Here, before we start, I'm gonna give a statement here. Statements made in these discussions which are not 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. All right, with that-
Thank you for that.
Yeah. With that out of the way, I'm pleased to have Charlie Giancarlo and Kevan Krysler, CEO and CFO of Pure Storage. Thank you so much for being here.
It's our pleasure.
Great. I thought a place to start for folks who may be newer to the story, Charlie, maybe give a high-level overview of Pure Storage, a little bit of the history of the company, and the solutions you offer to your customers.
Absolutely. Well, thank you for having us. So, we're actually about to celebrate our 15th birthday at the end of the month, actually October 1st. So we've been in place for 15 years. We started selling our first product about 12 years ago, so about 12 years in the market. And during that period of time, now we've grown to, you know, roughly $3 billion, you know, and a profitable public company. At that time, we had one core product that we sold into the data storage market. It was really the first of its kind, all-flash, high-performance block storage system. And we grew very quickly to, I think in just a couple, three years, went from zero to $600 million, so very rapid.
Over the last, you know, 10 years, really what we've been doing is continuing to follow a very focused strategy. That focused strategy was, first of all, targeted at flash storage. Our belief that flash storage will eventually be all storage, replacing hard disks as flash continues its historic catch-up, if you will, in price performance to hard disks. Secondly, unifying the storage market, which had been very, very fragmented. Even the leaders in the market would have a dozen different systems to address all of the different use cases for storage. You know, different protocols, different formats, block, file, object, high-performance systems, low-performance systems, small systems, big systems. They all have different software.
We are very focused on one software environment, which we call Purity, to be able to operate across the entire range of data storage. Third was our focus, right from day one, on simplicity, and what we mean by simplicity is very simple, less labor to operate. Storage should not be something that IT is focused on all the time, shouldn't require two weeks of training. You know, and in fact, directly from our founder, it shouldn't be any harder to set up than the new iPhone that you buy. iPhone had just come out the year before, took inspiration from that.
We remain very focused on simplicity, but now we're operating at the full enterprise level, bringing simplicity not to, not just to a single array, a single storage system, but frankly, the entire fleet, if you will, of storage systems inside an enterprise. Our view is that enterprise storage should look like what we all use these days. You don't carry external hard drives around with you. You use, you know, some cloud provider, whether that's OneDrive or Dropbox or Google Drive, whatever it may be, and you get to share all of that with your colleagues. We're doing the same thing in the enterprise market, making enterprise storage look like a cloud of storage rather than as individual arrays.
And this really frees up data from being captive in silos to individual applications and making it available for AI and other use cases. It also brings tremendous simplicity to the organization. It brings programmability to the organization, and it allows their developers to get immediate access to enterprise-defined storage environments, rather than having to construct each one mechanically. So basically, what we're trying to do is create an enterprise environment for storage, unify storage, consolidate it, and really make it look more like a cloud of storage rather than as hardware.
That's, that's a great overview, and looking forward to digging into some of those topics. But, maybe Kevan, for you, since you recently reported, maybe just a brief summary of the quarter, kind of key results and things you think are important to the story right now.
Yeah, I think the key messages from our latest Q2 earnings really is we outperformed on the top line in Q2 against our guide, as well as operating profit. In terms of our overall guide, we did reduce the TCV sales expectations of Evergreen//One. We can spend a little bit of time walking-
Yeah
... through that as well. Really, as a result of the larger deals for Evergreen//One, seeing a longer progression time associated with the very specific as-a-service larger deals. We're getting some great traction with our FlashArray//E, FlashBlade//E, as well as FlashArray//C solutions, really targeting the cost-sensitive workloads, and we're seeing that traction. Now, that did bring our product gross margin a little bit lower, and frankly, brought it into our sweet spot, which is, in my mind, the higher 60% range.
We do see and expect more traction and sustained traction with those solutions in the second half, and that's why we talked about, hey, for product gross margins for second half, we would see a moderate sequential decline as well, as a result of the increased expected market share of those solutions and sales of those solutions.
Yeah, that's great. And yeah, we'll definitely dig into some of those product dynamics. But maybe starting higher level, you know, as Flash continues to see improvements both in pricing and capabilities, I guess, where are we in that journey towards maybe 100% Flash? And I guess what are the levers to kind of see those market share, you know, gains-
Yeah
Over the next coming years?
So, you know, I predicted, starting last year, that in about five years' time, very few disk-based systems, I mean, obviously, there are individual disk replacements, but, disk-based systems would no longer be a strong part of the market, and I'm sticking to that timeframe, so got four years left in that prediction, and it's really quite simple. I think most of us at this conference probably believe in Moore's Law. It's been active for, you know, 40 years now, 50 years.
If you were to look at flash price performance versus hard disk, hard disk started out many orders of magnitude ahead, but you could follow it for over the last 30 years, and flash has continued to decline or improve, if you will, at a much more rapid rate than hard disk. And so we're getting very close to that crossover point. And because of our unique technology in terms of direct to flash, and I think if anyone who follows this understands this. Because of that unique technology, we are, you know, very much at the forefront of driving flash price performance.
So, you know, whether it's in large-scale, hyperscale environment that still are about 90% hard disk today, or whether that's in enterprise accounts that are probably about 50% hard disks today, we really see that coming to an end over the next several years.
Yeah, that's great. And maybe touching on competitive dynamics, can you just walk through how you think about the segmentation there? You know, you've talked about kind of higher-end workloads vs lighter workloads. And if you can touch on just the Pure differentiation, I think you've talked about Purity as being kind of a value prop. So maybe just touch on the competitive dynamics and your kind of positioning there.
Yeah, well, so at the most fundamental level, Pure is an R&D technology company. We'll invest, you know, somewhere 18%-20% of our revenues in R&D every year. Now, for any of you that have been following storage, it was considered to be a commodity, you know, a decade ago, and most of our competitors have treated it as such. Therefore, no differentiation, but we're cheaper. That would be the selling point. Consequently, most of our competitors spend less than 5% of their revenues on R&D and are not advancing forward, right? We, by comparison, you know, as I mentioned, we're the only ones that have a consolidated single software environment for all of our storage systems.
We have led in a number of different areas, whether that be Evergreen, whereby we consistently upgrade our products year after year, decade after decade, based on a subscription, so that a customer that bought a product 10 years ago, without a single additional dollar of CapEx, if you were to visit that product today in their data center, would look brand new. I mean, it would be identical to a product we sold last year. So every product, whether we sell it initially as CapEx or as a subscription, eventually becomes a subscription. We have 10x higher reliability than our... This is guaranteed compared to our competitors. We're somewhere in, we are at most one half and, as low as 1/5 the space, power, and cooling of our competitors.
And starting this year, we roll out as a non-disruptive software upgrade to the 50,000 arrays that already exist in our customer base, this capability, this fusion capability I spoke of, where all of the arrays operate as a cloud of storage rather than as individual arrays. So these are just a small number of the advancements. And so you can take my word for it, but if you don't, you can look at Gartner. We've been the top of their Magic Quadrant for, I believe now, 11 years running, and we continue to extend that lead after 11 years. That should tell you a little bit about how we compete vs our competition.
Yeah. I wanted to shift gears to the hyperscalers. Maybe touch on the market opportunities there. I know there's some ongoing conversations. I guess, what are the key milestones that need to hit to maybe get a deal through the finish line here?
Let me first describe why we would be of interest to a hyperscaler, and then go into, sort of where we are in that journey. As I mentioned a little bit earlier, about 90% of the storage inside of hyperscalers is still on hard disk, which may be a little bit hard to believe, given everything they do, but that's, yeah, give or take a few % points, that, that's where it is. Now, they do use SSDs for their high-performance environments, but as I mentioned before, we're literally the only company in the world that operates direct to flash. Every other company, whether it's a vendor of storage, or whether it's a computer system, or whether it's the hyperscalers, use SSDs because SSDs were designed to make flash look like a hard disk.
And that's exactly what it is. It's a, it's like using a microprocessor to look like a typewriter, right? And it diminishes the overall performance or, and/or price of flash relative to the core semiconductor, which it is. So, because of our software, we're able to make a flash compete much more effectively and earlier in the life cycle vs hard disk, and provide better performance, and compared to hard disk, 1/10 the space, power, and cooling. Now, if you think about the hyperscaler, what are they starting to run short on? They're all starting to run all these AI environments, and they're talking about nuclear power. You're building new, brand-new power stations in order to power AI. About 25% of data center capacity, whether that's space or power, is data storage.
If you reduce that by a factor of ten, you've reduced total data center power by 20%, and space by probably more than that. Not only do they get in terms of what we're able to provide them a solution that you know on a total cost of ownership basis is about the same price as what they're paying today for hard disk, but one that's more performant, more reliable, longer lasting, and gives them 20% of their data center back. Not a bad deal. Those are our discussions with the hyperscalers. We have a lead horse as you would imagine in any type of situation like this.
The lead horse is for design win. I want to be very clear because hyperscalers design data centers the way we design products. Takes them a year or two. They design every aspect of it, and then they stamp them out as they go forward as their next generation. We're competing against what would've been their more traditional design using hard disks for this. We probably those conversations and we're speaking to others as well, but in the case of the lead horse, we've been speaking to them for about a year now, consistently getting deeper and deeper into both details as well as testing and integration and co-engineering of what this looks like.
You know, hopefully, at the end of this, what we end up with is the design win, a firm understanding of all the terms and conditions that both sides need to operate by, a basic understanding of rollout schedule, and then we'll be able to characterize it for our investors as well.
That's great.
So the co-engineered solution is in testing.
Oh, yeah
... at this point in time, and we're clearing gates, as you would expect, associated with the testing. And we've been through a substantial amount of the commercials as well. So we're pretty deep on both fronts, both on the technology side as well as the commercial side.
Yeah, understood. I wanted to ask about Gen AI, big surprise. I guess, you know, can you help break down some of these opportunities, whether that's, you know, training, inference, you know, what are the most immediate opportunities in front of Pure? And I guess, how are you kind of positioning yourself in-
Yeah
... you know, in terms of Gen AI?
It's very exciting, but to characterize it, actually, I'll try to be concise here, but it's a complicated market. You have the training environment, but the training environment is very different between, let's say, what the hyperscalers do, what non-hyperscale, but, you know, let's say, leading-edge technologies do, technology companies do, or sovereign states, and then what enterprises might do, right? Scale is important. In the case of hyperscalers, they're gonna use their existing storage infrastructure, which, because it's so parallel, they can use their existing infrastructure. If we get in there, it'll be our infrastructure, but it won't be something that's designed special for AI.
In the case of, you know, large sovereign states and/or, you know, technology companies that are doing this, yeah, they do need some specialized storage for this that meets their needs. We're very strong at the large end scale and at the small end scale, little bit of gap in the middle that we're working on, right now. In the case of companies, they're gonna be at the smaller end of the scale, and they won't always be doing training. Maybe they'll do small-scale training. They'll be doing things, you know, that these days are called RAGs or NIM. But really, those are... physically, they're just AI, it's just a standard AI infrastructure consisting of storage, some GPUs, and a network.
We've been selling into the AI space for over five years. It's been along with NVIDIA, and so that's a natural for us. We're working on some vertical use cases with NVIDIA in this space, but I'll say it's also the traditional AI is rather mature, and we continue to sell in that environment. The whole RAG/NIM space is, I would say, still to be developed.
Yeah.
It's brand new. There's not, I would say, substantial activity yet in that space.
Got it. Understood. I did wanna ask about, I guess, just overall demand trends and, kind of, I guess, the macro question. You know, anything incrementally you could add in terms of just the spending environment and customer propensity right now to,
Yeah, I'd say the macro hasn't changed much, you know, since the beginning of the year. It's not great, it's not terrible. I don't see any movement in it, better or worse, per se. We did comment in our earnings call last two weeks ago, that we're seeing a little, we're seeing companies put more consideration into new operating expense commitments, so subscription, whether that's cloud, software, SaaS or in our case, our Evergreen// One subscription. And I think it's because of sticker shock from, you know, there's been some cloud shock lately from bills coming back as companies have gone to production levels in the cloud. There's also been, you know, price increases in both the SaaS and software worlds.
Yeah.
And I think that's caused customers to put more-
Mm
Consideration into new obligations.
Yeah, so I wanted to shift gears a bit to products and maybe starting with Evergreen// One. Kevan, you know, you highlighted that earlier. I guess maybe walk through kind of the adjustments you've made to the guidance there and kind of maybe what you're seeing and what that kind of implies, I guess, for the rest of the year in terms of adoption.
Yeah, sure. I mean, look, the Evergreen// One offering, which is our as-a-service offering, and it truly is as-a-service. So this service is consumed as-a-service, and it's backed by specific SLAs. You know, the momentum we had last year was, again, our growth was well over 100%. When we looked at this year, we continue to see solid momentum, and consumption of our as-a-service offering. But what we saw that was a little bit different this year was a delineation between deals under $5 million, and these deals would relate to all customers, our largest enterprise customers, commercial customers, PubSec customers, and deals that are over $5 million.
The momentum we're seeing under five million is very consistent with what we saw last year, so really no significant change. That's what our assumption was going into the year. What we're seeing on larger Evergreen//One as-a-service deals, and this is deals over five million, it's just that the progression of those deals are taking longer. We're not seeing them flip to CapEx. We're not seeing any difference in terms of the competitive environment. You know, especially as customers truly understand our service offering compared to our competitors, they quickly find out it's well beyond creative financing. It's truly backed by services and SLAs that frankly are unmatched.
But to Charlie's point, I think with these larger as-a-service offerings, they're spending time in terms of the cloud cost. They're now getting caught up in terms of looking at, "Hey, when we look at all our cloud costs, including Pure's as-a-service offering, is this an approach we want to take, or do we want to go back and think about a traditional purchase and investment in Pure's technology?" More C-suite conversations in terms of the investment. But look, we're still very bullish in terms of the as-a-service offering as a whole, including expansion with our top enterprise customers. We're just seeing the time-
Yeah
... progression, taking a little bit longer.
Yeah. I guess, do you have a longer-term view of kind of what the implications of that? You know, maybe talk to investors about the evolution of the revenue mix and the implications to the model as Evergreen kind of continues to gain adoption.
Well, sure. You know, when you think about an as-a-service model on day one, when you compare it to a traditional sale, you're not getting any revenue. You're not getting any product revenue. That revenue is going to come over time as a service is being consumed, and that's going to come through as our subscription revenue over a period of time. It's a fantastic solution in terms of value from our perspective and our customers' perspective. It's allowing us to have the direct engagement with our customers. But what we would expect over time is that you'd have more consistent growth, obviously, with our subscription revenue. Product revenue would only be reflecting the hardware portion associated with our traditional sales. So it wouldn't reflect any of the revenue associated with our as-a-service offerings.
Yeah. I wanted to ask about, I guess, Fusion. You know, I think there's been some updates there. Maybe if you can touch on some of the early customer interest so far-
Sure
... with that product line.
Right now, it's very strong, and we actually introduced Fusion two years ago, but at that time, it was cloud. It was based in the cloud, and it required a greenfield opportunity, meaning a new sale in a new environment. The change we're making now this year, and we had a number of customers in that area, but we learned a lot. In particular, customers wanted in their own infrastructure rather than hosted, but more importantly, they wanted it to be available and applicable to all of their current environment. So, what we're releasing in just the next couple of months is a software upgrade, as I mentioned, a software upgrade applicable to everything we've ever sold.
With that upgrade, the customers will now be able to get the advantages of Fusion, which is this idea. Imagine you had external hard drives on your PCs, and without you doing anything at all, except upgrading your software, all of a sudden, that became part of a cloud storage and data platform, so that you didn't have to buy anything else. Everything you had was now on, let's call it, Pure Drive. On Pure Drive, you know, everything you've ever stored now can be shared with your friends and family. It can be scaled without having to buy anything new, and you get more efficiency.
And also, now that data that's stored on any one array is available to any other application, such as maybe AI RAG environments, without having to do anything other than authorize it. It's a really big deal, and it, in my opinion, I think enterprise storage has been in the dark ages for a decade. Networking became virtualized 20, 30 years ago with the introduction of the Internet. Compute became virtualized, and so I was part of that generation. Kevan was part of the virtualization generation with VMware. That became virtualized 15 years ago. We're finally bringing storage into that environment of full virtualization in the enterprise environment, with the introduction of Fusion.
Yeah, that's great. I wanted to touch on go-to-market. I guess, you know, what, where are the incremental investments being made, particularly as, you know, you look at the hyperscaler opportunity, you look at the AI opportunity? How is that go-to-market kind of evolving here?
You know, it's interesting, our biggest investment in go-to-market is really on the enterprise side of things. We're still going down that journey of being a, you know, very good but niche, you know, having a great niche, to being a broad-scale provider of storage. And storage is a very conservative market, so penetrating the enterprise, getting larger and larger wallet share, you know, inside of any one enterprise is still the major place where we're investing. I would say the sales investment in hyperscaler is relatively small. It, you know, it doesn't take a big team. It's actually much more of an engineering investment, and last year, we put together a dedicated engineering team specifically on hyperscaler, and that continues to be an area of investment for us.
In the case of AI, you know, that investment is actually spread out right now across the company in a variety of different places. Yes, on the sales side, to really focus on those large-scale technology companies and sovereign nations in terms of selling, you know, really large-scale storage into those environments. But equally well on the enterprise side, where it's much less about scale and performance and much more about integration of application environments to make it easier for them to use AI for specific purposes, and then so those are separate investments. But I'd say, you know, the largest investment for us right now is really that acceleration in large-scale enterprise-
Yeah. Yeah.
In hyperscaler.
And on the engineering side in hyperscaler.
Right.
Yeah.
Right. Understood. Maybe comment on, I guess, the pricing environment right now. Kevan, I think you mentioned a little bit about, you know, the gross margins maybe dipping for the back half of the year. But maybe more high level, how you think about kind of pricing and kind of the sensitivity to the model to changes in the pricing environment?
Yeah, are you questioning that in connection with flash pricing?
Yeah
... as well? You know, when you think about the high-performance workloads, frankly, well, first of all, when we think about flash pricing overall, our long-term thesis continues to hold that it's going down over time, and I don't think that's changing at all. When we have a short-term increase in flash pricing, which we're seeing currently, that's helpful, frankly, for us on our high-performance solutions. Reason being is competitors who price on cost plus, we have an advantage in terms of our pricing, which is basically predicated on the value of our software and our software solution. So it gives us an advantage of elasticity when flash pricing is increasing.
Now, on the cost-sensitive workloads, it's a little bit different dynamic because on the cost-sensitive workloads, we're looking to take out hybrid disk and disk. So there's a price point that we're aggressively looking to take out and gain market share on, and that's where we're seeing a little bit of the pricing affect gross margin by our choice, and I think it's a positive. But that's the dynamic between the higher performance workloads-
Yeah
... versus what we see on the cost-sensitive workloads.
Yeah. I have a few more questions, but any questions from the audience? All right. So, you know, we talked about go-to-market, margins. I guess maybe just when you think about kind of the growth in OpEx, we talked a little bit about the investments into R&D and for the hyperscalers and go-to-market, but how do you think about kind of the sustainable, I guess, you know, incremental investments, and how do you manage kind of hitting your growth targets versus kind of, you know, managing margin expansion? Maybe just high-level philosophy there.
High-level philosophy is, first of all, that we're a growth company, and so we're investing in growth. That being said, that is not inconsistent or incompatible with our commitment to deliver, you know, approximately a point of additional operating margin every year. You know, the storage market is, you know, it is affected by a number of different things, including the economy, and so there'll be, you know, better years and worse years, but you know, regardless, we think we can, you know, consistently deliver another, you know, through efficiency and scale and so forth.
Yeah
... deliver an extra point, roughly speaking, to the bottom line every year, right? And that's what Kevan and I have committed to.
Yep. That's what we're lined up.
Yep, agreed. And then maybe on just capital allocation, you know, I think you have a buyback plan in place. Maybe, you know, how you think about the right level of cash versus returning, you know, cash back to shareholders versus maybe strategic investments, M&A. You know, how do you think about that balance?
Yeah, our philosophy hasn't changed. I think we're, you know, in terms of partially offsetting dilution, leveraging approximately 50% of our free cash flow to do that. You know, I think we're a little bit behind in terms of that plan, but our thesis is unchanged. In terms of what we're looking to do with 50%, our commitment of using 50% of our free cash flow to repurchase shares to partially offset dilution. And then with that, in combination with that last quarter, we also extended to the employee group withhold to cover. So we're basically paying with cash the tax portion of equity instruments that are vesting, which again is contributing to our objective of partially offsetting dilution.
Yeah.
Yeah.
And with respect to M&A, we continue to look to the market, you know, for everything from tuck-ins to anything that might be transformational to the company. We've obviously been, I'd say, well, I don't know if I'd call it conservative. You know, we're looking for good opportunities to add value to our shareholders. And, you know, the private markets are rich, you know, frankly. So while here and there we might find, you know, interesting technologies or groups, unless we feel that we can make it a solidly accretive to our organization and to our shareholders, you know, we're gonna be really thoughtful about this.
Yeah.
Maybe to close it off, Charlie, kind of your top priorities over the next couple of years and kind of, you know, your aspirations for the business here over the coming years.
In the medium to long term, you know, our aspiration is to be the number one player in storage, you know, end of story. To give you a sense of that, today, we are the number two supplier of all flash storage systems to the world, second only to Dell, and now only a few points away from Dell. Our growth, you know, continues to be fast. If you believe that flash will take over hard disk, that gives you a sense of where we could actually be.
If I were to look just a few years down the road, my expectation is, you know, we start penetrating the hyperscalers in a significant way. Flash starts making real inroads against disks in enterprise. Fusion, you know, that is the idea that enterprise storage should operate as a cloud, becomes, you know, the expectation of every customer, and we're number one in the market.
Yeah.
If pigs could fly-
There you go.
That's where we'd be.
Charlie, Kevan, thank you so much.
Thank you very much.
We appreciate it.
Yep. Thank you.