Good afternoon. Actually, it's probably still morning, but good morning, I'm Tim Arcuri. I'm the semiconductor analyst here at UBS, and for the next session, very pleased to have Western Digital, and we're very pleased to have both David Goeckeler, who's the CEO, and Wissam Jabre, who is the CFO. So before I start to ask questions, I'm gonna turn it over to Wissam to read some statements.
Thanks, Tim, and happy to be here. So we will be making forward-looking statements based on current assumptions and expectations, and I ask you to refer to our most recent annual report on Form 10-K and other filings with the SEC for more information on risks and uncertainties that could cause actual results to differ materially. We will also be making references to non-GAAP financials, and a reconciliation of our GAAP and non-GAAP results can be found on our website.
Great, thank you. So let's just start with the you know high-level topic of about a month ago, you made a big announcement of the business split after being immersed in a long, you know, strategic review process. Can you please help us sort of better understand how was the process, what elements led to the conclusion, and why was now a good moment to announce the spin?
Okay, great. Thanks, Tim, and we're very happy to be here. Thanks for, thanks for hosting us. So it was a big announcement. We're very happy to be here. We're very excited about the direction forward. Over the last three to four years, we've spent a lot of time really focusing on putting ourselves in a position to realize what we think the latent value in these franchises is. So we've done a lot of work on restructuring the business, bringing in a new team to lead the business, including Wissam here, investing in our balance sheet, getting some long-term overhang things out of the way, like a tax settlement. So we've basically been structuring the business to operate better and get the most out of it. And I think we've seen that in the downturn.
We've gone through a historic downturn, unfortunately, but in that downturn, we've seen all of the changes we've made in the business really paying off. We've outperformed our peers in each business, and that outperformance continues to expand as we go quarter over quarter. So we feel really good about where the businesses are. They're ready to stand on their own. They're executing extremely well. We went through a very thorough strategic review, basically very public. You know, we're going through a strategic review. We're gonna consider all options for how do we recognize the value in these two great franchises, and I think we ended up in a very good spot, which is we chose the, we chose the solution that we think is going to...
Given the time in the business and all the changes we've made in the business and the way they're executing, it is gonna let all of you, all of us, realize the true value of those businesses and let investors choose which or both franchise they want to invest in. They're different franchises from a capital return point of view, from an investment point of view, and so we understand that there are different profiles for those businesses. So why now? We went through a very robust process. A lot of people came to the table, a lot of ideas were discussed, and we chose the one that is fully within our control to execute. We have the businesses in a position where we feel like they're ready to stand on their own.
We have the teams around them that can execute those extremely well as independent companies, and the market is improving at the same time. You know, we'll talk about that, I'm sure. We've gone through a historic downturn. The worst is behind us. It's now just a question of how fast we recover. So we're very excited about that. We think we've got the right strategic direction forward, and we're very excited to go execute on that over the next nine to 12 months.
Can you talk about some of the changes, or some of the improvements in each of these assets that have occurred, having them under the same umbrella, such that when you re-separate them, that these assets are independently better than they were prior to the combination?
Yeah, it was really... You know, so a lot of the changes we've made over the last three to four years were about bringing a lot of focus to each individual franchise. They have to be. They solve the same problem at the end of the day, storing data, but they do it in very different ways, different customer buying patterns, different technology roadmaps. It's really important that each one is managed as an independent franchise. So we've done that, and we've got to the point now where we put that focus around them first as business units and brought in general managers, so we make sure that any dollar we're investing in the business, we get the best return out of that.
We then moved on and separated the operational assets kind of organized in a way where it's easy to kind of take these apart. There clearly were some synergies in the current model around how we go to market, about, you know, a little bit on the manufacturing and procurement side, but we think those are very manageable to overcome with the better execution and the better franchises we've just built over the last three years, and when you net it out, this is a better, this is a better position for value creation going forward.
I guess a question for Wissam. Number one, when do you think we can get more information in terms of the capitalization of these assets? Number two, sort of what's the guiding principle and the framework for sort of understanding how the cap structure for each of these assets would be constructed?
Yes. So, we still have a lot of work ahead of us to get there. And as we approach the date of the separation, we'll be hosting Investor Day to provide much more details and describe all of this. Of course. With respect to the guiding principles, you know, we aim to construct capital structures that would provide the ability for each of these independent, publicly traded companies to grow and invest on a sustainable basis for the long term.
And so we will be looking at the growth opportunities as well as the changes in profitability for each of the business through cycle to help us inform how best to think of this, as well as the capital allocation strategy for each of them being balanced based on the markets and the profiles of each of the businesses. And so, for instance, when you look at the hard drive business, it's a bit more resilient with respect to cash generation, with more opportunities to delever.
It's got a good growth profile. When you look at the flash business with much bigger need for capital, it's much more capital intensive. It's got much bigger needs to invest in technology. And so, you know, each of the businesses have a different profile, and those profiles will be the capital structure for each of them to ensure that they're successful.
So is it fair, just on paper, to just look at, you know, you have one major HDD peer? Is it fair to just take your drive business and say, "Okay, well, that business should have a similar balance sheet to that peer?" Shareholders, I would think, would want a, you know, dividend also for that business, because to your point, it can, you know, it can pay one. And you sort of apply that to the drive business, and then, you know, whatever's left over goes to the NAND business. And obviously, to your point, the NAND business cannot shoulder as much, if any, debt, as the, you know, drive business could. Is that a fair, just high-level framework for thinking about how these two businesses will be separated?
Yeah, I don't want to talk ahead of myself here. The way to think of it is, we understand very well the competitive landscape for each of the businesses, and we will be—and we aim to construct capital structures and companies to be independent, publicly traded, very well competitive in their own sectors.
Great. Okay. David, let's talk about SMR versus HAMR, and let's just talk about the broad environment of the technology roadmap and what customers want. You know, customers only care that they're getting a drive in their existing systems, and they care about the capacity and cost. So can you just talk about your decisions around the technology roadmap that you're pursuing?
Yeah, I mean, I think you got it right. I mean, I think. You know, we feel like we have a very big obligation to our customers. Storage is 29%-30% of the TAM of all data center spending, so it's a big segment. HDDs are, you know, 85%-90% of that storage. We don't expect that to change appreciably over any time that's forecastable. HDDs are, you know, provide a very compelling value proposition for the bulk of storage in the cloud. They've got a long future ahead of them. And what our job is in the industry is make sure we have a roadmap on the technology, that we can continue to deliver higher and higher capacity points at lower TCO.
And if we can do that, then our customers stay happy, and we're able to generate free cash flow, and our investors are happy, and that's kind of what we're doing. So we think very deeply about that technology roadmap that sits between those two things. But you're right, the customer wants the outcome, right? The customer wants the 28 TB drive or the 26 TB drive or the 35 TB drive when we get there, whatever it happens to be. What's on the inside is less important. And I think if you look at what's happened in the industry over the last three years, roadmaps have diverged quite a bit. The drives that each player in the industry is building are fairly different. We have put together a roadmap that has very.
A lot of levers of innovation in it over the last three years. So, we first introduced ePMR, so energy assist into the read/write head. So that was a big innovation, right? That gave us more, more runway on our ability to drive higher-density products. That technology is now fully commercialized. We produce it at very high yields, the right cost point to derive the right economics for us and the right outcome from our, our customers. We then commercialized OptiNAND. That gives us a control point in the drive where we can, like, put more software in the control plane, get more density out. That's what leads to UltraSMR, where, you know, SMR is a technology that's been around a long time. It always delivered about 10% more capacity on a drive. With UltraSMR, we get 20% more capacity.
We've been layering in these innovations to give us a long roadmap of ability to continue to deliver to customers, highly reliable, high-quality drives at higher and higher capacity points. That strategy has been playing out over the last three to five years. I think it's very well constructed, and the results are very clear. We are able to deliver the highest density drives in the industry right now, and we have a long roadmap up until 40 TB on that technology that's already been commercialized.... We're going to continue to run that play. You know, if you look at our profitability, we have the highest profitability in the industry, and that level of profitability has been expanding over the last five quarters. So the strategy is working, it's working extremely well.
You know, at some point, HAMR will be folded into that roadmap as another innovation point that drives us higher and higher. You know, when that technology gets to the point where it has the reliability and the ability to produce it at the right cost and all of those kinds of things, we'll fold it in. You know, our calculus many years ago was we needed these other innovations in the roadmap to get to that point, and I think that that calculus is playing out to be absolutely correct. You know, we'll continue to drive our roadmap forward. I think the big picture message is, HDD technology has a long way to go. We can continue to deliver a very compelling value proposition to the most sophisticated data center operators in the world.
And so we have a lot of confidence in the ability, in the HDD business to continue to grow, for us to continue to innovate and deliver a very good business. And, you know, again, to get back to your first question: Why split? Why now? Because we want that, we want that value to be recognized, and it's going to give everybody out there the ability to invest only in that business. You know, that is business is a different business from, you know, the amount of investment required, the capital structure of that business, the expectations of the investor community in that business is different than the NAND business.
And so that technology, very long roadmap, very good customer base, and, you know, we have enormous amount of confidence in the roadmap that we've put in place, and we're executing every quarter to deliver that value proposition to our customers.
So I guess, is it right to think that, I mean, you're trying to solve some very advanced physics and material science problems, and when these problems exist, you want to have as many levers as possible to optimize capacity? Is that sort of the, you know, overriding-
That's the way-
-philosophy
... you should think about it. I mean, the hard drive is a system, right? There's a lot of stuff inside of it where you can innovate. And the more innovation levers you have, the more ability you have, more optionality to drive higher and higher capacity points. You know, obviously, the read/write is part of it, the number of disks in it are part of it, the control plane and what you do at a software layer is part of it. There's all different components of that. And, you know, we've built a drive, and we've built that system where we've got a lot of different levers that we can continue to pull and drive that roadmap for a very long time. And you're right on, you know, one thing that's worth saying is, you know, this technology has very long commercialization cycles.
You're talking really deep physics, material science. These are things that take decades to play out and commercialize in the market. So having as many innovation levers as possible to be able to really instrument that roadmap to deliver a consistent level of capacity increases and TCO declines to your customers is extremely important because that's where we get the returns in the business, right?
Is there an intersection point where you'd say HAMR, for you, intersects the roadmap at a particular density?
Well, I mean, I think it all depends on... You know, HAMR is technology that's been in development for 20 or 25 years, right? And I think what we're watching right now is the final couple of years of getting to the starting line of when we start to transition to that technology. And once you get to that starting line, that technology transition is years and years it takes you before that's then the dominant technology. And so, you know, when we have confidence in the technology we have right now, that we can deliver this very strong value proposition to our customers, it delivers the right amount of profitability to us.
That technology has many generations to go. As HAMR can be folded into that roadmap and gets us better economics, then we'll do that. We're not at that point yet. We don't need that. We're not. The technology we have right now is very clearly commercialized and accepted by our customers. And like I said, the profitability is, you know, just look at the numbers, it's showing up.
In terms of the cost curve, is it fair to think, HDD cost curve, longer term, is going to be mid-teens? Do you think that we can sustain that?
As far as cost down?
Yes.
Yeah, that's probably a little aggressive, but it's, it's negative, right? It's, it's like high single digit, double digit down, and I think that that's plenty of cost downs to, you know, continue to deliver a better TCO and expand the, the market. I think storage is an interesting market. Drive the cost down, you drive elasticity, people store more data. Our customers tell us all the time, "We want to, we want to store more data. If you make it more economical, we can store more data."
The way we do that is we keep, we keep driving innovation that drives those costs down. And, like I said, good news is we're talking, you know, we're, we're talking about, you know, innovation that takes decades to commercialize. You know, we're approaching the starting line of, and we're past the starting line on things like ePMR and UltraSMR. We got things like HAMR coming up, and I think our ability to execute that roadmap in a very consistent way will lead to very good proposition for our customers, where it starts, and a very good outcome for our shareholders.
Since you are in both the NAND business and in the HDD business, are you seeing just given how much pricing has come down in NAND, are you seeing particular parts of the market where SSDs are pushing into segments of the market because NAND pricing has come down so much that it's beginning to, you know, cannibalize HDD faster because NAND pricing has actually come down so much?
No, we're not. I mean, NAND is a great technology, right? There's always new use cases for NAND. There's always ways you can innovate in NAND and build clever systems that you put in parts of the market that deliver great value propositions. We're huge proponents of NAND. That is not coming at the expense of HDDs. HDDs are still bulk storage. It's still the primary storage medium in the cloud, right? The part where it was a substitute was on the device side of it, and that substitution has been happening over the last 15 years, and we're kind of at the end of that period, right? Not very many people buy a laptop anymore that have a hard drive in it. That's all client SSDs. That's fantastic.
But if we look at just Exabyte, you know, the deterioration of Exabyte in this downturn, Nearline and Enterprise SSD are about the same. In fact, I would say in the NAND business, one of the surprising things, maybe to a lot of people in this downturn, is going into this downturn, Enterprise SSD was seen as the best market in the NAND business. Like, how fast are you going to get in there? How, you know, it's the part of the market that suffered the most in this downturn, right?
I mean, we've essentially lost three years of growth on Enterprise SSD compounded growth rate, if you just look at Exabytes deployed into Hyperscale data centers. So I don't see that mix changing appreciably. They're complementary technologies. They're both great technologies. They're both growing. There's no doubt the demand for NAND is growing a little bit faster than the demand for HDDs, but that's okay. They're both growing.
You think the right longer-term? We always try to zero in on what the mass capacity, you know, markets can actually grow at. What do you think the right long-term growth rate is for that market in terms of exabytes?
You know, I think it's mid- to high 20s, but it's a little bit TBD with AI. We don't know yet. I mean, AI is just being deployed. We're at the point of AI now where we're basically the cloud is gearing up to make all of us enabled and everything we use every day to have AI capabilities. And I think, I mean, I think that's the beautiful thing about the cloud distribution model of software. Like, the cloud brings all of this technology to all of us. We can all access it, and I think once that happens, we're going to see... I think it's going to be a tailwind for data growth in general.
I mean, one way to think about it is, like, there's an enormous amount of investment going on right now to automate the creation of data even more than it is today. So, you know, I think as you look at the market now, I think we're in a mid- to high-20s growth rate on exabytes. I think we'll see once we get this technology deployed. It's kind of a second wave beneficiary, is the way I would think about it, of all the AI deployment.
Yeah, I mean, the bear, the bear would say, and I, I have these discussions with investors all the time. The bear would say, "Well, if you look at the last five years, the, I mean, it's fluctuated massively, but the CAGR has been in the mid-teens. It's been 15%, 16% if you look at it over the past five years. So if we're in the mid-twenties, that sort of already embeds, yeah, I mean, you know, AI ought to, ought to accelerate the market, but, but that's kind of a long way from where we've been the past five years." That's what the bear would say, but-
Yeah, well, CAGRs are very dependent on the endpoint you pick, right? So if you pick, like, we're in a historic downturn. We're coming out of a pandemic, where there was a lot of inventory creation. We went into cost of capital going out for the first time in 15 or 20 years. There's, like, all kind of distortions on the market. If you pick that as the endpoint, you're probably going to come up with a CAGR that's lower than what if you draw a longer trend line from... I encourage people to draw a trend line from well before the pandemic through the pandemic. We're going to return to that kind of line. And I think it's very exciting for the industry right now. We're on. That recovery is on. The question is now, how fast is it going to happen?
We're seeing, you know, we're past the bottom in both businesses. NAND prices are increasing. NAND demand is in... You know, we first saw demand increasing, now we're seeing pricing increases, exact behaviors you want to see. In the HDD business, we're seeing demand come back. We have more visibility. We've talked about it now for, you know, a quarter or so. We expect sequential improvement as we move throughout the year. We're not back to where we need to be yet. Don't get me wrong, we're not happy with the numbers, but the trajectory is now in the right direction, and we're coming off a historic low position. So I don't think we completely know, especially in NAND, what the recovery is going to look like just yet, right? Because we're coming off a very low point.
But very clearly, in that industry, we, we need to see a significant recovery in pricing for the investment to come back to the business. And, you know, the, the overall demand picture isn't changing, right? We're not using technology less than we did before. We're using more technology than we did before. So it's not like we're fundamentally not going to need the technology in the future. So the trend-- we're going to return to the trend line. Question is, how fast? And to your point, what is the trend line? If you pick different points, you're going to come up with different numbers. But I would encourage people not to pick the last year as the ending point, and then go back a couple of years and draw a line between just those two.
Got it. It does seem you just, we haven't talked about NAND yet, but you did just mention NAND, and it does seem like things are coming back a little faster in NAND than they are in, than they are in HDD. And so I wanted to ask sort of where you think we are in the, in the, in NAND, and I know you did say that NAND inventories are at a, you know, four-year low. And I think there's still some confusion because we still see there's still, you know, $80 million worth of underutilization charges, yet inventories are at, you know, four-year lows. I understand that we're in the early phases of, you know, things getting better. But can you just talk about NAND?
It seems to me like the business actually is coming back pretty quickly, actually, if you look at pricing.
Oh, you want to talk about timing of underutilization charges a little, little bit? Because some of that is a timing issue.
Yeah, exactly. Some of the, what you're discussing, Tim, is timing issues. We've always, over the last several quarters, we've been focused on managing the supply of our products in a very dynamic way. We look at where the demand for the products are versus where the inventory is and how we manage also our cash flow. And so this is where, in many respects, that really paid off by the end of September, we were at the four-year low from a days of inventory in the flash business.
We've seen approximately $400 million reduction in inventory in that quarter alone. And so, as we sort of, as well, provided for in terms of visibility for this quarter, we continue to see that underutilization because we continue to manage the supply of the, basically, the fab outputs or the wafer starts in a dynamic way-
You're not cranking them back on.
Relative, relative to where we see the demand of the products. Yeah, the plan isn't really to sort of crank it back on completely. The plan is to continue to manage it in a dynamic way as as we see where inventory looks like-- what the inventory picture looks like and where the demand of the of our products is.
So let me talk about the pricing a little bit to your question. So we're in a point where the market has turned, and so there's a wide disparity of pricing because there's a... It's a big market, and it moves at different paces. So there's everything from markets where people just sell wafers, right? Which the price can change every day or even multiple times a day, to pricing for retail products, where price, again, and promotion happens constantly, to channel business, where you can change the price every week, to OEM business, where you negotiate it every quarter, to, you know, some customers have year-long contracts. Small number, but there are some of those kind.
So when the market starts to turn, you see the pace of each of these markets move at, you know-- they move at different rates because they're negotiated at different times. And so there's no doubt what the trajectory is now going in the right direction. The question is: How fast is it going to ripple through each market? And the further you go on that, the more liquid markets, they're moving very, very fast. And now the question is, is how fast is that going to translate into, let's say, the quarterly markets? And we don't know yet, but what we do know is we're coming off of a very low point. And so, they have to increase a significant amount to get back to what would be typical through-cycle margins.
You're gonna want to get back to typical through-cycle margins before you start to invest a lot of CapEx back into the business. So the prices have to come back a significant amount. And, you know, I think, I think peak to trough, they're down over 50%. So very clearly, just to get back to where you, where you were, you know, you had cost down, which is another, another, just another way to think about the market, but you got to come back a significant amount. And so we're now in that process. I think that process will play out over the next several quarters. We'll see how fast it goes, right? But in general, the recovery is on, and now it's a question of the magnitude of it.
So I just wanted to, in the last time here, I wanted to talk about your NAND roadmap, because-
Yes.
You've always done a really good job, and on a per wafer basis, you spend way less than what your peers do from a CapEx perspective, and yet you've always had very, very competitive costs. So, can you talk about what's the magic there? I mean, you have more of a... I kind of perceive it to be more of a brownfield approach versus a greenfield approach. You know, Samsung, for example, takes a greenfield approach. You take more of a brownfield approach.
In answering the question, can you talk about the transition from BiCS 6, which, you know, is a fairly small portion of your bits, and it seems like you're going to jump to BiCS 8, which optically you say, well, to go from a 112-layer to 3XX for BiCS 8 is a pretty big jump in terms of technology. So how do you help investors sort of mentally think about mitigating the risk?
2x, so 162-218. So look, but your question, the first part of your question, this is where the JV is so important. Like, we invest as the largest player in the market, right? Because Kioxia and ourselves, we have one R&D team. We have one technology roadmap. That technology is the foundation of our NAND business. If you get it wrong, if you get the cost wrong on that, it's very hard to make it up on, you know, on the product side of it. We've developed, I think, 17+ generations of NAND. I mean, Kioxia invented NAND. So we have a team that has been doing this for over 20 years, and that team has had a very explicit goal of minimizing the amount of CapEx required to get incremental output, right? That's a design goal when you go into this, right?
So it's not just: How do I get the most bits? It's: How do I get the most bits at the least amount of CapEx with the right cost down? You got to get all three of those variables correct. We spent a lot of time thinking about that and have it as explicit design goals. If you look at over, like, the last five years or so, you look at the numbers, to your point, we spend, like, a third less than the industry average on the amount of additional CapEx we put into the business.
That's a very, very good thing.... Right? That showed up a lot in the downturn, right? Our margins were significantly better than our peers. So why do we, why do we do that? Because it's a design goal. We've got a team that's been doing it for decades, and we're able to invest as if we're the largest player in the market because of the JV, right? That is the magic behind the JV. We get economies of scale of being much larger than we are, and that's why the JV is so important, and that's why the JV has gone on for 23 years, and that's why the JV continues to be very, very healthy.
Yeah, I guess my question was more my BiCS 6 to BiCS 8 question was, well, you know, BiCS 5 is the workhorse right now for still, you know-
Right.
most of the production.
That's right.
It seems it's a monumentally big leap to go from BiCS 5 being the, you know, bulk of your production, to then say, "Okay, I'm gonna, you know, go to BiCS 8," so.
So how are we thinking about that? So BiCS 8 performed extremely well, ahead of schedule, very great innovation, wafer bonding. There's a lot of great advantages of that. If we had more time, we could talk about. So usually, in the portfolio, you move the entire portfolio forward, node to node. Like, everything's on BiCS 5, you move everything to BiCS 6, then you move everything to BiCS 8. Because BiCS 8 is ready, and performing so well, we're making more strategic decisions about what part of the portfolio to take to BiCS6 and what part to just go from BiCS 5- BiCS 8. So you do that by product by product, by, you know, client SSD in a certain part of the portfolio, enterprise SSD products, products for gaming. Which ones are you gonna take to BiCS 6? Which one are you just gonna go to BiCS 8?
So it's more of a, as opposed to a whole, the entire cohort moving forward, kind of more of a hopscotch approach. You move some parts of the portfolio here, you take some parts of the portfolio up there.
Got it. Well, we're out of time. Thank you to both of you. Appreciate it.
All right, thank you. We appreciate your time.
Thank you for having us.