Seagate Technology Holdings plc (STX) Earnings Call Transcript & Summary

Hello. Good morning, and welcome to the ODC. And next up, we have a session with Seagate. And so just to let everyone know that doesn't know me, this is Mark Newman. In Bernstein, I cover global memory as well as global in-storage and delighted to be welcomed -- welcoming our next guest, Dave Mosley. He is the CEO of Seagate. And we are going to start off with some brief introductory remarks from Dave, and then we're going to go to Q&A. A reminder that there is -- we're going through the usual protocol, so we've got the pigeonholing. I will be monitoring. And you can enter a question in there, you can also vote on questions, and I'll be using that to monitor the questions and ask -- and try to ask any questions you put in. And with that, Dave, over to you for a few introductory remarks on yourself and Seagate, how's Seagate doing, and how's the year shaping up, the fiscal year shaping up so far?

Thanks, Mark. It's great to be here virtually with everyone. Wish we could catch up in different times, of course. As a reminder, I'll be making some forward-looking statements, and you can learn more about these statements in our SEC filings, which are on our website at www.seagate.com. But to your point, let me just frame this to highlight a few key aspects of the business with the times that we're in right now. I think everyone knows Seagate. We're a mass capacity solution vendor, legacy of innovation that goes back 40-some years now making disk drives for all kinds of different applications. Though the last few years has really dealt with this kind of unabated growth in mass capacity storage, that's the big secular trend towards hyperscale and even edge scale-out, which is kind of incredible, and it's now about 2/3 of our revenue and 3/4 of our exabyte shipments are actually pointed there. So no longer is it laptops and mission-critical drives now, it's more of this mass capacity stuff. So we pivoted over the last few years. We've done that with a strong technology road map, which we've been very thoughtful about and in how do we stage the right component tree, whether it's the recording heads or recording media or disk drives for that. All of our high-capacity HDDs now are based on a common platform approach that, well, that allowed us to get to 16 terabytes early and it will get us to 18. And even 20 terabytes, I think, provides us the ability to produce not only those drives, but also the midrange capacity points as well, 12, some 14s and things like that, and we can do that so we get better synergies inside of our manufacturing operations, better cost synergies, cost efficiencies, and we can go after the market a little bit more aggressively that way. This quarter, also, we're going to introduce a new technology, which is called HAMR, we've been talking about for a long, long time. It's a critical milestone really for the industry because this HAMR, heat-assisted magnetic recording technology, doesn't just -- it's not a simple extension of our old S-curve, it's actually a new S-curve itself and really allows capacity jumps of 10 terabytes or even more, so we should get to 30 and 40 and 50 terabytes some day, which is what the world needs to run data centers more and more efficiently. It also keeps the differentiation within the other technology solutions that are out there, and we can talk about that if you want to. This need to store more data is not only evident in the hyperscale, there's also a lot of stuff happening on the edge right now, and we believe there's a lot of edge applications for mass capacity storage, starting to break it down a little bit into micro edge and metro edge. And there's some natural use cases that are coming out, but it's really about decentralized data. Again, not everything is going to be centralized in one cloud. So we've just recently been talking about this new Lyve storage platform which is meant to get at that. All the complexity of data management in a distributed enterprise world, that's how we're thinking about it. And we would -- I would love to take questions about that. That's -- we believe that's a good growth vector in the future. Just to tie off to the current market environment, December quarter is off to a good start. The linearity is good. The end markets are tracking pretty well with our expectations and which is actually pretty different than the last couple of quarters that we have seen during COVID. But the COVID pandemic is by no means behind us from a supply chain perspective. There's still logistics issues that we have, so continuing to manage our business really carefully through these times and making sure we stay nimble in our supply chain. Our long-term outlook really hasn't changed. I mean it's all about this big growth in mass capacity storage. Data is going to explode because of IoT and AI and ML. All the things, smart cities, everything else people are talking about, all these trends are true in the public cloud, but they're also true in the private cloud and on-prem kind of investments, and they really fuel some of these mass capacity trends that we talked about, things like Lyve and whatnot. So all this long-term confidence in our business is illustrated by the fact that we just announced a dividend raise of 3% and a share buyback authorization at $3 billion additional. From my perspective, that's a confidence factor in our long-term free cash flow that we can generate in this business. I'm pretty excited about these times. I know they're tough for everyone, but we're reacting really well as a company. I'm happy to take your questions.

Thanks very much, Dave, for that. And I think I'll start first on the demand side. So totally agree with those comments about this, the data explosion that's happening right now, and there's a huge demand for storage devices via hard disk drives and SSDs as well. And of course, so you gave -- you're very well-positioned in that market. But there has been a bit of a weakness recently, partly perhaps due to a slight inventory buildup and part -- on the cloud side and partly on the end -- on on-premises status on the side because of the lockdowns and just the overall economic situation being a bit poorer there. So I would just like to get the latest view from you on that market, data center market, both cloud and on-prem. And as I recall, on the Q3 conference call, you seem to be mentioning a little bit of improvement in this December quarter. And did you want to see like -- is that meaningful improvement or is the real improvement further out? Or when should we expect that from the data center recovery, that's like. Or if you could talk about that demand picture, first of all, it would be great.

Yes, that's really good. So just to explain the dynamics, I think it happened to a lot of people this way. We ended up in the April time frame with factory dislocations around our supply chain. And that created a situation where people didn't know they were going to be able to get supply, so they probably took in too much. And then the July time frame, the true demand picture due to all the macro disruptions really came home to roost, and it really affected people on-prem. I mean our discussion at the time, we couldn't even get people on-prem in a lot of locations. So I don't know what letter to assign the recovery, but I think it's a gradual recovery. We are seeing signs of it. We were, at the end of last quarter, were seeing gradual recovery on-prem, even mission-critical and markets like that. I don't think it's profound, except for surveillance, which we talked about last quarter. We're starting to see that some of the smart city and smart factory and smart everything investments are happening as people do come back on-prem. Relative to the hyperscale and the, I'll call it, on-prem enterprise, it's actually interesting because the -- I think there was a big push into cloud applications at the early days of COVID. And this is -- I'm speaking in generalities, of course, but it happened globally like this. There were a lot of people who said, okay, I'll go to a cloud instance for an application so -- because I just can't get back on-prem or whatever migration I was going to do has just been accelerated. I think that creates all kinds of logistics issues for the cloud service providers. And also, it creates people who are doing any on-prem investments to really take a step back and question what they're doing. I think the world is settling out from some of those discussions. The cloud service providers are going to grow data big, I think, in the next few years, probably ahead of our plans pre-COVID. And then the data on-prem will go to some of the newer instances where there's better TCO trade-off, frankly, than whatever they were planning to buy before, which was maybe what was qualified before. I think people are taking a pause on that and then coming back when the time comes, if that helps your context.

Yes. So -- but do you think the first half of calendar '21 will be similar to first half of calendar '20? Or is it still going to be -- is it going to still be lagging a little bit, or is it going to be back to that level, do you think?

Yes. Some parts of the market, right, some parts of the market are stronger already. And I think, generally speaking, the first half of '21 will be stronger than the first half of '20. It was in Chinese New Year where we started seeing the first disruptions in calendar '20, so I think, in calendar '21, we're not really forecasting any of that. And talking to customers about what they're planning to install, they're giving us a lot better visibility to some of these trends that I just talked about being more solid.

Right. And then any other comments on demand on the other markets, the client drive markets? You mentioned a bit about surveillance. Maybe just wrap up on the kind of like the overall demand picture for client drives right now and how that's recovering.

Right. I think there's a lot of people who have talked about notebooks, there's not too many hard drives in notebooks anymore, but the demand is, I say, predictable and steady for us. Same thing on desktops, there's a movement afoot in some of the desktops where we're still present with hard drives to move to higher capacity points. So they're blended solutions. They might have an SSD and an HDD in them. So that's a good trend. And then the distribution channels are a lot stronger than they were just a while ago. I think most of that goes into either NAS boxes or some of these smart city, smart factory, video applications, so I think the desktop markets that have stabilized nicely are growing in exabytes. Consumer has been great. Consumer demand is strong. I think that some of that is work-from-home, game-from-home, learn-from-home, those kinds of things that other people have talked about quite a bit. So on the client side, I think -- we think it's very predictable what's going on, how much ramp-down there is in some markets and then how much stability there is in others.

Great. So if we could move over to the technology side, technology and product side. There's actually a question from the audience that I was also been asked, which is about how HAMR. So if you could share the latest on the HAMR products, how is that, the 20-terabyte HAMR product, that you talked about starting to ship within this calendar quarter, kind of Q4. Like how is that -- is that rollout on target? And also, any updates for the 24-terabyte for that ramp as well?

Yes. So we started talking about -- gosh, we started talking about 20 terabytes in 2013, started talking about earlier this year, calendar year, we had enough confidence that we were going to be able to go into some markets with the 20-terabyte HAMR product. The way I think about it right now is the volumes are small but growing through calendar '21. We should be able to penetrate other applications in '21 as well to make sure we get that learning. And the capacity point, since HAMR is capable of higher areal densities, we may even dial that up a little bit through the course of the year. I think we'll be watching that and working with our customers on exactly what they need. I'm not going to announce the 24-terabyte product just yet, but that's definitely the goal of HAMR, is to be able to go well beyond that. That's what we can do with components that are in our labs right now. We just have to build the right solutions for the customers. And these are tough applications, so we have to make sure that we get it right. I think the other opportunity in HAMR is obviously to start taking components out of some other capacity points, say, the 16s or the 12s that we're still shipping. And we'll stay focused on that as well, making sure we optimize the portfolio and blend the cost. If we can remove components from inside these capacity points, we're saving money. So all that stuff is the blend. We're going to be learning about real-time in '21.

But you're going straight from 20 to 24 because up until now, it's been every 2 terabytes. But is that -- after 20, is the next one 24 or is there also a 22-terabyte?

Well, I mean, it will be wherever we can land the areal density capability and the reliability, the quality that we need for the market segments. I mean -- but the goal is definitely, exactly to your point, is not to just incrementalize, it's to get bigger jumps. And that's why we've been investing in HAMR, and that's what we believe the promise of the technology is.

And then if you think about HAMR technology specifically, there's a bit of extra cost related to the laser. There's a little of extra cost to the HAMR products. So I think you've mentioned in the past, 20-terabyte is not really going to be that cost-competitive. I'm not sure if that's the right wording, but I think, effectively, I think you're saying that 20-terabyte, you're not getting the full economic benefit yet from HAMR because of the extra cost from the laser diode and not really getting the big benefit of areal density yet. What point -- at what point should we expect HAMR to really come into its own? Is that 24 or is it later?

Yes. Obviously, we don't know the answer to that yet, nor we're still calculating exactly how we're going to go to the market and we'll announce the products when we want to. But we're staying in touch with our customers about what technology opportunities that there are, and there will be people taking advantage of this throughout the course of the next year, not just on the learning front, but also, it will be a pretty good economics for them. I think when you get into the massive TCO shifts, you do have to get up into the mid-20s, frankly. And it's not only about the highest capacity point. Again, if we can save a disk and 2 heads in a 16, we'll look at doing that as well. So it's really across the whole portfolio, which is why we think that this platform play is so important. We can introduce HAMR into the same platform. The cost increases are really nominal. They're more tooling and things like that, that have to be amortized, and that's the direction we're getting on. Once we see we can accelerate that, we'll really step on the gas.

So -- okay. And then lastly, what portion -- within nearline, roughly what portion is 14-, 16- and 18-terabyte versus sub-14 today, roughly? Do you have any kind of estimate on that, just to get a sense of how much of your product is that kind of leading-edge 16 and 18 currently?

Yes. It's really -- it's interesting. We haven't really broken it out, and part of the reason that we don't is because we don't talk about individual customers. I think there are some customers who stayed on 12s and 14s much longer, even though we had 16s. And then we were able to use that 16-terabyte to qualify them as a 14 with fewer heads and disks, like we talked about. My sense is that the things that caused some trepidation in the market of old are now getting worked through such that, I think, globally, you'll start to see a lot more people at 16, 18, 20 over the next few years. There's probably going to be less of -- only a few customers are buying the highest capacity point. I think you'll see a big broadening of that, and that's opportunity for us as well.

So there's no -- in your mind, there's no risk to the 20-terabyte shipping, it's just really a question of when that becomes a significant commercial portion of revenue, which I assume that's the way into 2021?

Yes. We work deeply with the customers on all technology trends that we see. We plan these transitions, I'll say, 6 months to a year in advance with them, so they have pretty good visibility into what we're doing.

But can you say, are those 20-terabyte drives, have they started shipping yet or not yet?

So the HAMR drives, so we, like I said, we could do 20 terabytes with conventional as well if we want to. Yes. The HAMR drives, we said this quarter, and I'll stick with what we said in the past, we're going to ship them this quarter.

Okay. Right. That's good enough. So on -- you recently introduced the Lyve Rack and CORTX open-source software aimed at helping enterprises cost effectively store and manage mass data. Could you speak a bit about the strategy for the Lyve platform and how receptive are customers to that?

Yes. So I think people know that we make systems as well that are enterprise storage systems, call it a rack, if you want, that most of the mass capacity drives that live out in the world don't live with 1 or 2 of their friends, they live with hundreds or even thousands of their brother and right next to them. So we believe that, over time, our ability to understand what's going on with the drive and reduce cost the right way is a good trend. Now part of it is the hardware, and that gets a lot of attention, the rack itself, but part of it's also the software, understanding how to treat each one of the drives, their specific nuances, especially when you go into qualifying new drives. So say, for example, the dual actuator drives that we have, they're called MACH.2, those -- sometimes, the system may not handle those right, whereas if we can provide the software solution as well, we can garner even more value for the end customer. And that's the way we think about it. It's a good business extension. It largely shifts mass capacity for us, and we're optimizing around mass capacity and all the different trends that we're seeing there and love the partnerships that we have as well.

Got it. Okay. And then, actually, back on the demand question, I wanted to check what assumptions you're currently making to achieve the flattish revenue, FY '21 versus 2020.

Yes. I think if I tick down through the markets, which we already talked about a little bit, the biggest thing is cloud, making sure that the on-prem comes back and then making sure that the hyperscale trends are the ones that we've been forecasting because there's been so much disruption in the market to both of those things. We're talking to the customers, trying to understand the trends globally as well, what investments are they going to make. We're still fairly confident in all the numbers that we put together. And part of this is just to inform everyone of what we're seeing out there. The July plunge, if you will, in the market, was really a temporary thing around the demand and supply situation with the early days of COVID. It had nothing to do with long-term secular growth of mass capacity, which we believe is still there and probably being accelerated because of some of these trends that we've seen during COVID.

Right. Right. So on NAND flash, I think this is a big topic, right, about NAND cannibalization, cannibalizing hard disk drives. And I spent a lot of time covering memory, and I used to work at Samsung, so I'm quite familiar with the memory industry. And clearly, NAND-based SSDs, as you alluded to earlier, have pretty rapidly cannibalized some portions of the hard disk drive market, such as laptops and mission-critical, and starting to a bit in desktops as well. But just what's your view on NAND as a threat to hard disk drives in cloud data centers, considering overall like cost and energy efficiencies, NAND versus hard disk drives? Like what's your view of NAND versus hard disk drive as a potential threat to the HDD industry?

Yes. The -- I preface every comment by saying NAND is a great technology. It has a bright future. Actually, I think that if you architect the right solutions with NAND with DRAM, combinations of DRAM and NAND, you can do pretty amazing things. There may be even new memory technologies coming that find their own niches inside of the memory ecosystems. When it comes to the storage back end, I don't -- I really don't view it as an either/or proposition, I view it as a both proposition. And I think that's when -- we've seen NAND come of age to service a certain tier. I think the hard drive tier is a different economics altogether. It's a different problem set, actually. And if you were to invest in NAND to take over the exabytes of the hard drive, it's probably not a good economic proposition for you. That said, maybe somebody is going to try it. I don't really worry about it too much. I think there's a lot of other noise in the system about total cost of ownership, that you have an intercept 0.5 years or 7 years. That kind of assumes that hard drives won't stay still. They will stay still. They will get better in areal density capacity. They will get better in performance, and they will get better in power as well, significantly better. So from my perspective, there's a lot of great NAND investments out there in the world. We actually use a lot of NAND and move it with our brand to satisfy customer problems. I think that either/or proposition is not really well-informed.

And if you talk about more on that Seagate strategy in NAND flash and SSDs, like what is your strategy with regards to NAND supply? And where are you getting the supply from? How are you locking in those contracts? And maybe talk a bit more about your SSD product strategy a bit more.

Yes. We don't run the fabs. I mean like a lot of people that do run fabs, I mean, they have their own challenges because that's a huge scale business. It's such an interesting operational challenge to run a memory fab, as you know. So I look at that and say, I'm not sure I want that challenge. I mean the way we look at it is, how can we work with partners, design our own controllers or use merchant controllers, be as flexible as we possibly can, use our brand, use our channels to go out there and make money in the market. And the market's been really challenged over the last couple of years, especially for people who didn't control the NAND, but even for some people who did control the NAND. So we have to tread very carefully through that technology landscape, but I like the way we've improved our plan over the last few years and just stay small, nimble, make sure that we service what the customers need and take advantage of the fact that we're not necessarily tied down to be nimble.

So your NAND contracts, are they very long-term, sell it at a certain price, or is it all variable based on the current market dynamics?

Well, we have to negotiate all the time. So yes, they're -- we have different contractual agreements with different vendors, but we're always in negotiations based on what's going on in the market.

Because presumably, it's going to be something market-related. And so in a period of oversupply, with weak prices, actually Seagate is quite well-positioned because you get cheap NAND. But then vice versa, right, when there's tightness, then you're going to get high prices, and that's going to be a slight disadvantage for you versus, say, Western Digital. Western Digital is going to be getting it at cost, but when the price is below cost or pretty close to cost, then you're effectively better positioned than them. But when price goes up, obviously, that helps them, so it can go both ways.