The Transcripts, Transcripts
Western Digital Corporation (WDC) 51st Annual J.P. Morgan Global Technology, Media and Communications Conference Transcript
Western Digital Corporation (NASDAQ:WDC) 51st Annual J.P. Morgan Global Technology, Media and Communications Conference Call May 22, 2023 9:20 AM ET
Wissam Jabre – Chief Financial Officer
David Goeckeler – Chief Executive Officer
Conference Call Participants
Harlan Sur – J.P. Morgan
Good morning and thank you for attending J.P. Morgan’s 51st Annual Technology Media and Communications Conference. My name is Harlan Sur. I’m the Semiconductor Capital Equipment Analyst for the firm. Very pleased to have Dave Goeckeler, Chief Executive Officer; and Wissam Jabre, Chief Financial Officer of Western Digital here with us today. Let me turn it over to Wissam, who will read the Safe Harbor statement, then we’ll go ahead and kick off the Q&A.
Gentlemen, thank you for joining us today. And Wissam, let me turn it over to you.
Thanks, Harlan and good morning, everyone. Happy to be here. We’ll be making forward-looking statements and I ask you to refer to our SEC filings for the risks associated with these statements. 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. Thanks for that, Wissam. So before I jump over to discuss the business with Dave and you, I wanted to ask a couple of high-level questions about some of the recent headlines that investors have been interested in. So the first question is, while I understand you can’t comment on a few recent headlines and the strategic review that is ongoing. How about just reminding investors about the benefits of the JV with Kioxia as well as how the 2 companies differ from a go-to-market strategy and also from a supply chain perspective.
Sure, Harlan. So first of all, it’s great to be here. Thank you for the invitation. So the JV is an — it’s a wonderful relationship. Let me just start with that coming into the company now for 3-plus years. It’s been just fantastic to really get to know this JV. This JV has been in existence for 23, 24 years now. Very, very successful. Anything that last 23, 24 years is because it works very well and it provides a lot of benefit. I think the easiest way to think about the JV, if you think about the NAND business, you think about a wafer that goes through the process everything up until the wafer comes out of the fab is the purview of the JV.
So all of the — it starts with all of the R&D, our BiCS road map, we collaborate very, very closely. It’s like there’s one big engineering organization that’s defining our road map, our technology road map. Our technology has been very strong, very capital efficient for decades now. That’s a reflection of the fact that we’re both investing our combined might in the market together on one road map. So we get very large economies of scale out of that. Then of course, the manufacturing is done together at our fabs in Yokkaichi in Kitakami. And so all of that is the purview of the JV, the production of the R&D, just how all that works, all of the supply chain side of procuring everything is done as through the joint venture. But then, the wafer comes out of the fab and we each take our share and we go our own way.
Separate back ends, we’re peers in the market and we compete against each other and then selling that NAM. We make independent decisions of what markets we’re going to enter. Are going to be — we going to build client SSDs, enterprise SSDs which part of the market are we going to be in? Those are completely separately — separate and that’s not part of the JV. So the JV gives us this great scale position on our technology road map and the ability to drive that forward, gives us a great scale position on manufacturing, all that is very important. And then we take our wafers and we’re peers in the market at that point.
And my understanding is that you talked about the synergies from an R&D perspective, process technology development, manufacturing development. But my understanding is that there’s also some synergies even at the basic design level, right? There are certain fundamental building blocks that are common to every NAND chip that you produce. But even there, the teams collaborate on the foundational IP, right? That is also critical in determining ultimately the performance of the product as well.
Yes, absolutely. I mean I think there’s always different technology building blocks. There’s a research going on for many, many, many years before it then gets commercialized into a product. I mean, take what we just did with the BiCS6 product just announced, wafer bonding. So that’s a technology primitive that the teams have been working on. You decide when to bring it into the road map. We brought it into the road map and now commercialized it. So it’s a perfect example of where we get these economies of scale and we get this — if you would look at the R&D team working on NAND research, you would think it was 1 company. It works hand — hand in glove with each other and has worked that way for decades now.
Another point that we wanted to just bring forward in discussions with clients, post our earnings call, there are concerns around the potential violation of the terms of the credit agreement and debt covenants that cover your term loan A revolver and DDT. And you amended the covenants in December of last year, right? And that took into account a period of industry weakness. And similar to an amendment that your credit agreement, similar to the amendment that you made during the 2019 downturn as well. But given the severity of this down cycle by our estimates, the team could potentially exceed the 5x gross leverage ratio threshold in the September quarter.
Now, I would argue that the design win pipeline, future competitive position of WD is stronger than what it was when you renegotiated the amendment back in December. But in any case, what actions can the team take to renegotiate or further amend the terms of the agreement? And how do we think about the potential economic cost of such a move?
So Harlan, we — let me first start talking about our liquidity. We ended the quarter with $5.3 billion of liquidity between the cash on the balance sheet and the credit lines that are available to us. And as you mentioned, in December, we did renegotiate — we amended our credit agreement. And if we see the need for it to be amended again, we’ll be proactive about doing that going forward. The — we’re very comfortable with — actually, we feel good about where the portfolio is from a technology perspective as well as a lot of the cost that — actions that we’ve taken to take costs out of the system. That positions us really well for future — improving the future profitability of the business.
Perfect. Okay. Now moving on to some of the business-related discussions, sort of on the near to midterm demand environment, your earnings call was a couple of weeks ago, industry is in the midst of a downturn. Team has been disciplined under shipping end market consumption for 3, 4 consecutive quarters. Typically, after under shipping for this duration, you start to see some signs of excess inventory normalization and a move higher in shipments towards consumption. So maybe, Dave, help us understand how you’re seeing the current supply-demand environment in your flash and HDD businesses in the first half of this calendar year.
So I think — let me walk through it by business. So the consumer business has stabilized, I would say that’s probably a good word for it. It’s pretty consistent from quarter-to-quarter. When we go into a quarter, we kind of understand how it’s going to play out and it’s been that way now for certainly the last quarter or 2. So that market feels like it stabilized. It was kind of the first one into the downturn that we saw early last year.
The next one in was the PC OEM market. That started inventory correction mid last year. I think if you look at our last quarter, we saw actually good dynamics a little more — that market was a little better than we thought, especially in distribution than we thought going into the quarter. So it feels like that is through its inventory digestion at this point now shipping closer to end demand, to your point, for several quarters there, we were significantly under shipping end demand as customers had a very sharp inventory correction.
Other markets around that; gaming has been strong. That still feels — that’s still a great market. We have some great brands there, especially with WD Black. And then you get to the data center market which is still — it was kind of like the last into the inventory correction towards the end of last year. It’s still going on. It’s a bit idiosyncratic against with the big players, the big especially U.S. hyperscalers are so big that they can move markets all by themselves.
So there are different points in their inventory digestion but I still think we’re a couple of quarters away from getting through all of that.
So given all the dynamics, you did mention on the earnings call that you anticipate accelerating sequential Flash bit growth from this quarter into the second half of the year; so better second half shipments in Flash. From an HDD perspective, as you just mentioned, despite some of the improvements, I think, that you guys saw in China Cloud, U.S. cloud and hyperscale still uncertain. Some are spending, some are still working down inventories. Is that sort of the way to think about the second half demand profile for the team as better still a little bit of uncertainty in.
I think that’s right. I think that in Flash, I mean, in the consumer market is seasonally better half of the year, clearly. As we said, the PC OEMs feel like shipping more to true demand, data center, enterprise SSD is still going through a correction in the second half of the year. But when you add it all up, we see sequential growth going into the next quarter.
And capacity enterprise drives, again, it’s going to be — I think it gets better towards the end of the year. I think we still have a couple of quarters to go through here of demand normalizing in that market. And then China, on top of all of that, feels like there’s better activity there. It feels like the second half is going to be better given the activity we’re seeing from customers. So that’s a bit of a wildcard about how that plays out as well.
Any questions from the audience? If you do, just feel free to raise your hand. On the cloud HDD weakness and uncertainty, there seems to be a couple of concerns by investors, right? First concern is cloud HDD cannibalization by SSDs as flash pricing has dropped significantly, right, through this downturn. And then the second concern is the shift of more cloud CapEx dollars to support generative AI, compute infrastructure build-outs, right? And so love to get your thoughts on both of those investor concerns.
Yes. So the first one, the idea of enterprise SSD cannibalizing HDD. I mean if there was any time that was going to happen, it would happen right now. I mean there’s like underutilization of NAND supply in the world. NAND is in a very serious downturn. And I can tell you from our portfolio, we see more deterioration in enterprise SSD exabytes than we do capacity enterprise HDD exabytes. So I think that customers are very smart about this. They’re not going to design an entire hyperscale data center architecture based on what pricing is in a couple of quarters. They look out over the long term. The products are used in different use cases in the data center. That hasn’t really changed that much, right?
And again, if it was going to change, it’d be changing right now. And kind of the exact opposite is happening which is all the NAND providers are taking CapEx out of the system because the demand is not there, yet we’re still shipping hundreds and hundreds of exabytes every quarter of capacity enterprise HDDs. So I don’t see that long term. I think there’s pretty stable demand in the data center with how data centers are architected and the old saying maybe it’s not so old. The 2 things can be true. They are complementary technologies in the data center. Both markets are growing. I think the NAND exabytes are definitely growing a little bit faster than capacity enterprise but they’re both growing markets in the data center. So that’s a good place to be. Certainly, generative AI is going to be — generate more demand for enterprise SSDs. If you look at servers, going from 4 terabytes to 32 terabytes the way they’re configured. You’re going to see more demand for enterprise SSDs. That’s a part of the data center architecture where you didn’t have HDDs to my point here, right?
So I think from an HDD point of view, the AI algorithms have just shown that once again, the world has figured out new ways to monetize store data. So you have new ways to monetize store data to train engines degenerative AI engine that says marginally going forward, storing more data is important because I can monetize it in the future and that’s going to drive HDD demand. So I think it’s just another — it’s easier to track to say, hey, a server has got more terabytes of enterprise SSD on it, that’s clearly going to drive enterprise SSD demand.
Maybe one of the reasons why that TAM is growing faster; it’s also going to pull along the dimension for capacity enterprise hard drives in the data center as well because we’re all just generating more data, we’re storing more data. So again, they’re both good markets in the data center; they’re complementary.
And to your point and you guys can verify this because you guys have both technologies under 1 roof but we’re just looking at some of the industry data as it relates to this whole concern around SSD cannibalization of HDD because of the pricing declines. But by my calculations, even with the sharp pricing declines in flash, flash price per bit is still 7x higher, right? Than cloud-optimized HDD price per bit. Is that — is that sort of the range that you guys are seeing as well?
That’s the range we’re seeing as well. And in both businesses, there’s 1 thing we feel very good about our portfolio, we’re driving new technology in both businesses. And here we are in the NAND business, we’re right on the cusp of BiCS8 wafer bonding, much more productivity in that market on bits per wafer. At the same time, we’re driving new innovation in HDDs. EMR, OptiNAND, now we have SMR coming where you can get 20% more bits per drive than you could before, by making a software change. So I think in both markets, we’re still driving productivity and that’s a good thing. If you drive the cost of storage down and it increases the market. So that’s — both businesses are still doing that. And while they’re both doing that, they’ll both be good businesses in the data center.
Yes. I think the market tends to be more focused on the memory sort of flash side of the business and sort of understands the economics of cost per bit declines moving from 1 generation BiCS to the next-generation BiCS and 100 layers, 200 layers and so on. But my sense is that aerial density improvements in HDD are certainly keeping pace with that, right, as you could tell by the big — the continued big delta between the pricing in SSD versus pricing in HDD but we’ll get a little bit more into that. You’re anticipating a stronger BiC [ph] shipment profile in flash as we discussed in the second half of the year. Pricing is the other dynamic that will drive revenues and margins. From a supply side perspective, you guys are doing all the right things, right, cutting flash capacity utilizations by 30% back in January. Consolidation, lowering HDD capacity since second half of last year as well as cutting cash CapEx plans for fiscal ’23 by 50% versus your targets at the beginning of the year.
Your customer — your competitors are executing to similar supply cutbacks. But are competitors acting in a similar disciplined fashion from a pricing perspective?
Look, I mean, we’ll mainly keep our comments to our business but I think it’s no secret that the market is in a major cyclical downturn. And I think the industry is doing the things it takes to bring the market back in balance. All of us have to look at our own portfolios and make the best decision for our own business. We’re doing that certainly. We want to make sure we can match our supply and demand as best we can and not let inventory get out of control. And I think we’ve done a pretty good job of that.
We’ve got a relatively good inventory position and we plan to keep it that way by looking at where we have homes for the bits. And again, one of the things I think that the downturn has shown is the diversity of our portfolio is working. The fact that we have a consumer franchise. We have a strong position in client SSD. We have a strong position in mobile. We have a very strong position in gaming which is an emerging market and we’ve built out now, of course, the leg of the stool on enterprise SSD.
Now that’s a very depressed market right now, given the downturn but the product strategy is still there and our ability to basically mix across that entire portfolio to get the best return we can, that has worked. And I think that the numbers that are showing up in the downturn are showing our ability to generate the best margin possible given what the environment is.
Before we dive into the product portfolio performance leadership, any questions from the audience? So let’s start off with HDD. We just — the recent March quarter, market share rankings just came out. And boy, the WD team took some pretty significant share in capacity HDD, right? Your 22 terabyte CMR HDD, now the majority volume among your 20-terabyte plus offerings. Like I said, you gained some significant market share traction in the March quarter. And I think the expectation is that you’re going to continue to gain share for the remainder of this year.
Now you’re on target for full qualification of your 26 terabyte Ultra SMR drive this quarter. So what kind of reception have you been getting at your cloud customers with 26 terabyte UltraSMR. How much of an advantage is UltraSMR technology giving you? And what should we expect the volume ramp of that platform?
Okay. I’m going to — let’s talk about 2 things in HDD. I’m going to talk about the portfolio. We’ve also been doing a lot of work on the cost side of the business which is just as important. We saw him talk a little bit about that.
So first of all, what you’re seeing playing out is a strategy that was put in place 5, 6, 7 years ago which is we looked at the aerial density road map and realized, hey, we’re going to get to 20 terabytes and now we’re going to need some other things to get us up to 30. And to a point where HAMR will take over and be kind of the major technology. We believe that, that technology is going to take a little bit longer. So we put some things in the road map. So we started layering in these technologies. We layered in ePMR. That got us more aerial density per platter. We then layered in OptiNAND, same thing. Then we layered in UltraSMR. And these technologies you’re now seeing them play out in the market.
So the first driver that really shows up all together is in the 22 terabyte CMR drive. We talked about this past quarter, that became our — that above 20, that became our leading capacity point. So that’s something we’ve been planning for, for years. It’s actually playing out in the market. The portfolio is being well received from a share and margin perspective, that technology strategy is working.
Another layer on top of that, to your point, is UltraSMR. Now SMR is not a new technology. SMR has been around for actually a very long time. So it’s proven technology. We did 2 things. One is we — with OptiNAND, instead of SMRs traditionally get 10% more capacity per drive. With UltraSMR, you get 20% more capacity for drive. So that’s why the 22 CMR drive becomes a 26 UltraSMR drive. So that’s a unique position in the market. It was a very explicit strategy to like fill this gap and be able to continue to step wise this aerial density road map.
Again, the issue with SMR adoption has always been there has to be work on the host side. So the customer has to do some software work to adopt the technology. So what’s happened in the last year, 1.5 years is the big hyperscale players are now all doing the work to adopt SMR. So our SMR drive is in qualification at the largest data center operators in the world. Some of those qualifications will start to be done this quarter and we’ll start to ship in volume in the second half.
So the strategy that’s been put in place for years to kind of get us from 20 to 22 to 26 is now playing out and I think it’s being very well received by our customers. And it’s leading to good performance of the business and why we’re in a big cyclical downturn. It’s harder to see but we believe when we — when volumes return to normal, we are in a very, very strong position to offer our customers a very, very strong value proposition.
Now on the other side of that, we’ve been looking at our cost as a business. So maybe, Wissam, you can talk a little bit about that.
Yes. So on the cost side, for the hard drive business, we started taking action towards the end of fiscal ’22, the beginning of fiscal ’23. We restructured our manufacturing footprint, taking out around 40% of the capacity of the client manufacturing capacity.
In parallel, we continue to drive the fixed cost down across the board. And so we’re now — when you look at where we are today relative to where we ended fiscal ’22, our fixed cost in the hard drive business is around 10% to 15% lower. In fact, I don’t think it’s been that low for at least a decade. And when you look at what we delivered last quarter in terms of gross margin and you factor into that, the underutilization charges and some other charges were pretty much north of 28% on a reasonably low or relatively lower revenue.
And so as volumes come back, we expect that gross margin to trend back to higher levels at a faster pace.
Perfect. Dave, you mentioned HAMR, heat-assisted magnetic recording HDD technology. It’s been around for many, many years now under development by you and your competitor. Your competitor claims they’ll be ramping HAMR-based drives next year. WD team always has had a really good track record of being able to extend current generation drive technology and bringing new technologies to the market at a time when the market needs it and the economics to WD are optimized, right?
And so obviously, you just talked about your CMR architecture, your OptiNAND, your UltraSMR. These are all great technologies to continue to extend aerial density. But what’s the team’s strategy and adoption curve for HAMR-based HDDs?
Yes. I was in Japan about 3 weeks ago and I actually met with a gentleman that actually wrote the seminal paper on HAMR, right? And it turns out he published that paper in 2001. So — and that’s like it’s very — it’s great work. I mean but it just tells you the industry has been working on this for over 2 decades. And that’s not surprising when you’re talking — I mean HAMR is not a software product. I mean it’s material science, it’s physics. It’s like stuff that takes a long time to really make sure you can produce it.
And the issue is not just producing it and making it work. The issue is producing it at scale, being able to produce millions of drives through quarter, being able to produce a drive that has the reliability where somebody is going to put it in their data center and bet their business on it and it’s going to be there for 5 years. Think the industry feels good that this technology is going to be there and it’s going to get to that point. It’s still — we’re in the final stages of something that took a decade. So we’re still a year, year plus away from that.
The reality is the market right now is adopting SMR. That’s the next step of technology in the data center. And that’s where our portfolio is focused. We had a very explicit goal to like bring this technology to market and make it better exactly in this time frame because we didn’t believe that HAMR, not that our HAMR wasn’t going to be ready but just HAMR in general, the technology, the physics, the material science of it wasn’t quite going to be ready yet. And I think that’s turning out to be true. That turned out to be the correct bet.
Will it be there? Yes, it will be there but it’s still — we’re still a year away from 1.5 years or more away from that in volume where it becomes the major part of the technology. And I have every confidence in our engineering team. They’ve been working on this technology for several decades. When that technology is ready to be commercialized, we will be there with the right products.
Perfect. Any questions from the audience? Let’s pivot over to your flash technology. You and Kioxia and you mentioned this in some of your remarks, announced that you’re starting to productize your next-generation BiCS8 NAND flash technology. I assume it’s 200-plus layer 3D NAND architecture. There are some pretty innovative features and major architectural changes on BiCS8, right, versus your power generation technologies. Namely the move to periphery circuit bonded to memory array architecture, right? You’re actually bonding 2 wafers together. It seems like a very complex architecture but help us understand some of the advantages of the BiCS8 platform.
Yes. It is — we feel very good about it and it was a — it was an ambitious technology direction. And one that I think the industry knew was going to have to tackle at some point. I mean it was circuit next to the arrays to get under the array and both of those architectures have been — some of them are — that’s what’s in the market now but we had to figure out like how do we build the memory stack and the CMOS separately and then bond them together as opposed to building the CMOS and then stacking the memory on top of it.
That project is something that’s been in development for many, many years to the question you started out with. We’ve had R&D folks working on this for years and years and years and we got to the point where, okay, we feel like we can now commercialize this. We’ve got enough confidence in the technology and that’s what BiCS8 was all about and we feel really good about where it’s turned out. We were ahead of schedule on it. We’re producing it. It’s starting to go into commercialization in some early products. Clearly, we’re in a downturn. So when we need — when we ramp back up the fab, we’ll be ramping back up into BiCS8 and we feel like we’re in a very, very strong competitive position. It is a 200-plus layer product. Again, in true Kioxia-Western Digital fashion, we feel like we can — we know we can produce the same quality and the same density of product with fewer layers. That’s a good thing.
It’s a 218 layer product. But the key thing is we’ve got — 2 wafers. You build the CMOS separately from the memory bond them together. So what happens, you end up with much, much higher quality on each side of it and then when you bond them together. And I think people are going to see when we put all the specs out on it. It’s a very, very impressive product on the performance, on the density, on the productivity and on the capital efficiency.
Perfect. On the financials, your 3- to 5-year financial targets that you put up at your Investor Day last year, 7% to 9% revenue CAGR, 30%, 36% gross margin, 17%, 22% operating margins, cash CapEx intensity, 8% to 10%, right? Obviously, before we enter the current downturn question is, did the team actually factor in a downturn over the 3-year horizon back last May when you set your targets? And do you still see the targets as attainable by 2027?
So Harlan, naturally, we’re in a cyclical industry. So we do factor downturns into our modeling. And when you look at the 3- to 5-year targets we have, we’re still comfortable with those. The — we talked about during this cyclical downturn, taking action on the portfolio. On the hard drive side, we have a very strong technology portfolio that would help drive the business forward. On the flash side, we talked about the technology — the nodal technology transition and that would help us also to continue to drive the business forward in parallel, we’ve taken a lot of cost out of the system from an OpEx perspective as well as from a manufacturing cost perspective that would help position the business for a much more improved future profitability. And so there’s a lot of actions that the team have been taking during the severe cyclical downturn to position the business for longer-term success.
Is there anything that you’ve done from a development perspective, R&D perspective to try to accelerate the innovation, the product road map such that when you emerge from this downturn, not only do you have the most efficient cost structure but maybe even increase that technology advantage by half a step coming out of this downturn.
I mean, it’s what we just talked about. I mean the technology road map is a long thing, right? You don’t plan a technology road map just in the 2 or 3 quarters that you’re in a downturn. So it’s just continue to execute the plan we had put in place and make sure that we’re obviously adjusting the cost sides of the business to like — to deal with the current market realities. And when you do that, make sure you preserve the long-term competitiveness of the business. I mean in the NAND business, your foundational NAND, how good is the BiCS road map? I mean that’s the foundational technology. The foundation has to be strong and we feel like, again, we’re emerging — we’re going to emerge out of this downturn on to technology that is really differentiated, right? This idea that we have BiCS8. We’ve gotten past wafer bonding. We’ve got that figured out. We’ve got to commercialize.
Everybody else is going to have to go through that whole transition eventually to get to the same level of the ability to continue to drive the road map forward, make sure we preserve that, make sure — and obviously, there’s other BiCS versions that are in the works and so make sure we keep the investment on that road map and the same thing on the Drive business. Make sure that no matter what we do around the business, we keep the core technology moving forward on businesses where we have long-term investments to make sure that you always stay on the leading edge of competition and you can bring the best value proposition to our customers. And so we always want to protect that, whether we’re in a down cycle, mid-cycle up cycle, continue to invest in that foundational technology.
We’ve got just a couple of minutes here but I wanted some closing thoughts from you, Dave. What are some of the key messages that you want investors to take away from today? But what are the areas you think investors may underappreciate with regards to the WD story?
I think it’s a lot about what we just talked about. When I think about it, one is the foundational technology is very, very strong in both businesses. We’ve always had a very capital-efficient road map in NAND. That’s been something — sometimes people say to me, well, you must be falling behind because you’re spending so much less CapEx which is kind of an interesting thing which is not the case. We just have an engineering team and we have the scale of R&D investment because of the JV to make sure we can invest as much or more as anybody in the industry and that has worked for 20-plus years now and we feel like we’re in a very strong position there.
Same thing on the HDD business; the road map that we put in place over the last several years is working. ePMR, Opti NAND, UltraSMR, that is what all customers are adopting. You’re seeing it now show up in the numbers. When the next transition happens, we’ll be there with that as well. At the same time, we’ve spent a lot of time to adjust the cost structure of the business, the fundamental structure of the business, take cost out, especially on the HDD business. So that when the volumes normalize, profitability. We’re in a much stronger position to deliver profitability.
And then, the go-to-market synergies that we have, this idea that we can sell into the consumer market and reach every individual in the world with our products in the SanDisk brand and the WD brands and we go all the way from that to selling $1 billion worth of product to the largest customers in the world is a breadth of go-to-market and portfolio strategy that allows us to mix across that and get the best financial return no matter the market conditions we’re in.
Perfect. Dave, Wissam, thank you very much for joining. Thanks for your insights today.
Thanks for your time. Thanks, Harlan.
Thank you for having us.
End of Q&A