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

Okay. Great. I want to welcome everyone to our fireside chat with the management team from Pure Storage. My name is Mehdi Hosseini, Susquehanna analyst, covering technology hardware. It's with a great pleasure to host this call. We have the entire management team from Pure Storage on this call. We have Charlie, Kevan and also Paul from IR. We are going to review a set of questions and topics, including the just reported quarterly results. If there's any question from the audience, you can e-mail me direct, [email protected]. There's also a way for you to text me your question through this Zoom call. But before we get started, I want to go through some [ regulatory ] disclosure. 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 most recent SEC filing under forms 10-Q, 10-K and 8-K. So with that out of the way, I want to thank the Pure Storage team for being available.

Perhaps we could start with reviewing the recent quarter. And the question for Charlie and Kevan is, can you just give us an overview of what is driving the incremental weakness in demand? And how do you see the rest of the fiscal year trending?

Yes. Thank you, Mehdi, and good to see you all. Thank you for your time today. As we exited Q4, we saw a strong Q4 grew 14% year-over-year. And then for the full year of last year, we grew 26% year-over-year. So strong growth in what was our fiscal '23. As we look to fiscal '24 and as we exited the final month or so of FY '23, we definitely saw the pace of development of pipeline slow down a bit. And we think that's largely macroeconomic factors, it's companies, customers looking over their budgets and coming to terms with whatever new realities their budgets have been set out for, for this year. And so we did see a slowing in the development of pipeline. And with that, that really changed our view of what FY '24 would be on a risk-adjusted basis and therefore, our overall guide in that space. As we look across the industry, we're seeing others that -- other companies that don't operate necessarily on a -- as everybody knows, there was a supply chain pipeline constraints. Some companies are still shipping from backlog. For those companies that don't ship on backlog, I'm seeing similar things. Some of our competitors are announcing same, similar weakness. We're absolutely adjusting our budget as we go into this new year to adjust for that, [ Ergo ], our guide for 15% operating margin. And we're also modifying our training for our sales force to sell into a much more near-term cost-conscious customer base. So I think that should provide you the context of our view.

And maybe what I'll do -- yes, just to add a little bit more in terms of just discussion on the Q4 a little bit and what we saw going on specifically with Q4. And obviously, Charlie talked about our revenue growth in Q4. And what gave rise to that performance? Well, we continue to see some good strength across our U.S. commercial business. Obviously, international continues to perform, in particular, EMEA with some softness, if you will, with the U.K. The other thing that was quite interesting for us as we were working through Q4 is we saw a really good conversion and conversion trends of our advanced stage opportunities. And so that was still kind of aligning with expectations, which we view as a positive. And where we saw a slowdown, to Charlie's point, was really around these sales progression and these newer opportunities in early-stage opportunities. And what was interesting in terms of that dynamic is that the demand was actually quite healthy in terms of what we saw in Q4. And frankly, it was a little bit stronger than what we were expecting. But where the dynamics shifted is, to Charlie's point, the progression of those opportunities slowed substantially versus what we've seen. And when those opportunities are projected to close, moved out a lot further than generally what we've expected. And that's really what gave rise to our more moderate guide as we look out to next year.

Sure. And just a quick follow-up here, guys. I did this question every quarter since you're an IPO. Some of the Q-on-Q compares, especially embedded in your guide is different than what your peer group has reported. We've had Dell and HP reported last night. Is it just purely an algebra or wrong comparison? How would you describe your guide that implies like 30-some percent decline in product revenue to your peers that have a different overview.

Yes. Look, I think -- Charlie, maybe I can comment first and feel free to jump in. But obviously, our Q1 compare, it does provide some complexity in terms of what we saw last Q1 and last year versus what we're seeing this year. But it's really what we saw last year. So there's a couple of things going on. First of all, as we spoke about, Q1 of last year reflects a $60 million pull-in, if you will, that we are quite clear on in terms of in previous earnings, in terms of some enterprise customers who are looking for earlier delivery, we believe that, that request is really being driven by the supply chain constraints at the time. But obviously, that's impacting comparability and that was all product revenue. But even if you pull out that $60 million, the Q1 of last year really wasn't impacted by the macro backdrop at that point, and we reflected and drove 37% growth year-over-year. So our starting point on the Q1 compare for this year is a really tough compare. And then you layer on top of that the points that Charlie made really around the sales progression of these new and early-stage opportunities is really what's giving rise to the guide that we've reflected for Q1 of this year.

Okay. Got it. And then just quickly with Meta, I remember a decade ago, having exposure to Apple, smartphone was a curse and also a blessing. Now we have the same situation with [ CSV ]. In that context, what was Meta as a mix of your fiscal year '23 revenue? And how do you see that mix evolving in FY '24 and '25?

Yes. Let me start, Kevan. And in general, it's absolutely the case that the large tech companies, which are, that's a significant market for us, slowed down even more than the economy, and you've seen that commentary across the board from suppliers into that environment. And therefore, our -- the guide that we gave does not assume a big pickup in that in our FY '24. So that could be -- and with respect to Meta, in particular, given the frequency, if you will, of how Meta orders and when they take their shipments, we don't forecast Meta revenue until we've actually received in order. Kevan, if you want to give any more color on that?

No, I don't think so, Mehdi. I mean, obviously, when we first started having larger Meta opportunities, we gave some color in terms of what that looked like. We did that in particular, given back in FY '22 in Q3, where we saw a drop in product gross margins. I wanted to give some color to the drop we saw, which is really driven by the Meta opportunity. And then obviously, we updated from a qualitative perspective, the Meta business, as we navigated through this year. So you should have kind of a general guardrails in terms of the Meta business in order of magnitude, but then to Charlie's point, as we look out to FY '24, given that we didn't receive any orders in Q4 that would not be reflected in our current guide for FY '24.

I will -- I think it bears repeating because we had gotten a question from an investor. The quality of our Meta relationship remains very good. So there's been no change in the tone or tenor of the relationship.

Have you quantified how big Meta was in FY '23?

We have not. We've provided some color on guardrails, Mehdi, and folks should have, and I think they do have a general range of what that business was, obviously, less than 10% in terms of total revenue, but we haven't given specifics on that.

Sure. And just as just 1 final question as it relates to the earnings conference call, you introduced FlashBlade//E, and you said it would start becoming material in the second half of fiscal year '24. Is there any way you can help us understand how big it could be as in FY '24?

Look -- and I don't think we used the word material. We said moderate ramp, and we were very intentional with those words. Look, we're super excited about the offering that we're coming out with, and maybe Charlie can spend a few minutes on that in terms of the opportunity set in front of us. But look, with the macro backdrop, we wanted to build that in. We believe that the ramp of FlashBlade//E will be near or breaking what we did with FlashArray//C, which FlashArray//C was a record-breaking product in terms of ramp for us. And we do believe that FlashBlade//E will follow along that track as well. But Charlie, do you want to provide just some color and some backdrop in terms of the opportunity set, what we're looking for in terms of workloads and overall opportunity for that lens?

Absolutely, and the audience will see why we're excited about it. When Pure introduced the first all-flash array about 10 years ago, it took only about 5 years -- it took about 5 years before the entire what's called primary storage market became all-flash. And that represents about 40 or 50 cents -- 40% or 50% of the total storage market today in dollars. But it only represents 10% to 20% of the total storage market in terms of bits or bytes stored. So still today, more than 80% of all data is still stored on a hard disk. Well, FlashArray//E is the arguably second product after FlashArray//C to be able to compete head-to-head with what is called secondary tier disk storage. It's a lower price, lower performance tier. And up until now, flash has just been too expensive for that tier. But now really, we're opening up 80% of the bites that were stored in the world that up until now have not been -- have been too inexpensive for flash to be able to go after. And we believe we have at least a 3-year head-start on our competition here. So we do believe this is going to be 5 years. I would project that in 5 years from now, all hard disk storage will migrate -- that is new purchases in that area will migrate to flash. So it's a huge opportunity. We estimate $5 billion to $10 billion TAM, maybe for FlashBlade//E alone, let alone its successors that will go after ever less expensive on a per byte basis tiers of this storage.

Mehdi, I'll just want to probably continue [ to add ] a little bit more on FlashBlade//E with Charlie. We were very intentional about putting a dollar per gigabyte out, which is very much -- that's very -- we don't do that. And I'd love to provide some color on why that we were very intentional about doing that and the importance of that from our lens.

Yes. So look, as Kevan said, it's the first time ever, I've put out a price publicly, right? There are lots of reasons why you don't do that. The reason we did it is because there has been general disbelief that flash can compete with hard disk at a price level. And by identifying $0.20 per gigabyte, not only did we sort of set an upper level, which could be used competitively, et cetera, but we should turn -- take any nonbelievers to let them know right away that this can compete head-to-head with hard disk in the -- in both the enterprise, not to mention, in the large data center environment. So we're going direct for hard disk replacement over the next several years.

Sure. It seems to me that the funnel or backlog of opportunities over the next 2 to 3 years is significant. You also have a macro headwind in the near term. So why change your sales team given the longer-term opportunity? And the near term is macro, who knows, but why modify your sales team?

We're not -- I'm not quite sure the nature of the question. We're not modifying the sales team. We are modifying the training that we're providing to them in terms of selling into a -- the different -- customers look at cost from a variety of different lenses. And when money was cheap, they were looking at providing functionality and capability at a reasonable price, but not necessarily looking at the first 12 to 18 months for breakeven. As you get into a recessionary environment like this, the expectation of getting to cost effectiveness comes down, right? And so they're looking for solutions that makes sense in the next 12 to 18 months. And that -- so that training, which we used to do a lot years ago, but then as we went through the COVID period, that became less effective because customers were looking more for feature functionality and other capabilities. But now that's back in vogue, and so we're retraining the team. But we're not changing -- there's no fundamental -- we have the same sales management, we have the -- there's nothing other than normal sales management going on. What's the nature of the question, Mehdi?

Actually, you answered it. To me, what separates Pure is focused on growth and free cash flow margin, you introduced this target of 35% to 40% revenue growth plus free cash flow margin. So as you make some -- as you modify or enhance your OpEx, I think this target is the main driver. And that's why I ask when you have all these opportunities, but why think about moderating OpEx. So is that correct to assume that you have this target of 35% to 40% revenue growth plus free cash flow margin? Yes, right?

Yes. So why -- go ahead, Kevan.

Yes. No, look, Mehdi, it's fantastic. And obviously, with the continued strength of our subscription business that we're seeing with Evergreen//One, in particular, especially in this environment as well as Evergreen//Flex. And hopefully, we'll spend a little bit of time across our subscription business. But when you take a step back and reflect on FY '23, it was terrific in terms of our revenue growth as well as our operating margin expansion as well as our free cash flow expansion. And so we're pretty proud of that. Now as we look out to next year and the opportunity set in front of us, the opportunity set is -- remains fantastic, and we're quite excited about it. Now that opportunity set has a backdrop of a tightening economic environment that we're certainly considering. And that's why we've reflected that in our guide for top line revenue. But we're also very focused in terms of prioritizing profitable growth. We're committed to a 15% operating margin next year. Both Charlie and I are very aligned with that, that 15% provides a slight expansion when you account for the 2 points of tailwind in FY '23. But it also considers the fact that operating margins expanded at a much faster pace in FY '23 than we anticipated. So we're still well ahead of our expectations specific to operating margin performance and our commitment of 15% for FY '24. And specific to free cash flow margins, we expect that to continue to be strong, slightly outpace and be slightly above our operating margin, obviously, being affected somewhat by product sales and [indiscernible] product sales, given the growth objectives we've given for next year specific to revenue. So hopefully, that provides you a little bit of color, Mehdi, in terms of how we think...

Yes. Yes. And you added to your buyback budget, but if I would just go with this free cash flow margin trend, you were still accruing more cash. How should I think about priorities? Would you become even more aggressive with buyback? Or would you be able to look out for tuck-in acquisitions or doing both?

Yes. And I'll let Charlie hit the M&A side at the end here. But specific to our stock buyback and return of capital through the stock buyback program. Look, our primary business objective with the stock buybacks, which remains the same today, is really to help offset dilution. That's the primary business objective of using the share buyback program. We did commit a couple of years ago with Analyst Day that we would look to drive about 50% of our free cash flow towards stock buybacks. Obviously, given the strength of our free cash flows in FY '23, we're a little bit behind on that. But I think directionally, we'll still target that 50% target. The other thing to think about in terms of use of cash is we do have the maturity of our converts coming up, and we're quite fortunate, frankly, to have the strength of the balance sheet that we have. And so currently, our current thinking on that is that we'll use all, and if it's not all, it will just be a small portion of our cash to fund that could be maturity of the converts and maybe use a little bit of our revolver, but we'll make that call closer to maturity of the notes. And then Charlie, do you want to talk a little bit about M&A thinking?

Yes. As the audience knows, we've made a couple of acquisitions over the last 5 years. And the Portworx acquisition continues to perform for us, and we're very pleased with that. I would say that in general, however, we're blessed as a company with lots of opportunity for organic investment as FlashBlade//E demonstrates. It's the reason why we have a relatively -- still a relatively high R&D investment. We believe we have lots of opportunity to continue to expand our product line and our TAM through internal development. We -- as everybody knows, external M&A was quite expensive during the bubble. We're starting to see some of those expectations come down in the market. And if the right opportunity comes along, either for a tuck-in or a small acquisition that can add to our internal momentum around innovation, we certainly are always open to that. We do have an active set of activities out there looking at different opportunities. But one never knows when one of those is right at the right price and available for us to take. But we continue to consider M&A, and we'll do so as we can go forward.

Okay. So I'm going to switch gears and focus on industry trend. I'm getting a lot of questions -- still some questions on the [indiscernible] product revenue trend. But I think going back to the industry trend, you may answer or may help clarify some of the concerns or confusion. In that context, Charlie, help us understand unstructured data is growing much faster than structured, unstructured data cloud, structured data on-prem, just to put it in perspective. And this is where like, okay, merchant versus captive market, to what extent would the cloud service providers would continue to work with like Pure or merchant storage vendor versus building out their own storage solution?

Yes. Excellent question. The hyperscalers as opposed to, let's say, the largest SaaS vendors and a number of the cloud -- the non-hyperscale cloud service providers, the hyperscalers, in particular, develop their own storage software and use probably 90%-plus hard disk. And they've been able to do -- they've done very similar development to our major enterprise competitors. And what I mean by that is that each of them, both our competitors and the hyperscalers started out with software that was designed to run hard disks, manage hard disks, manage data on hard disks. And to the extent that they needed to deliver higher performance, they would slide in an SSD, which is a device that uses flash to mimic a hard disk and they would slide that in, but their software was still designed for hard disks overall. And what that meant, an SSD is like using a computer as a typewriter. And it's underperforming. It is underutilizing the [ inherent ] flash technology, which is a semiconductor. We are the -- literally the only company that actually writes our software to manage the flash itself at a system level. And that is why we've been consistently years ahead of our competition in the area of price performance, in the area of flash technology as well as being able to utilize some of its inherent advantages. We're able to provide Evergreen. We're able to provide nondisruptive upgrades with that software, which is unlike anything else in the storage industry. Well, the hyperscalers have the same handicap, if you will. They don't have the advanced IP necessary to leverage flash at its best, including at a price -- at a cost at its best. And we think that gives us an opportunity over the next several years to sell into that environment at a system level. Now if you go beyond the hyperscalers to cloud service providers and certainly to SaaS companies. Those are areas where they don't develop their own code at the system level other than managing their individual application environments. So this is something where we've already seen success. The interest in our fusion product, which allows for large data centers to manage their storage like the cloud, like the cloud hyperscalers manage theirs, has been very high, and we're working very closely with both cloud service providers, MSPs as well as very large private clouds to allow them to manage their storage and offer it as a service to their developers just like the cloud companies do. So I hope that answers your question. What it is, is that because we have this unique IP in managing flash at a system level to maximize its performance and its cost, that gives us the opportunity well ahead of the rest of the industry.

Just a quick follow-up. When you're saying 90% of the storage is HDD based, that is predominantly the cold near-line storage, right?

Well, yes, near-line storage would not be classified as cold. It could be -- you could think of it as warm. It's what's known as secondary tier. And there's several different layers of secondary tier, if you will. But it's slower. It's databases that don't have a higher performance characteristic to it. It is data that is being held for later use. But again, it went nearly all of whether it's AWS or Azure or even if you were to go into Meta Facebook in terms of the type of storage for pictures and movies and even use their data and post, that tends to be all hard disk.

Sure. Now you also made a reference on the earnings call the other night that you're focusing on replacing a 7200 RPM HDD. Is that also -- does that also include some of the near-line? Or...

Yes. So there used to be -- without getting too technical on this, the primary storage -- for primary storage, there used to be 15,000 RPM and 10,000 RPM hard disks to satisfy the high-performance storage environment. Those don't exist anymore. Those have largely been discontinued because of -- because flash is taken over the primary disk market. What's left is this so-called 7200 RPM disk drives, which are slower, obviously, but much less expensive. And what we're saying is it's the beginning of the end for that segment of the disk drive market that flash now is able to compete at a price level.

Sure. So if I just try to bring it all together in the near term, like what is implied in your fiscal -- in your April quarter guide, perhaps could there be some volatility in the near term? This -- you said 90% of the storage is still HDD. As we look into that mix coming down, there will be volatility. So just because 1 quarter you have a higher decline on a Q-over-Q basis doesn't mean that there is a change. Perhaps some of your peers have more exposure to the more expensive legacy product, maybe that would skew some of these compares, but it's still very early. I think what you said will have a more meaningful impact 1 or 2 years or longer term, not more so in the near term. Is that the right way to think about the near term and longer term?

Well, I suppose you -- like, any new product or any product going into a new category or a new segment always has a ramp. It doesn't turn it on a dime. It's not like coming out with a new, the latest iPhone where consumers say, okay, well, I won't buy 13, I'll buy 14 now in mass. In the enterprise segment, customers take time to change behavior. And so we're expecting to see a ramp for FlashBlade//E similar to what we saw with FlashArray//C, which was a ramp, but it was a very -- FlashBlade//C was our -- the fastest growth of a new product we had ever seen in the company. And we would expect to see something similar to FlashBlade//E, but it's still a ramp.

So when you -- just, Mehdi, when you talk about the product revenue seasonality implied for FY at '24, and I think that's what you're getting at, right? So as Charlie talked about the new contribution of revenue from FlashBlade//E, we're really contemplating that in the second half. So that obviously would be a factor in our consideration. Another factor -- so another consideration, Mehdi, is, again, we talked about the strong demand that we saw in Q4 of early and new stage pipeline, which again was quite positive. And, frankly, higher than what we were expecting. But again, as we looked at the slower sales progression of those opportunities, that takes us to Q2, Q3 time frame in terms of really looking at converting those opportunities. And obviously, that's contemplated in the seasonality when you think about our implied product. The last thing that I'd probably point to is that when we look at seasonality, overall, and it's probably more first half, second half type of view. I think FY '22 is kind of a data point for reference as well for consideration.

Got it. Okay. Just on the FlashBlade//E, how does it -- how does this product open the disk market? Is it -- is there any more analytical way of thinking, because I'm assuming the flash E compared to C uses [ QLC ], but my impression is that it's actually focusing on a more cost-sensitive workload. And any color here? And then how does it help gain share from the disk market?

Yes. Thank you, Mehdi. First of all, we're very pleased more than half now of our shipments of capacity, that is the total number of bytes that we ship are now on QLC. So we have QLC on our FlashArray//C as you mentioned. Also, all of our flat new flash blades now are on QLC technology. So the question is, okay, how is E different than FlashBlade//S, which we introduced last year, which also uses QLC? Well, FlashBlade//S was designed for high performance. It's the second generation of the technology that powers that Meta research supercluster, the AI supercluster. It is what we ship into machine learning environments and some of the highest performance environments that are out there. FlashBlade//E, which also uses QLC technology, but it's being designed to do several things that really lower the price on a per byte basis. First of all, it is a very high density, but it actually starts the opportunity for customers to buy FlashBlade//E, starts at about 4 petabytes. So it is a -- it's a large, very dense system, which means the ASPs are high even if the price per byte is relatively low. It is designed to eke every last bit of efficiency out of the QLC flash. And it's designed to leverage the common elements, so power supplies and the controllers to manage as much flash as possible so that the contribution of the common equipment is lowered on a per byte basis. So again, really designed for cost. We have -- because of our intellectual property, we're able to use a much greater amounts of the flash that's inherent in the chip. And that gets into a technology that I'd be very willing to do a separate tutorial on that for anyone that might be interested in it. But it does allow us to lead in the area of the low cost per byte. Now just -- to be clear, we expect over time to have similar margins in this area that we have with any of our products that are similar to our company as a whole. Obviously, because it's a new product, we won't see that until we get to a little bit of scale.

Okay. So a follow-up to this and has to do with cost of ownership to your customer or data center, custodian. You are using QLC. There's also ESG requirement. Does this change the trajectory of HDD penetration, like if HDD is still 90%, would it take a year, 2 years? Or this is more like a 3 to 5 year to reduce that to like 80% or 70%? How does this...

Understood. Well, you mentioned ESG component there. So even though we're introducing FlashBlade//E at roughly the same price as, let's say, a new hard disk system, it operates with 1/10 or 90% less space, power, cooling, labor and e-waste. That's an incredible ESG benefit. If you look at large data centers with, as we said, 80%, 90% of their data being stored on hard disk and that hard disk taking up 20% to 25% of data center power. If you reduce it by 90%, you're saving roughly 20% of total data center power. And that's a huge number. And so especially in Europe, we're getting incredible response to that. The U.S., somewhat less sensitive to this thing right now, that is to energy savings. Although the Meta deal was won because of the space and energy savings that we put into place because of how large and significant the amount of data was in that environment. But the ESG savings is absolutely incredible. Now that being said, you're right, that 80% or 90% is just not 1 tier disk. Some of it is more near line, meaning has a little bit more performance requirement. That will be replaced first. There's other hard disks that is, as you pointed out earlier, sort of cold storage that has a lower price performance. We're not quite there yet, but expect to get up there by next year. So it will be layer by layer, that's part of the ramp that's involved in all of this. But I truly believe that in 5 years, as I mentioned, not a single -- there will no longer be hard disk to purchase in this -- for this environment. So you can think of it as a ramp over 5 years. Relative to our size, however, it can be a significant revenue growth element for us over the next several years.

Sure. And then just on Meta, just a follow-up here. They're private service provider. They have their own [ social-generated ] revenue through ads, but what about public cloud, the AWS, Azure and Google? How is the opportunities there trending? And I'm looking at it from a private cloud to the public cloud, is there any differentiation here?

Well, the only real differentiation is that the hyperscalers have -- as I mentioned earlier, have developed their own storage software technology. They use hard disks, but -- and they use standard compute for their controllers. But other than that, they run all of their own software for hard disks. As I've also mentioned, they have not created their own intellectual property for really managing flash as a semiconductor. They use -- when they use flash, they use SSDs. So the same software that they developed for hard disks. We believe -- well, first of all, these companies are large enough that they can do anything they want, but it doesn't really shorten the period of time it might take them to develop the kind of intellectual property that we've developed over the last 10 years. So we do believe that it would take them some time to be able to really leverage flash to its best price performance. And even if one of them does it, we feel that there will be other -- some of the others that will perhaps prefer to get a head start by utilizing someone like us. So it's anybody's guess. Our conversations with them continue. And hopefully, we'll be able to convince one or more of them to work with -- that is better to work with us for time to market and for mature technology than to attempt to do it all on their own.

Let me approach it in a different way, and this is a question that's not on our list. Does the [ bustling ] make a difference, like I'm assuming that your blade or FlashArray works with PCIe, [indiscernible] and so forth. But there are some of the new AI applications that rely on a proprietary like a [ NVLink ]. As you look into the future, is this something that could work in your favor or create some challenge or it's not relevant?

It's always relevant. To give a sense, we were the first ones to deliver flash on what's called NVMe and that -- which was a significant improvement in overall performance of the system, and we support NVMe over its TCP, over fiber channel over what's called [ RoCE ]. So we support many of the different NVMe technologies. And there's always a new link being looked at to improve the speed of performance between the processor and storage. And as you know, that's something that is performance-based storage. It's something we care very deeply about, and so we're always on top of those. But it's like many other things, it's a new feature or capability. And as long as you stay on top of these, I don't view it as something that is either -- as long as we're at the forefront of it, I think it's something that will allow us to stay on top. I don't view any of them as a threat.

Got it. Okay. And then 2 follow-up, maybe more for Kevan. How should we think about the length of these software subscription services?

Multiyear. I mean the duration that we're seeing with our -- and this is whether it's Evergreen//One or Evergreen subscriptions that are attached to product sales are actually pretty consistent in terms of what we've seen. Now we also see a healthy amount of on-demand usage as well with Evergreen//One. But overall, when customers are committing with us on our Evergreen//One SLA offerings, it's multiyear and generally consistent with what we've seen on our CapEx sales that also attach Evergreen. Mehdi, there's also been some other questions, I think, really trying to understand the seasonality from Q4 to Q1. And so let me just provide a little bit of color because we've talked a lot about the Q1 to Q1 compare. And then obviously, the implied product revenue from Q4 to Q1 is steeper than what we've seen historically. And I want to give a little bit of perspective on that as well. And really, if we go back to our Q4 performance in [ print ], I mentioned that we did really well in terms of our conversion of our advanced stage pipeline that was quite consistent with our expectations, did not have any significant or unusual pull-ins outside of trends specific to Q1 into Q4. But really what's driving the more pronounced seasonality from Q4 to Q1 is exactly what Charlie and I have been walking through, which is we saw really strong demand for new and early-stage pipeline opportunities. But given that the sales progression of those opportunities have slowed obviously, that's impacting the implied view in terms of product revenue for Q1 and really creating a more pronounced Q4 to Q1 impact. And so I wanted to walk through that just to make sure that we were answering the audience questions on that piece as well.

No, thank you. No, I mean, I've received some of those questions, too. And then what's the revenue mix by the channel partner direct versus OEM versus distributors? And how would it change over time?

Well, first of all, we are only -- we sell -- 100% of our product is fulfilled through a partner. So there is no direct whatsoever. We use distribution internationally, but not in the -- not in North America and North America -- or rather U.S. and Canada, we don't use distribution. So -- but internationally, especially in the Asia Pacific region, that is almost all based on distribution and in resellers after that.

Okay. But those partners are not part of prolonging the sales cycle, right? It's just purely end customer that is prolonging the sales level?

Yes. I mean the amount of time it takes to go from through the partner is typically reasonably short. So no, there's no -- it's all customer-driven. Yes.

Okay. So we are a minute over time. Is there any other topic that we need -- that I didn't address or question you that you want to just comment on before we wrap it up?

Well, the only thing that I'd like to add is that we're as confident in our medium to long-term growth opportunity as we've ever been. There's a lot of concerning consternation in the market overall with some of the announcements that are coming out and the guides that are coming out, but the thing to remember about recessions is they always end, and we know this very well, and we're just tooling ourselves up to make sure that, a, we gain more ground during the recession than our -- than any of our competitors, but more importantly, that we're able to accelerate coming out of the recession as well. So we have our eyes fully focused on the future and the opportunity ahead of us.

Great. Well, Charlie, Kevan and Paul, thanks so much for giving this opportunity to host this fireside chat. For investors and the colleagues, if there's any follow-up question, hopefully not on the April quarter guide, send me an e-mail, and I wish everyone a great Friday and happy and great weekend. Thank you all.

Thank you, Mehdi.

Thank you.