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

Thank you very much for joining us. My name is Simon Leopold, Raymond James data infrastructure analyst here in person, thankfully, at our conference -- tech conference in New York. I'm pleased to have with us the management team from Pure Storage. We have, to my right, Charlie Giancarlo, CEO; Rob Lee, I believe you're now a CTO?

That's right.

I got that up to date; and Kevan Krysler, CFO. So I've prepared an outline of questions, so we're going to sort of follow a fireside chat format. If folks in the audience have questions, try to grab my attention. I've been asked to read the following disclaimer. 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. I almost have that.

You did that really well.

I've heard it a few times.

So I guess I want to sort of start out with the recognition that we've got some people in the [indiscernible] who know you well, some are new to the story, but I'd like to help everybody sort of get level set. So how do you like to introduce Pure Storage to a new prospective investor?

Yes. Simply put, well, [indiscernible] well aware that data continues to explode at an unprecedented rate and customers more and more, companies are trying to take advantage of that data. And quite simply, Pure Storage is the sole company that looks upon data storage and management as high technology. It's been promoted as a commodity by our competitors for literally for decades. We came in with a very different approach and really focused on investing in data storage as if it's high technology. It's what it's allowed us to be, it's 9 years in a row, a leader in the Gartner Magic Quadrant. It's what allowed us to be the sole company in data storage that continues to grow at double-digit rates and taking market share left and right. So.

And I guess the challenge we often receive from investor pushback is the storage market is an IT-linked market at best single-digit growth. Normally, you'll grow likely over 20% this calendar year. You're expecting, I believe your words were solid double-digit growth for the next fiscal year. And that just does not align with sort of the general sense of storage. So what makes you different from the other companies that are in this space?

Well, first of all, we're a market share taker. We operate in minimal a $50 billion market. We're less -- this year, we'll come in something shy of $3 billion. There's a lot of market share to take up. And being highly differentiated as we are, having in the tech industry the highest Net Promoter Scores and having a very differentiated model, which we can go into, obviously, in detail, just puts us -- it just gives us this tremendous opportunity to continue taking share in a large market.

And I want to maybe start the discussion from kind of a macro perspective and then drill down to the specifics of your company, but we've been asking everybody what's your take on the macro environment, the risk of a recession, inflation, interest rates. So when you think about running your business, how do these macro factors play into your thinking?

Well, the macro factors will have an effect directly upon individual customers that we're speaking to, right? And therefore, slow down deals, cause deals to be reassessed and so forth. But it doesn't change that the market is still in and around $50 billion. Whether it grows another couple of percent or shrinks a couple of percent, it's -- in a sense, as a share taker [indiscernible] is meaningless. Our opportunity is taking more share.

But the other thing I'd add on to that is flexibility, right? So we've really been ensuring that we are offering our customer base flexibility in terms of how they consume our technology. And that's where the Evergreen offerings really come into play. And so whether the Evergreen subscription is attached to CapEx, whether they decide that they want to purchase CapEx but want to consume and pay on a consumption model, we offer that model. We've got the full SLA model with Evergreen//One, which is advantageous. We've leveraged now pure financial services as well. So it's really being focused on across the spectrum offering flexibility to our customers depending on what their circumstances are. And I think that's been really helpful for us.

But you have acknowledged some lengthened decision cycles. Maybe help us understand what that means or where it's coming from.

Yes. So I'll redescribe what we've described in the last 2 quarters as to what we saw in the market. So we're on a -- we're offset by 1 month in terms of our quarters, right? So it's January, April, July, October quarter. So in our July quarter, we referenced the fact that we started to see, meaning in that last month primarily, we started to see lengthening of deal cycles, and then we saw it for all 3 months in the last quarter. So we are seeing lengthening of deal cycles, more checks controls inside the customer base, which is like in the sales cycle. What we've not seen is what we refer to as demand destruction. We haven't seen deals just being taken off the table.

And what's interesting -- let me just add on to that.

Sure.

What is interesting, and it's actually been a benefit to us as we are working through the additional conversations with customers as they're going through their own internal approval process, the concept of TCO is really becoming relevant to the customer. They're wanting to understand that in a lot more detail. That's an advantage for us. And we're actually seeing some wins when we get through that TCO conversation holistically in terms of moving us across the finish line to close some of these deals. So there's some benefits coming out of this as well that we're seeing.

So one thing I've observed in talking to investors is that the shift where we're talking a lot about macro and the recession and decision processes. And I suspect part of it is fatigue in talking about supply chain. We're all sick of it. I'm sick of it. So I apologize, but I have to check this box and ask you about the supply chain issues. So maybe if you can give us an update of how it's affected your business and what's your outlook for supply chain issues?

Well, as you mentioned, you might have a mix of people in terms of their familiarity with us. So throughout the entire COVID period, we've been able to continue shipping to customers with generally less than 3-week lead times. So while the supply chain, we've had every -- all of the same sets of challenges in terms of debookings of components and shifts in worldwide availability of contract manufacturers and late shipments and things caught up in the Port of Long Beach and all of that. We've had the same set of challenges as everybody else. And despite that, we've been able to ship constantly within roughly 3 weeks of lead time. So we've really navigated this really well from our customers' standpoint. And frankly, we've won a lot of business because of that. So -- and while things are getting better in the supply chain now, they're -- by no means is the supply chain back to where it was prior to COVID as when you read headlines about China in different cities shutting down as everyone should be aware. That being said, it is easier than it was 6 months ago, certainly easier than a year ago.

And the other line of question we've been getting related to the supply chain is the producers of the memory chips that go into technology, more predominantly for you, NAND-type chips talking about oversupply and price declines. And so that's created a stir of questions to us about how does that affect enterprise stores.

Right.

And on top of that, you've got a somewhat unique architecture. Keep in mind, we've got a financial audience, how do we explain what's the sort of puts and takes of this dynamic?

Well, this is a great opportunity for us actually. Now, right, NAND pricing is declining. And we've seen that before. And on average, it declines roughly if you take a long-term view of about 15% a year versus hard disks that declined maybe 5% per year on a cost per bit basis, plus or minus. The great opportunity for us this time around is now the price of flash is coming to the point where we, with our architecture, and we're many years ahead of the rest of the industry on a system basis can compete for what's called near-line disk or the 80% of bits that are not on flash, the 80% of enterprise bits that still exist on hard disk because of the price. We are now approaching that with what we can do in terms of delivery of what's called QLC technology, the next generation of flash after that market. So we're actually quite excited about this. The high tech, in general, is driven by price performance improvements and elasticity. That's what drives the high-tech market. And we now have elasticity to the 80% of bits in the enterprise. And when I say the enterprise, I mean both operating companies that have their own infrastructure as well as the cloud. Probably more like 90% of the bits in the cloud are on hard disk. And so it really opens up both of those opportunities for us.

And maybe this is a point to -- sure. Yes, go ahead, Brian, please. And I'll repeat the question for the webcast.

For the price performance perspective, is this the best that's ever been for Pure? If you were to look at price performance versus predominantly disk-based providers, what's kind of the order of magnitude of your advantage today versus maybe what it was 3 years ago? And is it opening up conversations with lots of customers like cloud providers?

And I'll repeat for the webcast. Basically, has the -- how is the price performance versus spinning disk changed versus 3 years ago?

Right. Well, so to be clear, when flash first came out in enterprise storage, we competed against disk there, but it was for primary storage, which was much faster disk in much more expensive environments, right? And so we could compete at that level. And today, about 100%, so all of primary storage now in the enterprise market is flash, right? And we have a continuously growing market share there. What -- there was this huge price delta between there and what is still secondary tier storage. And yes, the price of flash is getting -- and with our intellectual property is getting to the point where we can compete for that. And we think we're at least 2 to 3 years ahead of our competitors in that space.

And if I can jump in, I think the other thing to think about is as we look at the broader Tier 2 environments and workloads as we compete with the near-line disk opportunities, when we first started out with primary storage, it really was about price performance. As we look at this next set of opportunities, some of the other elements are becoming very, very impactful as well, right? The energy savings, the density, the footprint savings, where these workloads may not be as performance demanding, but they're large, right? They're large. They're messy. And so some of these other elements that feed into that TCO argument that Kevan was talking about start becoming very important. And this is an area where, Simon, back to your last question, where we've got a significant and structural advantage built in based on our differentiated technology. The fact that we can work with the run and chips where our competitive set is really limited to working with highly transformed and then consequently more expensive enterprise SSDs. Not only does that open up a tremendous amount of flexibility and agility to us in terms of navigating the NAND road maps and the different vendors in the space there, it -- also the economics don't necessarily travel on the same curves with the same volatility. And throughout that process, we've got a multiyear advantage and lead in taking the QLC technology and putting it into enterprise applications.

And this -- I'll just then add the same is true in the hyperscalers. They've not developed flash management technology as we have. We've been working on it for 10 years. And every new release of a new flash chip requires characterization, requires modifications to that software. So we believe that we have several years of opportunity to bring down the power space and cooling in a hyperscaler data center.

Question in the back row. Eric?

Yes. I wanted to follow up on what Brian asked. So maybe [indiscernible] on that point a little bit. So is it a 2023, 2024 for Pure where we're starting to hit that curve now?

Yes.

It is?

Yes.

Let me repeat the question. So basically, it's a time frame for when are we hitting [indiscernible] a market displacing more spinning disk. And the expectation is around '23, '24.

Around '23 as the beginning of it, right? Because again, it's not a -- it's a spectrum of applications there. And we'll hit the highest cost ones and slowly go after the lower cost ones.

And in fact, you were leading into part of what I was thinking. But in the -- in some of the cloud applications, is that lower power sort of the green light that's going to close that near term?

So for the first time last quarter -- we mentioned this in our call. For the first time last quarter, mostly in Europe but a couple in Asia as well. We started seeing deals close solely on the basis of the power savings.

And if you think about it, it's kind of natural to expect that as the acquisition cost of the storage media declines over time, some of the other costs that feed into the overall TCO equation start having a much larger weight, right? Said more simply, when you're spending $1 on storage and $0.05 to power, just picking some numbers, the power costs were there. They were part of the TCO equation. That's very different than, hey, if now it's $0.10, $0.15 on the storage. And now that power bill, instead of being $0.05 is actually $0.10 or $0.15 to go and power it.

$0.40 in France.

$0.40 if you can get it.

Yes. $0.48 in California, by the way.

Really? Wow. So I'm going to come back to some discussion on the power question because I do think that, that's relevant. But I wanted to ask a little bit about what I think about the cloud adoption question. And I'm not so much asking about you're interworking with the public cloud. But the IT sector is viewed negatively because theoretically, enterprises are migrating to the public cloud. And so who's buying on-prem if everything is going to public cloud? And I think the investment community is matured enough. We've recognized no, no, no. It's a hybrid. It's a multi-cloud. But how do you think about that ongoing trend because we're still seeing the growth rates of AWS, Azure, GCP are phenomenal as businesses. And so you're countering that. How do you think about that backdrop?

Well, we think about it in several ways. First of all, we want to make the public cloud a customer, right? So the best way is not to compete with the public cloud. It's actually to feed the public cloud. And we are working hard on that. We have one proof-of-concept with Meta. We have others with, for example, the SaaS players, ServiceNow, et cetera. So we still view that as an opportunity. Secondly, what we'd say is that as we're starting to see, although it's anecdotal, but repatriation because of it's -- the cloud is a great deal when you're in development because you don't have to buy your equipment before you start your development. You can pay as you go. Before you get into production, the amount of utilization is relatively small, right? So you don't have any CapEx. It's very flexible, so your developers can do what they want. The minute you get into production, now you've got high volumes, and that pay-for-use starts to get very, very high. Customers are finding out that production workloads in the cloud are very expensive. And we're starting to see some repatriation. And for other situations, there was a phrase that was used quite often before, lift and shift. You don't hear that much anymore. The economics of moving traditional workloads into the cloud just really aren't there for a variety of reasons. First of all, all of the applications need to be redesigned, refactored. And when an application is working, and if you're not going to get a lot of economic benefit out of it, why bother doing that. So we don't really see in the enterprise space a lot of migration of enterprise workloads. And of course, they continue to scale as well. So there's a variety of reasons why we think, first of all, hybrid is the way to go. We plan -- we're working to sell into the hyperscalers. And then finally, a lot of our technology, which we've not yet gotten into yet, is about creating a cloud operating model for our customers. So what do we mean by that? So storage has yet to be virtualized. What do I mean by that? Storage is in the enterprise environment is dedicated to workloads. And so you'll -- the typical environment is you sell an array, it works on these workloads. There's a different set of workloads over there, so you sell another array. Over across the room or on the other side of the country, a different workload, another -- these are all silos of data. There's no concept of spinning up a new application and just subscribing to a storage service. And what we're delivering now with Pure Fusion and PortWorx and to some extent, with Evergreen//One is the ability of enterprise customers to create pools of storage operated like a cloud, not just in a single data center, but across their data centers, and in fact, across their cloud providers as well where they just set up a set of policies and then a new application spins up and just calls -- through APIs calls upon the storage it needs. And so we're creating a lot of flexibility for customers across a multi-cloud environment.

And I guess maybe it'd be helpful to -- for the folks who are newer to the story to help get a better sense of the breadth of the portfolio. I remember when you first joined the company...

We had just built our second product.

Just built your second product, and you explained I don't want to get into sort of telling you guys revenue for each and every product, and...

That went over well, yes.

But I get it now so -- but you've got a much broader portfolio, different use cases. So maybe for the folks who are learning about help us understand the breadth of the portfolio.

Yes. Well, let me start and Rob, please fill in. So we have -- we benefit from the fact that we have a very consistent portfolio. We basically have 2 hardware architectures. One -- for those of you that have followed the industry, in general, we have one scale-up architecture, one scale-out architecture. But we have -- we leverage the same software for both. We leverage the same storage modules for both. And so it's just a question of whether the customer wants more performance with scale or whether they want low latency for the scale-up architecture. These are 2 very fundamental things for which we need those 2 architectures. Both of those architectures provide either and can be configured to provide either or both high performance and high capacity, low cost. So now many of our competitors have a half dozen or more products to fulfill different niches. We can fulfill it with just these 2 architectures, one software architecture and the different storage modules that we build for it. And so we go after, again, both the high performance and low cost. On the high-performance side, for example, we have our FlashBlade architecture. This is what feeds the Meta Research SuperCluster, the largest AI artificial intelligence supercomputer in the world. And we also sell them our what's called FlashArray//C, which is our low-cost general-purpose data store that's very low power, very low footprint to feed into that. And that's just one example. And then finally, we also have PortWorx, which is our software to support next-generation what's called cloud-native workloads. These are Kubernetes and container-based workloads that we do with our PortWorx. So that's on the product side. In addition to that, we have a variety of services. So we are unlike other -- all other hardware players, not just other storage players, but I don't care if you're in the networking industry or in the compute industry. When we sell a product, it never gets old. So what does that mean that when we sell a product, it doesn't get old? When we -- even if we sell the product for CapEx, it then comes with a subscription agreement afterwards. That subscription agreement allows that product to be upgraded on an ongoing basis, both hardware and software, without taking the customer's application environment down, without them having to change a thing. We actually can do an upgrade in less than an hour, even during working hours. And so -- and we just celebrated the sale of our first product 10 years ago that has been consistently upgraded. It looks like a product we sold last week because we just upgraded it again for about the third time last week. It looks like a brand-new product. It's very unusual. It's like driving to work in a car that's 3 years old. And when you arrive at work without stopping, it's a brand-new car. So it's a very unique model that we have, and it is like a SaaS service with equipment. We also have the opposite side of that, which is that with Evergreen//One, we will provide a customer an SLA. They will not buy a product at all. They subscribe to a class of service performance and capacity, and we manage everything behind that. It may be on their premise. It might be in a colo. It might be on a hyperscaler, but we provide storage as a service, and they only pay for what they use when they use it.

So I wanted to get back to power. You mentioned that, that was part of your script on the last earnings call. I think my note title was something effect of a power play. So remember that I picked up on that. But what I think is intriguing is none of this is new to Pure. It's changing in terms of the customer interactions. So I think, first, it would be helpful to sort of lay out some of the studies you've done and some of the numbers of how much less power do you consume and why, and then how is that customer engagement changed over time [indiscernible]?

I'll let Rob do it. He's done a lot of the work behind this, but identify both the disk to flash and the flash to flash.

Yes, absolutely. So Simon, in our initial ESG report, we outlined primarily a set of comparisons between our products and competitor all-flash products. And the net takeaway was, typically, we're seeing about a 5x improvement in power utilization or 80% reduction. And so where does that come from, right? Well, the primary drivers for that are, number one, far better data reduction. Number two, the density of the systems that we can go and ship. And really, the idea that with our direct flash technology, the integrated hardware software, we can just be far more efficient about delivering the bits of flash. And then number 3 is just other system efficiencies that we drive through the other parts of the integrated hardware and software. And so when you net that all out or multiply it together, that really gives you about a 5x advantage when you're looking at competitive all-flash systems because of the performance that we continue -- we're able to continue to deliver while driving those significant data reduction benefits, density benefits, so on and so forth. Now you shift that conversation in comparison over to disk, and your comparisons double, right? And so you're looking at really a 10:1 type of improvements there. And we've seen this both in our own analysis as well as customers coming back to us, right, and saying, "Hey, I made the switch to Pure. I might have made it for a variety of other reasons. But hey, these are the benefits I'm seeing in terms of significant power reduction, 75%, 80% reductions" and then meaningful cost reductions, right? We had a customer come back to us and say, "Hey, very much realized and can measure significant operational savings to the tune of [ 50% ] labor."

And are you seeing essentially this manifests itself more in terms of existing customers, recognizing they're using less power and saying, "Well, gee, I can deploy this more?" Or is it essentially a vehicle to land new customers?

Well, I think it's both. And getting back to the other part of your first question, Simon, I think that, to your point, none of these benefits are new, right? They're core to our technology. They have been since day 1. Now we invested in building set of technology upfront because you remember, when we started the firm, flash was 10 to 15x the price of disk. And so we needed a set of technologies to really close the gap between TCO and acquisition cost. What I'd say is that up until a couple of years ago, we were having the conversation about these benefits in purely dollar terms, purely TCO terms. And I think the environment around energy security, energy prices as well as awareness of sustainability efforts in ESG has shifted the units that we're talking about this in, but the core conversation isn't new. It's just we're talking about it in a different way.

So this sort of begs the question of getting an update on the competitive landscape. So we're coming off a period where one of your competitors reported last week, and storage sounded like it was going to be in decline in calendar year. Even the other players, the biggest player in the market is saying, "Well, storage is certainly better than servers, but it's slowing down." And so I sort of have this feeling that your competitors have painted a picture of, well, storage is not the worst thing, which just is not exciting.

Damning by faint praise.

Exactly. And I sort of feel like you've suffered for that perhaps unfairly. How do you think about what's going on in the competitive landscape? What are they doing differently?

I don't think you can support a high-tech business at less than 5% R&D a year. It's very simple at the end of the day, and they're falling behind competitively. They -- in the case of our largest competitor, they have a very high fraction of the market. And they [indiscernible] -- the GDPs are much more meaningful element certainly in their case. But I think for most of our competitors, including one that spends considerably more, but they're spending it in new areas rather than in storage, I just don't think -- if you believe that data storage is a high-tech market has to keep up with the ongoing advancement not only in processors and networking, but also in software. If you believe that, then you believe that technology is going to make a difference. And I think that's why we're different.

And I guess, historically, the subsector was always called hardware.

Yes.

I cringe when I'm introduced as the hardware analyst because that's ball bearings and steel pipe. So...

So I'll give you a sense. 95% of our engineers are software engineers, 95%.

And so that's -- I'd love to sort of get a perspective of the software aspect of your business. You talked about Evergreen, which part of that is the business arrangement for upgrading. But can you talk a little bit about the software aspects of the platforms?

Absolutely. So the Evergreen model is, in fact, software. I understand that we have our subscription financial element behind it, but it's enabled by the fact that we can do what are called nondisruptive upgrades on an ongoing basis. None of our competitors do nondisruptive upgrades, right? They all -- when you want to go to your bank to view your record at the banks over the weekend, and they say, "Sorry, the site will be down between 10 p.m. on Saturday and 6 a.m. on Sunday," it's largely because they're upgrading their storage systems. And they have to go through a disruptive upgrade. And by the way, sometimes they don't reopen at 6 a.m. on Sunday. Sometimes it takes longer because it's a very -- not only is it a labor base, but it is very, well, disruptive in a meaningful sense. It can be much longer than that. So the nondisruptive portion of it is something that is built into the core architecture of the product. And every time we do a new release, we have to be thinking about that nondisruptive portion of it. So that is a lot of software. But moving on to, say, PortWorx and as well as Pure Fusion creating that cloud operating model, that is all software. It's about automatic orchestration of large arrays of arrays, large -- and software, our Cloud Block Store in the hyperscalers is all orchestrated through software, but -- and it's not even software, it's SaaS. I mean, it's built through software, but we host that or rather we host it on Amazon. But that is what the customers then use to orchestrate all of their storage. And then finally, the last thing I'll mention is [ Pure One ]. So our customers manage their storage entirely through our SaaS environment. That allows them to have not just management, but increasingly active policy-based orchestration of their storage environment.

Yes, Simon, and I'd also just point back to our core array portfolio, right? If you ask me what really allows us to deliver on the price performance, power efficiency elements, it's the flash management software. It's a software running the storage controllers that we designed from the ground up for flash. Now you'll ask me, hey, so why do you build your own hardware? And the answer is quite simply that nobody else in the market has the software sophistication that we have to be able to go work with flash in this manner. And so we had to build our own hardware to go and support that. And so I think the fact that we're actually building hardware to support our software and to make those elements shine really goes to highlight how differentiated that is, the fact that nobody else has that software approach to support a vendor going and independently commoditizing that hardware really shows you that the IT and really the secret sauce, if you will, is in the software.

So believe it or not, we've run out of time. Because we had so much fun, it flew by. But I want to close with just one last question is if you sort of reflect on the investment community, what do you think is the least well appreciated out of Pure story in the stock?

I don't think the investment community fully understands the significance of Evergreen, which is our subscription-based services, right? Now I would say the analyst community is coming around, right? For many years, we suffered, well, isn't that support, right? Well, even a full SaaS service has support behind it, right? They also do their upgrades in the background, right? They have capital behind that, right? In our case, if you take apart our Evergreen, you will see that there's no difference between what we do and SaaS. Customer interacts to the service through a SaaS-based web portal. It's nondisruptive. It's constantly upgraded. It's always new. It just happens to sit on the customer premise.

That's great.

So I think -- and it's something that is growing for us. Now is what, 36%, 37% of our revenue, growing at over 30% a year, and because of that, by the way, we don't recognize in capital a resale of the same product, which our competitors do every 5 years or so, right? So that's built into the subscription line.

Great. Well, thank you very much, Kevan, Bob, Charlie. Folks, thanks for joining us with Pure Storage.