Everpure, Inc. ($P)

Good afternoon, everyone. Thank you for joining us here, day 2 of BofA's Global Tech Conference. I'm Wamsi Mohan. I cover IT hardware supply chain here for the bank. Delighted to welcome Everpure to the conference today. We have the CFO, Tarek Robbiati. Tarek, thank you so much for being here.

Thank you, Wamsi, for having me, and thank you to BofA for hosting us.

Yes, excellent. Well, before we get started, I do have to read forward-looking statements here. Statements made in these discussions, which are not statements of historical fact are forward-looking statements based on current expectations. Actual results could differ materially from those projected due to a number of factors, including those referenced in Everpure's most recent SEC filings on Form 10-Q, 10-K and 8-K. And with that, I think I'm getting the green light from Sandra back there. They jump into Q&A.

So Tarek, this has been a period of really strong growth across the board for most companies. You guys had a pretty good solid quarter 2. As you think about what has changed maybe in the -- in what your customer conversations are, maybe we can start there and then dive into more questions.

Yes. So unquestionably, the market has woken up to the need to modernize infrastructure in the day of -- in this new AI era. And we see a lot of demand across the board in the commercial segment, in enterprise customers and this globally, North America, EMEA and Asia Pacific. So it's very encouraging to see that, and this translated into very strong growth across our business, and we posted a very solid Q1 result, which came on the back of a very solid Q4 result. You know that we generate a lot of orders in a quarter. Those orders materialize in revenue when we ship and possibly in the quarter, sometimes in subsequent quarters. So we entered fiscal year '27. This is our first quarter with very, very strong momentum across the board, and it bodes well for the rest of the year.

Yes. So as you think about sort of the outperformance in the quarter, and you also guided like the second quarter higher, how do you think about the sustainability? Because you kind of left the back half sort of relatively unchanged. And obviously, in one world, it's prudent. I don't know if you were meaning to signal anything about back half or not, but curious to get your thoughts.

No. We manage the business on a couple of quarters forward-looking basis. I feel very good about Q2. We continue to solve for Q3, Q4. Our guide reflects what we started the year with. And you shouldn't be reading too much into a decel because we have time to solve for what is a perceived decel if you simply do the basic math between the full year guide, the Q1 actual and the Q2 guide. I mean, so far, the thing that we all need to agree on is the demand is very, very strong. And so even if you look at how much we grew, excluding pricing or so-called pull-ins, we grew very substantially. We grew 35.5% overall, but we said 1/3 of that growth came from pricing and pull-ins. And so even if you were to adjust for that, a good 22% growth comes from what our normal business would do normalizing for pricing and impacts. And that is a multiple of what the competition has done. I think probably with hindsight and hindsight as low as 2020, I would not talk about pull-ins anymore because no one did, but everybody had them. And so I put myself into a difficult position talking about pull-ins. I maybe have been a little bit too transparent and honest, and that prompts a lot of questions. I don't regret being honest and I'll never be dis honest. But I think you should ask the question about other players who have grown in single digits and who have very different levels of performance. What is the amount of growth that they had that comes from pricing and pull-ins? And that is pointing to them having negative growth. And that's what I want to highlight to you is that we are the market share taker, and that is not by accident, it's by design. And the way we address pricing was really very intentional. We did not raise prices quickly. We raised pricing last. We raised pricing less than the competition because we felt that there is much greater value created for shareholders by actually driving higher top line growth, acceleration of the top line growth with -- yes, you have expenses that come as well. But on the whole, we're creating much greater levels of operating leverage, and that's why you see that our guide for the full year is implying a very strong operating profit growth on about 22% top line growth on a full year basis. And by the way, in the space of a year, we doubled top line growth. So on high numbers, right? So we didn't start from a low base. But last year, when we started last fiscal year, we said 11% top line growth from the onset. We raised it to 15%. We ended up 15.5%. And now we started at 18% top line growth, and now we're saying 21.5% and counting.

Yes. Yes, it's nice to see the company get back about 20% top line growth, which the last few years has not been the case, but great to see the company reaccelerate growth in a kind of somewhat challenging supply chain environment, at least to execute on that. So I think, Tarek, you've noted that back half of this year, we should start to see some contribution from hyperscale customers. What are some of the gating factors? Or I mean, are we at a point now in the visibility in the back half where there's really -- it's just sort of marching according to schedule and there's delivery? Or are there other things that need to happen, a lot of heavy lifting needs to happen to be able to count on that revenue in the back half?

Yes. No, the revenue in the back half is based on customer commitments. So I don't think there is doubts around that, and I feel comfortable reiterating that point. And this is a function of when they want our solutions to be shipped to them, right? So we planned that way in advance. We have secured the NAND, and we're all fine. In terms of heavy lifting, there is always a lift. But I think right now, we have a formula that works. And the formula rests on 3 things. One is the NANDs, the NANDs have to be qualified and the hyperscalers procure the NAND through their own supply chains. The second thing is the other componentry that is non-NAND related that we procure just to facilitate the adoption, and there's not a lot of calories attached to that for us, but we do it to remove friction and to facilitate the adoption in a hyperscaler environment and make things simpler Wamsi. And then the third thing is our software. And on the whole, when you look at this, with this model, we're targeting 75% to 85% gross margins, and I feel comfortable about that range. On top of that, we offer support, and we offer spares occasionally when they need it. That all is goodness that translates into the overall revenue line from hyperscalers. But the model is set. We're going to go with this model where the lifting needs to take place is making sure that our solution fits in the production environment of a hyperscaler. And that is always a technical evaluation. Aside from the VAs being qualified, there is a lot of the software that needs to be tuned to function with their operating system. And that is what we're doing, and we are working to get into more and more hyperscalers with that. I'd say it's a matter of time. Remember, this is not an enterprise sales cycle. It's an engineering sales motion. And it's an engineering sales motion that takes the time it takes to essentially have a plane fly for the first time. That's what you have to remember, right? Because they're building a new environment for the next generation of data centers. Everything has to function at [ unison, ] servers, networking, storage with all the civil engineering around that. That is the design of a next-generation data center, and that's what we have to fit into.

Right. Yes. So on that point, right, like as I think you've had your initial customer, the very first sort of sales motion was different from the way that you structured the deals in subsequent time frames. And now you're at a point where you're suggesting that these hyperscaler deals will have margins of 75% to 85% -- is that sort of the way that we should think about, not that you're announcing any new hyperscalers, but as more hyperscalers come online, if and when they do, should we be expecting this to be the normal sort of way? Or are there other business models that you are contemplating? Because, again, as you said, right, this is like a plane taking off for the first time in some ways, like everyone has their own specific needs.

Yes, you're correct. But I think the business model is set. I mean, we had to establish a formula, right? And that's -- we have established a formula. And so you can expect 75% to 85%. It varies depending also on, for example, how large of a DFM module the customer would want, right? So if it's a 300 terabytes or 150 or 600 terabytes, it fluctuates a bit. But the range doesn't really change. So we feel very comfortable that we've created a replicable formula here. And we also believe that this formula has a broader appeal beyond hyperscalers onto Neo Clouds and also AI Titans and there are conversations with a dozen of companies who have worked fairly large and who have built their own storage management software in terms of compression, dedup, et cetera. They don't need that, that comes with our systems. They want the medium and software to manage the medium. That's where our hyperscaler solution comes in.

What's the -- like if -- there's so much value that you add in purity. There is so much of value that you add from sort of your know-how of not just how to manage and write to that NAND, but in the compression techniques and in the deduplication and creates your storage to be very efficient compared to a lot of other similar media players like that use similar media. So in some ways, if you're talking about maybe potentially some of these other potential customers who are not wanting that piece of it, how does the economics to you change in terms of not wanting a certain piece of the -- like pretty high-value software?

Yes. No, look, the economics change in a sense that when you buy a system and you reduce the price of the system to a price per terabyte, obviously, the richer the system, the higher the price per terabyte. And so you know that if you are only providing the medium and managing that medium, the software that manage the medium, the price per terabyte is a lot lower, but the volume is a lot higher. And so if we put things in perspective, we ship, and I think you know that, Wamsi, we ship on an annual basis for our core roughly 6 to 7 exabytes when we currently ship to hyperscalers customers, a multiple of that. And hyperscalers as a whole industry buys about 650 exabytes per year and counting that number is rising. And so you can see that you have a very different set of economics when you buy a big system or when you buy a -- essentially the medium and the software to manage that medium. But every customer is free to choose. So hyperscalers and AI Titans probably would -- some of the AI Titans build their own storage management software and all some do. Neo Cloud, very few of them do. They buy fully made systems, and that's a different world. That's where EXA comes in. We are very pleased with the progress we're making on EXA. EXA is really designed for massive parallel storage. And so effectively, if you really want a simple rule of thumb, you can have a data center where you have a few hundred GPUs, call it, 1,000 packs. We can cater to that in the enterprise world for our customers with FlashBlades. When you break above the 1,000 GPU and you go to 10,000 GPU and so on, now you're talking about a massive parallel storage system. That's where EXA comes in. And we have -- we're just at the very beginning with EXA. We did very well in the first quarter of the year. I think Charlie on the call told the Street that we generated about less than 10% of the bookings that we generated in the quarter that will materialize in revenue in Q2, Q3, Q4, and we're counting because EXA is really gaining traction, and we're very pleased with the progress we're making.

Okay. And just -- there seems to be some concern anyway from an investor perspective that some of these hyperscalers are trying the approach of do it themselves to sort of replicate to some, a subset of what you do, not exactly what you do, but a subset because there are environments where we don't need everything that you do. Have you seen any evidence of that? Have you -- like what's your point of view on that?

Sure. So nothing is immune to replication. -- the barriers to copying the know-how that we put together on mastering flash modules is high. the barrier is high, right? So it's been built since 2009. And since 2009, we've mastered multiple versions of the NAND that comes from the same provider, multiple generation of NAND from the same provider, multiple providers and every combination of that, right? So the software neutralizes those differences. And so it takes a long time to get to understand the intricacies of the NANDs. And what's interesting to observe is that we buy the NAND from the same providers that provide SSDs. And so they have not done that. They have not built the software to just go into competition with us. They prefer to build SSDs. So we help them in -- also in the mastery of the development of their NAND, and we extract more -- about 30% to 40% more performance from our flash modules than they do through their own SSDs. And that's very interesting for them, and it's why we are adding to their total market that they otherwise would not have.

Okay. That's super interesting. Obviously, we're working with this inflationary environment from a NAND perspective. How are you managing through that? It seems like everyone has pricing for value and everyone is taking price. You guys have raised pricing a number of times. As you think about, a, what's your view on how long this inflationary environment plays out? B, is it impacting supply for you? And lastly, like are we going to reach a point of where there is going to be demand destruction or we don't see that?

These are really $64,000 questions each. So I'm going to venture in a territory of opinion and not fact because it's not obvious. I have the belief that this cycle is longer. And the reason why I say the cycle is longer, meaning this rush for infrastructure and the purchase of components that comes with it is because the demand comes from very, very large players, right? So the cloud providers, if you add the totality of the CapEx spend on putting data centers up, it is only -- and I saw a great piece of research from [ DNA ] about this. it only points to 90% of their operating cash flow. So if you take the whole CapEx of every hyperscalers in the world and you add it all up, and it's pretty much 95% of the operating cash flow that they generate. And so they can afford it. And they can continue to put more and more data centers up. Then you have to add on top of that the AI Titans. And there's a lot of talk right now on IPOs and how much these companies are worth and how much they will raise. So again, there is the money to invest in this. But the path to return is not necessarily obvious for all these players. Having said that, they have core businesses that can fund an expansion, which makes me say, it's going to last for longer. So I don't see this environment changing before at least the first half of next year. Beyond that point, it's hard to tell. So how do we procure in this environment, which is the second part of your question, it's an allocation-driven market. And you either have the allocation or you don't. We feel very comfortable with the NAND. That's not the question for us. We can procure NANDs. Other components are tighter. So CPUs are tighter. DRAM is tighter. We just have to manage that, and we do. And I'm glad to see, for example, that CPUs prices have gone up. I'm a fan of [indiscernible] and he's a good friend, and we work really well with them, and I'm delighted to see that there is something there for them as well. But this cycle is there to stay. So now the third part of your question, at least for a while. The third part of your question is, do we see demand destruction. And that's why in our earnings, quite honestly, we debated this internally quite a bit. We see demand very, very strong. Linearity in Q1 was very strong. Linearity in Q2 is still very strong. So customers still buy stuff. We don't see evidence of demand destruction. However, when you're raising prices 70%, 80%, we raised prices in the first quarter, 70%. Competitors raised them even higher than what we did. You're testing levels of price elasticities that are unknown, right? So is this a new level that customers are accustomed to? It's a good question. What we're seeing, and this is very interesting, what we're seeing in our -- the things that we sell is there's a shift to quality. There's a shift to higher-end performance systems. And in our revenue, you see price impacting our revenue, but you also see mix shift impacting the revenue quite dramatically. And the units that we're selling at higher performance systems are higher than what we have ever sold so far.

Okay. Are people -- like when you think about budgets generally in IT, there's a certain amount of dollars that you have and prices go up, you're getting fewer exabytes of whatever you want. Are you seeing any evidence of people changing policies around storage, doing other things and stop gap versus saying, okay, well, we have a need to store x exabytes, and we can only buy half of that with the given pricing. So we're just deferring the demand and it's kind of reflecting in backlog, RPO, whatever metric you want to book. But like you've got some way of saying there's actually demand deferral, and we can count on this cycle being longer because we actually know that this customer here wanted more but they couldn't buy right now.

There is every flavor of no demand to super strong demand and everything in between, right? So you have the customers who sweat more their assets. That's a possibility. There are customers who are more selective about what they buy. That's the majority of what we see. You have customers who are buying more because they have needs to transform their infrastructure. And you have customers who say, okay, I'm going to experiment a new model like Evergreen//One because I still have my need, but I don't want to commit dollars to it. I want to buy an outcome. I want to buy an SLA. And that's where Evergreen//One com in. We had an incredibly strong performance with Evergreen//One in Q1. We grew our total contract value by about 73% for our Storage as-a-Service offerings. And so we can grow with the product by selling the high-end stuff. We can grow with subscription by selling Evergreen//One. And I think right now, we are 55% product, 45% subscription. And you saw the acceleration of our ARR, and you mentioned RPO. RPO grew 41%. ARR grew 19%, 300 basis points higher than the past Q4. So this is very solid for us from our vantage point. The good news is that we have a hedge model, right? That's effectively the strength of Everpure..

Yes. As you think about the progression of product gross margins in this environment, like what's the right framework to use for that? And as more hyperscaler demand comes in, how should we think about that based on different models that people might be thinking about?

So what I would say is for the rest of the year, you can expect product gross margin in totality to recover because you have the benefit of the mix and the contribution from hyperscaler revenue coming in the second half. If you exclude that contribution, we said that product gross margins, excluding hyperscaler contribution, would recover gradually over time. I was not definitive on timing. And the reason why I don't want to be definitive on timing is because it is -- we want to preserve our market share taking stance and it's commercially sensitive. So we left it at that.

Is there a target market share like that dynamic that you're thinking about in balancing what you're reinvesting in terms of pricing and being able to take share? If we think at the end of this year or next year, like are we talking about a point of share? Are we talking more? Like what's kind of like how do you think about balancing that?

So we've taken on average, a little bit more than 1 point of share per year for the past 13 years, right? But this is accelerating. And if you take 2025 figures, we have captured 40% of the growth in the industry, and we have captured 50% of the capacity of the industry in enterprise storage ex China. So I think our goal is to obviously get to #1. We're right now neck to neck with a couple of players. Well Dell #1, neck to neck with NetApp, and we are growing them. And so our goal will be to get to #1 in revenue terms or in value terms, if you prefer. And my personal goal is to sustain this performance durably as I love to scale companies and I love to scale Everpure. So I'm guiding to -- if you really look at our guide, it's a Rule of 40 guide for the full year. And so I'd like to see that performance sustained over multiple years. We hit it in Q4. We hit it in Q1. I'm guiding Rule of 40 for the full year, and we see it that way for this year, and we'll guide at the right time for next, but so far, so good.

Okay. Maybe just on the competitive landscape and the share gain that you're talking about. Do you -- historically, just to put this in context, and I should go back and look, but have you gained more share in inflationary or deflationary environment?

I don't know the answer to that. I would say -- well, we've gained -- let's say this in simple terms, the past years up to the late part of 2025 were deflationary in nature. And we've gained share in a deflationary market. Now we are in an inflationary market, and we're gaining share in an inflationary market by judging by the differential growth rates between us and the other players in the industry. So ultimately, I think the value that Everpure provides with a software-defined solution gives us a lot of ability to grow with a very simple solution from a hardware standpoint and all the sophistication residing in the software. And ultimately, customers love Everpure for the simple reason that we are 10x more reliable than the next player, thanks to our software architecture and our simplified engineering. And that counts a lot when you think about market share gains over the long run. reliability and performance are the 2 things that stand out.

Yes. No, absolutely. Unfortunately, I know we're coming up on time over here. So it's a very quick 30 minutes. But Tarek, would love to like maybe just get some closing remarks from you in terms of what should investors be most excited about Everpure's opportunity over here? I mean there seems to be this reacceleration of growth. You have potential new hyperscaler opportunities. You mentioned Neo Cloud. So a lot to like sort of look forward to, but would love to hear your perspective on that.

I think if you look at the software space and you look at Rule 50, Rule 40, there are not many companies who are at that level of performance or at least let me rephrase that. There aren't many companies of scale that are at that level of performance. So we are accelerating revenue. We doubled the growth rate. Last year, we did $3.66 billion of revenue. We're guiding to 22% above that. This fiscal year, first quarter, a lot happening. We have a lot of tailwinds, and we're in a unique position in a sector that is very much at the forefront of AI. And so there's a lot of things to be excited about. We have to continue to execute and deliver, and we'll do everything we can to do so and satisfy our shareholders.

Excellent. Well, thank you so much, Tarek. I really appreciate you taking the time, and thank you for being here.

Thank you for having me, Wamsi.

Thank you so much. Thank you.