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

Well, thank you, everyone, for joining us today at day 1 of Bank of America's tech conference for 2024. I'm Wamsi Mohan. I cover IT hardware and supply chain here at the bank. We're delighted to have Seagate with us today. From Seagate, we have Gianluca Romano, who is Seagate's CFO. He's been CFO since 2019. He's had a long distinguished career in finance spanning over 20 years. So we're delighted to have him here with us today. We also have, from IR, Shayne Hudson sitting here. If you got questions for a follow-up later at [ Rod Cooper ] as well. So feel free to grab their cards on the way out if you need to have any follow-up questions. But welcome, Gianluca. Thank you so much for doing this. Really appreciate it.

Thank you very much.

Well, so to dive right in, maybe let's start at the high level from a demand perspective and talk about what you're seeing on sort of cloud enterprise in China, which seem to be major drivers for your business.

Very good. Before we start, let me remind everyone that we'll be making forward-looking statements today, and you can learn more about the risk associated with those statements on our website. So demand is getting stronger, especially in the cloud segment. So if you look at our revenue progression since September, which was the bottom of the cycle, the improvement in December and March and what we have guided for the June quarter is coming mainly from the cloud segment. There are some improvements also in the enterprise OEM, some small improvement in China in general. But really, the majority of the recovery is coming from cloud. I think in another maybe couple of quarters, we will see the second step of the recovery that is actually probably driven by China and the big enterprise customers.

Okay. Great. So when you talk about a couple of quarters more for China recovery, are you seeing any green shoots over there that's giving you that confidence?

Yes, we have seen some improvement, especially in data and image application is still not very strong. So we think it's just the post part of the recovery. This is why I'm thinking we will see a much better situation in China, but also in worldwide enterprise OEM in a couple of quarters, so through the end of this calendar year. And that will drive, in my opinion, the second step of the upcycle. The first step is cloud. And of course, is already happening. It will get stronger in the next few quarters, but it's already happening. The second step in the recovery is actually from those other 2 segments that are very important to hard disk and to Seagate, in particular.

And when you talk about the strength that you've seen in cloud already manifesting, can you talk about what the underlying drivers of that are? Is it basically the willingness or wanting to migrate to higher capacities? Is it more about an attach with AI? What are some of the reasons why you're seeing the strength in cloud come back? And I know part of it is inventory as well.

Well, I think there are different drivers. The first one is more the macroeconomic situation. 1 year ago, 2 years ago, the cloud customers were managing their business a little bit differently and they were trying to raise utilization more than no increased the number of exabytes that they wanted to install. They have consumed a lot of the inventory that now they had. So we are now in a different trend. I would say AI is another big driver. As you know, those big customers, they focus mainly on creating the infrastructure for AI. And so being able to generate data. Until you don't generate data, we don't really need a device to do the storage of data. So the focus was on other components for 3, 4 quarters at least. Right now, the businesses are starting to use AI, and there is infrastructure will support that application. Of course, they need more storage, so it's just the beginning, in my opinion. But we have visibility that some of our drives are actually going into data center but they are being created for -- mainly for AI.

Okay. Excellent. I wanted to talk to you a little bit about exabyte growth going into sort of the next few quarters. How are you thinking about that in totality? And maybe if you could break that out into cloud and non-cloud.

Yes. We know -- made a few comments on the rest of the calendar year. We said we see revenue increasing sequentially through the June quarter that we guided and then September and December. We also said we expect [ '25 ] to be a good year for storage. I would say the majority of the revenue increase is driven by more volume. Of course, there is also the pricing impact that is positive to us, but the majority of the increase is, of course, coming from volumes. So you will see a sequential increase in exabyte that we sell every quarter.

And when you think about the magnitude of that, any way to quantify sort of what that sequential exabyte increase might look like?

Well, we don't give precise numbers also because those numbers can change a little bit during the quarter. But what we said is we have a certain capacity installed. Right now, we have opportunity to grow revenue using that capacity. We also said when you look at the end of this calendar year, we think supply will be fairly tight, and you could have a little bit of unbalance between supply and demand. If demand continues to grow, and if the China business and enterprise OEM business start to go into the real up cycle, I think probably capacity will be not enough for the demand.

What would you say then from a pricing trajectory perspective? It seems that already pricing is quite favorable, both at cloud customers and maybe it's going to start to pick up even more so for enterprise customers based on your comments. How should we think about the pricing trajectory here in the next couple of quarters?

Yes, we see a few good news in general. Pricing. We have started to increase pricing more than a year ago, even during the worst part of the down cycle. And we have been very consistent. Quarter after quarter, we increased pricing in almost every segment. Of course, every quarter is a bit different, where we target the increase, but we have been very consistent. We have not increased pricing in a way that is really strongly impacting our customers. We have always done a little bit of increase, not very high increases but very consistent quarter-over-quarter. And actually, we have told our customers what was our strategy, so they can put that in their plan and they have a much lower impact. So this is continuing. So I always receive the question, when are you stopping. There is no reason to stop, right, because we should continue into the next several quarters until we see this strong demand and actually demand eventually being higher than supply. So we will continue that. I think the industry is going in the same direction where a certain level of profitability is needed to manage correctly this business and to also support the big supply chain that there is in the hard disk. So this is one positive compared to a few quarters ago. Second is the mix. Of course, the mix is moving more and more into the mass capacity right now, mainly in the cloud. But as I said before, we expect improvement and volume increase and revenue increase also from the enterprise OEM and from the video and image application. And the cost. When you start increasing volume, the underutilization costs start to abate, and that is, of course, positive to our gross margin. We actually said in the last couple of quarters, that we were expecting to achieve that gross margin target range of 30% to 33% at a level that was way lower than the prior peak. And we indicated about a level of revenue of $2.2 billion. I would say, looking at what I can see in the quarter right now, I think we could be in the range already in this quarter. So it's a major improvement compared to even our internal expectations in terms of the speed of this report. And of course, with a better gross margin, we also see an EPS that is higher than what we guided. I think we'll be fairly close to the top of the range. We guided $0.70 plus or minus $0.20. So I assume we'll be fairly close to the top of the guidance region. So this is good, of course, for the short term, but is, I think, even better for the long term because revenue will continue to grow. And of course, we think that will help continue to improve in profitability and, of course, the bottom line that is always a focus.

So if we just step back for a second and think about that trajectory based on your comments around pricing and already achieving gross margins in that range and heading into a more even supply-constrained environment in the back half of the year, how high can this gross margin really go? Because you've got just -- you're talking about being in the range of almost $1 billion lower than your peak revenue that you saw a few years ago. So as China recovers and maybe gets closer to $1 billion a quarter, what are -- and you get enterprise pickup plus cloud sustains, can we imagine a world where we're sitting at 36%, 37%, 38% gross margins?

Well, we don't guide gross margin. So far in time, I would say, we had a target range. And of course, this was not including the positive impact of HAMR. So we are -- now we think we are back into that range, as you said, at a much, much lower level of revenue. Revenue is increasing. So this will help on the cost side. As I said before, our strategy on the pricing is not changing. So we will continue to be very consistent with some increases every quarter. And the mix. Of course, every quarter is a bit different, but the mix is clearly moving more into mass capacity and less on the legacy. So all those things together would -- will actually improve our profitability. We can see that fairly clearly through the end of the calendar year.

Okay. That's great. Maybe just to follow up on your comment about HAMR, right? So I think you made some comments recently talking about some of the qualifications that your lead customer getting pushed out into the September quarter. Can you just provide some more color around that? What are some of the issues that is creating the shift in qualification? And when we think about that in the context of what you just said about the current quarter, how much of that upside is the lack of HAMR maybe in the quarter versus other drivers of upside in the quarter?

I would say the change in the HAMR for the quarter is very immaterial because we were not supposed to shift a high volume of HAMR units. So when you ship in a quarter, I don't know, 12 million units, 200,000, 300,000 of one specific product is not really changing your P&L. So I would say HAMR is not really impact in the short term. I would say from a revenue perspective, we can replace EMR with PMR drives. So we don't think we are losing any revenues there. We just replace HAMR with a different product. Of course, HAMR has higher capacity. Or a gross margin is, of course, positive for Seagate. So we will have that impact a little bit later in the year compared to what we were expecting because qualification has shifted from the June quarter into the September quarter. I would say the good news is we believe we have learned enough from this qualification process that we can start the majority of our cloud customers qualification in the September quarter. So we will not wait any longer. Starting next quarter, we will have the majority of U.S. cloud and China cloud starting the quarter. We think maybe the [ court ] could be a little bit longer than a normal PMR drive, looking at what has been the recent qualification process. So last week, I said, probably good to estimate about 3 quarters to go through that court. And you will see a major ramp in volume of HAMR, probably around the mid of calendar '25. And at that point, we will have several cloud customer qualified, some also non-cloud customers qualified. And therefore, the pricing dynamics will be different. Demand will be very strong for a product that has the highest capacity in the industry. So we are just shifting a little bit to the benefit of HAMR out in time.

Can you talk about what are some of the underlying things, the learnings that you're finding and what were some of the constraints or the changes that you had to make?

So we had one major issue in the March quarter due to a component. We now have several vendor for that component. So we just switch to a different vendor, and we restarted the process. Unfortunately, that took out another 2 or 3 months. But I am glad to say that was the result that we have seen until now has been very positive. So that issue has been solved. We have test internally and the customer has tested externally. So now, the results are very consistent that, that component issue has been solved. Now during that period of time, we also look at our internal manufacturing processes. This is actually to improve our internal lead, that is back to your gross margin comment, how can you make this product even more profitable. Well, one point is to make it more efficient into manufacturing. So we went through some change in manufacturing. The customers also require the change into the firmware. So we are doing those 2 changes. And the court will be pushed from the June quarter to the September quarter. But for us, considering where we are from a competitive landscape of pushing out 2 or 3 months compared to the benefit that we can get is a no-brainer. So I don't know, we just push that out and we will get the full benefit later.

Would you say that -- I mean, I know you just spoke about sort of the issues that were in the March quarter, those were mechanical component issues that you had called out on the call. Can you just clarify if there are any other sort of more recording related pieces of the drive that are still maybe a little bit in question around the qualification?

No. I would say no. That component defect created a contamination to the drive, of course, that was impacting other components. But when we replace that defective components, the problems did not show up again. So we are very confident that, that problem is not going to show in the next batch of court.

Got it. And then when you originally thought about HAMR maybe last year, you thought this first half, you would be shipping about 1 million units or so. Now we're maybe shipping a couple of hundred thousand. If we think about the full year this year, is it reasonable to think that we're somewhere in that 1 million unit range then for 2 quarters of shipments?

But I would say, basically, everything has been shifted out in time. So what we were expecting to ship in March and then in June is basically going into the September and December quarter. So of course, we will be more precise later. But yes, there is a little bit of shift in the time of the quarter than in the time of the revenue related to HAMR. But as I said before, it's not really a major impact from a revenue point of view because we can replace with PMR drives. And again, the volume was not huge enough to really change the P&L structure, and we will get the benefit later.

And can you just talk about sort of gross margins in the context of HAMR, too, because you said it's accretive. When you say it's accretive, is it accretive to mass capacity? Is it accretive to overall company? And at what level would you say -- I mean, when you said next year, we'll be shipping substantial volumes of HAMR by mid next year. How should we think about that in terms of mass capacity drives for the industry of, call it, 55 million, you guys are shipping close to 20 million drives or so. What percent of that would you consider a substantial when you think about the ramp of HAMR and the margin overlay with that? Sorry, too many questions in one.

Well, in terms of the gross margin, we look mainly at a cost per terabyte. And then, of course, the gross margin will depend from how we price the drive and how strong is the demand in that specific quarter and know how many customers are qualified. Of course, in the first part of the ramp, we have one main customer that is qualified. So the dynamic is very different when you have 1 and when you have 4, 5, 6, 7. So we have more customers. As I said before, we need to wait to go through the other court, and this is why we will ramp high volume in the mid of '25 when we have a lot of customers qualified. Ramping a lot right now with just one customer from a financial standpoint, of course, is not the best thing to do. So I think the benefit of HAMR, you will see the benefit of HAMR sequentially higher and higher. Most of all it's because it's higher volume; and second, because you have more customers and with more customers, more demand for the same product, of course, will have a positive impact to the pricing. So no, we don't specify what is our objective for HAMR gross margin, but we always said gross margin will be accretive to the company. And of course, because it's the highest drive, we'll have the lowest cost per terabyte, and then we will see exactly how well we can price it. We think we can price very well, but let's see when we are there.

Okay. Great. And I wanted to just go back to exabyte growth for a second because in the March quarter, we saw some of the largest differential between you and your primary competitor on exabyte growth. And I'm just kind of curious what were some of the dynamics that kind of created that? And how do you see that maybe divergence close or sustain? Or how do you think about it?

I would say, of course, we don't have the same market share with all the customers. So it depends which customers are ramping in a specific quarter. You can see a little bit of change in overall market share. I would also say that no, we were very focused on HAMR and that possibly resulted in a little bit of Seagate being laid to market with a 28-terabyte SMR. Today, we are qualified with multiple customers. So I think this -- you will see the realignment starting probably next quarter. But no, we could be a little bit late by maybe a quarter on that product. And of course, there was a few issues we had on HAMR court to have a little bit of focus on the short term because short term is not our priority, but we are ready with the same product with a lot of customers. So I think starting in the September quarter, we will see a good realignment. And of course, now depends then from customers. So who is ramping, who is refurbishing data center could drive some changes between the -- from the 2 main competitors.

How would you say, as you're working on one primary customer right now for qualifications, but as you qualify more cloud customers, how do you think the appetite for getting the highest capacity drive is within your cloud customers? And what is driving that underlying behavior? Because you just said not having the highest capacity drive created a fairly different like exabyte growth rate. So if you have the highest capacity drive, should we expect the inverse to happen where you really have significantly higher exabyte growth?

Absolutely. No, even if Seagate focus is not market share, it's more profitability. But when you have the product that has the best performance in this case, the best capacity, you see some change in market share. This is normal. Now for customers, capacity is extremely important. Think about if you already have a data center and in this data center, you have this drive with 20 terabyte. So you have a cost for that location and the cost is not only the hard disk, but is the building, is the power, is now the entire infrastructure component and the labor cost, people that are there. So that location has a cost. And of course, based on how many exabytes you have in that location, you have a certain revenue. Now if you can change the hard disk and you go from 20 terabyte to 40 terabytes, we have the same cost for the location, the same labor cost, the same rent cost, the same power cost, the same cost of the infrastructure, but you can double the exabyte in the location. So your revenue can be double. So it's not only -- when we say TCO, is not really a cost per terabyte that our customers are just looking at. Of course, this is also important. But they can also -- with the same cost, they can generate much more revenue. That is a huge benefit for them for the location where they already have. And of course, when you build a new location, you want to get as many exabytes you can for that cost of the new location. So capacity is extremely important. Of course, in the data center, there are many components that are important and in certain cases, like we have seen in the last 1.5 years, there was a big focus on certain components that need to be used for AI application. So that was a priority for a while. But finally, our customers are selling storage. So they need to optimize the storage, the cost of the storage and the revenue that they can get from storage. So I think capacity is very, very important. And that's why I think not only with AI, but there are many applications that will continue to generate an increase in data creation like autonomous driving, but is still very limited and the Internet of Things and smart cities, machine learning, all those applications will continue to generate more data. And that data has to be stored in a hard disk. And that will generate revenue for our customers.

Okay. That's great. Well, if I could just ask you around the TCO comment that you made, right, like there are certain all-flash companies that kind of talk about, oh, the TCO on all flash is so much more superior, hence, the HDD industry is going to go away. Well, clearly, we don't subscribe to that view. But the question I have is what have your customers told you about the relative TCO when it comes to using all-flash-based array versus HDD-based array?

Well, I would say if you look at the last couple of years during the down cycle, the NAND cost and the all-flash array costs went down a lot, more than 50% lower. And if you go into a data center, you will see no change on how they are managing their storage. 90% of the storage is on a hard disk, 10% is on NAND. The reason why it's 90-10 is because the storage is on hard disk, but when they need to use the data for certain applications, they move the data from hard disk to NAND. They do the compute part and the analytics part and then the saving back. So despite the huge decline in cost of NAND, we didn't see any change. And now that NAND price is going up, I know, 10%, 15%, 20% quarter-over-quarter, we have not seen any change any. So positive were negative, we don't see the change. The reality is the architecture of the data center is already optimized between hard disk and NAND. They use as 2 different components, not as a 2 storage component, but as one storage component that is hard disk and another component that is used for the utilization of the application, [ petty ] as SSD. So we have not seen any change from the delta price or data cost of NAND. I think in future with HAMR, the cost per terabyte will continue to decline. The NAND cost will also continue to decline. But of course, the impact of moving from [ planner ] to vertical is now much, much lower. Before, you could double the gigabyte per wafer, so reduce a lot the cost per gigabyte. Right now, when you go from 128 to 168 layers, you don't double anymore. And more and more, you go up in layers, the impact to your gigabyte per wafer and then the cost per gigabyte is lower. HAMR is just as a beginning. So going from 30 terabyte to 40 terabyte to 50 terabytes will strongly decline the cost per terabyte. But independently from the cost, we have not seen -- even if some positive from this down cycle is that we had the opportunity to test this theory that the data center are using the components, the 2 components differently. And I think this has been proved.

Yes. No, that's a great point. We don't have a lot more time, unfortunately, but I did want to ask you 2 more questions, if we could. One is just your primary competitors talking about separating out at HDD and NAND businesses. Do you think that has an impact to profitability of the industry? Or do you see any big change that might be material to the industry because of this split?

In general, I don't expect any change. I think today, the NAND business as they are this business are managed in 2 business units, 2 separate business units. So I think they are fairly independent, one to the other. So the fact that they are in 2 different companies should not create any difference. Of course, finally depend on people that are managing the business because they decide the strategy. But again, I don't see any reason why the strategy tomorrow should be different than today.

Okay. Great. And again, like I think it'd be irresponsible to not ask you a free cash flow question. So maybe if you could comment on sort of what is the annual free cash flow generation that you see for Seagate, if you think about quarterly revenue run rates going from $2 billion to $3 billion over time, like what does that do to your free cash flow generation?

Yes. We don't really guide free cash flow. But even during the worst part of the down cycle, we were always able to generate positive cash. So I would say that is our bottom free cash flow. I think when revenues start to go up, profitability is better. Of course, bottom line is stronger, you will see an improvement in free cash flow sequentially. We'll follow basically the business. Of course, it's always working capital from one quarter to the other or the CapEx. But in general, you will see a similar increase in free cash flow, as you can see in the business. The CapEx now for the time being, we have been, I would say, fairly good in managing the last couple of years. Our target has not changed. So we are still targeting between 4% and 6% of revenue for our CapEx. So again, that will not be a negative impact to our free cash flow.

Great. Well, we're just about out of time. So appreciate everyone coming here. And thank you, Gianluca, so much for all the time you spent and for the real update you've given us.

Thank you very much.

Thank you.