Everpure, Inc. (P) Earnings Call Transcript & Summary

Great. Thanks, everybody, for being patient and sticking around. Welcome to Day 3 in the final day. I'm David Vogt, the UBS, Hardware and Networking Analyst, and we're excited to have Pure Storage with us today. From the company, we have Rob Lee, Chief Technology Officer in the audience, milling about is Paul Ziots. So if you have any other questions, you can reach out to Paul after. Just quickly before we get started, let me read the pure statement, those statements made in these discussions, which are non-statements of historical fact are forward-looking statements based upon current expectations. Actual results could differ materially from those projected, due to a number of factors, including those referenced in Pure Storage's most recent SEC filings, on Forms 10-Q, 10-K and 8-K. So with that out of the way, thank you for joining us.

Thanks for having me, Paul -- David.

So let's pivot a little bit from what we were going to talk about. So obviously, you guys reported earnings last night. And I think, from my feedback last night from investors and from the call and from talking to the company last night, I think investors want to kind of understand the demand environment kind of -- I know that's not necessarily your direct purview day-to-day, but let's start there. Like what are you seeing from a product versus subscription perspective? Because I think that surprised some investors last night clearly, and what do you attribute sort of that mix or that dynamic to today? I know on the call, you talked about maybe it's not macro, but we heard from other companies. It certainly sounds like maybe macro might be a contributing factor. So maybe we'll just start there.

Yes, absolutely. So when we think about the demand environment, first of all, we see robust demand. We see robust demand, generally across the Board, we called out a couple of areas in the portfolio, certainly FlashBlade, record performance. the E Family has been tremendous. And then certainly, a lot of the dialogue and really the focus has been just the significant outperformance, even of our elevated expectations around Evergreen//One, right? Our subscription and consumption services. And so -- and then, I think the question becomes, hey, so what do we believe is driving the, I would say, excess or increase in demand for the subscription and consumption services, relative to perhaps the more traditional buying models. And I think you have a couple of factors there. I think number one is, as we've said really over the last several years, we believe this is where we're able to deliver the most value to customers through those services. But it's taken some time. It's taken some time to educate customers to that value. It's taken some time to educate our sales teams, our partners, in terms of how to articulate those values. How to sell a service rather than a product sale. So you're starting to see some of that come through. Number 2 is customers are starting to see the value, right? And so you're getting customers electing for those services rather than the product sale. And then number three, certainly, and we've seen this in prior periods, where there are tighter conditions. There's a set of customers that because of cash constraints and things of that nature, really prefer the optionality that the consumption services offer. So when you kind of net that out, we're definitely seeing increased demand in terms of the mix for the subscription consumption services relative to the product. And I think that's really been the new part of the dialogue, certainly that we've had last night.

So have you -- has the company -- I know you've been stressing the value of this consumption model for customers for quite some time. Has that -- has the conversation changed, as we've moved through, let's say, this calendar year given the macro because I think you disclosed last night, obviously, you had expectations for what that would look like within the overall revenue pie. And obviously, it's considerably outperformed. So as we move through the year, have those conversations changed, and where we're sitting today, just given the general macro, I mean, would you expect that sort of dynamic to be relatively consistent as we move forward? Not talking about guidance, but just sort of a buying pattern behavior going forward. Is that a reasonable proximity for the new normal effectively, at least in the near term?

Yes. So if we look at how have those conversations progressed over the year, certainly getting stronger, again, partially because realization of the value we're delivering, but also we keep making the services better, right? If you look at how we've innovated in the Evergreen//One service over this year. We've added a number of additional SLAs and really guarantees and [ teeth ] to the service that customers are seeing, "hey, this being able to offload this risk and depend on Pure for this, gives me a ton of value, paid power and rack, right?" The idea that, "hey, we're going to go not just deliver you the storage, but we're going to go take care of our power and space requirements as a part of that." Again, so we've been improving that service that helps. Customer realization that definitely helps. And so then I think the first part of your question was, "hey, so what did we see, as we navigated throughout the year? Why are we having this conversation now, as far as big inflection points?" So if we take you back to the beginning of the year, right, when we kind of looked at the full year fiscal kind of outlook, a couple of key assumptions that went into that, right? #1 was outlook in the macro just overall. And the #2 was our expectations for a strong Evergreen One performance throughout the year. If we look at how that's played out throughout the year, the macro assumptions really not necessarily changed right? We kind of said, look, don't expect huge improvement, don't expect huge worsening. That's kind of how it's played out. So really a nonfactor. If we look at Evergreen//One performance, right, we had contemplated a strong performance from Evergreen//One in the original guide. As we go through the first half of the year in Q2, it became clear that we are already seeing strength in excess of what contemplated at the beginning of the year, that's where we kind of said, look, we're now expecting about a 1 to 2-point headwind, against the full year guidance, given the strength we saw in the Q2. As we got through Q3 and a little bit of Q4, what we saw was strength even well in excess of that elevated level of expectations. And that's now where we're having the dialogue of saying, "hey, we're now thinking that the full year kind of revenue impact is going to be more like a 3-point impact." Again, a great thing from the point of view of growth of Evergreen//One services, but certainly bringing the near-term impacts into the conversation.

But does that mean -- so obviously, I think customers are incredibly excited about consuming Evergreen//One and the consumption model, is there a risk, that becomes a more prominent part of the selling motion, and that creates a bit of a revenue challenge just generically, like obviously, that revenue will build and stack as you add customers to it. But have customers kind of learn the value they appreciate what you offer and maybe CapEx becomes less of a priority in the intermediate term or the medium term?

So I think the question is, "hey, is customers see the value, do we see that potentially growing, that would be great, right? And we do see -- and then I think the other part of the question, I didn't...

But at the expense of product, in terms of traditional CapEx.

And as it grows and ramp up, the rate of growth accelerates, yes, you are going to see that near-term impact. But again, I take you back to strategically, right? This is where we really want to focus on driving the maximum value through those services. I think the other part of your question, the previous question was, "hey, so what are we seeing now outlook? And staying away from specific outlook and specific comments on next year, we are seeing strong momentum, right? And that's kind of reflected in the comments, we made last night and really our outlook for the rest of the year. And so the momentum continues.

And then maybe just sticking with sort of demand trends and pricing. I know there was a conversation about pricing being a little bit more competitive than maybe I think people had anticipated. We've heard that from other companies that the market, a bit more challenging. What do you -- is that macro that you're attributing it to, a handful of vendors just being a little bit more aggressive? Like how are you thinking about -- I know you talked about being competitive to win business in certain parts of the market, where you're not particularly dominant today or strong in today. Just the natural sort of dynamic in the marketplace or anything sort of crop up that maybe was a contributing factor to a more aggressive pricing backdrop?

No, nothing specific to call out. I mean Storage has always been a competitive space. We've got a large incumbent competitive set that is trying to hang on to their installed base. And so it's always been like that, nothing specific to call out. . Other than to your point, we are going to get aggressive, specifically in the areas of disc takeout, what we're pursuing with the e-family. And so, if you look at kind of the overall pricing environment, gross product margins, certainly, we're running a little bit hot there now. Major benefit -- that's really the major benefit of that is from that is coming from our Purity Software Flash Management benefits. That gives us the flexibility, and we're going to go use that to get aggressive, specifically in the E family, where we're competing for that disk footprint alone in Flash.

So when you think about being competitive total cost of ownership, this take out that model, I think Kevin in the past has talked about, potentially using, you just referenced using gross margin as a tool effectively, to be more competitive, so when you think about sort of running this business over the medium term and the long term, obviously, product gross margins. I would imagine the messages should trend back towards kind of the target model. Is that the right way to think about where gross margins on the product side at least should end up, over the medium term, 72%, 73%, 74% is probably not optimal from your perspective when you factor in sort of the growth outlook for the company?

Yes. So to your point, 72%, 73%, 74%, we're definitely running hot there, right? Now we are going to be aggressive and that's going to bring that down a little bit in the E family. That said, right, if we look at the mix of business, we look at where E is, it's going to take some time, right? It's going to take some time for that to grow as a meaningful percentage. If we look at the rest of the business, where I'm not winning on price, I'm not losing negative to lose a deal on price. It's not going to benefit me to erode that, right? And that's again, that's really just coming through from the value I'm getting from the Purity Flash Management. So we do want to kind of bring that down. We want to do that and open up more market with the E family, but it's going to take some time for them to have a material effect on the blended.

So you think -- I'm going to stick on the E family for a quick second too. Given the launch of the E family, whether it's Blade or Array, maybe can you kind of qualitatively talk to the feedback right now? I know you referenced it was relatively strong, since launch. Where are you from a total cost of ownership perspective? And relative to what maybe the more mass market hyperscalers are looking for? Can you maybe just kind of give us a sense for how you think that progresses from a price perspective, relative to legacy disk systems? Like are we at that inflection point, where the conversations are much more robust than 6, 9 months ago? Or is it still sort of on the come story in that market?

Yes. So let me answer that question from the point of view of Enterprise Systems in the Enterprise Market and then we're getting to hyperscale, right? So if we look at the Enterprise Market, where we're really competing against hard disk, OEM vendors. We're at acquisition cost parity today?

And that's the nearline guys.

And there is the nearline guy, exactly. So the enterprise customers that are buying kind of looking at, "hey, do I go with a nearline OEM product or do you go to the Purity family. On an acquisition cost basis, we're going to go be at parity today. We're there, right? On a TCO basis, where you're now factoring the benefits of power, space, labor savings, longevity, the fact that you're not replacing equipment all the time, very advantaged there. Depends on specifics, but could be like 20%, 30%, 50% advantaged. And so that's the enterprise guys. If you look at where we are with the hyperscalers. The hyperscalers, obviously have a much more optimized cost structure. They've built those environments very specifically, a little bit further out. But that's where our road map specifically driven by Flash density and technology, as we go from the 75 terabyte modules, which we just shipped to the 150s next year. That's really where you're going to start to see that crossover point.

Got it. And maybe just one final question on the quarter. Obviously, you referenced a fairly large $41 million contract, that will get shipped and recognized in fiscal '25. Can you kind of just walk through qualitatively, obviously, the road map to win that transaction and ultimately, why the customer prefers more measure scheduled shipments. Yes, measured schedule shipments sort of dynamic as opposed to book and ship effectively?

Yes, absolutely. To your point, the majority -- the vast majority of our product sales really are book and ship, in matter of days and weeks. With this particular customer, as you'd imagine, a deal of this magnitude, we've been working for some time, definitely had contemplated in the forecast, really looking at back half of the second half of this year. Really, what happened is as we worked that deal got to the kind of closing stages of that, working configurations and getting to that level of detail, became clear that this was part of a staged 5G rollout, right, from this telco customer. I mean lots of other logistics involved, "hey, I've got buildings coming into place at certain times. And so, not uncommon for the Telco industry, but again, just deal this magnitude and [indiscernible] for us. And so that's really why we called it out, nothing ...

And was that a new Pure customer? Was that an incumbent relationship? Or is that a basically a competitive win from an existing vendor? Or this is a new use case for that particular customer? And it's obviously 5G related? Just how did that come to be?

It was an existing customer. So, familiar with the Pure technology. I would say every win is competitive, right, but definitely existing customer.

Got it. All right. So obviously, you are the CTO, maybe I want to spend a little bit on the technology road map. I think at your Analyst Meeting/Industry Event over the summer, you talked about the road map in terms of why your scale, your road map more of the growth dynamics in terms of your ability to grow faster or add capacity faster in a more cost-effective way, relative to some of your competitive peers in the industry should help you win out over the next couple of years. Can you maybe walk us through sort of the competitive advantages, as you see today with sort of your use of run-in, limited use of DRAM, obviously, how the operating system Purity kind of plays into that? And what the road map looks like into next year and beyond, not specific about numbers, but and plus the use of QLC and how that kind of plays into your strength and your competitive advantage going forward?

Yes, absolutely. So I think this comes down to really IP -- software IP that we built going back to the beginning of the company. We took a very different approach, than everybody else in the market, when looking at Flash. We saw Flash for what it is, which is a very different media than disk, requires a different approach to go user, requires a different software approach. Every other vendor in the market kind of said, look, I've got software that's written for disk, the hardware manufacturers, the SSD manufacturers. They're going to go make Flash look like disk. I can just go slot that in. That's kind of the easiest path forward. And that progressed for a while, right? And in the early days, I would say, we had advantages through our approach. In terms of efficiency, performance, reliability and so on and so forth. Now what's happened is that over time, those advantages -- the magnitude of those advantages has just increased, increased, increased, increased. And some of the question is, okay, well, why is that, right? Normally, you would expect in technology differentiation gap to compress, not increase. What's happened is that as the memory manufacturer, raw NAND, right, has become more driven by the consumer market. It's pushed the road maps down a path of denser and denser chips, that are harder and harder to work with, right? We've built the software from day 1, to go work with that Flash natively. That allows us to make much more use out of the same chip than anybody else can get...

Is less complex, the read, write, reading?

Well, and we can get more use out of it, right? We can get more -- we can put more data on the same chip than SSD can. And so what's happened is, over time, as the chips have become more complex, the SSDs are having to do more and more work internally just to keep up. And frankly, they're at a breaking point. They're just bursting at the scenes in terms of complexity, the amount of work that's required. And to your point, the resources that are required to do that, the DRAM, the processing power and so on and so forth. A good data point or a point of validation for this is, if you look at -- on the market, you look at 15.36 terabyte SSDs, you go to Amazon, Newegg, whatever go buy one of the SSDs. If you tear it open, you look at the chips in there. You may find the same chips that are on our 18.3 terabyte DFM. You're going to find the same number of chips. Well, then how is it that the SSD is offering 15.36 terabytes and we're offering 18.3 terabytes, it's because of that software advantage. We're able to squeeze out a 20% more storage out of the same number of chips. That gap, that advantage is just increasing with every generation of NAND.

Does that resonate with your customers right now? So obviously, you've taken a considerable amount of share, since the introduction of the strategy. And I know what Jon-Carlos talked about, I think he said publicly like we want to double share again effectively, go from, I think, correct me if I'm wrong, all like 7% share to like 14% share or somewhat of magnitude. What do you see is the impediment to achieving that, given sort of the technological sort of road map that you see over the next couple of years? Is it maybe NAND prices increase and that maybe makes that gap a little bit less relevant effectively, where customers are like, wait a second, maybe we're not going to take out disk, and we'll continue to go with a lower cost solution. Just, what are sort of the hiccups or the hurdles that you're worried about of achieving that?

I would say the main hurdle or the main challenges, something we talked about earlier in the conversation, which is storage has always been a competitive space, right? Taking share, means winning it from a competitor, as competitors are going to do everything we can to go and protect that. Now that said, right, we don't believe that it can match up to us in terms of product innovation, in terms of value. And so -- but that's going to be your main hurdle. I would say, if we were having this conversation 2 years ago, I would say the other hurdle is, we didn't at that time, have as complete a portfolio or set of offerings, right? We can go into an enterprise customer and say, hey, for this database workload, this AI workload, or this thing, we can be the best thing for this workload. And we might win that workload. But without being able to go in and say, "Hey, look, we can sweep the floor we can satisfy your needs. It's a little bit of a barrier to becoming a strategic vendor, right? We're now at the point where we've got that platform position. And we believe that's going to open up a greater wallet share in these larger enterprises.

So what are you seeing are sort of the primary use cases for the success of the e-Series right now? Like when you go into a customer, what are you solving for them? Obviously, mass capacity, lower cost, but where are you seeing sort of that traction? Is it just strictly disk replacement or maybe a customer might have been a little bit more amenable to a slightly more robust solution or offering and the E makes more sense at this point? Just kind of maybe walk us through that.

Yes. So maybe 2 parts to that question. What are we solving for them in the major use cases? I think what we're solving for them is exactly what you mentioned, right? A much smaller footprint, much easier to manage, much lower power consumption, all the benefits of Flash at the cost of Disk. If you think about it, there's no reason why anybody out there says, "I would prefer Disk over Flash, even if it was more, the only reason that this exist is has historically the cost. And so that's really what we're solving for them and the headache of managing the failures and so on and so forth. And so that's really been the feedback. Now I think the first part of the question is, hey, major use cases, where are we seeing the traction? Where are we seeing it deployed? I would say that a number of use cases, really broad-based, right? Certainly, bulk data, kind of content repositories, medical imaging, a whole host of things, data protection, archiving. And it's not surprising, right? If you think about it, the disk environment that we're going to replace, these have historically been the large parts of the estate that have just been shoved in the corner and kind of less cared for, most in need of modernization. Now when you look at it through the lens of technologies like analytics and AI and "hey, how do I go get value from that data that used to be a cost center." But now I can actually really value from them [indiscernible]. Well, there's no way I'm going to be able to do that if it's stuck on slow spinning rust. Well, now it's got a vial outlet to go and modernize that solution, get the benefits of Flash, save a bunch of power, space and get the performance. That's really the set of conversations we're having.

Maybe if we can spin up, no pun intended, to more higher use cases or more robust use cases. So obviously, every company this week has been talking about, in some way, shape or form their AI strategy Obviously, it's a little bit different for you, right, in the sense that you've got a long-standing relationship with a hyperscaler for a different capacity, right, for different capacity than I think what people in the Storage industry have done historically, like I don't think Dell and NetApp have a position that you have had. Maybe can you kind of talk to where you see the opportunity set next? Like you talked about some small AI wins last night on the call or not small, but a number of AI wins on the call. Where are you seeing the most traction in AI? Because it sounded like me, it was not more -- it wasn't more like ChatGPT, open AI kind of dynamics to start. It's more machine learning or autonomous vehicle learning, et cetera. So maybe kind of walk through what you're seeing today and how do you think the model or the industry evolves for you?

Yes, absolutely. And so when we think about the AI opportunity, broad strokes, right? I think about it in terms of the training opportunities the inference and deployment of the AI technology and then just the broader data flows. And how does the data come in and out of the system? How is it an all of the kind of downstream effects of that. And so in the trading environment, I'd say that we do see more coming from maybe the more traditional AI use cases. So self-driving cars, I gave an example of, radiology, things of that nature. I think a lot of the large language model training is happening in the public clouds. But we do see training infrastructure. That's really been historically where we've played in the AI market. We're now seeing wins in the other 2 spaces of AI opportunities as well though, right? In the inference environments, where you've got real-time data that's coming in, operational data that needs to be connected to these inferencing systems, being able to serve those, whether it's a Flasharray, being able to set up the applications that are kind of tying this all together. We're seeing that driving a number of Port Works wins because it's just so highly aligned the cloud-native ecosystem. And then as well in the broader data environments, right? So we had a win, for example, with a company that's using E, leveraging a fair amount of E to go and modernize their historical data so that they can use that to bring into their training of ongoing models, and I think this customer is in the kind of the security space, right? And so you're seeing wins across all 3 of those categories. And so look, I mean, I think it's overall, we're well positioned for the AI trend. But I'd also remind you that it's an important use case, but overall, it's a not a dominant segment for us.

Got it. Let me just make sure we don't have anything from the audience. I kind of touched on it earlier, but I still think, maybe, we'll just go back to it. I know we have a couple of minutes left. And there's more that I want to ask. But actually, a couple of questions came in on this. Let's go back to subscription and product for a second. There's been some debate that even if we make the adjustments for subscription being much stronger and that $41 million contract getting pushed out into 2025, it still suggests that maybe there was a little bit more macro softness or demand softness than maybe like it sounded then on the call. Like is there anything else kind of under the surface that maybe impacted the quarter that we didn't touch on?

Nothing I'd call out because, again, if we go back to kind of entering the year, full year kind of outlook, even contemplating Evergreen//One performance at that point, we kind of said, look, mid- to high-single digits, call it, 7, 7.5 points. Against that, we've said, look, Evergreen//One just tremendous 3 points there. You've got 1.5 points impact of the Telco deal, and you kind of stack that up 3 points from Evergreen//One, 1.5 points and then the 2.5 resultant, you're kind of into that 7-point range. So right in the fairway of what we saw.

Got it. I'd be remiss if I didn't ask about Meta. Any thoughts, updates qualitatively kind of how you're thinking about the relationship and kind of proof points, touch points with them, haven't heard a lot in the numbers as of late. Just would love to get an update on what you're seeing in hearing from Meta?

Yes, relationship is great. We continue to work very closely with them, and they're getting great value out of the solutions that we've deployed there. And to be clear, we haven't commented and given an update on the specific deployment, which they called out in the RSC environment. We do have sales to other parts of Meta and not just the kind of -- not just the corporate piece, but just like we don't comment on sales to other specific customers, we're not kind of providing commentary on sales to other parts of it. But relationship is great, and we're very, very close to them.

Got it. Maybe just one final question for me on AI. So when we talk to the hyperscaler is a little bit tough to get information out of them, as you can imagine, they have stressed to us repeatedly. And this is the 4 U.S. hyperscalers, maybe we can throw in a little bit like Tier 2 Cloud guys as well, that performance, but cost keeps coming up over and over getting, I think you showed that slide at the Investor Day, right, like $0.05 per gigabyte of storage. Are they amenable to the total cost of ownership conversation today, meaning to get to $0.05 or some metric that's even close to $0.05, it seems like a little bit of a challenge. Now I would say, since Jon-Carlos said this to over the summer. It just seems like a bit of a stretch in the near term. But does the AI -- the growth in AI maybe change that conversation given how important the AI investments are to them. And so this is not just general purpose workload applications. This is much more robust, requires a better solution than what they have there. And so maybe that $0.05 per gigabyte isn't the right metric. Maybe the gap is more narrow today than it was 9 months ago. I mean I know that's more of a theoretical question, but it seems like that would encourage them to be much more amenable to working with someone in your capacity in a much deeper relationship.

Yes. So what I'd say is in the hyperscalers just like in any firm, you're going to have -- it's not just one type of storage, right? They're going to have multiple different types. And certainly, the AI effect is going to increase performance pressure on some amount of that storage. And to your point, that's going to lessen the focus on getting to that exact [indiscernible] , it becomes a price performance conversation, as opposed an absolute price conversation. You have that effect. But we're really focused on the broader picture, right? If you look at just the total amount of capacity deployed their total spend on storage, if you will, the vast majority of that is still going to sit in bulk storage, sitting on disk. That's where that kind of $0.05 mark, which we called out was. And really focused on getting to that market. And that's where we believe our road map right, where we have clearly in our sights really gets us there. And if you look at the evolving and really very productive conversations, we're having with the hyperscalers, it's really centered around the TCO benefits we can deliver against that bulk storage. Said more simply, if you can hit that mark against the bulk storage, you can go deliver more performance, up the stack.

Got it. I think we're out of time. Rob, I want to thank you for joining us, Paul, thank you for joining us. I'm sure you have a full day of meetings. So thank you again, and hope to see you again next year. And thanks, everyone, for sticking around on the last day, and have a great safe trip back, and we'll talk soon.

Great. Thanks, David.

Great. Thanks, Rob.