Western Digital Corporation (WDC) Earnings Call Transcript & Summary

Okay. Great. My name is Mark Delaney, and I cover Western Digital for Goldman Sachs. I'm very pleased today to have with us Mike Cordano, the President and COO of Western Digital; and Bob Eulau, the CFO. As quick background, Western Digital designs and manufactures memory and hard drive devices and solutions, roughly 50% of revenue coming from flash and from hard disk drives, annual revenue of roughly $17 billion.

Maybe I thought, to start, we could cover something that's very topical, the unfortunate health situation that's underway in China. And for either of you, maybe just discuss how does WD play in this business operationally in China given the situation is going on? Is there any impact to business conditions that you can share with us?

Yes. So first and foremost, we've been kind of focused on employee health and wellness. So we've instituted a number of things, both in our factories, which we have 2 in China, which continue to operate, but are really about making sure we keep our employees in the best possible conditions. We've also done some things to limit travel, as you might expect, as others have done. But from the standpoint of operational activity, from a supply standpoint, we have 2 factories in China. Those we're operating through the Chinese New Year. They continue to operate. Although we see some small perturbations that you might expect but nothing that meaningfully impacts our operations in the current period. Flipping to the demand side, sort of similarly situated. It's a little too early to tell. And our resolution on that is too early. But as we see things coming in now, nothing that meaningfully impacts our view of the current quarter reporting. But that's continuing to evolve, and we have to continue to see how things go.

Yes. Just I guess, one thing before we get too far along, we'll obviously make some forward-looking statements, and I would encourage you to review our SEC filings, 10-Q, 10-K. And when we refer to numbers, most likely, we'll be referring to non-GAAP numbers. And again, I encourage you to look at our reconciliations to the GAAP numbers. With that, we'll keep rolling.

No. Thank you for the disclaimer. On the NAND business, nice that we're at the early start of an upturn. Maybe you can talk about why WD thinks we're in a NAND upturn now and how sustainable the company thinks it may be.

Yes. I think the first indicator is the inventory levels, both at the manufacturer level and then down channel. So we really -- and we said this on the call, we think inventory is in a good position across the industry. Then it's kind of a simple calculation from there as we see it, we see a bit of supply growth on the year, our view is in the very low 30s. We see the demand side being in the mid-30s and maybe a little higher than that. So the juxtaposition of the 2 kind of gives you a view of where we think the year will be. And certainly, we see the markets, meaning our customers, convicted that that's the case as well. And obviously, we talked about where we saw pricing trends on the call, and things have been kind of moving up sequentially in most instances over the last quarter or 2.

One of the factors that's contributing to the upturn, in particular for WD, is some improved momentum in its all-state drives, and the company spoke to a hyperscale win that it realized in the last quarter. Can you talk about what led to that win? And what that may mean for your market share in enterprise SSDs?

Yes. We've been talking about enterprise SSD as a strategic area -- growth area for some time. What we showed in the back half of the year and the calendar quarter just reported was substantial sequential growth at a particular hyperscaler. The optimism about this year, we talked about the call, we expect to double our revenue year-on-year. That will be driven largely by our NVMe product portfolio, both at hyperscalers and at OEM accounts. So both our breadth of customers, so the numbers in which we'll be engaging with, and the numbers of products that we'll bring into the market this calendar year will continue to grow. And it's all based on a single NVMe silicon and firmware architecture that we're now beginning to get leverage into the market on.

One of the things that happened during the last NAND upturn is that the supply-demand situation became so tight that we actually saw some switching of customers from SSDs back into hard drives. Do you think pricing gets that tight again this upturn? And how do you think about managing your business as you think about some of those dynamics?

Well, I think it's too early to tell on where the top of the cycle will be. So I wouldn't want to predict that. But I think it's important to note that I think customers are very smart. They understand what the margin rate for somebody like ourselves needs to be in flash. And when they plan their product portfolio, they don't plan it for pricing on flash at the trough pricing. So they're anticipating more normalized pricing. Where it goes at the high end and where availability go is sort of hard to tell. But I think customers are very smart about how they think about their deployments, whether it be PC, mobile customers or enterprise customers.

Okay. Bob, I wanted to go to you on the margins in the NAND business. The company guided to the NAND gross margin being 35% to 40% in the second half of calendar '20. Maybe you can help contextualize what that number means. Is that more of a mid-cycle number or more of a peak type of a number?

Well, I mean as Mike said, we don't really know where this cycle is going. We tried to give some indication for the calendar year because we do think that there's going to be a continued demand strengthening in the second half of the year. And we had actually said that a quarter ago. Now we're seeing that actually get pulled in a little bit forward. So I think 2020 is shaping up to be a good year.

I wanted to touch on some of the chip mix topics, and QLC or X4 is something that WD has put some R&D into. Maybe talk about how meaningful QLC is as a percentage of your business today, and how do you see that evolving over time?

Yes. I think in this calendar year, we're ramping up, but more in the consumer and client SSD segments. And certainly, as you sort of project out into like 2025, it's going to be much more meaningful and will kind of cross all of our segments. So it's something that we think is strategic, and it's about the right technology at the right time into the end market. So it's something we're quite committed to.

One of the storage OEM announced a QLC-based all-flash array. Now you said, for WD in this year, QLC is more of a consumer-grade product. But 2, 3 years down the road, how important do you think QLC could mean? And does QLC potentially disrupt parts of the cold storage market that are currently using your own hard drives?

Yes. Actually, no. When you think about cost declines over, let's call it, the next decade, including the deployment of QLC, we're talking about 15% per year cost down on flash. Similar cost down per year on disk. And that's roughly, in order of magnitude, separating. So those 2 -- in the data center, those 2 tiers will coexist sort of as far as we can see it. Certainly, we think there's advantages to QLC but not as a displacement technology for the capacity enterprise segment. What that may or may not do for 10K and 15K hard drives, which we no longer produce, different story.

And sticking with the NAND product road map, WD and Kioxia recently announced BiCS5. Maybe just talk about the ramp time line for that and what do they mean for your business.

Yes. We're going to be ramping BiCS5 throughout this calendar year. We talked about that. We're at 112 layers versus others. Really, what we were trying to do is the right cost performance tradeoff. We think we made the optimal one. We got some XY shrinks. So when you look at wafer to wafer, we picked up about 40%. So it was -- of course, the layer count moving from 96 to 112, but we saw a benefit in the linear direction, in the XY-dimension, so that gives us that 40%. So it's a lower CapEx node for us. So we see a nice sort of cost optimization -- cost performance optimization with this node.

Maybe to tie in some of those product announcements together with how investors should think about cost per bit decline, I mean you have QLC, you have BiCS5. I mean Bob, help us think about what annual cost per bit declines could be in the NAND segment.

Yes. So -- and we've been saying for a while, on a long-term basis, we're expecting about a 15% cost decline per year. That's what we've experienced the last few quarters, and it won't be linear, but that's generally what we're expecting over a long period.

In terms of the competitive landscape, the company said on its last earnings call it doesn't expect YMTC, which is the main Chinese company that's focused on the end market. WD said they don't expect YMTC to be a big factor in the market for the next 2 to 3 years. Maybe help us understand how WD came to that assessment.

Yes. I think it's 2 things. One is its demonstrated performance at the node. They're trying to ramp 64 today. They've announced an intention to go to 128, that's unproven in terms of -- it's a difficult thing to do, and it's unproven. And then it's just the amount of total capacity they have available. So the combination of those 2 things, to us, says that, at least in the next 2 to 3 years, it's not going to be a meaningful impact.

Do you think YMTC could be disruptive, even if it was just at the 64-layer technology node?

Well, first of all, they don't have a lot of capacity to be that disruptive. But for us, in addition to just looking at all bits in the market as the same, a big part of why enterprise SSD and other investments we're making -- you'll hear us talk about quality of revenue. And what do we mean by that? There are certain end markets that have a higher margin, better growth rate and, frankly, lower-margin volatility. So if you look at peak-to-trough gross margin, those are segments we're aggressively pursuing: enterprise SSD, performance part of mobile with UFS and, frankly, the performance part of client SSD, all being examples motivate us forth. And to the extent we are successful, which we've made those investments starting in 2017, we're really in a great position coming into this year. So as we go through this calendar year, the diversification of our revenue streams into those higher-quality revenue segments and the diversification of customers really puts us in a better position to insulate us from any particular low-end involvement.

Maybe we can switch gears to the hard drive business, and the company gave a positive outlook for its data center-driven hard drive business, so-called nearline hard drives. Maybe just talk about what's driving the pickup this year in the nearline space and how sustainable do you think it may be.

Yes. I think we talked about we see strength in the hyperscale space right through the first half of this calendar year. We're seeing some indications that lead us to believe that might continue right through the full calendar year. Obviously, we want a little more time to get a more clear resolution in the back half of the year but the continued data growth and deployments, as we see them, and that group of customers continues to mature operationally. So our ability to have kind of a more clear view of what they plan on doing has improved year-on-year as well.

It kind of dovetails nicely into my next question as -- the customer base there is potentially maturing. But the nearline hard drive business has historically been very lumpy. And the pickups there are 4, 5 quarters long. And then there's a contraction period of 1 to 4 quarters, depending on the cycle. What's WD's expectation for this business over not just maybe this year but longer term? And do you think it stays a very volatile segment?

Yes. I think it will be somewhat lumpy by nature. But I do think, both in terms of the size and scale of that market, the diversity of customers in that market gives it some sort of normalizing factors relative to volatility. So to the extent our operational model continues to change -- we've talked about on calls in the past, we have some sort of, we'll call it, logistics arrangements with certain customers that allow us to sort of understand what their intentions are because they want to guarantee certain flexibility of supply. So that's a service that we actually get paid for. So our ability to sort of project out, have better visibility, not only in terms of their forecasting but the direct engagement and the nature of the relationship, is maturing as well. So that gives us a little better sense for it. But I think that, combined with the size and scale and diversity of that market, is going to allow it to be a little less lumpy as we go into the future.

Okay. Pricing in the nearline space have been a little bit more challenging than typical for parts of 2019. But last quarter, WD saw an improvement in the pricing situation in the hard drive space. What led to that improvement in pricing?

Yes. I think generally, for us, we have not seen -- we wouldn't depict the pricing environment as negative or aggressive. We think it's been fairly stable as we've seen it. You saw our results in terms of our reported ASPs as well as our margins. So for us, it was more about getting our costs where we wanted it through the KL consolidation and other things. When we're sort of in that leadership position, it allows us to sort of manage our outcome and manage the portfolio in a consistent way. So pricing, in general, we don't see as particularly aggressive.

Bob, maybe I can ask you a margin question on hard drives and specifically related to nearline hard drives. And understand that the company doesn't give exact margins by product area, but our view has historically been that the nearline space in the past has carried high 30s to even 40%-plus gross margins for WD. As this becomes the bigger percentage of the hard drive business over time and all of the hardware companies focus more on that business, do you think margins on a like-for-like basis in the nearline space stay at the types of levels that they've been at historically?

Yes. We think as we continue to invest, we continue to deliver quality products and lead in 18-terabyte, 20-terabyte, we definitely think we'll be able to sustain those margins. I mean it's a business where we have to reinvest as well. When the test times are longer, there's more capital that we have to deploy in that particular business. But we're very excited about how we're doing with capacity enterprise. And as you said, as the mix skews more and more towards capacity enterprise, that will drive our overall hard drive margins up as well.

The company recently launched some new products around 16- and 18-terabyte conventional recording that use partial energy-assist. Maybe talk a little bit about the customer adoption and road map for those products this year.

Yes. So we talked about it. We're going to be shipping for revenue in the current quarter, although small, we'll see the ramp being in -- more earnest in calendar Q2 and then right through the year. So we've got a number of qualification schedule, each of the customers kind of go in their own time, but we would expect that product to be qualified across the totality of the hyperscalers as we progress through the year. So we feel very good about it. The product is well positioned. We talked about a single platform, being able to get 18, 16 and 20 SMR. That makes the qualification process for customers even easier. So we're trying to again thinking about total cost of ownership and making things more simple. This is a broader platform. We get it with a single qualification, multiple capacity points.

WD was gaining share in the nearline space on an exabyte basis. For most of calendar 2019, with the strength of your 14-terabyte product, and the company's talked about on its earnings call and on different investor presentations, a little bit of normalization last quarter in that space. Help us think about how you expect your exabyte share to trend maybe in the second half of calendar '20 when you have more time with this new product.

Yes. We talked about it on the call. So obviously, we peaked at 50% -- 57% exabyte share in our fiscal Q1 or calendar Q3. That was sort of above where we would expect to operate over time. We're sort of in this notion of right through calendar year '20, sort of 55% plus or minus a little bit. So something -- I think I talked about in the call, 53% to 55% is where we'd expect to be in the first half, and then we'll see where it goes for the second half.

Can you maybe talk about the longer-term product road map in hard drives and different types of energy-assist? And 18 terabyte is a partial energy-assist technology that WD has introduced. How far do you think partial energy-assist can scale because you got 20, 22, 24 terabytes before you would need a full MAMR or HAMR.

Yes. Although I wouldn't get into the exact generations. But certainly, this was our first, and we've talked about 2 or 3 generations that we think will have an advantage from an areal density and head disk interface standpoint. Beyond that, will it be, let's call it, MAMR derivatives or will it be at HAMR. We're working on both. And to the extent it's either one we're comfortable, we're prepared for it. Really, it's about for us, right technology, right time. We see lots of advantages with our versions of energy-assist, in at least in these next few generations, and then we'll see where it goes from there.

When WD is doing R&D today and you have an R&D budget you have to allocate, do you have to say, okay, this dollar is going to HAMR and this one is going to MAMR? Or is there shared R&D for energy-assist technology more broadly?