So why don't we go ahead and get started? So I'm Aaron Rakers. I'm the IT hardware and semiconductor analyst here at Wells Fargo. I appreciate Wissam Jabre, the Executive Vice President and CFO of Western Digital, joining us here this afternoon to talk a little bit about hard disk drives, talk a little bit about flash. To say that you haven't been busy over the past year plus is probably an understatement. You've had a few things going on with regard to what's going on with Western Digital and, you know, spinning out the flash business and what was announced in the most recent quarter. So first of all, I think you've got a piece of paper there.
You want to read the safe harbor, and then I'll jump right into questions, and we'll go from there.
Thanks so much, Aaron. Happy to be here, and thanks for having me. Good afternoon, everyone. Let me first start with reading the safe harbor statement. 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 our 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 the reconciliation of our GAAP and non-GAAP results can be found on our website.
Perfect. Good job. So both your businesses have gone through some, you know, down cycle dynamics here in the last handful of quarters. You know, I was just looking at some metrics before we got on stage. You know, hard disk drives have come off, and, you know, what I wanted to really kind of get at with this discussion is really understanding some of the foundational things that Western Digital's put in place throughout the midst of this downturn. We'll dive into both hard disk drives and flash, but what I really want to understand is how we've gotten, you know, those foundational things to get to the point where, you know, the company's going to, you know, separate the two entities as we look in the next year. So maybe we start on hard disk drives.
You know, can you talk a little bit about some of the things that you've done over the last year or so to really kind of realign the business, you know, in the midst of the downturn that we've seen?
Of course. Well, let me first start by saying I'm thrilled that we announced with our earnings the completion of our strategic review. Basically, we also announced the separation of the flash business and the creation of two independent publicly traded companies that would be leading in the storage industry, the HDD and the flash businesses. That will basically allow us to unlock the value, which was the purpose of our strategic review, and leading to that, we had been doing a lot of work over the last few years to align the two businesses.
So, first, when David Goeckeler, our current CEO, joined, he put in place the two business units focused on hard drive and flash. And then we reprioritized the capital allocation to delever the balance sheet. So since then, we've reduced around $2.7 billion of debt. And with the beginning of the cyclical downturn, we've taken the opportunity to reduce quite a bit of cost from our system, whether on the fixed cost side or on the OpEx side.
So, when we look on the hard drive business, for instance, we've taken out 40% of the client manufacturing capacity, reducing the fixed cost related to manufacturing by around 15% if we compare it relative to where we ended fiscal 2022. In parallel, we've also reduced the CapEx investment in both the hard drive and flash. And so that, on the CapEx side, we've reduced overall in fiscal 2023, around 35% relative to fiscal 2022. This year, we've taken also down our cash CapEx significantly to continue to focus on cash generation and ROI.
In parallel, we've taken a lot of operating expenses out of the company, where if we compare, for instance, the most recent September quarter, where we ended at the non-GAAP OpEx at $555 million, we compare that to the fiscal Q4 of 2022, which was at $760 million. You can see that we've taken quite a bit of OpEx down from the total company to basically continue to manage the profitability as well as cash generation. Over the same time period also, we've settled a tax dispute that would basically help clarify things.
And so we've been really focused on taking action to strengthen both businesses and allow both businesses to stand on their own in a very sustainable way. On the flash side, we've also reduced the CapEx quite a bit in fiscal 2023, and we continue to do so in fiscal 2024. But we also continue to develop new technologies. We've announced our BiCS8 technology that would be ramping in the future. So, there's a lot of actions that were taken to strengthen really both businesses.
Yeah. Yeah, that's good. So maybe I'm going to stick with hard disk drives, and we're going to go to flash. But, you know, maybe just, you know, double-clicking a bit on the hard disk drive business. So, you know, I was looking back at the last four quarters. I think you've recorded, I think it's somewhere around $300 million of underutilization costs. Last quarter was $83 million. You've talked, I think this quarter's guide is somewhere around $40 million. You know, I guess what I'm trying to get at is how do I think about that number as we start to maybe think about the semblance of a recovery? Do we look out the next couple of quarters and that underutilization is kind of lifted out of the model? Is that a fair assessment, or how are you thinking about that?
So the underutilization charges reflect basically the period costs associated with the underutilized asset. The way we manage the business is we really focus on inventory management as well as margin, gross margin, management, in addition to obviously talked about the fixed cost structure that we've also taken out of the system. And so as utilization improves, underutilization charges decline. The way to think of it over the next couple of quarters, as we continue to see improved, gradually improved, demand, we would expect the underutilization charges to gradually decline.
Yep. And, and, and I guess what I'm ultimately getting to, I'll just stop beating around the bush, is what is a normalized gross margin for the hard disk drive business? Like, and, and the way that I'm thinking about it, I think a lot of investors ask me about it, is that you've got a hard disk drive business that, you know, is, is down considerably relative to what it was peak, call it early 2022 kind of time frame-ish. And you've at the same time taken, you just mentioned 15% of cost structure out of, you know, the fixed cost organization. So do I think about, one, what is normalized gross margin? And, and two, should I not think that you get to that normalized, quote-unquote, gross margin at 15%-20% lower revenue or bit shipments? How do you think about that normalization?
So, with the reduced or the lower fixed cost structure, as well as the bigger focus on cost and efficiencies, I would expect us to reach quote-unquote normalized levels of gross margin at the lower revenue relative to, let's say, the previous-
Yep
... peaks. Now, when we get to the target level, our target for the hard disk drive business, gross margin is still at 31%-34%. And if you see the performance of the business over the last few quarters, and you adjust for the underutilization charges, you can see that the actions we've taken on the cost side have already started to show results. Obviously, they won't be fully visible until all the underutilization charges are sort of disappear. But nonetheless, we're starting to see good traction on the gross margin side.
You can assume that I've done that math, right? And it... You're pushing 30% almost now, last quarter, when you adjust that gross margin.
That's-
So there's nothing that structurally keeps you from being 31, 34, lifting out the rest of, you know, just, just sheer volume, basically, absorption.
That's correct, Aaron. And we've taken a lot, we continue to take also actions as we work with our customers on the build-to-order type of model that would allow us to have a bit more visibility in advance to be able to manage our supply chain and our manufacturing capacity in a much more efficient way.
Kind of going back a little bit, I want to put this out there. Just curious, is it like last quarter, I mean, are you starting to see some demand stabilization in hard disk drives? Is that kind of how we thought about the guidance into this current quarter?
I think that's fair, and, as we said on the earnings call, we've started to see some signs of improvements, and we would expect a continuous improvement for the rest of the fiscal year.
Okay. I'm going to keep working on this P&L, and we'll go to flash, trust me. But fixed cost structure, reduction, everything else, I... How do I think about the OpEx structure of a hard disk drive business? Like, do I look at your closest peer and say: Hey, you should be similar to that, you know, on a straight-up HDD business? Because ultimately, what people are in the process now of thinking about is how does it look separating the two businesses out? And I think everybody's trying to understand that P&L. So how do you think about the normalized OpEx of the HDD business?
Yeah, I understand that we, we all want to sort of get to a point where it's very clear what, what the HDD business would look like versus what the, flash business looks like. But it's... Look, it's, it's early, early days, and we're, we still have a lot of work ahead of us, to, to get to that point with respect to OpEx. And not only OpEx, of course, capital structures-
Mm-hmm
... and allocation, et cetera. And as we get closer to the date of the separation, we would be hosting Investor Day to provide a lot of details for everybody to be able to model properly the businesses and get a clear picture of what we're dealing with. Back to the OpEx question, though. The way to think of it is, at the company level, we've taken down our OpEx since the end of fiscal 2022-
Yep
... by around, roughly speaking, 25%-26%.... you know, we ended Q1 at around $555 million for the total company, which was probably a little bit on the, on the lower side. I think, as we normalize for the, for, over the next few quarters, I would expect OpEx to sort of be in the sort of $582 million-$600 million range.
Okay.
Depending, of course, on what the. As we continue to see improvement in the top line and profitability.
Yep.
Of course, we won't be increasing our operating expenses faster than revenue, but, as we see improvements, that would be probably a good number to use.
And then on the HDD business, the capital intensity is pretty nominal, right? I mean, we wouldn't expect to see any, you know, low single digit kind of CapEx intensity. I mean, is there anything that we should be thinking about as far as investments on the CapEx side there?
I think—well, our target is 4%-6%. I think we will-
In total.
It's, it's in total for the hard drive, for the hard drive.
Yep
... business. So, I'll, you know, I, I would expect the, at least in the short to medium term, to be on the lower end of that range, because we've, we have, the capacity we need for the next-
One final question on the hard disk drive side, I promise. Your closest competitor, you know, that should maybe take a step back. I mean, this hard disk drive industry is a duopoly, right? I mean, oligopoly, you could argue, you know, Toshiba is still a competitor, but it, it's very competitively narrow. One of your closest competitors did extend the depreciation and life of their equipment in their hard disk drive, obviously, the hard disk drive business. So, have you? Do you think about that? Is there anything to consider there on the gross margin as well?
So, we've always been conservative on our accounting, and if you refer to our most recent 10-K, I think our depreciation lives for the vast majority of the machinery, software, furniture, et cetera, is around from the 2-7 years, 2-7 range.
Okay. Shifting over to the flash business. You know, maybe I'm gonna start with just kind of the question of, you know, remind me again. I think in some of your slide decks and stuff, you've mentioned how much actual cumulative footprint you've got in your portion of the JV. I think it was nine- it was $18 billion and $19 billion, it might be $20 billion. Remind us again of how big of a capital investment you have in that, in the JV operations.
I think since inception, if my memory serves me right, the investment that's been put in place is more than $20 billion, closer to probably $21 billion. It's a massive investment. It's highly capital intensive. But also the JV offers, the joint venture, offers quite a bit of great scale in terms of manufacturing and technology development.
Just maybe for the audience, remind us again how the JV actually structurally works, right? You make joint decisions, you get your portion of the wafers out, you make those decisions across the portfolio, et cetera.
Yeah. So, the way the joint venture... By the way, this joint venture has been in place for 23+ years, so it's not recent. This is probably one of the longest joint ventures in technology. And the way it works is that, the technology development happens at the joint venture for both partners, and the manufacturing is... And both partners benefit from the scale of the manufacturing, with basically each partner gets 50, approximately 50% of the output.
Anything that happens within the joint venture, from the development of the technology to the manufacturing, until the wafers come out of the fab, are pretty much shared 50/50, with each partner responsible for the 50% of their capacity, financially speaking. Once the wafers leave the fab, each of the partners have different. Like, for instance, I can speak for our company. We have our own product portfolio, we have our own backend assembly and test capacity, where the wafers go, they get turned into finished products, and then each of the partners have their own go-to-market strategy, marketing, distribution, et cetera, totally separate from each other.
Yep. And can you talk a little bit about what you've done, you know, similar to the hard disk drive business, but the, you know, capacity footprint, how you've realigned that in the midst of the cyclical downturn that we've been dealing with? And, you know, maybe also, you know, help us appreciate, you know, what you thought about in terms of the flash business showing signs of recovery in the current kind of quarter progression.
Sure. So, yeah, one of the things we looked at, and the way we sort of manage our business, is as we were in the depths of the cyclical downturn, we were looking at where our inventory is, where the demand of our products is relative to where the supply coming out of the fab was. And, in January of 2023, we've taken the decision to reduce our wafer starts, at the time, by 30% to basically manage that inventory and cash flow situation. And, since then, we've been managing in a very dynamic way.... in a proactive way.
We look at where the supply of our product is and where we think inventory is headed, and we basically manage the supply of our products accordingly. This has been quite successful for us. When you look at where we ended in the September quarter, we pretty much our inventory situation for the flash business was our days of inventory were at the four-year low.
Yep.
So we've been managing very, very closely, obviously through proactive supply management, but also, we had a record bit shipment quarter in Q3, in calendar Q3, which helped with getting to that low days of inventory. And so with that, you know, we're seeing an improved environment from a demand perspective, as well as the price erosion has slowed down. And so there's, as we said also on our earnings call, we're seeing improved pricing environment over the next few quarters in the flash business.
Yep, and how do I think about you take out 30% of your bit output, right? Your wafer starts. How do I think about that coming back on, you know, the progression or what, you know, just kind of walk me through kind of the dynamics that would start to bring that capacity back on for you.
Yeah. So, you know, as I said earlier, Aaron, we do look at this in a very dynamic way. We make these decisions week to week, month to month, quarter to quarter. So as we always look at where the demand of our product is versus where the supply is, and then we basically adjust the factory or the fab utilization accordingly. So, there's also, within that, also you need to factor in over time the node transitions as we-
Right
... start moving from BiCS5 to BiCS6. Right. So these are all things that we consider as we manage the business to make sure we manage the inventory and the supply of our products in a real-time basis.
So, to get to the P&L impact of that dynamic, I think, you know, this most recent quarter you talked about, I think, a $70-$80 million a quarter kind of underutilization. The progression of that as we start to see bit shipments improve, you've got supply side, you know, still probably, you know, realizing the effect of the output, you know, the wafer reductions that not just you, the industry have implemented. You know, how quickly could I start to see that $70, $80 million a quarter headwind reported in the P&L related to underutilization start to come out?
If, if we continue to see improvement, improvement in demand, you know, I would expect, our underutilization charges over time to, to decline. It's a bit too early for me to, to talk, much more in detail about numbers, given that, as I said, this is something that we manage, in a dynamic way. But, you know, if, and when we see the continued, improvement in demand, and profitability in our business, we should, we would see the, underutilization charges, decline.
The, the, you know, if I look at the next couple of quarters, I still see obviously underutilization charges, maybe slightly lower than where we are now, but I wouldn't expect it to be, at least based on what I see today, no.
Right. You know, you mentioned it briefly there, but maybe walk us through where we're at as far as the progression of process nodes or the layer count, you know, BiCS5, BiCS6, BiCS8, which is, you know, I'll touch on here in a little bit, architecturally different. Just maybe level set us on where we're at as far as that progression through Western Digital's flash business.
Of course. So, the latest node we announced was the BiCS8 node, which is the follow-on to BiCS6. Today, most of our volume manufacturing is on BiCS5. BiCS6 is our 218-layer node. Look, what we've shown, I think, over the years is that we have a very capital-efficient technology. The way the technology is developed is based on looking at capital efficiency as well as cost reduction, having both of these in mind. Our engineering teams look at both of these and have both of these in mind as they develop the technology.
We've seen really from a capital efficiency perspective, probably we're near the top, if not at the top. When you look at, for instance, the BiCS8 technology, which is the 200+ layers technology, you can see not only the capital efficiency continuing, but you can also see the cost as well as the density improvement as we move from one technology to the next. There's always the talk about the number of layers and
Yep
... the perception that the higher number of layers means a better technology. However, when you look at over the last few quarters and the performance of our, for instance, our gross margin... we've shown that we are very competitive. We've pretty much outperformed many of our peers, and that's related to sort of the number of layers and the manufacturing costs associated with those.
So, BiCS8, you know, is architecturally, it's a hybrid bonded architecture, right? You're like, I think you and obviously your JV partner are one of the, you know, the first out there not to actually do the periphery and the CMOS together. It's actually a separate process. So the cost effectiveness of that, you know, do we stay on the same cost curve as you cut over to BiCS8? And what's the am I not to think that there's a lot more capital intensity involved in that manufacturing processes changing?
BiCS8 is based on what we call the chip bonded on array, which,
Right.
Which basically is made up of really two wafers that are bonded on each other. One wafer that has the CMOS, the circuitry, and one wafer that has the memory array.
Yep.
They're bonded on each other. So yes, there is more capital intensity involved in the manufacturing process. However, the bit outputs more than makes up for that. So there's also higher bit count, meaning you get more bits out of every wafer as you manufacture these. That's sort of one. Two, there's also the potential of reusing the CMOS wafer. And so that in some ways allows for better capital efficiency, because you really have to deploy that toolset once, and then you can reuse it for sort of newer or more advanced memory arrays in the future. And so BiCS8 is, as I said, the 200+ there. It is cost competitive.
We do expect to remain on that sort of cost curve, which is we've always targeted our cost reductions in the mid-teens %, over time, between as we move from one technology node to the other, and I don't see today as BiCS8 being different.
When should we start to think about BiCS8 from a materiality perspective on volume? I mean, I don't know if you're going to say when you think crossover might happen, but, you know, how do you think about the ramp of BiCS8?
Yeah, I think it's a bit too early to speak about crossover, but we are, we would be and we are in the process of converting our product portfolio overall to BiCS8. And it all depends really on the demand and profitability of the business. As we see demand and profitability improve, it could be, we could start seeing BiCS8 ramp as early as the second half of next year.
Second half next year. Second half calendar year?
Yeah. It all depends, but again, it all depends on the demand and the profitability of the business.
So I got a few other questions I want to try and, like, really get through here real quick. So, yeah, maybe I'm going to bounce around. I apologize, Wissam, for this, but-
It's fine.
You know, the, you know, one of the things that we've been, we've been saying, we've been reporting, is that we're clearly seeing NAND flash pricing turning higher. Like, spot pricing on average looks like it's up 40% so far this quarter. I mean, how would you characterize the, the, you know, have you seen signs that, that, "Hey, NAND flash pricing has definitely improved faster?
So, as we said on our earnings call, we are seeing an improved pricing environment, and we anticipate that we would continue to see that for the rest of our fiscal year.
Okay. That was a good try by me to ask about specific price. So I'm gonna have another try. Capital structure, I think I know the answer to this, but I'm gonna ask anyway. So everybody's trying to figure out how we kind of think about the debt for the combined company. You know, I guess the way that I've thought about it a little bit is if I take your closest peer, and I could assume that normalized, they would run typically 2-3 times debt to EBITDA. I think about the hard disk drive business. Is that a good proxy in your mind? Am I crazy in thinking that way?
Is there any kind of framing you can help me kind of, like, level set, like, here's how I should think about the capital structure at all?
So, look, there's still a lot of work for us to do before we talk about specific numbers, as we go through the work we need to do to get ready for the separation. I can maybe talk a little bit about the guiding principles, which basically our aim is to create two independent, publicly traded companies that are competitive in their own markets, that are sustainable, that can basically sustain long-term growth and long-term success. And so each of the businesses will be, you know, we would look at the profile of each of the business from a profitability and growth perspective.
Based on that, obviously, those capital structures would be optimized.
Okay. As you look to turn positive free cash flow, how do you think about using positive free cash flow? Is it all about deleverage still? Is there any other thoughts in mind, I mean, as we move toward separation?
I mean, we do have, we want to continue to optimize our own capital structure as it stands today. We did, we did issue a convertible notes shortly after earnings. That would help us address the. Well, that was done really to address the notes that that mature in February 2024. We also have the delayed draw term loan that we drew down in August to cover part of the IRS the settlement payment that's due at the end of June. So we would want to basically cover, cover these thing, these liabilities. And then, for the rest, we would continue to optimize our, our cash structure and just manage it.
So in the four or so minutes we've got left, one of the common questions I get, and again, bouncing around here, but hard disk drive-wise, I get a lot of questions around HAMR versus energy assist. And you guys, this last quarter, talked about having a 40 TB energy assist-based, you know, platform in the market. I'm not sure exactly if you gave a timeframe of that, but how do I think about that? Because to me, that's an important kind of delineation between the competitive landscape that I've never seen historically in hard disk drives, different technologies being deployed. You've also said that you're not against using HAMR either. So can you— Why am I not concerned that, you know, your areal density curve on hard disk drives could fall behind that of your competitors?
So, when you look at what we've done historically, and we continue to do in terms of executing to our technology and product roadmap in the hard drive business, we've continued—we've continuously improved the capacity points, so from one generation to the next, from one generation to the next. When you look at, for instance, the most recent. Today, we still have the technology leadership with our 26 TB SMR product, which, by the way, is almost 30% higher than where we were last year this time with the 20 TB. We've also announced a couple of weeks ago the 28 TB SMR.
The point is, we will continue to improve from one generation to the next, based on the energy assist PMR, as well as technologies that have been proven over time that are very reliable, and we continue to innovate on them. We see basically a path to get into the high 30 TB, 40 TB, following our current roadmap that we've executed successfully. We also have a HAMR development, and it will be added to our product portfolio in due course.
Okay. And then the final, real quick question I'm going to ask you is that, similar on the product side, you know, enterprise SSDs has been one area where, you know, I feel like I've reported on time and time again with Western Digital, is that you're just not you don't show up as a big competitor in that market, but it looks like a lot of opportunity. You know, I think at one point, you know, a handful of quarters ago, you talked about having, I think it was two or three different cloud hyperscale customers designed in for the NVMe drives. Where do we stand on you participating in enterprise SSDs, in that competitive landscape in flash?
Yeah, I mean, we do have these products qualified. The unfortunate situation is that the enterprise, the cloud side of the SSD market has been really down quite a bit and hasn't yet recovered. We've seen on the flash side recovery or stabilization and recovery to some extent in the consumer and client space, but we haven't yet seen it on the cloud side. As we start, you know, when, as and when we see that recovery, we should start seeing some more shipments of our enterprise SSD product.
Okay. I'll end it on this open-ended question, which is, you know, you think about the questions I've asked, you think about your discussions you've had with investors. You know, is there anything you'd say like, "Hey, you know, people aren't digging enough here, or not, you know, not paying enough attention to this part of the Western Digital story," that maybe you'd leave us to kind of think about as we walk out?
Look, I think, look, I'm very excited about the completion of our strategic review and the announcement of the separation of the flash business into a separate company. Over the last few quarters, as much as we were going through a cyclical downturn, we've taken the opportunity to really enhance the health of both of the businesses, and it's starting to show in our results. You can see, on the hard drive side, with all the fixed cost actions we've taken, the lower CapEx, we're seeing good traction on gross margins. On the flash side, it just shows, I think the cyclical downturn really showed the strength of the product portfolio from a consumer and client perspective that helped sustain also higher gross margins than many of our peers.
We've exited the September quarter with a very healthy inventory position. We continue to manage the business in a very conservative way from a cash flow perspective. And so, both businesses are well poised to be able to stand up on their own.
Then the timing of this, just, and we'll end here, is the timing, you know, the progression is analysts say at some point, you know, unpack this thing, you know, some point in 2024, mid-2024. Is that-
What we said at the time is that we plan to execute on this before the end of, by the end of calendar 2024.
Right.
I expect basically out of this flash Investor Day, ahead of that.
Okay, perfect. Thank you so much. I appreciate you joining us.
Thank you so much.