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

All right. We're going to get started here at 10:45. So welcome to day-2 of the TMT Conference. My name is Erik Woodring. I lead the hardware coverage here at Morgan Stanley. Before I get into our speaker, let me just read this disclosure. Please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your sales representative. So I'm delighted to be joined today by Seagate's CFO, Gianluca Romano, kind of constant on the -- at the TMT Conference, been around Seagate for over half a decade now. But Gianluca, thank you for joining us today.

Thank you. So before we start, I will be making forward-looking statements today and you can learn more about the risks associated with those statements on our website. You see enough.

So let's start from the top. It's been about 5 weeks since you reported December quarter earnings, you guided to about 10% sequential revenue decline for the March quarter, and we'll get into some of the factors behind that. We're about 2/3 of the way through the quarter now. Can you just maybe give us an update on how the quarter is shaping up versus your expectations? Any of the kind of underlying drivers of the quarter changing from a revenue perspective, and we'll go from there?

Yes. The quarter is shaping out as we were expecting. As you know, we have some limitation on the supply. So that limited a little bit our opportunity to have a higher revenue in the March quarter. So it's coming out as we were expecting, I would say, at the midpoint of our guidance range in terms of revenue. The problem, as you know, has been solved, so we will not have the supply impact in the future quarters. So we are anxious to move after the March quarter.

Okay. Perfect. And then how about just the margin side of things? You've talked about gross margin expansion in the March quarter. Is that still how we should be thinking about profitability? And kind of any guardrails you'd be willing to put around, how much sequential expansion we could be thinking about gross margins this quarter?

Yes, we said in the earnings release, we expect gross margin to be slightly better in the March quarter compared to December, now despite the lower volume. Of course, when you have lower volume, you have a negative impact on your cost side. But we also said the cloud is actually growing in March compared to December. So mix is going in the right direction. The pricing action is still going in the right direction. So when you put all together, plus and minuses, we still believe we will have a slight better gross margin.

Okay. And then maybe last one on the quarter, just OpEx, kind of $290 million quarterly run rate, nothing changed...

I think OpEx, maybe will come out a little bit lower.

I'm going to circle back to maybe some things around the June quarter, but let's stay on March, which is -- you and your competitor guided to sequential revenue declines in the March quarter. A big kind of point of feed or pushback that I get is that elicits concerns about the cycle, so to speak. What gives you enough confidence to kind of sit here and say, really what you're seeing, what Seagate is seeing in the March quarter is transitory, this will dissipate after we've seen multiple quarters of sequential nearline growth. I think that's kind of one of the key concerns on investors' mind?

Well, I would say, for us, cloud is still growing in the March quarter compared to December. So we have not seen a change in the trend. The other segments, of course, are more impacted by this lower supply and maybe we could have done even better in the cloud, but it's still growing sequentially. And then when you go longer, as we discussed at the earnings release, we have, of course, the visibility in terms of the build-to-order, especially in the nearline space, and this gives us confidence for an estimate of a better sequential revenue through the calendar year and also, of course, a better profitability.

Perfect. So I realize that if you want to, you can, but you might not want to guide for the June quarter quite yet. But do you have any better visibility now that we're into March about how we should be thinking about that kind of $200 million of customer demand that you were unable to fulfill in the March quarter because of the supply issue? How do we think about the recapture of that $200 million? Any kind of more solidified ways in thinking about that for us?

I think the majority of that demand is still there. So we'll probably move into future quarters. As you know, especially in the cloud space, demand is above supply. So say, supply is more the driver for the volume that we ship and I think the industry ship in that segment. So no, again, we are positive looking at our build-to-order plan that we can grow sequentially through the calendar year, mainly because with the transition to ML, we will be able to produce more exabytes with the same unit, and that will give us opportunity to take little bit more of a demand between now and the end of the calendar year.

And maybe to that point, again, looking beyond kind of March and June, again, to reiterate, you've guided to sequential growth and kind of revenues, margins, earnings power through calendar '25. I realize build-to-order is probably the primary factor that impacts that. I guess my 2 questions are, are there any other factors that you can speak to that help give you the confidence to come out and kind of guide that way loosely as we sit here in early March? And second, when it comes to build-to-order, can you help us all understand kind of like how much of your capacity shift is really built-to-order that you have visibility on through, say, calendar '25 or however you can frame it for us?

Yes. I would say, the confidence is coming from the fact that we see the demand, and we see our opportunity to produce more exabytes sequentially, again, without adding unit capacity, but just with the transition from high capacity PMR to HAMR and with that transition, we will be able to produce a little bit more exabyte and this is giving us the confidence to give indication of what we expect for the next 3 or 4 quarters. In the short term, you have a higher level of build-to-order, of course, and then when you go a bit longer, that amount tends to decline, but it's still a fairly high level. So the confidence, I would say, is high on the calendar '25.

Okay.

Calendar '26, we will discuss more maybe in a quarter or 2.

Okay. So you're talking about calendar '26 and the middle of calendar '25?

Yes.

Okay. Another kind of big point of pushback or investor concern is just thinking about the longevity of the cycle, right? In the 2019, and I'm really focusing on the nearline market here, first and foremost, that 2019 cycle lasted about 13 quarters. There's maybe a pause halfway through related to COVID. And you can make maybe some arguments that it got toppy towards the end. Obviously, that's what led into the subsequent down cycle. Right now, again, we're 6 quarters through kind of the inflection of the bottom. Where are we on this cycle? Does this cycle because of build-to-order, because of what you're hearing from your customers from a demand standpoint? Is there more longevity to this cycle than prior cycles? Just kind of given the supply demand imbalance, why would this cycle look different maybe?

Well, in theory, you are right. No, we are not taking all the demand that is available today as an industry because supply is below demand. And so in theory, this demand just shifts out in time and keeps the cycle going longer. I would say, in general, all the technology industries have some kind of cycles. But the fact that you can push with demand out in time should reduce at least the level of the cycle. So you push little bit of the peak, you reduce also the trough of the cycle.

Right. And how do your customers respond to that? And really what I'm getting at is, if I'm a customer and I say, I need this amount of capacity, sure, you can get it through build-to-order. But if I can't get a now of it, is there a level of frustration that you hear coming from customers? Or how do they respond to maybe some of the actions that the industry has taken to, again, you elongate the cycle, focus on profitability and cash flow?

I don't see a lot of frustration. I think demand is above supply, but is not so much higher than drive frustration. And we are part of the CapEx of our customers. So they hard disc drive to install in data center that will be used also in the future. You have a part of what they buy that are used to refresh current data center and they're used basically immediately and some that are used for new data centers, so they need to build the data center, install the drive and then have the utilization rate coming up. So I don't see frustration. I think, a good environment in terms of supply-demand. And so far, we are not concerned about reaction from customers because there is not enough supply in the industry. I think it's just a matter of pushing some of that demand a little bit out in time.

Okay. And I realize that a couple of days of market volatility doesn't necessarily impact the way any customer thinks and obviously build-to-order helps to protect you. But when you speak to customers or when they speak to you, is the mindset shifting at all for them when it comes to data center build-out, when it comes to the need for -- addressing the needs of storage growth?

No, I don't think that. No, what is happening or recently happened in the market is due to a change in view of how data will grow and how high will be the need for data storage, I think are completely different issues. And data is growing very rapidly. AI will be an additional driver to that growth on top of all the other applications that you already have. So I don't think there is any relation between the 2.

Okay. Okay. Let's shift the conversation to pricing. Again, with this kind of build-to-order approach, you've been able to take pricing for several quarters now. On a like-for-like basis, realizing there will be mix shift towards higher capacities, but on a like-for-like basis, do we still think we're capturing sequential growth in pricing in calendar '25? And if I were to say that low single digit sequential, is that a fair characterization of how to think about the pricing environment?

Well, we are not changing our pricing strategy. As you know, we started almost 2 years ago at this point, to increase slightly our pricing every time we were renegotiating a build-to-order and this is still going on. So I don't see a reason to change this trend. The industry is very focused in improving profitability. Profitability was fairly low about 2 years ago and so the industry refocused. And right now, we are getting the benefit from that action and focus on producing only what is needed. Right now, with the demand, which is above supply, you can produce all what you can and you're still little bit short. So no, I don't see any reason for the change. As we said in the past, we don't go for very high price increase, are actually fairly small, not very impactful to our customers. And this is what we want. Now we don't want to impact them. We want them to be successful, but we also want to improve our profitability.

Right. If we think about kind of going back to the cycle and your point on kind of supply factors here, I think the peak for you guys is about 150 to 160 exabytes per quarter. Dave has said publicly mass capacity exabytes can grow something like mid-20%. My question really is, is it possible to achieve that rate of growth with kind of capacity constraints and really just driving exabyte growth primarily through mix shift? Like is there a dampening effect to that? Or is there enough visibility and your customers are willing to take on these higher capacity drives, I realize HAMR has an impact on the story, such that that's still a fair way of thinking about mass capacity exabyte growth?

Yes. I think this model of mid-20% of growth for nearline is a good model. It is a good way to think about the industry in the next 2, 3, 5 years. Technology can drive that growth or very close to that growth, especially with HMAR. HAMR will allow us to go from 3 terabyte to 4 terabyte to 5 terabyte and continue on that growth. So we can -- depending on how you move your mix, how much of your units you move on the highest capacity drive, you can generate a good increase in exabytes that we produce with the same units. So we think we will be fairly well aligned. Now every year is a bit different. If you have -- if you look at the last 2 years, the industry grew much more than 25%, but of course, it's part of the cycle.

Okay. Let's shift over to everybody's favorite topic. Obviously, you anticipate this one HAMR. So you've already qualified your first cloud customer, where do we stand with qualification for the remaining major hyperscalers? And any hiccups you've encountered during the process as you've worked with them that would kind of elongate the process as you previously described it?

No. The qualification is going well. As we said, we will have multiple customers qualified by the mid of this calendar year, so just a few more months. And then we will ramp our HAMR volume, and you will see a much higher volume of HAMR product even at different capacity of HAMR products starting the second part of calendar '25 and of course, even more in '26. We also said that by the end of this calendar year, we will have 4 terabyte per disk product, so potentially 40 terabyte drive available to start qualification. So -- but it is a bit of compression between the 30 terabyte and the 40 terabyte, mainly because the 30 terabyte was qualified little bit later than what we were expecting. But we never stopped working on the 40 terabytes. So there are different teams working on different platforms. So right now, we have this transition that is happening fairly rapidly. Probably in the future, we will not go so fast. But for this year and next year, you will have a fairly quick transition.

Right. And so -- and we'll get back to kind of that shift in [aerial] density, but again, a point that I hear back is, sure you've qualified HAMR with your first cloud customer. It's still kind of low levels of adoption, I get questions about yields. And so maybe my question for you is what are you hearing from your customers, be them qualified or going through the qualification process about the potential adoption of HAMR drives? Do you start small and build over time? Do they need dual sourcing? These are all the kind of questions I get, and I'd love to maybe reframe...

For them, it is a better TCO if they can buy a higher capacity drive. So HAMR is the highest capacity drive, so they want to qualify HAMR technology, not because of the technology but because it's a higher capacity drive. And they want to move as fast as they can to buy the highest capacity drive that is possible and that is available in the market. The point is how much we have available and when we ramp? And because now we have a certain advantage on this technology, we want to qualify a certain number of customers before we ramp high volume. And we don't want to have the risk that we ramp the product and then remain in inventory for 2, 3, 4 months. So we have the time to do it in a good way. Of course, qualification is the main focus, but we already qualified in all the important segments. We are qualified in the cloud. We are qualified in enterprise OEM. We are even qualified in video and image applications. So we have the right configuration for all those segments, and we just need to go through a normal call with customers. And then because we already have built-to-order in place, the only change is the mix. So when they are qualified on HAMR, they can buy a certain volume that we will produce on HAMR and what we don't produce, they need to buy on PMR.

So let's get back to the technology innovation point that you were talking about. So we've qualified, correct me if I'm wrong, but low 30 terabyte disk drives. Again, can you kind of go through the time line for all of us? I know at earnings, you talked about 36 terabytes, but the time line from low 30 to 36 to 40 and beyond, how do you think about that either from a shipment or qualification timing perspective?

So we have qualified the low 30s, and we are sampling 36. And we also said very quickly by the end of this calendar year, we will have the 40 terabyte out to start for. So as I said before, very compressed schedule. Customers are qualifying different capacity depending when they go through the call time. You will see this even with a 4-terabyte per disk. So we are not jumping from 30 terabyte to 40 terabyte to 50 terabyte, with our product in the middle. But the platform -- the only change in the platform is from 3 terabyte to 4 terabyte to 5. But when you have the 3 terabyte per disk and you have like a year or 2 years' time before you go to the new platform, you have development on that platform. So you go from 30 to 32 to 36 and then you move to 40. This is why we call that 3-plus or 4-plus or 5-plus because there are multiple product -- multiple capacity on basically the same product. So we are doing this with a 3-plus, and we will do the same with the 4-plus.

Okay. Okay. Let's bring the topic of AI into the fold here. You've talked about your mass capacity business being an enabler and a beneficiary of the AI trends. I think the concern would be mix shift to SSD, just questions about power consumption, performance, IO speeds. How do I think about AI in the context of that mid-20% mass capacity CAGR? Is that incremental just -- I know it's hard to parse between an AI workload and non-AI workload. But how do we think about the incrementalism of AI relative to how you've kind of already told us the world looks like from a capacity growth standpoint?

Well, as you said, it's difficult for us to understand where the data is coming from, the data that is going into our hard risk because AI or a different application, they require the same kind of hard disc drive. So we don't ship a certain drive specifically to AI where we can separate the 2 is actually the same drive. They are very high capacity drives. We believe that knowing that mid-20%, there is, of course, a good volume coming from AI, but that percentage could go up if the adoption of AI is really strong, and it's a little bit happening faster than what we are expecting. Now we are maybe a little bit prudent on the time of the adoption. We are not prudent on the impact that AI will have on storage. But we have seen in the past with new applications, there was an expectation that the new application immediately is adopted by everyone and actually, it was not the case. So we are a little bit more prudent on time, but the impact will be huge. And the storage that will be associated to AI, and that will end up in -- on a hard disk is also very, very important.

Okay. I want to make sure that we bring a number of different topics here into the fold outside of just kind of capacity growth. So first one, gross margins, obviously, a critical part of the story. They're at a 10-year high. You're kind of guiding to gross margin, you kind of -- you are guiding to gross margin expansion through the remainder of calendar year '25. What is helping you get to these gross margins at your lower revenue rates? I know that you've told us that you can do that, but what are the underlying specifics of how you're able to accomplish that? Maybe let's start there.

I would say, mix is very important, not differently than in the past when you move up in capacity and you move your volume to that higher capacity, you have the opportunity to increase profitability. But I think, of course, the pricing action that we started almost 2 years ago is going on, is continuing. So the combination of the change in mix and the improved pricing is, of course, driving a lot of that gross margin improvement. And now that we start to get ready in term of -- full in terms of supply, we don't have any more the unused capacity additional cost that we suffer for a certain number of quarters in the past. So all this is actually going in the direction of generating better profitability. And we see this continuing, at least for this calendar year where we have visibility.

I think, when you say this calendar year, you're just not willing to talk to calendar '26 quite yet?

Yes, we will wait until we have more -- a better visibility.

Yes. And maybe I'll ask a question beyond just calendar '25 and give it a go and see what you have to say. Just when we take all these factors, mix, pricing, supply, yields, underutilization costs, internal cost efforts, how do we all think about the incremental room for gross margin upside? Your old target or your kind of pre-existing gross margin target is 30% to 33%. Is there a new target that we should be thinking about for gross margins at this point? Obviously, your competitors come out and kind of stake their claim in the ground? How should we think about it for Seagate?

I think our objective is continuous improvement. We don't have a specific target. We want to continue to improve our profitability. We have a good understanding of what other components in the different applications can generate in terms of profitability. We think we are a very important components in the data centers. And so we deserve a better profitability, and we are still continuing to go in that direction.

I'm going to pose a hypothetical to you before we get into kind of the last handful of questions, which is, let's say, the cycle turns and things turned a little nasty, I'm not saying that will happen, but hypothetically, work with me here for a second. If there were potentially share losses or anything to that degree, how would you respond to that? And I think I know the answer, but I just want to kind of confirm, which is, historically, there might have been some action taken around pricing to regain that share. Would your focus remain on maximizing profitability and cash flow? Is that kind of like the change -- I want to call it a change in mentality, but I don't want to put words in your mouth. Is that how we would think about it?

Well, the short answer is yes. The build-to-order should give us visibility eventually when the volume in the cycle will start to decline. When we have visibility, we can prepare our manufacturing, reduce our volume, avoid to enter into the lower part of the cycle with a lot of WIP because when you have a lot of products in your WIP, you need to move them out and try to move product to customers that are already full, of course, has an impact on pricing. So this is why for us, it's important to as a build-to-order is a visibility not only for the up cycle, but even for when the cycle eventually turns. So we can reduce manufacturing in advance, produce less volume, but keep the pricing and manage the cycle in a different way.

Right. Okay. Let's go back to OpEx. I know you touched this. So March quarter slightly low $290 million. How do we think about beyond March? Can you remain below $290 million? I know there is some variable comp that was stepping up. So as we just think about maybe the remainder of calendar year '25, what's the OpEx run -- quarterly OpEx run rate we should think about?

Yes. I think between $280 million and $290 million is a good range.

Easy enough. Something that is changing in the world is just the global minimum tax and what company is going to be forced to pay. Historically, you have a very low effective tax rate. Help us understand where that tax rate has to go in light of global minimum tax? I assume fiscal '26 is when we maybe need to think about it starting. But what does that rate go to? Or what's -- how should we all be contextualizing this?

Yes. So the global minimum tax is 15%. So as you said, starting fiscal '26, now even Seagate will be subject to that global minimum tax, so we expect to align more or less to that tax rate.

Okay. Super helpful. And then let's move to capital allocation. You said the goal is to get to kind of $5 billion of gross debt, then we can kind of think about buybacks returning to the model. Two questions, has that viewpoint changed at all? I assume not, but making sure I'm checking on that. And then second, how do we think about the timing of that kind of delevering and then the potential of buyback stock?

Yes. We did a big step for this new level of debt in January. We reduced our debt by about $500 million. So we still have about $200 million to go. Probably next quarter or the following quarter, we will take care of that debt. And probably in the second part of fiscal '26, we will look at restarting the share buyback.

Okay.

In general, during the up cycle, we also have dividend increase. As you know, during the down cycle, we protect the dividend during the up cycle, we tend to increase the dividend. So I do not exclude that. Usually at the October Board meeting is where we go through the request and authorization for increased dividends. So that, of course, assuming the cycle is continuing as we believe, could happen again.

Okay. And any thoughts you can share on maybe longer-term leverage targets? So getting to $5 billion is kind of near term? Where would you like to bring leverage whether that the absolute value of debt or leverage ratio?

Yes. Maybe we will discuss that little bit more at the Analyst Day. In the past, we have not had a specific target in terms of leverage. Of course, if you look at our EBITDA level, today and the run rate of EBITDA compared to $5 billion of debt is a very low leverage and probably will go even lower in the near future. But we will think it will be more about having a target. But again, today, when we think about the $5 billion is a number that we believe we can manage in any part of the cycle during the up cycle, during the down cycle, and therefore, it's more a dollar target than a leverage target.

Okay. Before we end, I want to make sure we kind of touch on one of the recent proposed acquisitions of Intevac, a key supplier of yours. Can you just walk us through kind of the rationale there and what the benefits of that deal would bring from you and any implications for the broader HDD landscape?

Yes. I don't think there are implications for the broader HDD space. That is the supplier that was mainly working with us for a certain number of years. So it was making sense to us to integrate that operation inside Seagate.

How about -- I referred to the industry as an oligopoly, but there's clearly 2 leading players. If the HDD industry has kind of structurally become more profitable and predictable, is there the potential for any newcomers or new entrants into this market? How do you think...

I would say, it's very difficult to enter into the hard disk market. First of all, the production, the manufacturing is complicated. You have ad production that is in the form of wafer fabrication. You have the media production, you have the substrate production. You have a lot of components that you need to buy through the supply chain and technology has evolved. So if someone is trying to enter into the hard disc and develop a PMR drive, they probably don't go very far and to develop HAMR takes many years. So I think it's very, very complicated.

Okay. Okay. Maybe with the remaining time, again, we went through a lot, but maybe what's the closing message you want to send everyone right now as we think about Seagate? Anything that might be underappreciated or not fully understood by the market?

I don't know if it's not understood, but I think it's a reality that the industry has changed compared to several years ago. I think there is a clear focus on improving profitability that was really way too low in the past and that the technology is important, that hard disc is a main component, especially in the high capacity applications, especially in the cloud, in the data center, even for other segments, but I would say, mainly for those segments. And therefore, it is a component that is extremely important to our customers because finally, our customers are selling storage to all of us and to all our companies and to government agencies and they cannot do it without having hard disk.

Perfect. We'll leave it there. Thank you so much, Gianluca.

Thank you very much, Erik.