All right, good morning, everyone. Thanks for being here, kicking off the conference with us with Pure Storage. I'm very happy to introduce Rob Lee, who's the Chief Technology Officer for Pure Storage. Also have Paul Ziots, Head of Investor Relations, who some of you may know. And you guys are becoming a regular here, so we appreciate that. Kind of like a house William Blair name. A lot of the folks here own it on the private wealth side as well. Before we begin, I require to inform you that a complete list of research disclosures or potential conflicts of interest is available on our website at williamblair.com.
From Pure Statements made in these discussions, which are not statements of historical fact, Forward-Looking Statements are based upon current expectations. Actual results could differ materially from those projected due to a number of factors, including those referenced in Pure's most recent SEC Filings on forms 10-K, 10-Q, and 8-K. All right, with that out of the way, Rob, thanks for being here with us. We're just going to do Fireside Chat Format. Like I said, I didn't introduce myself. I'm Jason Ader from William Blair. Nice to see a lot of familiar faces. We're going to get rolling here, and we'll have some time for questions. Following this session, we'll be up in Adler Room for the breakout. To start things off, Rob, can you just tell us what you see are the big opportunities that Pure is focused on at the moment?
Yeah, absolutely. Thanks for having us back, Jason. Look, there's a number of opportunities we're focused on at Pure as we think about growth, as we think about the future of the next couple of years. I think number one is the set of opportunities that still sit in the enterprise for us. If you look at what we've done over the last couple of years, expanding the portfolio, now driving into the disk replacement ranges in terms of the storage tiers, really being able to go and address the full suite of enterprise needs. We've still got a lot of growth left in the core enterprise business F usion.
As we've discussed over the last couple of quarters, in Charlie's Remarks, as well as quite a bit on the calls, is a significant Strategic Initiative that's geared towards further expanding our footprint in the enterprise and building that network effect that's really going to help us grow and expand within those Enterprise Accounts. I think AI, clearly a topic that's top of mind for Investors and Customers alike. We're now really the only company that's got the Full Suite of Offerings for AI, from smallest research and kind of Experimental Environments to the largest Scale Environments in the world. Hyperscalers, right? We've had a robust dialogue with this community about our Hyperscaler opportunity over the last year and a half, really driving from the initial pursuit of that opportunity to announcing the design win and now progressing steadily towards production.
A lot of opportunities across the board, but those would be the key ones that we're focused on here.
Okay, great. We'll drill down on each of those. Maybe we'll start with the Hyperscalers, probably the one that you get asked the most about. We'll get that out of the way. Maybe just to kind of frame for the audience here and for those on the webcast, what have the Hyperscalers done historically with storage? Why do they care about what you guys are doing?
Yeah, so the hyperscaler environments, as we've discussed before, are very different than the enterprise environments in a number of ways. Number one, they tend to design their entire Infrastructure Stack, Compute, Networking, Storage, Software, all entirely integrated. They tend not to buy off-the-shelf systems and then just deploy fit-for-purpose products. They design those stacks wholly together. The other material way that Hyperscalers have differed from the Enterprise is somewhat paradoxically, they've lagged the enterprise in terms of mainstream Flash adoption. They've been able to, through R&D, through their optimized designs, they've been able to drive a longer lifetime and, or not service lifetime, but they've been able to stay on Hard Disk Drive Technology for a much longer period of time because of those optimized designs that they've put into place over the last dozens of years. What's happening, though, is a couple of things.
One is the Disk Drive Technology really is, in terms of a Roadmap, in terms of Rate of Improvement, is coming to an end, and they recognize that. The ability to get further performance, further Capacity and Density Improvements really is tailing off. Number two, they are pressed for power, power and space. This is an ideal moment. This is kind of a point in time where Hyperscalers, while like the enterprises several years ago, are recognizing, hey, it's time to move to Flash. They also recognize that there are inherent disadvantages with the most straightforward approach, if you will, of using Commodity off-the-shelf SSDs. Recognize that with a direct-to-Flash approach that Pure has with our Direct Flash Technology, really that's the only approach that's going to give them the efficiencies, the reliability that they're going to go need.
That is really what has brought us to the table in a lot of these opportunities, the prospect of helping the Hyperscalers navigate the transition from Hard Disk Drive Technology to Flash, being able to do that and get the reliability, the performance, and the economics they need. Number three, at a time when power and space are in such high demand because of GPU Buildouts.
Okay, great. Thanks for that background. Now, on the design win with Meta, can you talk about how they're going to be deploying your product? Maybe some sense of who you're competing with in that process and then what ultimately drove their decision to work with you guys. We'll talk a little more about the actual Deployment Process and the Licensing, but maybe just start out with kind of the genesis of the win and how that came together.
Sure, absolutely. I'll hit those in reverse order. We've been working with Meta. Meta has been a customer on the IT Side, on the AI Research side for a number of years, and we've had some dialogue with the Financial Community about that over the years. As far as the larger scale Hyperscaler kind of Production Environment Engagement, that started about two, two and a half years ago. Started with initial discussions around the technology, DirectFlash , our DirectFlash Technology, that is what that would be capable of in terms of giving them Cost Savings, Power and Space Savings in their Production Cloud Environment. A lot of Bidirectional discussions looking at TCO Analysis, looking at Configurations. This is Engineer-to-Engineer Modeling of, hey, if we were to integrate this technology in a variety of ways, what would that mean for the environment?
Through that process, our teams identified very easy-to-consume integration of our DirectFlash Technology, really the software, into their Software Stack. We identified a way to deliver the technology in a way that did not require them to fundamentally redesign their software that really could be more of a drop-in replacement for the Storage Technology they previously used. From there, progressed towards some testing, some initial POC Testing of the technology, them wanting to see, hey, what is DirectFlash capable of? How does this fit within their stack? Moving all the way towards what we call a design win, really, which is that gate that they go through in their design process, before which they're evaluating technology, seeing, hey, what is it capable of? Could this fit my need?
To some point in the Engineering Process, you kind of stamp a plan and say, all right, this is my plan of record. I'm moving forward with this. You still have work to do to go integrate that and test it and validate it, but that becomes the plan of record. That was really the process that we followed with Meta. I would say that one of the biggest drivers in terms of their motivation to work with us really was initially driven by a long-term view of TCO and Cost Savings, but significantly accelerated by a view of the Power and Energy and Space Savings that our technology was going to go forward.
Just for those unaware, what is the difference between what you guys are able to deliver on the Power and Cost Side, the power and space versus the traditional SSDs?
Yeah, no, that's a really good point. Stepping back, one of the significant differentiators within Pure's Flash Technology is we don't work with SSDs, right? We build on Flash-based Systems. We build off of the same NAND Flash that is in your Tablets and Phones and Laptops. The difference is we source the Raw Chips. We don't use SSDs. Instead, we build our own Drives, and we move a lot of the work and the inefficiency that typically sits within an SSD into our higher-level Software. If we go back in time, SSDs were essentially an Industry Technology coping Mechanism. Flash came on the Market. All of the software in the world was built to write to Hard Disk Drives . Hard Disk Drives behave very differently than Flash.
SSDs were kind of that coping mechanism to be that translation layer to make Flash look like and behave like Hard Drives for all the software in the world. Like any translation layer, like any sort of go-between middleman, if you will, there's inefficiency introduced. There's inefficiency measured in Performance, Reliability, and ultimately Density and Cost Savings. Our technology removes that middleman and at large scale has very, very significant benefits. To put some numbers on it, our typical drives, we're able to get about 30% more utilization out of the NAND Flash than SSDs. If you go to Amazon and you buy a 15.36 TB SSD and you were to tear it open, you'd find the same number of NAND Chips, same chips in a lot of cases, as our 18.3 TB drives. And that extra 25%, 30% comes from the efficiencies that we drive.
For the Hyperscalers, we're taking that technology and those efficiencies and packaging it in much denser ways. We're able to ship in the same form factor up to five, sometimes 10 times as much usable storage as they could achieve with SSDs, resulting in commensurate 5-10x improvements, reduction in space and power requirements. When you put this in the context of folks that are looking at, do I need to build N more Data Centers? Do I need to go refurbish this many more Nuclear Reactors? That amount of power and space savings becomes very meaningful.
Okay, can you talk about where you are in the process with Meta in terms of the deployment and then how the Licensing works? Because probably a lot of people are thinking that you're just shipping Hardware into these guys. And in fact, you're not actually going to be recognizing any revenue from Hardware, correct?
Yeah, let me hit both pieces of that. In terms of where we are in process, I mentioned before, that gate, that design win where we become a plan of record, we've passed that gate. We are part of that plan of record. There is additional Testing, Validation, and Ramp to Scale that is ahead of us. We're in the process of that Production Validation Testing, as we discussed in last week's call. This is really the stage of testing where they're bringing the entire solution together, inclusive of Pure's Technology, their technology, their software in a production-ready, essentially form to do pretty close to Final Validation. We would expect, as anticipated at the end of the year, we didn't expect the one to two exabytes of the solution to ship later this year, as we expected going into this year.
That would represent further progression of that Validation Testing into smaller scale pilots all the way to production deployments, which we would expect closer to our fiscal 2027, as we discussed coming into the year. As far as what does the business model and the license structure look like, you're right, Jason, we would not be anticipating providing the Hardware directly to Meta. Instead, allowing them to leverage their Supply Chain. We'd actually look at the Business Model here as more being centered around Technology License or Royalty, if you will, that would be associated with the delivery of the solution and accompanied with a small maintenance support contract. We've gotten a lot of questions, hey, how should I expect this to be recognized? We would expect that royalty to be generally recognized under product revenue associated and co-timed with the shipment of the solution.
The support we'd expect to come in over time over the duration of that support contract.
Okay, so you get the Hardware and the Software Licensing, basically, plus the support over the lifespan.
Yep.
Okay. So then the next obvious question is, what percentage of the solution kind of goes to you guys versus to the Hardware? Have you talked about that at all?
In terms of the overall value of that Stack? We haven't broken out specifically. I think we'll be in a position to speak a little bit more as we ship the anticipated one to two exabytes later in the year. I think we'll be in a better position to break down more specifically kind of the financials and how that looks. Net-net, I think the way I'd look at valuing the solution is, look, at the end of the day, the alternatives that the Hyperscalers have, and I think this is a question you had earlier in terms of who are we competing with, there's no other vendors at the table, right? Let me just put that out there, right? What we're competing with is their alternatives in terms of moving forward their designs. What are those alternatives?
Number one is the Hyperscalers attempting to build a similar technology themselves, which we do not see at the moment happening. They have tried in the past and generally have not progressed in those efforts. Number two is staying with Hard Disk Drive Technology , essentially continuing to do what they have always done. Number three would be attempting to move to SSDs and kind of the value second TCO that they get from doing that. Really, those are the three options or alternate options on the table. When we think about the value to Pure of that solution, generally, I would be thinking about it as essentially the Total Cost of Qwnership all in of the SSD-based Systems and the Hard Disk Drive-based Systems are going to form somewhat of a floor and a ceiling. We are going to float somewhere in between that.
If you look at what that total TCO is of the Hard Disk Drive-based Systems, you apply a bit of a premium for the simplicity, the reliability, the longer service lifetimes we can go drive, and then back out some of the Hardware Elements, you're going to get pretty close to what we'd view as the value of the solution.
I mean, I guess it's somewhat novel though to sort of figure out for you guys together with the customer to figure out what's the value of the software in the full solution, right? I mean, that's sort of somewhat new.
At the end of the day, for the Hyperscalers, they're looking at really, they're looking at what are their options on the table, right? It's continue with status quo. It's move to better technology with a significant cost premium. Or if there's a better option that we can provide that solves for acute issues that they have elsewhere in terms of power, space, and the associated Cost Savings of that, it's all a balance there, right? You've got to float somewhere in between there in terms of what the value is to the customer.
Okay, great. What have you guys said in terms of the impact on the margins or anything around the business model for you guys in terms of the long-term impact, assuming this ramps up the way you want it to?
Yeah. So I mean, the primary thing we're looking at is, and the primary lens we're looking at it through is, we expect this to be very Operating Margin accretive, right? In the expectation of the Licensing Model that we outlined, I think you can make some assumptions in terms of what that translates to in terms of Gross Margins as well. But our primary lens is looking at this through an OM accretion lens.
Great. I want to open it up to questions in a minute, but one thing I did want to have you address, which seems like it's been a big trend in the market, is Storage as a Service Models that you guys have helped pioneer. I know that's been kind of a growth area of the business. Can you just talk about what Storage as a Service is, what some of the history there is? Because it's not a new concept, but it seems like in the last, we'll call it three, four years, it started to really see meaningful adoption. What's the long-term kind of picture there? Do you think in five to ten years, most enterprise customers are going to buy that way, or do you think it will still be heavily weighted towards CapEx?
Yeah, it's a great question. Let me start by saying, I think if I look across the industry, storage as a service, a lot of vendors use that as a fancy term for a dressed-up lease with some teasing C's, right? The way that we approach Storage as a Service and really driven by our Evergreen//One and Evergreen Flex offerings is really centered around an SLA-based and Outcome-based Sale as opposed to a Product-based Sale, right? What does that mean? In a traditional product sale, typically you go to a customer, they'd have some sense of what their needs are. Hey, I've got this application. I've got this growth plan. I've got this set of performance requirements.
You then have the customer sit down with our Sales Personnel, our Partner Sales, go through an exercise to map those requirements to a specific set of products, apply some conservatism buffer, like, well, if I think I'm going to grow 20% year on year, what if I grow faster than that? Let me overbuy a little bit. Maybe I need 200 TB. Okay, let's call that 250. Well, but I don't want to go to finance if I'm wrong. And so let's go buy 300 because getting a PO approved is a pain. That process is typical with a traditional Product Sale and Purchase. In an Evergreen//One or SLA-based sale or Storage as a Service-based Sale, we meet the customer at step one. Hey, you've got a sense of an application, your performance needs, what your capacity needs are today. We'll go meet those requirements.
As your needs change, we'll go add to that environment, right? If you grow faster than that, great, we'll come and we'll add to that environment. If your application needs more performance over time, we can move that application to a higher Performance Storage offering all under that same contract. It is really a mindset shift from the customer taking the upfront risk, if you will, of locking, kind of taking their needs, identifying the products, and then going and running and operating that versus in a storage as a service sale, us meeting the customer at their needs, and then us adapting the footprint to meet those needs over time. From a customer point of view, a tremendous amount of flexibility, tremendous amount of optionality that we give in terms of that model.
What we have seen historically over the last couple of years is as the amount of uncertainty increases in the market, whether it is macro, whether it is IT Budget related, that flexibility, the value of that optionality becomes much higher. We have seen, certainly as we came through the post-COVID kind of recessionary anticipatory environments, a little bit of a tailwind to our as a service offerings. As customers, I think, paid renewed attention to the value of that optionality.
Okay, great. I want to ask one last question, then I'll see if anybody in the audience has a question, which is just the enterprise storage market. You talked about your expansion there, and obviously that's been the engine of growth for the business since the beginning. What is the outlook for kind of just enterprise storage? Is it still a growing market? I know you guys expect to take share, but what is the ultimate kind of base that you're working with, right, in terms of if that market will actually grow over time?
Yeah, so a couple of things. Yeah, the enterprise storage market is and continues to be a growing market. It's certainly measured by bit capacity. It's growing quite well on a dollar basis. Generally, we would view it the next couple of years, low to mid-single digit growth for the market overall. Jason, to your point, I think within the enterprise storage market, we've got a ton of share to go take, right? I think we're well positioned to do that in a way that the company really hasn't been in the past. If I look at the breadth of my portfolio, if I look at my ability now to go into an enterprise and have a much more strategic conversation about, hey, how can I meet all of your needs, right? Soup to Nuts from AI to Archive from block file object.
How can I do this on one consistent technology? How can I do this with Pure Fusion in a way that not just modernizes your storage infrastructure, but allows you to modernize your operations, automate and save on labor? I think we're well positioned to grow our share in that enterprise market. I think that's going to be a big driver of the company's growth over the coming years. Obviously, incremental to that would be the Hyperscaler.
Okay, we've got about five minutes left. Does anybody in the audience have a question?
Yeah. So for the transcript, so the question essentially was, is there anything unique about Meta that would make them more likely to adopt our solution versus other Hyperscalers? And look, what I'd say is each of the hyperscaler firms, as I look at the top four, five, even top ten, all have their differences in one way or another. That said, stepping back from it, nothing terribly unique, right? Since announcing the design win, our discussions with other top Hyperscalers have advanced as we anticipated. What we have seen from these advancing discussions is the methodology, how we are integrating our technology into Meta's Software Stack and their Technology Stack is generally, we believe, generally very similar to the way that other Hyperscalers would look to integrate the technology.
We would expect that we can take a lot of the learnings through the early stages of the Meta engagement and take those to our other Hyperscaler prospects and allow us to progress through that Technology Selection phase a little bit faster.
What has been, I guess, who do you compete against in something like Meta? What has been your win rate, I guess? Has that changed over time?
Yeah. The question is, what would we be competing with at a place like Meta and the Hyperscalers I seem here speaking with? What's our Win Rate been over time? We have just announced our first design win. We are progressing with early stages with others. Look forward to hopefully at some point speaking with you about further design wins. We are really early on in this process and really defining this opportunity. I think Meta has been, we have been out in front with Meta. This has really been the first time the Hyperscalers en masse have given serious consideration to, I would say, shifting a large part of their footprint over to Flash. In terms of what we would be competing with, again, there is really not another competitor, a third-party competitor at the table.
We would be competing with their own internal efforts, either to develop similar technology or to stay with their existing technology set, extending the kind of designs built around Hard Disk Drives or utilizing SSD-based designs.
Are you mixing SLC and QLC Technology into a Single Storage Sevice?
The question is, are we mixing SLC and QLC technology into a single storage device? We certainly have that capability. We have parts of the portfolio that use those capabilities. What's unique about our technology is because our software controls how each bit of Flash works down to the die level, down to the block level, down to the individual cell level. We have a lot of flexibility in how we use the same media. We do have the flexibility on the same physical device to differentially control how individual Chips, Planes, Blocks, et cetera, behave, whether it's QLC, TLC, SLC mode.
Just to wrap up, Rob, can you talk about the different layers of storage? You guys on the earnings call talked about how Meta and other Hyperscalers have different layers of storage. What does that mean?
Yeah. At a high level, to maybe oversimplify a little bit, when you look at these Hyperscaler environments, very unlike enterprise environments, right? They do not design infrastructure stacks per application. There are simply too many of them, right? They have got like hundreds of applications. What they typically do is they will design horizontally. They will design for high performance storage needs, lower performance, more cost sensitive, all the way down to archival, almost content storage, cat videos, that sort of thing, right? Everything in between. One of the things that really has been driving their interest is over time, they see all of these tiers of storage moving to Flash.
What Pure and really the DirectFlash Technology integration offers them is a single unified architecture that would allow them to deploy our technology across each of those tiers of storage from, again, the highest performance demanding environments all the way down to lower performance, but the most cost sensitive environments and everything in between.
Okay. With that, we'll wrap it up. We got a.
This presentation has now finished. Please check back shortly for the archive.