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

All right. Perfect. So 10:45, so let's get started here. Welcome to day 3 of the TMT Conference. My name is Erik Woodring. I lead the hardware coverage here at Morgan Stanley. I'm delighted to be joined by Seagate's CFO, Gianluca Romano. You've been a mainstay at this conference for a number of years. So thank you for joining us today.

Thank you, Erik.

But before we start, the quick disclosure that you're all used to. Please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures for important disclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. So with that, why don't we start at the top, focus on kind of what has happened since earnings. It's been about 5 weeks since you reported December quarter earnings. You guided to 6% revenue -- sequential revenue growth in the March quarter. You talked about a lot of that coming from the big U.S. cloud customers. We're 2/3 of the way through the quarter. So -- just how is the quarter playing out relative to your expectations? And have any of the underlying drivers of March quarter performance changed?

Thank you, Erik. Before I answer, let me remind everyone, that we will be making forward-looking statements today and you can learn more about the risk associated with those statements on our website. Yes, the quarter is coming out as we expected. So we said during our earning release, we have increased demand in the nearline space, in particular in this quarter from the cloud customers. So now we have confidence in our midpoint of the guidance. That also, as you know, imply a good improvement in our gross margin sequentially from the December quarter that was already a good upside from the prior quarter. So we are finally putting together a certain number of quarters of improvement in term of revenue and profitability. And we also said that in the earning release now and I'm still confident today that we will see another improvement in the next quarter. So again, sequentially, finally, the business is improving sequentially and in term of profitability, it's not only the mix, it is also the pricing actions that we have taken and the cost actions that we have taken many quarters ago and that finally have been reflected in the financial statements. If I look at a year ago, the March quarter a year ago, I think our gross margin was in the 19%. Right now, if you look at how we guided, this means 6%, 7% higher in just few quarters. A good part of that is coming from the cost actions but the other good part is coming from the pricing actions that we have taken and that we are continuing to take. So this quarter, the improving -- the improvement sequential is mainly coming from pricing and mix. As you know, the March quarter is seasonally low for some segments like video and image application or the legacy part. So the increase in nearline is not only offsetting the seasonal decline for those segments but it's also adding another, about $100 million, which is what we guided at the midpoint of our guidance, sequential improvement. So it's a fairly strong quarter. But as I said, we expect next quarter to be even stronger. On the HAMR side, we are making good progress with our core. Now we discussed in the past, we have 1 major qualification with a cloud customer and we have a few other [ qual ] outside the cloud segment. With the cloud customers, we are starting this month, what should be our last batch of drives going into test. And usually, these tests last between 6 and 8 weeks. So we are almost to the end. In the non-cloud space, [ qual ] is a little bit shorter. So I think we will have one of our top customers, not in the cloud space, of course but outside the cloud space that could be qualified in the next week or two. So good progress, good confidence in the technology. And based on those timings, we will have ample time during the June quarter to ship now high volume of HAMR units.

Okay. Perfect. So we'll dig into all of that. Just to maybe final double-click in terms of the margin side of things. You had guided, as you alluded to, kind of 6 to 7 points of gross margin expansion year-over-year in March, you had $50 million of -- or you told us about $50 million of underutilization charges. Anything different as you sit here today about that thought process behind underutilization or gross margins, that trajectory in March?

No, I think -- well, I would say linearity this quarter has been a little bit better than what we were expecting but we still think the midpoint of the guidance is where we will end. Profitability and all the pricing actions that we are putting in place are coming out as we were expecting also. So I think gross margin, operating margin and EPS are coming out close to the midpoint or at the midpoint, we will know better in 3 or 4 weeks. But so far, this is -- I think we guided in the right level.

Okay. Perfect. So let's maybe focus on some of the big themes, we'll obviously get to HAMR. But if we maybe just start with your major cloud customers, obviously went through a number of quarters of sequential declines and challenges as they had built up inventory. You've talked about that inventory correction being done, obviously, an important step to the recovery that you're alluding to, you've also, however, framed the recovery as, call it, gradual or moderate. And so can you maybe talk to us after such a kind of steep decline, why -- what are the headwinds or what are the different factors that would get you to kind of qualify the recovery as gradual and not steeper. What are the most significant headwinds kind of offsetting that cleaner inventory dynamic?

Yes. Now, there are a few items. I would say where we are recovering is mainly in the cloud space, nearline and cloud in particular, not only U.S. but now we see a little bit stronger demand even outside U.S., mainly in China. However, the other segments and China, in general, is still very depressed. So to have full recovery, you need to have all the segments, all the regions coming back to a strong demand as it was before the down cycle. Right now, we still see a part of the demand to come back outside U.S. So it will take a little bit of time. And we also need to understand better when this increasing CapEx from our cloud customers will be more focused on storage and maybe less focused on preparing the infrastructure for AI. So it's a matter of time. Now AI will help to generate more and more data, that data will be stored in the -- mainly in the cloud but also on-prem. And all those data will be stored on hard disk but exactly the timing of when that will happen is still a little bit unclear. So we are a little bit maybe prudent with our forecast on the recovery. But right now, this is the visibility that we have. But it's still a good recovery, especially when you consider the difference in the cost structure and the different pricing levels, now even at lower revenue, we can generate a very good profitability.

Yes, Okay. Perfect. And then maybe just double-clicking on China. Obviously, a lot of moving pieces in that market, whether it's macro or even micro across cloud, via enterprise, legacy. Can you just maybe walk us through the visibility you have into China? Again, just beyond the March quarter and some of those moving pieces, how we might need to think about the moving pieces in China beyond just March.

Yes. So the cloud space is actually getting stronger. So if you look our customer like Alibaba, Tencent or ByteDance and Baidu are actually increasing demand. So this is a very good sign. And this is a contribution to the overall cloud improvement. The other segments, they are not improving too much. Video and image application is fairly flat at this point. And March quarter is seasonally low. So we also will see a decline in the current quarter but then will improve through the rest of the calendar year. But in general, I would say China is still fairly weak with the exception of the cloud space. Now they -- the China government, looks like, now they are trying to stimulate the economy as they can. But it will take time and it's difficult right now to forecast how much of that incentive will really result in an increase in the storage business. It will happen. But I would say this is part of our view of the recovery that is not a big jump but it's more gradual recovery because we think that part of the world will take a little bit of time to come back to the level of demand that we had before the down cycle.

Right. Okay. And then maybe just to put some numbers around this. I know on a prior earnings call, you had been asked about reaching 100 exabytes of nearline capacity ships by the end of this calendar year. You did about 65 exabytes in the end of last calendar year, calendar '23. What -- how do you think about that ramp? How do you see the shape of the nearline recovery progressing towards that kind of like that vantage point at the end of calendar 2024?

Yes, 100 exabytes is a big number. We are improving quarter after quarter. We did not guide the volume for the next 3 or 4 quarters. So 100 exabytes right now seems to be a fairly high number for this calendar year. But what I think is, every quarter will be a little bit higher and now if it's not at the end of this calendar year, we'll be probably be in the first part of the next calendar year. But again, the recovery could be a little bit stronger than what we're seeing today. It depends what happen in Asia.

Yes. And I wanted just to go back to the comment that you made about cloud CapEx and AI and the difference in kind of maybe wallet share within that dollar of cloud CapEx budget, every incremental dollar. Historically, there is a very close relationship, obviously, between your business and cloud CapEx. Is there any insight that you have into kind of how long this trend will or might last until we see a bit of a mean reversion back to the true investments in traditional infrastructure that would obviously benefit someone like yourself?

I think it's timing. Well, first of all, cloud demand is growing. So that increase in CapEx is actually resulting in increasing demand for storage. So that's already a good news. But I would say the better news is, those big customers are preparing their infrastructure for the AI application and for other applications that will generate a lot of data. So we are, let's say, not at the front of this spending. We are more at the end of this spending when the application is actually working, is actually used by people, companies, government, when everyone will start using those applications, that is when data is created. Right now, the focus is on preparing the infrastructure so that companies and government and every one of us will be able to use AI, in a way that AI generates something that has a value, that for a person, can be a new video or for a company can be a new manufacturing flow, for a government can be a smart city. All those things will have a big value and will be stored and they will be stored in our disk.

Okay. Let's turn to pricing. A very hot conversation with investors today. We have seen some slowing of those deflationary pressures and pricing. You've taken pricing actions with your customers, obviously. If we just set aside the HAMR dynamic for a moment, how do we think about the sustainability of the pricing actions that you've been taking? And if you can kind of weave into that, both any push back -- any areas of pushback that you've been getting? And then obviously, build to order is something that you've talked about recently. How is that contributing, again, sustainability to this pricing dynamic?

Well, I would say, in theory, it's more difficult to raise price when demand is low than when demand is starting to grow. So I would say, we did a very good effort in the last several quarters to start increasing pricing when the demand was not so strong. Now demand is getting stronger, especially in the nearline space. So we want to continue to have some price increase quarter after quarter. What we have done, our strategy was not to have a major increase in one quarter but to increase a few points every quarter so that the impact to our customer is not huge and is fairly predictable. So we have done that for more than a year at this point. We were discussing before about the gross margin improvement. A big part of that is coming from that pricing action. So there is no reason why we should stop right now. I think demand is getting a little bit stronger. And I think now the industry is aligned in getting better profitability, now profitability is improving but is not at all at a level that it should be for a very high-tech industry like hard disk. So I think we will continue with our actions and the build to order is also very important. When we have more predictable volume, then we can better plan our manufacturing, we can get better results from the cost side of the equation. But still, this is not taking out the focus from the pricing action.

Right. And something you mentioned that I think is important. It's just, when you talk about kind of the level of industry profitability today, historically, you've talked about working with partners to make sure that this industry stays healthy. Would you characterize -- would you say we're entering a new period where market participants are acting more rationally, simply because of that profitability dynamic and wanting to make sure that things get healthier beyond here? Is that a fair way of characterizing the industry?

In my opinion, now the industry changed more than 2 years ago, I would say, even before this down cycle. The industry was focusing much more on profitability than try to gain a few points of market share. And now you can see that during the second part of 2020 or calendar 2021, you saw a major improvement in gross margin and profitability in general, not only for Seagate but in general for the industry. So there was already a change in the focus from the industry. Then the down cycle happened a little bit in a unexpected way, unexpected time. So that means the industry had inventory, and we had to move that inventory. So I think maybe for a couple of quarters, there was a little bit of impact on pricing but as soon as that extra inventory was depleted, the industry went back into a fairly consistent price increase and it started more than a year ago. And you can see improvement from Seagate and from our competitors in terms of profitability, fairly consistent quarter after quarter. Of course, even the cost actions are important to drive that improvement. But I would say the industry has changed the focus compared to what was maybe 5 years ago or before, before the major consolidation and maybe different, different opinion on what should be the right level of market share for the different companies.

Okay. Just to quickly kind of touch upon the ramp that you've alluded to and maybe related to production is, if we look back in time, I think it was 100 -- roughly 165 exabytes shipped quarterly, was kind of peak, peak production capacity or peak capacity. What kind of capacity do you have today from a production standpoint? And at what point is there where you have to start to requalify lines and hire more people? And how long does that take? And ultimately, I'm kind of getting at, is there any risk of kind of supply shortages as we look out, again, beyond just the quarter here.

Right now, we have about 125 exabytes that we could produce. The difference between the 165 and the 125 exabytes, there are 2 major differences. One is, there are some equipments that we have put off-line. We still have those equipments. So possibly, we can put them back online and increase capacity. But we also have -- unfortunately, we also reduced our workforce by maybe 15,000 people in the last 2 years. So if we need to go back to the prior level of capacity, we will need to hire not the same number of people but now few thousands and that takes time. So I would say we can go quickly to the 125 exabytes per quarter, above that will take little bit of time. It's still possible but it will take a little bit of time. Of course, depends also from the mix. So when I say 125 or 165 exabytes, then it depends exactly what we have to build, it could be a little bit higher or lower but that's more or less in that range. So again, it depend how the market will evolve based on our expectation. Now we said we expect a CAGR of about 25% for the next 3 to 5 years. So you can calculate more or less when we will be at full capacity with the current capacity available and when we could be at full capacity even going back to the prior level. I would say, so far, the cloud space is coming back fairly strongly but we still need to wait for the Asia part to recover before we eventually go into that discussion on when and how to increase capacity. I'd say for now, we have enough.

Okay. Okay. So let's shift to everybody's favorite topic, which is HAMR, obviously. You said publicly you expect to ship about 1 million units in the first half of '24, predominantly with 1 customer. We've talked about expecting to have the majority of U.S. cloud -- major cloud players qualified by calendar year-end. Can you just talk about where we are in that qualification process and then how long it takes to go from qualification to revenue generation and therefore, the visibility that you have into the HAMR ramp?

Yes. Usually, to qualify even today with a PMR drive, it takes at least 6 months. And today, we are actually running qualification in parallel between our last PMR drive, which is a 24 terabyte or 28 terabytes in the SMR version and our HAMR drive. So we are doing a lot of qualification that is taking a lot of our resources. So usually, it's 6 months, I would say, HAMR is a little bit longer. I would say in the longer term, the qualification time will be the same. But when you have a new technology, of course, we expect customers to take a bit longer to do an extra test and extra [ qual ] to be sure before starting to buy that level of units that you were talking about. So the focus in this calendar year is qualify a certain number of major customers in the cloud space, in the enterprise OEM space. But that is the focus, more than ramping high volume. The main focus is, let's create demand. To create demand, you need to have a certain number of big customers qualified. Otherwise, they cannot order even they would like to order but they can't. So that's the focus for the calendar '24. Calendar '25 is where we expect to ramp much higher volume of HAMR. And to that point, I think supply-demand, now let's see what happen with the supply-demand balance at that point. I think demand will be very strong. Let's see how much we can really ramp. That should help with the pricing, should help with the overall profitability of the specific product.

Okay. Let's talk about maybe the margin side or the price and cost side of HAMR drives because obviously, there's a clear TCO benefit for your customers. But how do we think about for you, the bill of materials of the HAMR drive, of a 3 terabyte per disk HAMR drive versus a low 20 terabyte, say, PMR drive. You've talked about leveraging the common platform, what incremental components, processes go into HAMR? And how does that -- how should we then think about the bill of materials and kind of the costs associated with HAMR relative to PMR? Like is the BOM, 10%, 20% more expensive? How should we think about that?

Yes. We have never quantify exactly the costs, because, of course, we don't want to say exactly what is the cost of an HAMR drive. But what we said is the bill of material is very similar, in the sense that a PMR drive 20 terabytes and HAMR drive 30 terabytes or even 40 terabytes, will have the same number of disk and the same number of heads. And the platform in general is very similar. Now Dave, I think, spent some time during last earning release to explain the commonality between our current 24-terabyte drive and HAMR, that HAMR is not only the 30 terabytes but will be the same for the 40 and the 50 terabytes. So there are some extra costs on HAMR, of course. One is the laser. We also announced at our last earning release that now we can produce our own laser that will be a good improvement from a cost side. And then the substrate for the media is still a glass substrate, similar to what we use today but a little bit more expensive. And those are the major differences. So overall, I would say the important for us is to go through the manufacturing learning curve so that we bring this product to a mature level fairly quickly. And when the yield are the same, the difference in cost will be not huge.

Right. And then just in the near term, in terms of extra cost maybe associated with yields, is there a way that you can help us think about what some of those extra costs might be near term that eventually flush out of this system as you ramp capacity and production.

No. As I said before, I think we will ramp high volume mainly in the calendar '25. So possibly this calendar year, the HAMR cost will still not be optimized just because the volume will not be huge. However, the impact to the financial statement of that not fully optimized drive is not huge because the volume is not enormous. So again, you will see a major benefit from HAMR, more and more we ramp, bigger will be the benefit. It will take some time. I think our profitability will increase and will improve every quarter but you will see a major improvement in the calendar '25 compared to the current calendar '24. And this will be true for everyone that will move to HAMR. They need to go through the curve. They need to go through the manufacturing learning curve. They need to go through the ramp. All those things have cost and -- but despite those costs that we already have. And as I said, we are also qualifying a CMR drive. So we are doing all those things in parallel. Despite all this, our margin is strongly improving. So I would say [indiscernible] This is why I'm confident, it will continue to improve, some because of pricing and some because those additional costs will go away.

And how do we think about pricing for HAMR drives? Meaning, again, we talked about the TCO benefit your customers, your customers get. How do we just think about maybe the deflationary pressure on price per exabyte in regards to HAMR because obviously, that can mean a lot for margins but maybe you share some of those profits with your customers. Obviously, you have early customers where you might give discounts. So just holistically, how should we be thinking about the pricing trends for HAMR, just standalone?

Well, my opinion is the TCO for customers is huge, even with no price decline. Now if you think about building a new data center and you can choose between a 20 terabyte or a 40 terabyte, your overall cost outside of storage is basically us. You can have the same exabyte enough of the space with half of the power consumption and half of the people. So there are a lot of costs. But even if the pure price per terabyte in terms of storage remains the same, customer will have a major, major cost reduction. Or if you have already a data center and you replace a 20-terabyte drive with 40-terabyte drive, your overall cost of the infrastructure remains the same. You don't need more people. You don't need more space. You don't need more power but we have doubled the exabytes. So your revenue will double from that location. So -- but it's an enormous opportunity for our customers even without considering a reduction in pricing. The reason why we want to qualify more customers is because we need to create a certain level of demand that is aligned to our supply. And that is the way that you keep the pricing strong as we have today with the CMR product. And as I said, it's not that we consider our pricing today as our high level, we think it's just we are in a phase of increasing pricing and we'll continue to do it.

Okay. Something that has been important, if we move kind of towards the cost side of this business is, you talked about reaching 30% gross margins with 20% less revenue versus the prior peak cycle. My questions are twofold. One is, how does HAMR impact that view? And the second is, how has your cost structure kind of structurally changed? How should we think about those 2 impacting the path back to 30% to 33% target range?

Yes. HAMR will be accretive to that target, will be accretive to any quarter where we ship a good volume of HAMR. So it will be our top product. This is why we want to use HAMR not only for the cloud space but in future, we want to use HAMR technology for all the enterprise OEM space and for the rest of the mass capacity. So we think basically every drive above 16 or 20 terabytes for us should be built using HAMR technology because that is a driver that give us the best opportunity to improve profitability. So when we gave that 30% to 33% range in terms of profitability that was not considering the impact of -- the positive impact of HAMR. So we need to now get some time so that we can qualify those new customers and we can ramp high volume but the high volume of HAMR will be accretive to that range that you mentioned.

Okay. And then just help us think through the relationship between margins and free cash flow conversion. Obviously, you've talked about improving margins. Is that how we should kind of think about the directionality and the ramp of free cash flow as well for you guys? Are there any more one-off actions that you need to take as we look forward? Help us walk through that.

No, we are very pleased with our free cash flow, how we manage free cash flow through the entire down cycle. Every quarter, we were able to generate positive free cash flow. I think this is a competitive advantage that we have demonstrated through the entire 7 quarters of down cycle. So we will continue to generate positive free cash flow. For us, dividend is extremely important. So we have protected our dividend through all the down cycle. And of course, we will continue to focus on the dividend. We also want to focus on reducing a little bit our debt level. So when free cash flow will be above the dividend level, we will probably use part of that free cash flow to reduce our debt. And when we arrive at the right level of debt, we will restart doing share buyback. So we have not changed at all our capital allocation strategy. Everything has to happen at the right time. So at the right time, we will go back and do exactly what we were doing before the down cycle, that was a very strong shareholder return.

And as we think about that reduction of debt, is there a certain target? Is it a absolute level of debt? Is it a certain leverage ratio that you're trying to get back to? How should we all think about that?

I would say the first step will be probably to reduce debt to about $5 billion. We entered the down cycle with $6.2 billion and that has been proven to be too high. And during the down cycle, we reduced that debt to $5.7 billion, possibly still a little bit too high. So I would like to go back down to about $5 billion. And of course, then depends at that point, how much higher revenue will be, how much stronger EBITDA will be, we can consider if that is the right level or if we want to be even more prudent and reduce it a little bit more. But I would say that, that first step is probably around $5 billion.

Okay. And with the remaining time we have, I want to kind of touch on comments you made earlier and kind of the long-term drivers of this industry. You mentioned -- we talked about AI and the impact on creating data eventually and the need to obviously invest in HDDs. Do you view AI as almost like a rising tide lifts all boats type of dynamic where there's a benefit to HDD, there's a benefit to flash? Or like the debate in the market is, you might need more warm and hot storage versus cold storage. How does Seagate think about that dynamic, again, when it comes to these kind of new future workloads?

Well, we think those new application are a benefit for everyone. Everyone that is in this space. So of course, it's a good benefit for companies that are producing GPUs, is a good benefit for companies that participate into the generative part of AI. The creation of something new, you don't need an hard disk, you need other components. You need flash, you need GPU, you need DRAM, you need many other components. That is to generate something new but to store what you have created you need hard disk. So there are different phases but it's a benefit for everyone. So we are in the second phase. We are in the storage phase, pure storage phase. So that will happen a little bit later and will be an incremental benefit to the recovery of the business that we are starting to see right now.

Right. Okay. Perfect. And then bigger picture, we talk a lot about areal density and the gains that you can get from that. You're just at the start of the 30-terabyte HAMR ramp. We've talked about getting up to 50 terabyte. Do you believe that the kind of gains that you can make in areal density continue to -- will continue to give HDDs this sustained advantage over flash and SSD in the data center? Or is there anything else that the disk drive industry needs to do as a whole to maybe fend off that potential disruption?

Absolutely. Now, if you look at the cost per terabyte on the hard disk side, especially if you have HAMR, you have huge opportunities because the bill of material between a 30 terabyte, a 40 terabyte, 50 terabyte and future even higher capacity is exactly the same. And I think with the time, we will be -- we will become even more efficient in manufacturing. But at least will be the same unit cost for a much higher capacity. So the cost per terabyte will decline a lot. I think on the NAND space, is more difficult because the majority of the cost reduction in NAND already happened when NAND moved from planar to vertical. And then now fourth generation was 32 layers and then 64 and then 96. The more you progress, the less is the benefit in term of more gigabyte per wafer, so in term of cost reduction. So the cost curve is very different. The NAND cost curve declined a lot, maybe 7, 8 years ago and now is slowing down. HAMR is just starting now. So it's different times, different cost curves. So I think hard disk will be even more competitive than what it has been in the last 3 or 4 years.

Perfect. That's a great place to end. So Gianluca, thank you very much for coming today.

Thank you.