped the customer achieve its sustainability requirements with better total cost of ownership through rack space consolidation and more efficient power and cooling.
In Q1, we delivered enhancements to each of our first-party cloud storage services, increasing their value to the customers, and expanding our technology leadership. Our natively integrated services enable customers to discover, deploy, and manage storage seamlessly alongside other native services. This integration puts us in front of customers at the point of workload design, ties to hyperscaler committed spend, benefits from hyperscaler seller incentives, and reduces procurement and operational friction.
In addition to a seamless customer experience, our cloud storage services deliver the performance, cost efficiency, data protection, cybersecurity, multi-protocol support, and hybrid multi-cloud capabilities of ONTAP. Our native integration and advantages over other cloud storage services are driving strong win rates in enterprise critical workloads. Our highly differentiated first-party and marketplace cloud storage services continue to deliver rapid growth, increasing 33% from Q1 a year ago.
In the first quarter, Public Cloud services again served as a strong engine for new customer acquisition. A leading energy company migrated from its on-premises VMware environment to VMware Cloud on AWS and FSx for NetApp ONTAP. After experiencing ONTAP’s rich feature set, this first-time NetApp customer decided to expand its use of FSxN for AI and other workloads. A global technology provider migrated from a competitor’s on-premises environment to NetApp block storage in the cloud for its database environment. The customer chose NetApp for our ability to seamlessly move data across multiple clouds, as well as our robust data protection capabilities.
NetApp helped enhance the customer’s data security, disaster recovery posture, and operational efficiency. We are already in discussions to expand their usage of our cloud services into new areas. Just as we’ve empowered enterprises to harness public and hybrid cloud environments, we are now helping them achieve faster time to value in their AI journeys. Enterprises are beginning to explore and deploy inferencing workloads, shifting demand from proofs of concept to transformational initiatives, which require comprehensive analysis of datasets across the enterprise.
Achieving business outcomes from AI investments is our customers’ and our priority. Unlike training environments, successful enterprise AI deployments require more than just performance and scale. They demand unified data management, enterprise-grade data protection, production scale reliability, hybrid multi-cloud integration, and the ability to consolidate data without compromising privacy or security.
Our AI solutions are designed to truly unify enterprise data, delivering outstanding performance and scale while also seamlessly meeting all these critical requirements. This positions us exceptionally well to address the evolving needs of our customers. In Q1, we closed approximately 125 AI infrastructure and data lake modernization deals across various geographies, industries, and use cases. We further expanded our AI ecosystem in Q1. We introduced the AIPod Mini with Intel to address the cost and complexity of deploying AI at the department and team level.
Additionally, we completed the NetApp reference architecture for NVIDIA Cloud Partners, providing ultra-high performance with the richness of NetApp’s hyperscaler-proven data management. This collaboration allows AI service providers to deliver scalable, high-performance, enterprise-grade data services by combining ONTAP’s advanced data management with NVIDIA’s compute and networking platforms, delivering consistent governance, multi-tenancy, and security across hybrid and multi-cloud environments.
Our conversations with customers about AI infrastructure have intensified with organizations recognizing the critical role that data plays. Notable AI wins in Q1 included a leading automaker that chose NetApp to efficiently manage the massive data volumes essential for training advanced AI models and for building an autonomous vehicle software stack. NetApp’s data governance, security, and auto-tiering capabilities were key factors in this customer selecting us as the data infrastructure for their NVIDIA SuperPOD deployment.
Another example showcases our capabilities for service providers. A large sovereign cloud provider selected NetApp for their AI-as-a-service platform to support large LLM providers and AI development companies. Critical factors in this win included our certification with NVIDIA Cloud Partners, confirming compatibility and performance for AI workloads, and our secure multi-tenancy, delivering storage and data isolation with assured high-performance SLAs for each tenant in a shared environment.
Looking ahead, we remain committed to executing our innovation roadmap and are building on a position of strength. We have taken the number one position in the all-flash market by helping customers modernize with cutting-edge data infrastructure and industry-leading cyber resilience for storage.
Our highly differentiated cloud services enable hybrid and multicloud transformations. And we are well-positioned and growing in the emerging area of enterprise AI. I am excited about the advantages we bring to customers in these critical areas.
Our commitment to delivering customer value with the industry’s best products and cloud services is unwavering, our focused strategy is delivering share gains, and our teams are energized and executing with discipline. I hope to see you all this October at our NetApp Insight Customer Conference, where we will unveil more pioneering innovation for the era of data and intelligence.
I’ll now turn it over to Wissam.
Wissam Georges Jabre — Chief Financial Officer and Executive Vice President
Thanks, George. And good afternoon, everyone. As a reminder, all numbers discussed are non-GAAP unless otherwise noted. We delivered a strong quarter, exceeding the midpoint of our revenue and EPS guidance ranges. As expected, first quarter revenue was driven by strength in the Americas enterprise, offsetting softness in US public sector and EMEA.
Total revenue for Q1 came in above the midpoint of our guidance at $1.56 billion, up 1% year-over-year. Excluding the divested Spot business, which generated $23 million of revenue in Q1 2025, total revenue was up 3% year-on-year. The effect of foreign currency exchange rates was favorable to revenue growth by approximately 1 percentage point year-on-year, though it was immaterial relative to guidance.
Q1 Hybrid Cloud revenue of $1.4 billion was up 1% year-over-year driven by Support revenue of $647 million growing 3%, and Professional Services revenue of $97 million growing 18%. Keystone, our differentiated storage-as-a-service offering, which was up approximately 80% compared to the same period last year, drove the growth in Professional Services revenue. Product revenue of $654 million was down 2% year-over-year. Public Cloud revenue of $161 million increased 1% year-over-year. Excluding Spot, Public Cloud revenue was up 18% year-over-year.
We ended Q1 with $4.53 billion in deferred revenue, up 9% year-over-year and 6% year-over-year in constant currency. Q1 Remaining Performance Obligations were $4.94 billion, up 11%. Unbilled RPO, a key indicator of future Keystone revenue, was $415 million, up 40% year-over-year.
Q1 consolidated gross margin improved by 1.6 percentage points sequentially to 71.1%. Hybrid Cloud gross margin was 70%, up 1.6 percentage points sequentially due to favorable mix of highly profitable support revenue. Product gross margin was 54%. Our recurring support business continues to be highly profitable, with gross margin of 92.3%, and Professional Services gross margin was 29.9%, driven by Keystone.
Public Cloud gross margin was 80.1%, up 80 basis points sequentially and 9 percentage points year-over-year. Our Public Cloud business was at the high-end of the 75%-80% long-term target range in Q1. We are confident that Public Cloud gross margin will continue to improve and are increasing the long-term gross margin target range for this business to 80%-85%.
Operating expenses of $707 million were down 1% year-over-year despite the unfavorable impact of foreign currency exchange rates, and flat from Q4 2025. Q1 operating profit was $401 million, operating margin was 25.7% and diluted EPS was $1.55, all aligned with our expectations.
We had a strong cash quarter with Q1 records for both cash flow from operations of $673 million and free cash flow of $620 million. This was mainly driven by working capital improvement.
During the first quarter, we redeemed the senior notes due in June 2025 for $757 million, which
Comprised principal and interest. We also returned $404 million of capital to our shareholders with $300 million in share repurchases and $104 million paid in dividends of $0.52 per share.
Q1 diluted share count of 203 million was down 9 million shares, or 4% year-over-year. Our balance sheet remains strong. We ended the quarter with $3.3 billion in cash and short-term
Investments and $2.5 billion in total debt, resulting in a net cash position of approximately $840 million.
Now turning to guidance, starting with Q2. We expect revenue to be $1.69 billion, plus or minus $75 million. This implies 2% growth year-over-year at the midpoint. Excluding the divested Spot business from the year-ago comparison, our revenue guidance implies 3% growth. We expect Q2 consolidated gross margin to be 71%, plus or minus 0.5%, and operating margin to be in the
Range of 28% to 29%. Diluted EPS is expected to be between $1.84 and $1.94 with a midpoint of $1.89.
Turning now to full year 2026. We are pleased with the solid start to the year and remain confident in our strong portfolio, as well as our ability to execute in a dynamic environment. We are reiterating our full-year guidance and expect fiscal year 2026 total revenue to be between $6.625
Billion and $6.875 billion, which at the $6.75 billion midpoint reflects 3% growth year-over-year.
Excluding Spot, revenue guidance implies 4% growth year-over-year. We expect diluted EPS in the range of $7.60 to $7.90, for a midpoint of $7.75.
In closing, as we look to the remainder of fiscal year 2026, we are focused on executing our strategy, capitalizing on our growing opportunities, enhancing profitability and free cash flow, and consistently delivering value to our customers and shareholders.
I’ll now turn the call over to Kris for Q&A.
Kris Newton — Vice President – Corporate Communications and Investor Relations
Thanks, Wissam. Operator, let’s begin the Q&A.
Questions and Answers:
Operator
Our first question comes from the line of Krish Sankar with TD Cowen. Please go ahead.
Krish Sankar
Yeah, hi. Thanks for taking my question. Wissam, the first question is for you. Your all-flash revenue growth has decelerated from double digits to just 5% now. Is it pricing or demand or something else? And along the same path, the product gross margins are also down even when compared to similar all-flash revenue levels around $900 million. Is this due to NAND pricing or what is driving it? And how do you think about product gross margins going forward? I then have a long-term follow-up with George.
Wissam Georges Jabre
Yeah. Thanks, Krish. Look, the dynamics that drove the all-flash revenue were very much what we had said earlier in the quarter when we guided, they’re very much similar to what we also said in our prepared remarks. We did experience some softness in the US public sector and to a much lesser extent, EMEA, while it was offset by the Americas that sort of drove the dynamics on seeing all-flash growing slightly lower than previous quarters, but this was already sort of anticipated by us. What I would point out is when you look at fiscal ’25, all-flash grew by mid-teens, and we just saw some dynamics in the Q1. We anticipate this to basically not continue for the rest of the year.
When I look at the second part of your question, in relation to the product gross margin, we had a few dynamics also on the product gross margin. And if you look at it sequentially, we did see costs improved marginally, but we saw mix being unfavorable. We had anticipated a slightly richer mix of high-performance flash versus capacity flash, so that sort of drove that. But when you look at the product margin on a year-over-year basis, it’s mostly the vast majority of it is cost-related. We did see a much bigger uptick in flash cost in Q1 ’26 versus Q1 ’25. So I would say more than 5 percentage points of the 5.9% were driven by the cost dynamic. The rest is very much product and customer mix. And if I look at the rest of the year, we expect product margin to improve from here. I expect the product margin to continue to improve gradually for the rest of the year. We should be operating in our long-term range, which is mid-to-high 50% for the rest of the year. And so, that’s where I think Q1 was very much at the low point in terms of the product margin.
Krish Sankar
Thank you and it’s very helpful to hear that. And George, I’d make a longer-term question for you. It’s kind of very encouraging to see enterprises adopting AI and deploying inference workloads. It seems like they’ve gone from proof-of-concept to production. But can you give us a sense of some of the architecture these customers are following? Are they buying more storage or more capacity or more software attached? Any color there on enterprise adoption would be helpful. Thanks a lot, George.
George Kurian
We saw three types of wins in the quarter. As we noted in our prepared remarks, we had north of 125 wins for AI in the quarter as compared to 50 a year ago, so strong momentum. The three types of use cases were data lakes, which was about 20% of the total number. There was training, which is either fine-tuning a large model or taking a smaller model and customizing it for your enterprise or building a sovereign cloud model for model training. We thought that was about 45% of the total number, and the remainder were RAG and agentic AI use cases. For data lakes, it’s typically a combination of flash-based storage for the hot tables and hot data and a large archive with object storage. For the other two, they are typically all-flash storage.
Krish Sankar
Got it. Thank you very much, George. Very helpful.
George Kurian
Thank you.
Operator
And our next question comes from the line of Mehdi Hosseini with Susquehanna International Group, please.
Mehdi Hosseini
Yes. Thanks for taking my questions. Two follow ups for me. First, George, does the availability of 128 terabyte QLC-based NAND impacting your ability to provide search solutions, especially for the AI applications?
And then number two, how should we think about the seasonality into the January quarter? I’m not asking for your guide, but given that it — perhaps compares are getting easy, so should seasonality still be a factor looking into January quarter? Thank you.
George Kurian
On the first question, listen, we have a broad range of NAND technologies available to customers all the way from 15 terabyte super high-performance drives to 60 terabyte QLC drives, and we are working on bigger drives form factors. So, we don’t feel like we are gated on any of those. I think, as I said in my previous remarks, we use the right configuration for whether it’s a really high-performance use case or more of a capacity-based use case or an archival use case.
With regard to the second question, the seasonality, listen, we are very confident of the outlook for the year. We are one quarter in. We performed better than we expected at the start of the quarter, and we saw strength, particularly in the Americas commercial business and in several other parts of the world. I think as we look through the rest of the year, we are focused on executing it quarter at a time. There’s still a decent amount of uncertainty in the external landscape, and we’ll guide the year as we see visibility to the full year. So, we feel good at the end of one quarter, and we’ll tell you more at the end of the second quarter.
Mehdi Hosseini
Thank you.
Operator
Our next question comes from Erik Woodring with Morgan Stanley.
Erik W. Woodring
Hey, guys. Thank you very much for taking my questions. I’m looking forward to working together with you guys. Maybe just to start, George. I’d love to get your viewpoint as you kind of have conversations with your customers or look at your pipeline. When you think about strength in Americas commercial versus weakness in public in EMEA, is that expected to persist? Do you see any of your kind of end market customers maybe inflecting in the second half? For example, could SMB get better in the second half here in America? Would just love to know kind of underlying that guide by end market, how you’re thinking about these ‘ the strengths or weaknesses or accelerations and decelerations of these different end markets?
And then I’ve a quick follow-up. Thanks.
George Kurian
We had a really strong quarter in the Americas broadly defined as not public sector. This included the largest enterprises. It included the mid-market customer segment as well, excellent execution, strong demand patterns, and a large number of competitive wins. US public sector was very weak in the quarter. I think we were still awaiting budgets to be deployed to agencies. We saw a change in the budgets allocated to different types of agencies that we have moved resources to where the budgets are.
Q2 is typically a strong budget spending quarter for US public sector, and it’s a smaller part of our overall business in the second half of the fiscal year. In Europe, we executed well in most countries. We saw softness in UKI and some parts of the large enterprise in Germany. But overall, a lot of our European teams executed well. And in Asia-Pacific, again, outside of Australia and New Zealand, our teams have performed well. So, the spots of weakness were localized, but pronounced in those areas.
Erik W. Woodring
Okay. Super helpful. Thank you. Thank you for that, George. And then maybe Wissam, great to work together again. I would just love to better understand as you think about the gross margin guide for your Public Cloud business. It’s now 80% to 85% versus the 75% to 80% before. Kind of what has changed that influences that view. And clearly, you see that as sustainable. But if you could just unpackage that a little bit to help us understand why that continues to improve, and what influenced that change, that would be helpful. Thanks so much.
Wissam Georges Jabre
Yeah. Thanks, Erik, and good to hear your voice. I look forward to working together. Yeah, on the Public Cloud front, look, the business itself has been improving gross margin very steadily and actually at a very fast pace. And this past quarter, we just hit the high end of the prior long-term target range. And looking at where the business is headed, we feel comfortable that 80% to 85% is a good target range for it going forward, and that’s why we raised it. The few things I would say that contribute to that are, one, we’re seeing a depreciation roll-off for some of the initial installed hardware, so that sort of helps with the margin.
In addition, there’s more software content in the revenue. And so, both of these dynamics will help sort of continue to improve the gross margin going forward. As we think to the rest of the year, and I think I said that also on our call last quarter, we anticipate Public Cloud gross margin to continue to gradually improve from here. And so I hope this helps unpack it for you.
Erik W. Woodring
Yeah. Thank you, guys, and best of luck.
Wissam Georges Jabre
Thank you.
George Kurian
Thanks, Erik.
Operator
Our next question comes from the line of Samik Chatterjee with JPMorgan.
Samik Chatterjee
Hi. Thanks for taking my question. I guess, George, just following up on your earlier comments about the activity or engagement you’re seeing on the AI side. With your sort of first couple of quarters of the year, you’ve settled into this sort of nice run rate of AI wins. If you can give any color in terms of the magnitude or size of those deals with your customers? And should we expect that consistent level of new wins as you go through the year, or is the pipeline telling you that maybe AI wins do accelerate as you go through the rest of this fiscal year? Any color on that on both of those fronts, please. And I have a follow-up.
George Kurian
The wins are quite wide. You could see smaller environments being stood up like an AI center of excellence, proof-of-concept for inferencing or RAG. And then it could be much larger scale. For example, we have very large-scale RAG environments in some of the large financial services firms, as well as we have won some big model training environments, which are very large so that you can put a lot of capacity against a large number of GPUs.
We are also beginning to have some significant wins in the AI-as-a-service provider. So, it’s a wide range. I think we are excited about the fact that we have more than doubled the number of AI wins year-on-year. We are continuing to engage more clients around the world. We are bringing more technologies to that part of the market as we head into our November INSIGHT conference and we’ll tell you more. We should expect that to grow through the course of the year at a steady clip.
Samik Chatterjee
Got it. Got it. And then just on the guide for the full year here. I mean, it does imply, if you grow 3% ex Spot in fiscal 2Q, you do have a back half that needs to be around mid-single-digit growth. And I think last quarter, you did reference you’re working with some customers on large data center modernization deals as well. So maybe what’s the visibility in terms of that acceleration from the first half into the second half at this point? And does it include any large deals that are in the pipeline to sort of come through to enable you to accelerate to that modest number? Thank you.
George Kurian
We feel good about the comments we made in the prior call. We are working on a number of different opportunities. I think, Samik, we are one quarter in, in a pretty dynamic environment. We want to take it a quarter at a time, and should we have another strong quarter, we will tell you more about the full year at the end of Q2. So I feel good about the year. All of the things that we’re working on we’re executing well. We had the softness in US public sector start the year. We’re going to see a second quarter through and then the rest of the year, the headwind from USPS continues to be much less.
In terms of large deals, listen, we’ve had several large deals this quarter. We continue to work several more deals through the course of the year, and we feel confident about those.
Samik Chatterjee
Okay. Great. Thank you. Thanks for taking my questions.
George Kurian
Thank you.
Operator
Next question is from Wamsi Mohan with Bank of America.
Wamsi Mohan
Hi, yes, thank you. Can you just talk a little bit about some of the variables in that gross margin guide for Q2 that might push you lower on a quarter-on-quarter basis. I mean you guided 70.5% to 71.5%, which at the low end would be 60 bps lower. But if you expect product gross margins up quarter-on-quarter, and you’ve just taken up your range for Public Cloud gross margins and support is pretty stable, how do you even get to sequentially down gross margins in that range, or is there anything else that I’m not taking into account?
Wissam Georges Jabre
Hey thanks, Wamsi. Just to clarify, we ended Q1 with a gross margin of 71.1% and we’re guiding basically flattish. I think at the midpoint is like 10 basis points lower. And so, it’s just basically a fluctuation, if you like. But when you think for the various dynamics, we expect the — I would say, directionally, we expect the margins for the various components to improve, either stay in line or to improve from here. But then when you look at the mix of revenue, product — typically product revenue is a bigger component when it gets to Q2 and sort of with the mix of revenue itself, you’d expect some headwinds as well. And so, there isn’t anything fundamentally problematic. Actually, we’re optimistic and we’re confident that margins, as I said, for the various components of the revenue should be flat to up from here. It’s really the mix of the revenue itself, if that helps.
George Kurian
Yeah. Hey, Wamsi, I think — I wouldn’t read too much into the sequential. I think the big moving part within Q1 and Q2 is really the percentage of the total revenue that’s product. In Q1, it was lower.
Wamsi Mohan
Okay.
George Kurian
And so Q2, it will be probably higher. So that’s where it can get, right?
Wamsi Mohan
Okay. That’s helpful. And then just on operating leverage, right? If I just think about this roughly flattish gross margins, but quarter-on-quarter, your operating margin is up maybe similar to how it was last year. So your operating leverage is improving because last year you had gross margin improvement, too. So, should we expect that operating leverage benefit to continue in the back half of the year as well? Thank you.
Wissam Georges Jabre
Yeah, Wamsi, I would anticipate operating leverage to continue in the back half of the year. The reason, we’re the — probably the only dynamic you’re seeing with respect to Q2 and operating leverage is that gross margin differential. If you recall, last year, product margin was in the 60% range and much higher than what we’ve experienced of late.
Wamsi Mohan
Okay. Thanks, Wissam.
Wissam Georges Jabre
You’re welcome.
Operator
Your next question comes from the line of Tim Long with Barclays.
Tim Long
Thank you. Two as well, if I can. First, maybe just talk about the comment about the installed base, 45% of the systems on all-flash. Could you talk a little bit about what NetApp is doing to try to convert the rest of that installed base, and what that would mean for financials and margins, revenue opportunity, as well as the competition you would see there?
And then secondly, if you could just touch on Keystone, another very impressive quarter. I get it’s small. But could you just talk a little bit about what you’re expecting mid to long term from Keystone, given how successful the Public Cloud business has been, you termed corollaries [Phonetic] nd what Keystone could be to the on-prem business? Thank you.
George Kurian
Yeah. Thank you for your questions. With regard to — I’ll take the Keystone question first. Keystone has done well. It’s an as-a-service model. There are more and more clients who are comfortable with an as-a-service model. And so they are probably looking to harmonize their IT operating model for infrastructure to more of an as-a-service. We also see clients who are either doing transitional work going from one data center to another, or going from a data center to the cloud, where they want to have a bridge between our on-prem technology and our cloud technology that will buy a Keystone service for the transitional period. So I’m excited about that.
We are also using Keystone to help in competitive situations, where the customer may have a capital commitment in another vendor’s technology and wants to kind of not have to pay two capital investments at the same time. Ultimately, our view is we give customers choice and we’re not going to try to force them into one model. But we are going to continue to make Keystone as ever better offering just like we make our products.
With regard to your other question about the installed base, listen, we have a very, very large installed base of systems. And we continue to sell our flash technologies to customers who are not existing customers of NetApp and customers who are existing customers of NetApp. They refresh their technologies when it is time for them to refresh, and we have an extremely high win rate in those refreshes because the software layer is essentially entirely consistent, the operational layer is entirely consistent. At the same time, if you look at the number of flash systems that we sell a quarter, you can impute the fact that we are growing 1% in our installed base means we are winning a large number of customers outside our installed base. So we have both selling motions. We are providing clients the right choice for their right use case. Hard drive-based solutions are still the cheaper solutions for things like backup or cold storage. And so it has a place in our portfolio.
Tim Long
Okay. Thank you.
Operator
Your next question comes from the line of Simon Leopold with Raymond James.
Simon Leopold
Thank you for taking the question. I wanted to try to get a better sense of the longer-term AI product strategy. In other words, I know you don’t want to front run your INSIGHT event for October. But what I’m trying to get a better sense of is the readiness of your portfolio to support enterprise initiatives, and whether you’ve got more products in the pipeline to perhaps move up the value chain to support initiatives for AI at the enterprise. Thank you.
George Kurian
There are three or four technologies that clients look for from a data infrastructure provider to them to help them in their AI initiatives. The first, of course, is high-performance, high-scale data storage for model training or inferencing for agentic environments as well as for data lakes. We feel very good about our solutions for that.
The second is solutions that customers who want to use the cloud with equivalent storage capabilities so that they can use cloud-based tools rather than on-prem tools. We have made good progress. We have more work going on with the hyperscalers, and we’ll tell you more about that at our INSIGHT customer conference.
The third area from a technology standpoint that we have advantages and we are providing a lot more kind of capabilities moving forward are things that allow clients to actually manage their data more efficiently. This could be they want to be able to search the data. They want to organize it so that they can align certain datasets to certain volumes to certain models and keep track of it. They want to implement guardrails and access controls. They want to automate the vectorization and the RAG readiness of their data. We have a lot of those capabilities coming at our INSIGHT conference.
And then the last piece is ecosystem, right? So we already have reference architectures with NVIDIA. We are working on reference architectures more and more of them with the hyperscaler tools, and you’ll see more about announcements around that at our customer conference. So, good stuff today, but a lot more to come.
Simon Leopold
Thank you.
Operator
Our next question is from Ari Terjanian with Cleveland Research.
Ari Terjanian
Hi, all. Thanks for taking the questions. Good to hear from everybody. Just a couple from my end. First, just on competitive landscape, like really great to see the leadership position in IDC all-flash. But, I mean, it seems like today there are some different results, one of your peers raising their outlook. So just any update on the all-flash kind of competitive landscape and what you’re seeing out there?
And then, second, just in terms of pricing, I know there was some pricing actions taken last year that would flow through the model this year. Any updates just on the take rates of the price increases that you’re seeing with your customers? Thank you so much.
George Kurian
I’ll take the first one on the competitive landscape. Listen, it’s always been a competitive environment. We are more exposed to some of the environments that or markets that have had challenges like US public sector than some of our competitors. But, listen, we get up every day and go and compete. We feel good about our competitive position. We did not see any specific adverse kind of pattern in the quarter, and we’re going to go out there and post another strong quarter and that will take care of the competitive landscape.
With regard to pricing, listen, we have not taken pricing action related to the tariffs. We’re in a wait-and-see mode. In terms of pricing actions that we took as we introduced our new systems, those systems have performed well and they have done well in the mix. I think with regard to the gross margin, product gross margin commentary, it’s primarily related to the cost increases from a year ago. And so, we feel good about the mix between highperformance flash and capacity flash, that our high-performance flash systems have grown for two consecutive quarters nicely in the mix. And we’re just going to keep working on it. I think with regard to the outlook for the rest of the year, listen, that the cost compared get more benign through the rest of the year, which gives us confidence that our gross margin should return to the original ranges that we had shared.
Ari Terjanian
Thanks.
Operator
Our next question is from the line of Doug Vogt with UBS.
David Vogt
Hi, it’s David Vogt. Hey, Wissam. Hey, George. I don’t want to beat a dead horse, but can you go back to, George, the competitive landscape? You talked about performative high-end storage being a little bit weaker than you thought and shifted a little bit more to capacity storage. Was that in any particular region, vertical? I know EMEA and public sector was a little bit softer.
And then, Wissam, I’ll give you my second question. You mentioned, obviously, mixed supply chain — excuse me, components have an issue in gross margin. Can you share where we are from a pre-buy perspective? I mean, is the balance sheet effectively no longer carrying excess inventory that you purchased previously? And so, how should we think about your strategy with pre-buys from a component perspective going forward? Thanks.
Wissam Georges Jabre
Yeah. Hi, David, let me take the — both questions. On the first part of the question, my comment was relative to what we guided and relative to our expectations. I wouldn’t say necessarily there is weakness in one versus the other. It’s just that the mix that we had anticipated is slightly different, that contributed to a little bit of pressure on the product margin. So–
George Kurian
High-performance flash actually outgrew capacity flash year-on-year.
Wissam Georges Jabre
Yeah. Exactly. And with respect to the question on product margin, look, as we said earlier, it’s the — I expect it to improve from here. And part of it, of course, has to do with the cost. And we do have certain volumes with locked-up pricing that gives us confidence that we anticipate to see some improvement from here. So I don’t know if this helps. With respect to the part of the question related to the components in our supply chain, we don’t see any issues there. I mean there’s nothing really problematic.
David Vogt
Great. Thanks, guys.
Operator
Your next question is from Ananda Baruah [Technical Issues]
Ananda Baruah
Hey, guys. Yeah, thanks for taking the question. Really appreciate it. George, just going back to AI, you mentioned model training wins. I think you said that are very large. And you also mentioned AI-as-a-service, which sounds like neoclouds. Please clarify if it’s something bigger than just that. But, I guess, the question is, for the last couple of years, it sounds — and RAG is still a focus as you’ve spoken to, but it sounds like you might be expanding your TAM. And you talked about reference designs with the hyperscalers as well. So, are you seeing TAM expansion here? And if so, what’s a useful way to think about kind of magnitude of TAM expansion in the context of the metrics that you provided the last couple of years post the Analyst Day? Thanks.
And I have a quick follow-up here. Thanks.
George Kurian
Yeah. I think, first of all, we have several kind of large-scale sovereign providers who are users of our technology for their enterprise cloud business. And as many of them have stepped into the AI business, we needed to get certified as an NVIDIA NCP provider. And we have met those certification requirements, and it allows us to bring technology that is familiar to them for a broad range of use cases for AI-as-a-service use cases, where all our advantage around multi-tenancy hybrid operating models and so on play in. But also for model training, especially for countries that have sovereign models, we are seeing good progress. At this point, we still believe that the data storage opportunity is predominantly in the enterprise inferencing part of the world, but we are going to be opportunistic and compete where we see opportunities or other use cases as well, and we feel good about our prospects there.
Ananda Baruah
That’s super helpful explanation. And I guess, just a quick follow-up in that context is, just on the RAG front, are you seeing reasoning agents impact the RAG pipeline? Is it seemingly would grow the TAM? Are you seeing sort of, I guess, like engagement push-out as a result of that as well? Do you see any context there that impact on your business and your pipeline, would be great? That’s it. Thanks..
George Kurian
Yeah. I think broadly speaking, the biggest challenge that clients deal with in getting AI to be applicable for their enterprise use cases is really around organizing their data tagging in so that they can have the right controls around it, being able to then kind of apply vectorization, knowledge graphs, and other things to that data so that it’s ready for use in an AI pipeline. And so we have a lot of tools that we are bringing to market, and many tools that we already have that will help us provide extra value to clients because we have this gigantic installed base of unstructured data that clients are anxious to mine and that we can monetize with these tools. So that’s the first thing.
The second part of inferencing is, people are using more kind of reasoning models where the model goes to storage, request data, searches, works that data, and then comes back to the storage again. There are some clever capabilities that we have built that allows for that functionality to be more efficient, as well as to avoid having to go to the original data source multiple times. And so we’ll share more about that at our user conference that allows those reasoning models to be a lot more effective and fast at getting to the answer.
Ananda Baruah
Thanks a lot. Appreciate it.
George Kurian
Thank you.
Operator
And our final question comes from the line of Asiya Merchant with Citigroup.
Asiya Merchant
Great. Thanks for squeezing me in here. Just how do you guys think about your hypervisor offerings? I understand anything you can comment on that on as it relates to Nutanix or the disruptions that we’re still continuing to see due to VMware licensing. How do you think about your product offerings on this hyperconverged infrastructure? Thank you.
George Kurian
We support a broad range of hypervisors, and we will continue to support as many hypervisors as our customers want, right? And so those could be on-prem hypervisors like the ones that you mentioned that we will work to support if we don’t already support, as well as cloud hypervisors where people are using the cloud providers, the hyperscalers own tools. And so, there is a wide range and we’re happy to share more of that with you. Our general view of the world is we believe to operate with as many of them as clients want.
Asiya Merchant
Okay. Look forward to hearing about that at your INSIGHT. Thank you.
George Kurian
Thank you.
Kris Newton
All right. Well, thank you, Asiya. I’m going to pass it over to George for some final remarks.
George Kurian
Thank you, Kris. We delivered a strong start to fiscal year 2026, building on a position of strength. We’ve taken the number one position in the all-flash market by helping customers modernize with cutting-edge data infrastructure and industry-leading cyber resilience for storage. Our highly differentiated cloud services enable hybrid and multi-cloud transformations, and we are well-positioned and growing in the emerging area of enterprise AI. Thank you for joining us today, and I hope to see you at NetApp INSIGHT in October.
Operator
Thank you again for joining us today. This does conclude today’s presentation. You may now disconnect.