Good day. Welcome to the Pure Storage First Quarter Fiscal 2025 Financial Results Conference Call. Today's conference is being recorded. All lines will be muted during the presentation portion of the call, with an opportunity for questions and answers at the end. At this time, I'd like to turn the call over to Paul Ziots, Vice President of Investor Relations. Please go ahead.

Thank you. Good afternoon, everyone, and welcome to Pure's first quarter fiscal year 2025 earnings conference call. On the call, we have Charlie Giancarlo, Chief Executive Officer; Kevan Krysler, Chief Financial Officer; and Rob Lee, Chief Technology Officer. Following Charlie's and Kevan's prepared remarks, we will take questions. Our press release was issued after the close of the market and is posted on our website where this call is being simultaneously webcast. The slides that accompany this webcast can be downloaded at investor.purestorage.com. On this call today, we will make forward-looking statements, which are subject to various risks and uncertainties. These include statements regarding our financial outlook and operations, our strategy, technology and its advantages, our current and new product offerings, and competitive, industry and economic trends. Any forward-looking statements that we make are based on facts and assumptions as of today, and we undertake no obligation to update them. Our actual results may differ materially from the results forecasted, and reported results should not be considered as an indication of future performance. A discussion of some of the risks and uncertainties relating to our business is contained in our filings with the SEC, and we refer you to these public filings. During this call, all financial metrics and associated growth rates are non-GAAP measures other than revenue, remaining performance obligations, or RPO, and cash and investments. Reconciliations to the most directly comparable GAAP measures are provided in our earnings press release and slides. This call is being broadcast live on the Pure Storage Investor Relations website and is being recorded for playback purposes. An archive of the webcast will be available on the IR website and is the property of Pure Storage. Our second quarter fiscal 2025 quiet period begins at the close of business Friday, July 19, 2024. With that, I'll turn it over to Charlie.

Thank you, Paul. Good afternoon, everyone, and welcome to our Q1 fiscal 2025 earnings call. Thank you for joining us today. We are pleased with our Q1 performance, returning to double-digit revenue growth for the quarter. Our highly differentiated platform strategy continues to demonstrate success and resonates with customers. The recent advances in AI have opened up multiple opportunities for Pure in several market segments. Of greatest interest to the media and financial analysts has been the high-performance data storage market for large public or private GPU farms. A second opportunity is providing specialized storage for enterprise inference engine or RAG environments. The third opportunity, which we believe to be the largest in the long term, is upgrading all enterprise storage to perform as a storage cloud, simplifying data access and management, and eliminating data silos, enabling easier data access for AI. Pure is seeing early success in all three of these AI-based opportunities, and we can address them all with our unified Purity platform. Unlike other vendors, we do not require different operating system software to address different storage needs. AI inevitably draws customers' attention to the fragmented state of their data caused by their disparate data storage infrastructure. Data stored on widely diverse platforms, with different operating and management systems, which are siloed and individually managed, are unable to feed real-time data to AI inference engines. The Pure Storage platform strategy provides a unified and integrated data storage and delivery system across customers' various data environments. It facilitates seamless management and data access across data centers and the cloud, with simplified universal policies and management. We will be announcing significant new advances to this platform strategy next month at our Accelerate customer and partner conference. Our platform vision played a crucial role in securing several strategic enterprise deals this quarter. The ease of use of our platform, and notable interest in saving power and reducing environmental impact, led to a notable win with a managed service provider specializing in high-performance computing. Their accelerated environment for both large language models and inferencing delivers top-tier AI infrastructure and training solutions for their financial services, energy, and life sciences customers. Enterprise and International market segments were strong this last quarter. Our E family continues its strong growth and was also a key enabler in our discussions with hyperscalers. FlashBlade had a record Q1, including in AI workloads. We are seeing broader adoption across geographies, including both prospects and existing customers. FlashBlade's ability to span the price-performance spectrum, from the highest sustained performance required for AI, to low-cost applications such as backup, is incredibly compelling. We continue to make significant progress in penetrating every area of online data storage with our Purity and DirectFlash technology, both in the enterprise and in the cloud. The quantity and quality of our discussions with hyperscalers have advanced considerably this past quarter. Hyperscalers have a broad range of storage environments, including high-performance storage based on SSDs, multiple levels of lower-cost HDD-based nearline storage, and tape-based offline storage. We are in a unique position to provide our Purity and DirectFlash technology for both their high performance and their nearline environments, which make up the majority of their storage purchases. Our most advanced engagements now include both testing and commercial discussions. As such, we continue to believe we will see a design win this year. Pure Storage technology brings multiple advantages to hyperscaler infrastructure. Data storage is either first or second in power and space consumption in data centers. First, Purity and DirectFlash reduce the power, space, and cooling requirements for hyperscale data storage by a factor of 10 or more. Second, Pure reduces the need for sophisticated caching and other technologies that hyperscalers use to compensate for the relatively low performance of hard drives. Third, Pure's technology improves server performance by accelerating data delivery. Finally, Pure's technology significantly improves both the reliability and longevity of their storage, thereby significantly reducing costs. Increased energy usage is a major issue and cost for both hyperscalers and enterprise data centers. This is even more important as introducing AI in data centers promises to consume ever greater amounts of power and cooling. Pure Storage can dramatically reduce the power usage in existing data centers by upwards of 20%, freeing significant power and cooling for AI workloads; another reason why hyperscalers are interested in Pure technology. In the Enterprise, we are seeing continued momentum and opening new opportunities with our cloud operating model. Enterprises want their data centers to operate the same way as cloud companies operate theirs. Customers should be able to automate and manage storage as virtualized clouds of storage, whether located on-premises or in the cloud. Delivering a complete cloud operating model, Evergreen//One allows our customers to consume storage as a service, based entirely on guaranteed service level agreements, enabling them to store their data whenever and wherever they want with guaranteed reliability and performance. It provides customers with hands-free storage services where and when they need it, managed entirely through our Pure1 management portal. Additionally, Hybrid Cloud has now become the standard design practice by enterprises. They now expect the same cloud experience of self-service, flexibility, and agility from their private data center infrastructure. It's no longer only about price, performance, and features. Pure Fusion solves the complexity of traditional IT and storage by joining storage arrays into virtual storage pools. Managers can manage their entire fleet of arrays by policy and set up custom Storage Services for both IT and developers. Pure Fusion combines enterprise storage with a cloud operating model, cloud agility and scalability, and enables easier access to real-time data stores for applications such as AI. Hear more about Pure Fusion at next month's Pure Accelerate event. Moving on to the market as a whole, we have not seen a significant change in the overall macro environment or customers' intentions to buy. While we have great enthusiasm for our opportunities in AI, spending on AI may put pressure on other parts of IT budgets. We believe that the storage market will fare relatively well in this IT economy, but have yet to see a major inflection. Overall, we believe that we are well positioned in all segments we compete in, and we will continue to gain share across our markets. Our leadership position is now clearly demonstrated by our competitors' increased fervor to mimic our strategy and our messaging. We also believe that long-term secular trends for data storage are no longer based on the expectation of commoditized storage, but rather on high-technology data storage systems, which run very much in our favor. We will discuss these long-term trends in more detail at our upcoming financial analyst session at Accelerate. With that, I'll turn it over to Kevan.

Thank you, Charlie. We saw a strong start to our year in Q1, as our financial performance outperformed across both revenue and profitability. Revenue grew a healthy 18% and operating profit of $100 million resulted in a new record high for a Q1. Specifically, we were very pleased that we returned to strong double-digit revenue growth in Q1. Two key drivers of our revenue growth this quarter were: one, sales to new and existing enterprise customers across our entire data storage platform; and two, strong customer demand for our FlashBlade solutions, including FlashBlade//E. In Q1, subscription services annual recurring revenue, or ARR, was healthy, growing 25% to over $1.4 billion. As we mentioned previously, subscription services ARR excludes non-cancelable Evergreen subscription contracts where the effective service date has not started. Including non-cancelable subscription contracts where the effective service date has not started, subscription services ARR at the end of Q1 grew 26%. Total RPO, which includes both subscription services and product orders, grew 27% year-over-year in Q1 to $2.3 billion. As a reminder, product orders within RPO include a non-cancelable telco order in Q3 FY '24, and orders relating to a Fortune 500 financial services company in Q4 FY '24. RPO associated solely with our subscription service offerings at the end of Q1 grew 24%. Q1 year-over-year subscription RPO growth was affected by several large Evergreen//One opportunities that closed during Q1 of last year. Although we expect larger Evergreen//One opportunities to close as we progress through the year, no large opportunities closed during Q1. Total contract value, or TCV, sales for our storage-as-a-service offerings during Q1 were $56 million. We saw strong demand and pipeline build, including large opportunities for our storage-as-a-service offerings during Q1. Consistent with our original FY '25 forecast, we expect 50% growth of our storage-as-a-service offerings, including Evergreen//One and Evergreen//Flex, achieving $600 million in TCV sales. U.S. revenue for Q1 was $489 million and International revenue was $204 million. Our new customer acquisition grew by 262 customers during Q1, and we now serve 61% of the Fortune 500. Product and subscription services gross margins both contributed to strong total gross margin strength of 73.9% in Q1. Product gross margin was 72.8%, and subscription services gross margin was 74.9%. While reflecting strong gross margins, we are aggressively competing and winning customers' secondary and lower storage tiers with our E family solutions and FlashArray//C. Operating profit and margin strength of 14.5% were positively impacted by revenue outperformance, strong gross margin performance, and disciplined investing. Our headcount decreased slightly to approximately 5,500 employees at the end of the quarter. Pure's balance sheet and liquidity remain very strong, including $1.7 billion in cash and investments at the end of Q1. Cash flow from operations during the quarter was $222 million, and capital expenditures were $49 million, representing approximately 7% of revenue. Factors contributing to capital expenditures included test equipment supporting our engineering teams for new innovations, continued build-out of our new headquarters, and infrastructure supporting our Evergreen//One storage-as-a-service sales. Our objective of partially offsetting dilution using share repurchases remains, though during Q1, we did not repurchase shares of stock because of trading restrictions. Beginning in June, we will also fund withholding taxes due on employee equity awards by net share withholding, which will also reduce share dilution. We have approximately $395 million on our existing repurchase authorizations. Now turning to our guidance. For Q2, we expect revenue of $755 million, and expect that operating profit will be $125 million, or operating margin of 16.6%. Turning to our annual guidance for FY '25. We are reiterating our FY '25 revenue and operating margin guidance. We are pleased with the strong start to the year, with both top line and operating profit outperforming. Also, while storage spending for AI is still in its early stages, we believe that we are well positioned as demand for data storage accelerates. At the same time, we are also remaining mindful of the macro spending environment. In closing, we are pleased with the strong start to the year. This is an unprecedented time for Pure as we are well positioned to participate in substantial and sustained growth opportunities, whether driven by AI or our pursuit of replacing the vast majority of data storage with Pure Flash for all customer workloads, including hyperscalers' bulk storage. We look forward to seeing many of you at our product and technology-focused financial analyst meeting at Accelerate in Las Vegas on June 20th. With that, I will turn it back to Paul for Q&A.

Thanks, Kevan. Before we begin the Q&A session, I'll ask you to please limit yourselves to one question with one part so we can get to as many people as possible. If you have additional questions, we kindly ask that you please rejoin the queue, and we'll be happy to take those additional questions as time allows. Operator, let's get started.

Thank you. Our first question comes from Amit Daryanani from Evercore. Please go ahead. Your line is open.

Good afternoon, guys. I guess, Charlie, I'm hoping you could just talk a bit more about the cloud opportunity for Pure going forward. I think Meta had a white paper out recently that they talked about using some other technologies. Do you think the opportunity is bigger in AI versus non-AI? Just anything more on the cloud side would be helpful.

Thank you, Amit. We maintain a clear discipline when discussing our three segments: AI, cloud, and hyperscaler. Each time we mention an AI win, whether it’s in an enterprise, a GPU cloud, or within a hyperscaler, we categorize it as an AI win. When we refer to a design win in a hyperscaler, we're talking specifically about the standard storage they utilize for their own computing and for customer storage in their hyperscaler cloud environments. When we mention cloud, we highlight our capabilities to function like a cloud within enterprises and our software, such as Cloud Block Store and Portworx, that enables customers to achieve a multi-cloud approach to data. Regarding Meta, we have sold into their GPU environment. They operate various AI environments, and we have a presence in most of the Meta AI environments we know of. We categorize that as AI wins contributing to our overall AI success. Design wins in hyperscalers relate to their standard storage, while our cloud achievements involve both Cloud Block Store and Portworx, alongside our arrays functioning across various customer data centers and integrating with hyperscalers. As for your question, I wanted to clarify our terminology. We believe AI offers immediate opportunities in machine learning, prompting customers to upgrade their entire data storage systems. This represents a significant opportunity for us to enhance capabilities and transform storage into a cloud of storage rather than just individual arrays, which we see as a substantial long-term opportunity. Additionally, we believe there are considerable prospects to engage with hyperscalers regarding their traditional storage systems, whether in customer environments or in their own storage, which is an exciting opportunity for us.

Yeah, Amit, and this is Rob. Just to jump on to that, to help you think about size, scale, and just some context behind it. As we think about the hyperscaler environments in particular, there's been a lot of focus and attention paid to the AI deployments and build-outs, but the larger general-purpose bulk storage environments that Charlie speaks of, we believe consist, even today, of 80% to 90% of their total storage build-outs, which are sitting on disk, are being deployed on disk, and are going to serve these general-purpose environments. And that's really the opportunity we see as we refer to the hyperscalers. To some extent, as Charlie said, the AI opportunity is one that we do well in as well. I think the other thing that's going on in these hyperscaler environments is, because there's so much focus being placed on AI build-out, it's presenting a real challenge when it comes to power and space. And so that is becoming clear in our advancing discussions with these hyperscaler firms. It's becoming evident that power availability, securing just the operational limitations to maintain the storage footprint they have today on disk is becoming a real challenge for them and one that we're very hopeful to help them address.

Thank you, Amit. Next question, please?

Our next question comes from Meta Marshall from Morgan Stanley. Please go ahead. Your line is open.

Great. Thanks so much. I just wanted to get any commentary you could provide on just how NAND prices increase, just on how you guys are thinking about gross margins. Clearly, no impact this quarter, but just any expected impact over the calendar year? Thanks.

Thanks, Meta. As you recall, we operate in a dynamic pricing environment; storage, in general, is in a dynamic pricing environment. That is to say that customers buy storage episodically. When they do, they want to negotiate. We negotiate in a competitive environment against players who tend to price on a cost-plus basis. As such, ASPs, if you will, for on a per gigabyte basis, tend to fluctuate with the underlying commodity. So, net-net of all that, NAND prices generally don't affect our gross margin as much as they do with the market as a whole. With prices rising, I think that's somewhat of a tailwind in general, although customers generally have a set budget. So, putting that all together, it shouldn't affect our or the industry's gross margins significantly.

Yeah, in a meaningful way. And just following on to Charlie's point, look, we're really pleased with our gross margin performance in Q1. We did see product gross margins decline slightly sequentially, though remaining strong. While at the same time, we are aggressively competing and winning customer secondary and lower-storage tiers with our E family solutions and FlashArray//C. But the flash pricing volatility again just highlights the differentiated advantages of our Purity software and DirectFlash technology. And as a reminder, QLC flash represents the majority of the flash we consume today, really providing a meaningful sustained cost advantage both against TLC flash and commercial SSDs using QLC flash.

Thank you, Meta. Next question, please?

Our next question comes from Howard Ma from Guggenheim Securities. Please go ahead. Your line is open.

Thanks. Good afternoon, everyone, and thanks for taking the question. I guess for Charlie or for Kevan, how would you guys characterize the demand environment relative to last fiscal year and then relative to a few months ago? Would you say that IT spending optimizations are largely behind us now and you're starting to see a material uptick in demand or not necessarily? And then, just one quick follow-up. On AI inferencing specifically or AI inferencing-driven demand, are you starting to see enough of an uptick at least in pipeline? I know it's not in the numbers yet, but is it enough to call out, or is that still too early? Thank you.

Howard, we're going to take your first question because we're going one question of one part, but we'll get back to it if we can if there's time for the second one.

So, I think that I did state in my prepared remarks is that certainly it's an improvement over last year, but I can't say we've seen a major inflection yet. So, starting to pick up around Q4, I can't say it's changed much since Q4. I do think AI has caused customers to take a second look as to how they are going to be spending their money this year. But overall, I would say it's a modest recovery from last year, but I haven't seen a significant inflection point.

Thank you, Howard. Next question, please?

Our next question comes from Pinjalim Bora from JPMorgan. Please go ahead. Your line is open.

Hey, guys. This is Noah on for Pinjalim. Thanks for taking the question. You talked about a potential design win at a hyperscaler by the year-end. Can you just help us understand how those discussions have evolved so far? You mentioned in your prepared remarks that the quantity and quality of discussions with hyperscalers have advanced considerably this past quarter. So, any other color you can provide there would be helpful. Thank you.

Yes. What I meant to convey with that remark is that we are having conversations with multiple hyperscalers. When we talk about hyperscalers, and this may go back some quarters, we're really speaking about the top 10 hyperscalers. So, it goes beyond just the three public clouds, if you will, to the other, the FAANGs, et cetera. The quantity, meaning the number of those players that we speak to and the quality of those conversations have improved. I mentioned as well that we're now not just experiencing testing in some of those environments, but some commercial discussions as well. All those are positive signs, if you will, that lead us to believe in a design win this year.

Thank you, Noah. Next question, please?

Our next question comes from Wamsi Mohan from Bank of America. Please go ahead. Your line is open.

Hi. Thank you so much for taking my question. This is Nadi on for Wamsi. I saw that in the quarter your headcount declined slightly. Could you explain why that is the case? And are there any implications on how the sales capacity will be like moving forward?

Yeah. Well, we continue to invest in sales capacity, given the opportunities that we see in front of us. What I would say on the number, which does tend to fluctuate quarter to quarter, is we have been increasingly focused on two things that work against a consistently higher number, and that is that we're working on overall quality of new customer adds and secondly on putting more focus, if you will, than in the past on expansions in existing accounts. So, I wouldn't read too much into it. I do expect that with over 12,500 customers now, we're in the back half, if you will, of customer acquisition, and we should start to see moderating those numbers.

Yeah, just a quick follow on to that, too. Look, the customer acquisition count was consistent with what we saw a year ago in Q1. We're also pleased, to Charlie's point, about the quality of new customers acquired as we continue to increase our penetration of Fortune 500 customers.

Thank you, Nadi. Next question, please?

Our next question comes from Mike Cikos from Needham. Please go ahead. Your line is open.

Thanks for the time, guys. I just wanted to circle up on One and Flex growth just to get a temperature, but how did adoption or growth of One and Flex in the quarter track versus your internal expectations? Is there any way to think about customers which are adopting whether it's primarily coming from existing customers or new logo lens? Anything there would be beneficial.

Charlie, why don't you take adoption first, your view on adoption. I'll hit the comparability.

Yeah. The Q1 adoption was perhaps a bit lower than we might have expected from a Pure seasonality standpoint. But I would say that typically, first of all, Evergreen//One, interestingly, is adopted pretty much equally by both commercial — what we call commercial enterprises midmarket as well as by large enterprises. In the average quarter, certainly all through last year, we had a mixture of commercial and large deals in each of the quarters, which drove a lot of the number. This quarter was characterized by not having a particularly large deal in Evergreen//One compared with last year. So, we think of it as being just an aberration. As we mentioned, we're still expecting around $600 million for Evergreen//One this year.

Yeah, I'll double down on that a little bit. Just in terms of utilization on Evergreen//One, that continues to be very strong in terms of deployment of our infrastructure, following a very strong year last year. But as a reminder, we began providing an annual view of TCV sales for storage-as-a-service offerings to help provide insights into our annual revenue growth expectations, and providing quarterly updates to monitor progress against our annual expectations against the $600 million in TCV sales that Charlie alluded to. In Q1, our TCV sales for storage-as-a-service offerings were $56 million, which is a bit of a slower start to the year than we would have liked. However, as we progressed throughout Q1, we saw strong demand and pipeline build, including large opportunities for storage-as-a-service offerings. Consistent with our original FY '25 forecast, we continue to expect 50% growth of our storage-as-a-service offerings, and we'll continue to provide quarterly updates as we progress through the year.

Thank you, Mike. Next question, please?

Our next question comes from Asiya Merchant from Citigroup. Please go ahead. Your line is open.

Thank you for taking my question. Great results. In the last quarter, when you discussed the fiscal '25 guidance, I recall you mentioned some insights regarding product versus subscription services. Could you elaborate on how we should view product revenues? They seemed a bit stronger than my predictions, while subscription services appeared to be slightly lower. It would be helpful if you could clarify the outlook for both categories. Thank you.

Sure. Well, if you think about a transaction, any individual transaction, it will be generally one or the other. In the case of a CapEx transaction, there is a subscription component to it, but it's much smaller. In the case of an Evergreen//One transaction, there'll be no immediate revenue associated with it, but of course, a higher amount of subscription on an annual basis. From our standpoint, we want to provide the customer whichever part they prefer. While we have some incentives for the sales force to sell the subscription, we certainly don't want to lose a deal with a customer that wants to buy on a CapEx basis. And so, all else being equal, if we see slower uptake on the subscription, we'll still win the deal on CapEx; we'll see a somewhat overachievement on the revenue line in the immediate term. Conversely, if we see greater uptake on Evergreen//One, we'll see a lower effect on revenue in the immediate term. So that's the right way to think about it. We've given a rough number before that about 70% of the value of a subscription deal would go to CapEx or a CapEx deal.

Yeah. I think just taking a step back from what we discussed last quarter to kick off the fiscal year, I think it is better to look at this from a total revenue standpoint in terms of the bridging. That bridging, assuming we achieve the expectations of $600 million in TCV sales for our subscription-as-a-service, would put the total revenue around mid-teens growth, and that thesis has not changed. I think when you start looking specifically at product revenue, I would tell you that we're quite pleased with what we saw in Q1 from a product revenue growth rate perspective.

Thank you, Asiya. Next question, please?

Our next question comes from Nehal Chokshi from Northland Capital Markets. Please go ahead. Your line is open.

Yeah. Thank you. And strong results. Congratulations. Not sure if you covered this, but given the strong results, above-guidance results, why are you not changing your fiscal year guidance parameters both on the top- and bottom-line?

Yeah. Thanks for the question. And yes, we are definitely pleased with the strong start to the year, as you point out with top line and operating profit outperforming. But overall, at this stage, we believe reiterating our annual guide is appropriate. When you think about it from a revenue lens, Q1 performance provides us with increasing confidence for the remainder of the year, but considerations also include that Q1 is generally our seasonally slowest quarter in the year. Also, as we previously discussed, TCV sales expectations for our storage-as-a-service offerings have an impact on our annual revenue guide expectations, and we'll want to see how that plays out as well. When we think about it from an operating profit lens, the strength was driven by a few key factors, including revenue outperformance, gross margin strength, and while our hiring is strong, we're also seeing some financial benefit due to workforce alignment adjustments during Q4 of last year. While reiterating our annual operating margin guide, we're also planning to increase investments in some key areas in engineering and go to market, specific to AI and the opportunities we see specific with hyperscalers, providing our Purity and DirectFlash technology for both their high performance and bulk storage environments.

Thank you, Nehal. Next question, please?

Our next question comes from Simon Leopold from Raymond James. Please go ahead. Your line is open.

Thanks for taking the question. I don't want to ask such a backwards-looking question, but essentially, the 10% customer disclosure for the prior fiscal year makes me wonder what's your customer concentration like now? And how do you think about how that concentration might evolve over the next year or so? Thank you.

Yeah, thanks for the question. We did disclose in our 10-K that we filed last year that we had a 10% customer. Mix in terms of concentration remains relatively consistent, and we did not have a 10% customer in terms of revenue concentration this quarter.

And your outlook?

We don't provide outlook on concentration.

Thank you, Simon. Next question, please?

Our next question comes from David Vogt from UBS. Please go ahead. Your line is open.

Great. Thank you, guys. And this is for both Kevan and Giancarlo. Maybe you guys can talk about the TCV sort of trajectory that you expect this year? Obviously, you disclosed a number of $56 million in your prepared remarks. I'm just trying to think about how the timing of this sort of ratable or the subscription business starts to flow into the business as we move through the rest of this fiscal year. Is there the expectation that there is a sort of an acceleration of demand as maybe the storage market gets a bit healthier as we move into the latter half of calendar '24 into early '25? Thanks.

I want to clarify because your question was somewhat unclear. Regarding the $56 million related to new Total Contract Value, we are forecasting $600 million for the full year and are committed to that estimate. We believe demand will pick up, particularly as we think any slowdowns in the first quarter are more about timing. In terms of overall revenue for the year, we now have total subscriptions from Evergreen//One, Evergreen//Forever, Evergreen//Flex, as well as our Portworx products and Cloud Block Store, which is currently in the 40% range. We expect it to approach 50% this year and continue to rise above 50% next year and beyond. We are a seasonal business, and we anticipate both product and service sales to grow throughout the year. As such, we expect the share of subscription revenue to keep increasing.

When we're thinking about TCV sales for storage-as-a-service offering, we've talked about contract durations associated with those TCV sales, and we haven't had any meaningful change around three years is what we've disclosed.

Perfect. That was my follow-up. Thanks, Kevan.

Thank you, David. Next question, please?

Our next question comes from Aaron Rakers from Wells Fargo. Please go ahead. Your line is open.

Yeah. Thanks for taking the question. I apologize if I missed this earlier. I'm jumping in on a few things here tonight. I'm curious about the AI discussion. There seems to be just inflection around AI infrastructure, maybe vectoring for AI training, driving a lot of demand for increased capacity points for SSDs and just flash storage. I'm curious if you can give us any kind of updated thoughts on the roadmap of your DFM and whether or not you're starting to see some kind of an acceleration or inflection of just demand for these higher and higher capacity points?

You bet. Let me start. I'll talk about the demand for AI in training environments. As we look at training environments, we have a good understanding of what the average training environment looks like. The revenue associated with an AI training environment, after accounting for the costs of GPUs, networking, and racks, is in the 5% to 10% range. This may be slightly less than market expectations, but we are confident that the total storage related to AI training in the near term falls within this range. One reason for this is the high cost of GPUs. Over the last 12 months, we estimate that about $1 billion was spent on storage specifically for AI training environments. While this might be somewhat lower than anticipated, our data on this topic is strong. This provides a sense of the current scale. I'm going to pass it to Rob.

Go ahead.

I'm going to pass it to Rob on the current status of our technology.

Yeah, Aaron, this is Rob. And I think the specific to the first part of your question, asking about the training environments, these large-scale build-outs. What are we seeing in terms of the demand environment for SSDs and in particular, the larger capacity SSDs? Certainly, most of these high-performance training environments, whether they're provided by us or built by others, either the customer or other vendors, are generally residing on SSDs. We are seeing some of the larger SSDs—larger-capacity SSDs being deployed in some of these very specific environments, specifically because of the efficiencies and power limitations that exist in these environments. The number of GPUs that are being built out, frankly, are sucking up all the power out there. Most of these customers are struggling to power these build-outs, and the efficiencies that the larger-capacity drives enable are a great validation point for our long-term strategy. Now, on the flip side of this, what allows us to provide these large SSDs with all of the commensurate benefits in terms of performance, power savings, and efficiency is our DirectFlash software. Without that DirectFlash software, the large-capacity commodity SSDs simply can't provide these benefits. As a result, they're inherently less efficient, they come with significant trade-offs and they're very, very difficult to use, which is why we see them only being deployed in very specific environments.

Aaron, if you'd like to do a follow-up, we'd love to have you get back in queue. I think we might be able to get to it today. So, thank you. Next question, please?

Our next question comes from Mehdi Hosseini from Susquehanna International Group. Please go ahead. Your line is open.

Yes, thanks for taking the question. Given your Q1 performance, especially with revenue mix and your fiscal year, it seems to me that product revenue should be flat for fiscal year '25 and the growth primarily driven by services revenue, and given the higher gross margin associated with services, that should help with some gross margin expansion. Is that the right frame of mind for the fiscal year?

Yeah, I think directionally you're thinking about that right.

Thank you, Mehdi. Next question, please?

Our next question comes from Eric Martinuzzi from Lake Street. Please go ahead. Your line is open.

Yeah. I took note of the incremental increase there in your Fortune 500 customers, four new Fortune 500 customers. So, curious to know across your larger new customer growth, are you seeing any commonality in the use cases for why these new customers are showing up at your door?

Yeah. Honestly, it's the fact that we are able to provide them a solution that is, first of all, more consistent across the wide variety of use cases they have, block, file, and object, that's managed by a common operating environment and the same single management system. We're able to do so with a product that's simpler, more reliable, and takes less space, power, and cooling. Adding those things together, it's a pretty good value proposition that's hard to beat.

But there wasn't anything like the AI training environment that opened the door or something like that?

I'd say there are multiple entrances, and yes, sometimes it might be the AI environment that gets the customer interested, but honestly, it really is the new positioning of the company to be able to satisfy a broad range, which helps in the AI environment because one way of building your storage environment for AI is to make a copy of all the data you think you might want to use and put it onto a brand new array connected to your AI environment. What we're proposing is don't bother doing that. Upgrade the arrays that you have in place and make them available and accessible to your AI application inference environment. It's a strategy that costs less, enables more access to real-time data, and simplifies their environment. Again, I think it's the total value proposition, and you're right, the entrance in might open up through different doors or windows in the customers' environment, but I wouldn't say there's one big theme out of this.

Thank you, Eric. Next question, please?

Our next question comes from Krish Sankar from TD Cowen. Please go ahead. Your line is open.

Thank you. Just a question on the AI and hyperscaler opportunities. Charlie, I understand you'd look at them a little bit differently. But thinking back to the rev recognition you guys had with Meta, it seems like each one had a little bit different margin profile in it. So, just curious when you start looking at these new opportunities, particularly in the hyperscaler side replacing disks, do you have any sense of what the financial model for Pure will look like there? I imagine it's going to be a much different dynamic than just selling arrays into enterprises. So, I know it is early days, but anything high level you can talk about how dealing with these much larger customers with different business models than what you're currently seeing would play into the current framework of Pure? Thank you.

Yeah. Thanks. I'm going to take the two parts separately: AI and then hyperscaler. In the case of AI, we've now largely made changes internally to the company where that'll flow pretty much at the normal gross margins of the company. They may be a bit lower, but overall not going to affect the margins of the company as a whole. I would say that on the hyperscaler front, we're still evaluating different ways of working with these hyperscalers, and they won't be the same. They'll be different. Until we actually get to the point where we understand exactly how they want to buy and how we can provide, we can't really say right now, but once we do, we'll certainly have a model that we provide you all with at the earliest possible date.

Thank you, Tim. It looks like there is one more question. Asiya, thank you very much for getting back in queue. This is your follow-up question.

Okay, great. Thank you again. I apologize if it's been asked, but the upside in the gross margin on both, if you can just unpack that and how we should think about it in the outer quarters? Clearly, Q1 was soft, and you're expecting revenues to increase from here on just seasonally as well. How should we think about product gross margins as well as subscription gross margin in the outer quarters? Thank you.

Yeah. This is Kevan. When we think about the strength that we're seeing on gross margins, both product and subscription, I would say that we wouldn't be expecting any meaningful changes as we progress through the year. I do think that we've commented on being aggressive in pricing, with our price-performance solutions, whether that's our E family or FlashArray//C, and we'll continue to do that. But overall, we would not expect significant or meaningful changes, one way or the other.

Thank you, Asiya. Before we conclude, I think Charlie has some concluding remarks.

Yeah. Thank you all for joining us again on today's earnings call. I think as we've discussed, our platform strategy continues to lead the industry's transformation by enabling businesses to embrace this change with AI with a unified, versatile, energy-efficient platform that really allows them to look at their data centers and cloud environments as a single cloud of storage. I do look forward to meeting with you all, our customers, partners, and investors at our annual Accelerate conference in Las Vegas next month. We're going to be also on some roadshows throughout the globe where you'll be able to learn more about the investments we're making to continue to lead innovation in the industry and hear more about the Pure Storage platform. So, hope to see you all next month. Thank you so much for joining, and talk to you later.

That concludes the Pure Storage first quarter fiscal 2025 financial results conference call. Thank you for your participation. You may now disconnect your lines.