Pure Storage: Product Innovation Driving Growth With Profitability And Market Share Gains

Pure Storage - Pulling away from the legacy pack

There are many reasons to focus on Pure Storage (NYSE:PSTG) as an investment at this point. Of course the shares recently reacted to a quarter that was significantly better than anticipated. The reality of that, though, is beating a very de-risked forecast is probably not enough to recommend share purchase after a 40% appreciation in just a few days. The quarter did feature better sales performance than anticipated, particularly in the US enterprise, and for the company’s EvergreenOne SaaS storage offering. Overall, the demand environment hasn’t gotten better, but Pure’s field sales staff has apparently gotten better at dealing with macro issues such as multiple approvals.

One reason, that may not be immediately apparent, is that Pure is an NVIDIA (NVDA) partner and has developed AI ready infrastructure which is called AIRI. Just how much revenue is being generated by AIRI has not yet been discussed-it has been available for some time now, so I am not entirely sure if its availability alone would warrant a further purchase recommendation. That said, some of Pure’s likely growth will be fostered by use cases related to what I might describe as legacy AI-demand from those use cases has been important for the company for some time now.

What I think should be a focus of investors are a couple of product innovations that really have the potential to change the competitive environment in which the company has existed for some time. I acknowledge that often new products, or new use cases are overhyped and don’t deliver on the promises made by company managements and further endorsed by analysts. Yes, I have been guilty on occasion of overenthusiasm. But I believe that the introduction and potential broad acceptance of both Pure’s FlashBlade/E and its Cloud Block Store can be game changers, the former, more so, than the latter.

I have followed Pure essentially since its inception. Actually I first wrote about the company as long ago as 2016. I wrote then about Flash ending the days of spinning disc, and Pure’s market share gains. Those have been durable themes, and are more so today, than they were even 7 years ago. I can’t say that the annual percentage appreciation has been all that I might have expected; still the shares are up, in part due to the recent spike which has taken the share price close to the all-time high, while leaving the valuation at moderate levels.

In all that time Pure has never been able to compete across the universe of potential storage use cases and this has limited its growth although it has achieved a CAGR of a bit less than 22% over that span. I expect FlashBlade/E and other product innovations to add several hundred basis points to the CAGR over the next few years and I expect this to be profitable growth with free cash flow margins continuing to increase.

With FlashBlade/E, Pure now has a consolidated set of storage offerings that are, of course, all flash and span the gamut of price/performance steps in the enterprise storage continuum. As a result of this new product, the company CEO made a prediction that no more spinning disc would be sold within 5 years. Is that hyperbole?

I have seen lots of revolutionary technology in my life. I am old enough to remember when there were propellers on most planes and when Stowger switches and rotary dial phones were cutting edge technologies. When I first became familiar with enterprise storage, young men, attired in white shirts and ties were deployed on roller skates to change magnetic media on spinning discs in data centers. With FlashBlade/E the evolution has seemingly reached a point that was only a distant dream 50 years ago-it is that revolutionary, at least in my opinion.

These days, investors are fixated on the business opportunities of generative AI and well they should. The latest article that I published on SA benefitted from an app based on ChatGPT that SA has made available to contributors. The app reads articles and produces summary bullet points. It did so for my Nutanix article essentially in real time, and it has done so for this article as well. The points were created based on article comprehension, accuracy of themes and context. At some level it seemed a miracle.

And in a different way, FlashBlade/E seems similarly revolutionary, at least I expect that to be the case for those responsible for managing massive data centers who are searching for budget dollars. Just how fast flash storage arrays will replace spinning disc would be a guess. But the knell has certainly sounded, and it is likely to be part of the Pure investment case for some time into the future.

Reviewing Pure’s Quarter and its forecast

Pure Storage reported the results of its fiscal Q1 on May 31, 2023. The results were considerably above prior expectations, although to be fair, those expectations, at least in terms of published estimates, had been muted due to very conservative guidance the company had provided 3 months ago. The guidance for Q2 was increased noticeably, although the company’s full year guidance for revenue growth has not been changed in terms of its range. That said, almost certainly the consensus revenue forecast for FY ’24 will go from about $2.9 billion, which was growth of 6%, to revenues of $3-$3.1 billion or growth of just less than 10%.

Pure generated $590 million in revenues this past quarter, noticeably greater than the prior consensus estimate which had been about $560 million. The company’s non-GAAP EPS was $0.08 compared to the prior consensus of $0.04. The company is projecting revenue for the current quarter of $680 million, compared to the prior analyst consensus of $658 million. Its projection of $90 million in operating income for the quarter works out to about $0.28-$0.29, compared to a prior consensus of $.24. As mentioned the company did not choose to raise its full year revenue guide, although most analysts are increasing their estimates. For the full year, non-GAAP EPS estimates are now around $1.40 EPS, compared to prior consensus expectations of about $1.30.

Pure had what can only be called an extraordinary quarter for its Evergreen subscription offering, especially considering the current environment. Pure has a hybrid revenue model which distinguishes it significantly from its competitors. Subscription revenues are now 48% of the total. Subscription revenues grew by 28% year on year, almost offsetting the decline in product revenues. Subscription ARR grew by 29% to around $1.2 billion, and the RPO balance (basically the subscription backlog) grew by around 30%.

The overall revenue performance is actually better than the headlines suggest. Last year enterprise storage was still in short supply. Pure had the ability to deliver product because of its relations with supply chain partners and was able to ship $60 million of product earlier than had been expected. And even absent that $60 million, its product revenue grew at 37% well over trend rates. On the other side of the ledger, Pure’s subscription bookings are recognized ratably. Thus in a quarter like this, where subscription bookings were quite strong, reported revenues do not completely encompass that strength which is seen in the strong growth in the backlog metric, RPO balances.

Pure does not produce a metric for total storage bookings. But it seems likely, that despite the declines other competitors have seen in their sales of storage solutions, overall bookings of storage measured on an apples to apples basis rose for the company. The storage space sees many market share claims and that has been true since as long as there has been a storage space. But it would be hard to dispute that Pure’s share gains have been consistent, and have lately accelerated.

The company booked the largest international transaction in its history last quarter, an 8 figure deal based on the company Cloud Block Store technology. Cloud Block Store was announced more than 2 years ago. I will discuss that offering later in the article because it has a substantial, underappreciated potential. It is a technology that allows users to mitigate some of the elevated storage bills that have developed with the movement of many workloads to the cloud.

The company announced that it had made the first revenue shipments of FlashBlade/E. It also indicated that its forecast for the balance of the fiscal year had only a modest revenue ramp for FlashBlade/E. It is typical for major users to qualify and evaluate new technologies at some length before placing orders. And it seems only prudent to forecast that Pure should forecast a modest ramp for FlashBlade/E. But in considering the possibility for upside to the current forecast, I think some more rapid acceptance of the technology cannot be totally ruled out.

FlashBlade/E - What is it and why is it important

I would prefer not to leave readers with an impression that a single product is the key to the Pure investment story. Pure has a lengthy catalog of storage arrays, and consumption models. So do its competitors. One issue, however, is that with FlashBlade/E Pure now has a complete lineup of storage arrays for structured and unstructured data at all the points along the price/performance curve. While almost all large enterprises have multiple storage partners, many of those enterprises want to develop a preferred relationship with a single vendor. For years now, Pure’s user base could not consider the company as a full line supplier; they had to procure lower performance storage from vendors such as Dell and NetApp who have offered spinning disc and hybrid arrays for secondary storage.

With FlashBlade/E that changes; Pure can satisfy all of the storage requirements of an enterprise with a unified, all flash product line using the latest and most efficient architecture available as well as an all-encompassing operating system.

While industry analysts and stakeholders have been talking about the end of spinning disc for more than a decade, FlashBlade/E arrays are the first to actually offer cost/performance that makes it feasible to replace spinning disc for secondary storage use cases. It is a tough environment in which to persuade users to commit capital to replace an existing asset. But Pure does have its EvergreenOne SaaS consumption alternative. And FlashBlade/E has some stunning cost of ownership advantages that resonate even in this environment. Overall, the new technology uses 10% of the power and space as the hard drives it is designed to replace, with 10% of the manual processes and 10% of the heat dissipation of legacy technologies.

The company indicated that early interest in the product is “off the charts,” which is not terribly surprising given its TCO advantages and its lack of current competition. That doesn’t mean that FlashBlade/E and the slightly older FlashArray/C are going to replace all of the spinning disc installations in the next few years. It does, however, suggest that the basic growth rate assumed for Pure in most analysis will be exceeded. At this point it would be nice to provide some specific quantification-but I certainly do not have that kind of insight and I am not sure if even Pure management would want to try to forecast the demand ramp for the technology. Most 3^rd party market analysis no longer tracks the usage of HDD used as network attached storage in an enterprise environment. But the potential replacement market is quite large, especially given Pure’s current size.

Cloud Block Store and other product related demand drivers

The Pure investment case has always been based on market share gains and that remains the case. Most of those gains are based on sustainable product advantages when compared to competitors. There are all kinds of claims about performance and market share in the storage space. I am never going to settle those kinds of assertions. Often the conflicts as to share and growth arise from the computational methodology. Legacy competitors such as Dell (DELL), NetApp (NTAP) and Hewlett Packard Enterprise (HPE) are in the process of moving many of their customers from spinning disc and hybrid arrays to all flash configurations. Of course those sales are counted in most market share and growth calculations. But in terms of share of installed storage, Pure’s share gains have been consistent and they are all pretty much a function of offering more advanced products; Pure charges more for its arrays-hence its 72% non-GAAP gross margin, but customers enjoy significant TCO advantages because of savings in terms of energy usage, footprint, and the manual labor needed to operate. Pure apparently has more reliable flash simply because of its direct flash technology.

While its competitors do offer some SaaS consumption models at this point, Pure’s Evergreen subscription continues to be a significant competitive differentiator. Cloud Block Store is probably the ultimate manifestation of the Evergreen storage concept. Hyperscalers have offered their own storage for years. Pure's Cloud Block Store runs on AWS (AMZN) and Microsoft (MSFT) Azure. It winds up using less storage because of data compression, thin provisioning and what is known as deduplication. It facilitates data mobility between on-prem and different clouds. And the company offers a service called Terraform that helps in the process of automating storage and cloud workflows.

While Pure’s Block Store has been available for a couple of years at this point, it is just now gaining some significant traction with users. Last quarter, the company booked the largest history for its block store service since the offering was made available. In the past, and really up till this point, one of the investment arguments against Pure has been the migration of workloads to the cloud. It is hard to suggest that such a migration isn’t continuing. But the Block Store offering essentially allows Pure to compete successfully for cloud based storage. Block Store is one way users have to optimize their cloud storage spending without cutting back on functionality. The Cloud Store agreement is based on usage, and that has been a factor in reducing hyper scaler cloud spending, one of the principal goals of many enterprises in the current environment.

Generative AI is of course a technology getting lots of attention these days as well it should, given the revolutionary changes it is likely to foster over the next decade and beyond. Storage will certainly see demand growth because of generative AI, but the reality is that big data, which is required to animate traditional AI use cases is going to be a substantial beneficiary as well. FlashBlade, the flagship offering of Pure has a high-performance parallel architecture, and this is very useful for AI projects where performance is necessary to provide users with a satisfactory experience.

I think it worth noting, just as an example, the results and the product innovations announced by GitLab (GTLB) this week. What GitLab is delivering to its customers is not really a generative AI solution, but a solution built around AI capabilities that automate components of the software development process. And these use case often call for the capabilities inherent in Pure’s FlashBlade arrays.

Pure’s storage arrays shows up in projects that need massive storage as part of a machine learning paradigm. Meta (META) has been a significant Pure customer in recent years, although given that company’s cost containment strategy, any business with Pure this year is more problematic. But Pure does work with Meta in their AI research supercluster (RSC).

Competition in the storage space - There’s a lot of it and it is a risk for investors

When many investors think of the enterprise storage space, they think about fungibility. Okay, they may not think of it in that term, but they believe that all storage is the same and price wars can readily result as vendors vie for market share. After all, everything related to storage is simply a matter of capacity, latency and price. That is one reason, in addition to storage demand cyclicality, why Pure shares have never commanded a premium multiple. Price wars were a consistent backdrop in the era in which NetApp vied with EMC, which was one of the more aggressive competitors in the IT space-EMC was run as an organization in which the mantra of taking no prisoners prevailed, but even back then competition was never quite so simple as who had the lowest net price.

In my view, based on anecdotal checks and other 3^rd party research, fungibility is probably at a low ebb in enterprise storage these days. Just for completeness I have linked here to a presentation by G2 with regards to Pure’s leading competitors. I have also linked here to a direct comparison between Pure and Dell. While I almost invariably try to use 3^rd party analysis to look at competition for a potential recommendation, most of the time the reviews are so anodyne as to be useless for any investor. That is not the case here-the reviews posted about Dell by Gartner are about as negative as it gets, while Pure is given kudos.

There are other substantial competitors in the space; I think until Pure came along, NetApp had been considered as the leader in technology. That is self-evidently no longer the case. The review linked here is not as one sided as the review posted by Gartner, but clearly identifies Pure’s technology as leading the enterprise storage space.

While all storage vendors offer flash arrays these days, I think it is fair to say that Pure has typically led the industry in adopting the latest technology and continues to offer notable advantages for users. Pure offers direct flash modules which have increasing advantages over what are called Solid State Drives. There is considerable literature available on the differences in technology; I am not likely the best writer to explain those differences, but overall they suggest that Pure will continue to enjoy technology advantages within the flash market enabling it to gain share for the foreseeable future. I have linked to Pure’s primer on the difference between the traditional approach and DirectFlash. Most readers can skip the link but it does explain why Pure is different and why not all storage is fungible. At the end of the day, this difference goes a long way in explaining why Pure has been and seems likely to continue to gain share in the enterprise storage space.

I expect that the advent of FlashBlade/E will have a significant “drag” effect on the totality of the Pure product line. There is strong customer preference for working with a single, primary storage vendor. This has been a Dell advantage for years. And now, with Pure having flash arrays for all price/performance requirements of a customer’s storage estate as some describe it, there is likely to be a material change in that relationship which should translate into a sustainable increase in Pure’s growth rate.

Pure’s Business Model: Realizing the advantages of scale

It can be somewhat confusing to consider Pure’s business model in terms of a single quarter. The company has two basic revenue streams; i.e. product revenue and subscription revenue. When the relative contribution of those parts shows a substantial change, and that was the case last quarter, the expense metrics can look upside down. In the first quarter of FY’23, product revenues were 65% of the total, and this past quarter, product revenues were 52% of the total. Self-evidently, when product revenues occur, all of them are recognized upfront; the non-GAAP gross margin on product revenues was 70% last quarter, actually a little higher than non-GAAP product gross margins in the year-earlier period.

When Pure sells a subscription, self-evidently, not all of the revenue is recognized upfront, but all of the costs do hit the P&L in the quarter in which they were incurred. Pure’s subscription gross margin actually rose year on year last quarter. In this just reported Q1, subscription gross margins were 74% which compares to 71% in the prior year. Over time, Prime’s EvergreenOne will have better operating margins then product sales; the renewal rate of EvergreenOne subscriptions has been very high which is not terribly surprising seeing that upgrades are continually provided without additional cost-that indeed, is the reason for calling the offering Evergreen. But in a quarter in which subscription sales grew by 28% year on year, and by almost 6% sequentially, with a net growth of ARR of almost 10% sequentially, opex cost ratios will not be representative of the company’s longer term model.

Just for the record, the company reported non-GAAP expense ratios as follows: Sales and Marketing-36.5%; Research and Development-25%; General and Administrative-12%. The company obviously has a focus on research and development in line with a strategy in which market share gains are primarily a function of leading technologies in the flash space. On a sequential basis, operating expense went from $406 million in Q4 to $415 million last quarter.

Overall, non-GAAP operating margins were 3% last quarter, they are forecast to be 13% this quarter and 15% for the full year. These projections include the revenue mix shift seen in Q1. That mix shift will have a significant impact on the trend of operating margins over the balance of the year; so too will the ramp of FlashBlade/E.

Beyond this fiscal year, when revenue mix changes are less likely to be a factor and the ramp of new product revenues will be clearer, I expect non-GAAP operating margins to show sustained progress.

The company had a free cash flow margin last quarter of 21%. Free cash flow is quite variable in a single quarter and there is a notable seasonal fluctuation as well. The company doesn’t, itself, project a free cash flow margin; I am projecting a free cash flow margin of 29% over the next 4 quarters of 29%.

Pure’s valuation and recapitulating the case to buy the shares

Before looking at the valuation, I want to consider SBC and dilution. Just for the record, Pure does use SBC. Last quarter SBC was 13% of revenues, slightly above the 11% of revenues in the year earlier quarter when revenues were inflated by the company’s ability to fulfill orders in a timely fashion despite supply chain problems for the industry as a whole. I look at dilution as a more realistic valuation metric than SBC. Pure has been buying back shares at a rate almost equal to the dilution from SBC grants. Last quarter, outstanding shares rose by 0.6% sequentially. I have projected an outstanding share count for the next 4 quarters of 315 million shares. The company repaid the $575 million of previously outstanding convertible debt. Pure’s EV/S ratio is now just greater than 3X. Its valuation, considering the combination of free cashflow margin and revenue growth is more than 40% below average for my estimate of its growth cohort in the low 20% range.

Many stocks are seeing dramatic tailwinds because of investor enthusiasm with regards to the prospects for generative AI, and even for more traditional AI workloads. While Pure shares have appreciated 40% since just before its earnings release, the earnings reported were significantly greater than what had been projected, and the outlook for this current quarter was also increased. Raising guidance in this environment was a significant surprise. So too, was the performance of the company’s subscription business which is now 48% of total revenues, more profitable than its product business, and which grew almost 30% last quarter. That actually obscured the apples to apples performance of Pure in this last quarter.

Pure and Nvidia have had a partnership for years. They co-developed an offering called AIRI which is preconfigured for AI workloads. I wouldn’t buy the shares just because of the Nvidia partnership; it exists, it is likely to be a growth driver but try not to be dazzled by the AI hype in making portfolio picks.

More important to me was the introduction of FlashBlade/E, the first flash storage that is competitive with spinning disc in terms of price for non-Tier1 storage workloads. The company suggested that spinning disc would no longer be in the market in 5 years. Pure’s opportunity with this product is substantial and underappreciated, I believe. Pure’s Cloud Block Storage which is deployed on AWS and Azure is also an underappreciated competitive factor. Pure, prudently, has set its full year forecast including only a modest contribution from FlashBlade/E in 2H. On the other hand the CEO described the interest in the product as “off the charts.” It is hardly surprising that an offering that requires 10% of the footprint, 10% of the power, dissipates 10% of the heat and requires 10% of the manual labor compared to spinning disc is seeing such strong interest.

As mentioned, I have used a 3 year CAGR projection for Pure in the low 20% range. But its product advantages, and its opportunity to gain share could readily support a significantly higher growth rate forecast. And Pure has already become a significant cash generator, and its subscription business will increasingly contribute to free cash flow margins.

I expect Pure to generate significant positive alpha in the coming year.