Kick us off now, and the clock's coming now, but welcome back, everyone. I'm Tom O'Malley, Semi and Semi-Cap analyst here at Barclays. A pleasure to have Gianluca Romano here, CFO at Seagate. Thank you very much for joining us.
Thank you for inviting us, and before we start, I want everyone to know that we will make forward-looking statements today, and you can learn more about the risk associated with those statements on our website.
Perfect. All right. So why don't we just jump right into it? So, you were at a conference last week. You gave some updates on both the HAMR front and then also some updates on the March quarter. Maybe just remind those who are here in case they didn't see kind of what that update was.
Yeah. I would say the very good news for us is that we qualified HAMR in the cloud space with the customer that, you know, we have worked for a fairly long time to find the right configuration. And finally, we found the right configuration for the cloud space. As you know, we qualified earlier different segments like the Video and Image Applications, the on-prem data center. But finally, we were also able to qualify HAMR technology in the cloud. And so from now on, I think, the possible doubts that people could have on the technology not being able to be a good technology for the cloud space are possibly gone. The second part of the ask refers to a lower manufacturing volume that we will have available in the March quarter.
So, as you know, we have increased our volume because demand was very strong, especially in the Nearline, not adding new capacity but putting back in place the capacity we had before the down cycle, but we had not many equipment that we took out from manufacturing, and then we started to pull them back, and because demand was very strong, we were also trying to anticipate our WIP as much as possible in order to go to finished goods and then being sold. Now, the cycle time for our disk is. It's fairly long and complex.
So we tried to accelerate. A year ago, we were selling like 55 exabytes in the Nearline space. Last quarter, we sold 110. So it's not easy to continue to grow and, you know, get the, the WIP into finished goods and then out in time to meet demand. So in this process, some of the equipment took a little bit longer than we were considering. And so basically, we have a little bit of lower build WIP at the beginning of our manufacturing line. So this WIP, this lower WIP, will end up in finished goods and then out in the March quarter. So there is no impact in the current quarter, but there will be no impact in future quarters. But the March quarter will have about, you know, up to $200 million of impact.
Got it. Helpful. Let me walk through both of those. The first one is the qualification of large customer. You guys have been asked on every earnings call for the past year about when this is gonna occur. There were some issues to start. Is the qualification just the end of a trial of that product and then saying, t his works. We're gonna buy more? And would you expect the kind of linear ramp that you had originally kind of expected? It was.
Yes.
Coming up pretty quickly?
Yeah. Absolutely. So this configuration is the one that we already shipped to other customers in the cloud segments to now start the qualification, and we will qualify, you know, several customers in U.S. and outside the U.S. between now and mid of calendar 2025, so mainly in the next couple of quarters, and, you know, we said last week we will increase HAMR volume much more, you know, starting that point in time. Right now, we will be serving one major cloud customer and few other customers in different segments, but, you know, the volume in those segments is obviously much smaller, but starting, you know, mid of calendar 2025, we will have, you know, a certain number of customers qualified that are high-volume customers, so we will increase the production. We'll increase the ramp.
And, you know, I think it's just a matter of time and a matter of how much we can ramp. But especially in the cloud space, you will see, at least from Seagate, the vast majority of customers moving into HAMR for possibly the majority of the entire of their volume, in the future.
Does the qualification of the large customer help in any way of the subsequent qualification of the other hyperscalers? I'm sure each have individual things they need to work through in order to get that qualification. But can you learn from customer A and say, hey, here are the issues with B and C, and that may help that ramp? Whereas, like, into next year, you could see those customers maybe come on a little bit faster.
Yeah. I would say this configuration is, is really, really strong. And this is why we started to ship those products to our other customers even before getting the final call from the, let's say, leading cloud customer. In our internal test, that configuration was very, very good. And so we didn't even need to wait for, you know, the end of the call with the customer. We just start shipping to the other customers. So, the first customer that we fully qualified is not an easy one. And it's good. It's good. No, now, now we know we have a very strong product. And, as you said, every customer is a little bit different, but the configuration will not change. And we will use that to qualify the other customers.
Okay. And then shifting to the second part of the initial commentary just on the $200 million. So you spent some time on the last earnings call talking about moving to that 160 exabytes point where you've been very moderate about how you're gonna service the market. There's clearly a dynamic that's working for you, and we'll get to pricing in a little bit. But just can you describe to me you're bumping up against that point, the capacity point, and you're saying, hey, it's, it's work in process moving to finished good? Like, what changed in terms of when you were at earnings today? Is it just the product not moving as quickly or not being able to get it to the customer?
Yeah. It's just the WIP. You know, the movement of the WIP. You know, we have, as I said before, we have accelerated as much as we could the WIP to match demand. And you know, counting on those equipment to build more product and not creating a reduction in the WIP. Unfortunately, those equipment took a little bit longer. You know, when you take equipment out of manufacturing, it's not so easy to pull them back. You need to recalibrate. You need to requalify. Sometimes you need to requalify the entire line. It took a little bit longer than what we were expecting. And so for a certain period of time, we didn't have full production. So we created a little bit of lower WIP in a certain part of the manufacturing.
That WIP will end up in finished goods and then out in the March quarter.
I think something that you did that was really interesting during this period of time too is obviously there was some restructuring of the organization. You're getting to production levels that you were at prior with a different manufacturing base in terms of headcount. Is there a view that you need to in order to prevent, like, this happening in the future to continue to scale a bit more on that side? Or do you feel comfortable kind of at the levels where it's more equipment that?
No.
Okay.
We don't need new capacity. You know, I think it's good, and it's very financially beneficial for us to put back equipment that we already have. So this is what we are doing with those equipments and with less people than we had before the down cycle. We will be in that range of between 150 and 160 exabytes per quarter. Starting from that, the increase in exabyte will happen through mix. So moving from PMR.
Mm-hmm.
Into HAMR. Now, PMR today is, let's say, 24 terabytes when you go in the high-capacity CMR product. HAMR started 30. So the more we move units into HAMR, and more HAMR will grow from 30 to 32 to 34, 36, 40, etc., the more we can produce exabyte with the same heads and media because finally we produce heads and media. So we don't think we need to produce more heads and media. We just need to produce more of the HAMR mix. So with the same manufacturing footprint and a similar cost, we get many more exabytes.
Perfect. Do you wanna touch on that? But just to close the loop on the $200 million, obviously the business is very heavily focused Nearline now in terms of exabytes. If you look at where that impact is coming, is that on the Nearline cloud side, Nearline enterprise side, VIA side, Legacy side? Or should we think about it as the exposure of the cut is effectively pro forma of what the businesses are in your total?
Yeah. Those are common equipment, so will impact pro forma, basically all the segments. I would say the equipment that are only related to HAMR have not been impacted because we never took them out of manufacturing. But also, HAMR used some of the common equipment. But, you know, that's more or less impacting proportionally of the business.
Super helpful. And then in terms of the profitability impact of the $200 million, clearly there is absorption overhead that is impacted when you have a lower top line. But anything else that we should be thinking about below the top line, that would impact the gross margins which have been stellar over the last couple of quarters?
Yeah. I don't expect that to happen, going from December. And December will be better than September. And, you know, we have done this for several quarters. I think, now in March, of course, we will have a little bit of an impact from the lower volume, but we will have higher HAMR volume compared to December. And this is, of course, positive. We will have possibly more Mass Capacity mix because in the March quarter, usually you have lower seasonality in the Legacy part. Unfortunately, also in the Video and Image Applications, that is a good gross margin segment, but, you know, much more impactful in the Legacy. So the mix will help. HAMR volume will help.
So, I'm still positive that gross margin, despite the lower level in production and volume sold, I think gross margin can still improve in the March quarter.
Super helpful. I wanna touch now on the point that you were briefly moving towards before is just the mix of HAMR and your product portfolio into 2025. Do you think that your capacity point at HAMR when you first start shipping you obviously have a couple different customers is going to be at that 30, that 30 mark? Or will you see, you know, cutdown drives using the HAMR technology? So you'll see a blended average similar to, like, your PMR portfolio which is in the 20s right now? Or do you think it'll be skewing a little higher just because of the hyperscale customers?
Yeah. At the beginning, we'll be mainly focused on the cloud customers, so the highest capacity we can produce.
Yep.
Now we start with the 3 terabyte per disk, 10 disk. So it's 30 terabytes. But for this product, now when we go, you know, in the following quarter, you will see part of the volume that we ship being even higher capacity. So the 30 terabyte will become a 32, a 34, a 36 before we go to the new product that will be based on 4 terabytes per disk or start at 40 terabytes. So it's always a progression. Now, our areal density improvement is happening between the 3 terabyte and the 4 terabytes.
Mm-hmm.
So you will see in the next few quarters, you know, what we sell today at three terabyte per disk will be 3.1, 3.2, 3.3. So there is always R&D improvement in the current product. So we have opportunity to scale up also the volume. Right now, the focus is in cloud because they consume the, you know, higher volume. And they are the ones that are very interested in, in buying high-capacity drives. High-capacity drives is where we get a better profitability in general. So, of course, the focus is there. Then when you go in future, let's say after we have the four terabytes per disk and we have ramped enough volume for the cloud, then we can start using HAMR technology for lower capacity drives.
So once we have a 40 terabyte drive based on 10 disks and 20 heads, we can start looking at 20 terabyte drives with 5 disks and 10 heads. So financially, it will also be, I think, you know, a, a good outcome. But we need to have enough volume. So we'll not be in calendar 2025. Possibly it will be a little bit later.
Yeah. Perfect segue. So, you, you've spent time trying to help investors size the HAMR impact through kind of calendar year 2024. Are you willing to kind of think about when you may see HAMR crossover in calendar year 2025 or give a guess at the size maybe into next year?
Yeah. We don't want to give too many information at this point. I think it was important at the beginning to drive that confidence that HAMR is a product that can work in any segment. And, you know, as I said, the cloud customers want to have the highest capacity, so they will move more and more into HAMR depending also on when is the time of their call and what is the volume that we produce that is available for them. But I, I would say if you go out in time, you know, maybe two years from now, I don't see why a cloud customer should not try to move all the volume into the HAMR high-capacity drive. Also because, you know, two years from now, we will not be selling 30 terabyte drives. We'll be selling maybe 40 terabyte drives and, you know, higher capacity.
So the benefit for them to move into that new technology will be, you know, bigger and bigger.
Yep. Can you talk about the pricing side of the business? So I think that something that was surprising to me at first was just how powerful it was for you guys to moderate some of the CapEx and then say to customers, "Loo k, we're providing you increasing value year over year. You should be paying for that." And it, it really showed up right away in the financials. How much left do you have in your ability to go to customers and say, "Look, we're gonna continue to charge a higher price?" And how much of that is dependent on your ability to offer higher capacity drives over and over again?
You know, the cloud demand, the Nearline demand, but in particular the cloud demand is still very, very strong. I would say it's probably not more difficult right now to have that pricing discussion that it was five quarters ago where they were buying 55 exabytes.
Mm-hmm.
So now they're buying much more. So the capacity, as you see, you know, when we have a little bit of an impact in our manufacturing, we lose volume. We lose real revenue because now, you know, they want all what we can produce. So I don't think we are, you know, near the end of this. It's, I think it's a very good strategy because we were not very aggressive with the price increase. You know, we increased our price in a way that was predictable, was not very impactful because it was not, you know, a 10%, 15% price increase every quarter. It was a, you know, a fairly small percentage but very consistent every quarter or every time they negotiate a new contract.
Of course, if you negotiate a contract that is a little bit longer in time, you know, you fix that price for a while. But at the end of the contract, when you renegotiate, you know that you will have a price increase. But it's not something that I think is disruptive for those huge customers. And it's very, you know, beneficial to us. It's not too disruptive for them. I think it's a good solution. We produce what they need. They need to tell us what they need. And there is a price. If they want more, they need to pay a little bit more. And that is possibly, you know, one of the impact I can see right now for the March quarter. Now, losing that volume means losing basically the upside. And that upside is usually at a very good price.
But as I said before, this time this is just one quarter impact. I still think profitability will be able to improve. So it could have been better for sure, but it will still be, I think, a fairly good quarter.
Yeah. The $200 million had went in potentially gross margins up. That's a pretty good operating model that you're able to overcome that. So I guess if I was to look at this next year and look at the different pricing dynamics as well, I think volume is also a big factor. So you've talked about bumping up against this 160 exabyte point. So I think two things that I think about often, which is there's clearly an inflection point at some point coming here for the drive business with AI. And then there's also your ability to try to size that market and offer your customers a certain amount where it's not overflooding the market, not spending too much capacity. So too much CapEx, excuse me.
So when you look at kind of 2025, is this the year that we're gonna start to see an inflection in HDDs, in AI? Is that at the enterprise? Is that at the cloud? Just maybe your initial thoughts on where that's headed and how do you adjust your capacity to match that potential demand driver?
Yeah. As I said before, we are not planning to increase units. So this is also another incentive for customers to move to the HAMR high-capacity drives because they can buy more exabytes with the same number of units. Now, if they stay on.
Yeah.
24-terabyte CMR or 28-terabyte SMR with the same number of units, they cover much less of their exabyte need. So if they move to HAMR because HAMR, you know, I, I know you know, but HAMR has both CMR and SMR. So depending what, what is their infrastructure, you know, they can get a much higher level of exabyte. And, so I think this is the reason why they are so interested in, you know, going with the call and, and then moving up in capacity. AI is very important to hard disk. It's one of the applications that, you know, is growing very, very quickly. In the hard disk, I see two benefits from AI. The first one is retention period. You know, any, you know, company or individual or government entity, they all, all have rules in term of retention.
Now, when you start using AI, you want to have more data because AI gives you a better result if you can access more data. So their retention period is getting longer. And that, of course, is very good for hard disk. The second part that is very good and will be very good more in the future, retention we can already see right now. The second part I think we will start to see in the next several quarters. But the more and more people and businesses are learning how to use AI, the more they get valuable data and valuable result from AI. That data has to be stored and is getting stored into the hard disk. So the first part is very clear, already happening. Second part takes time. You know, we all tend to think that things happen very quickly.
It takes a little bit longer. And so there is a certain quantity of data generated by AI that gets stored. But I think it's still very small compared to what will be in, you know, six months from now, a year from now. The more and more we all use AI and big companies use AI, the more data will be created. And that data is very valuable and will get stored.
I look at the Nearline side. Cloud has been at least slightly more dominant than enterprise in that business in terms of mix. If you look at AI, is there one area that you think you'd start to see that penetration first? I mean, clearly in cloud, that's where you've made your bread and butter in the Nearline business, but enterprise, there's a thought that you would see more inferencing potentially at the enterprise. Is there a way you would think about where you would see that grow?
Yeah. I would say for us, it's not so important where the data is stored because we serve on-prem data center or, you know, the big public cloud. The important is that the application is used and the application generates a lot of valuable data so we can store it in our hard disk. We don't need a special hard disk for AI. So for us, it's kind of business as usual, but there is an application that generates data that, you know, before was not existent. So that is, of course, an upside to the data growth in the world and, of course, in the data storage. Now, you see, there are different phases of AI, and that maybe can create some confusion. But hard disk is very active at the beginning of the application because hard disk has all the data.
When you generate the new data, hard disk is not very active. So you move the data that you need into a different infrastructure where you have the DRAM, the HBM, the NAND, the GPU, the CPU. In that environment, you run AI and you generate data. When you generate data and that data is valuable to you, that data goes back to the hard disk. So if you just look at the AI infrastructure where the new data is generated, you say, "Oh, but it's not a lot of hard disk." Exactly. We are not there. We are before and we are after because we are data storage. We don't generate new data, but we are data storage. So for us, it's very important that this new infrastructure where we are not much present is actually working very well.
We are happy when DC is developing because this is where the new data is generated and then coming back to the hard disk.
So when you hear the flash player say, oh, this is a new domain, flash is taking share in the data center, you say, hey, let's look at the stage we are in the AI cycle right now. In a year, two years, that's where you're gonna see more hard disk drives.
Yeah. I would say, depends where you look. You know, if you just look in this environment where the new data is generated, you don't see hard disk or very little hard disk. No, you don't need it because you transfer the data from hard disk into this environment that needs a different performance. But then the data is going back. So we are if you look at the entire space of that, you know, where AI data ends up, of course, hard disk is one of the top components. If you just look at one part of the infrastructure when the data is actually generated, we are really not there. But we don't need to be there. We want to be data storage. So we want to have all the data before, and we want to have all the data at the end.
Super helpful. All right. Pivoting to VIA, you talked on the last call about some China stimulus potentially helping in that area of the business. Could you remind me, have you disclosed kind of how big China is as a portion of that business? And then, in your expectations, you kind of mentioned some better seasonality, which helps gross margin given that business into March. What are your expectations? Does that business still able to grow kind of in the future? Obviously, you have a Legacy business that's declining. You have a Nearline business that's growing quite nicely. But within that, where's VIA kind of moving?
Yeah. I would say VIA Image Application was fairly good last June quarter, better than what we were expecting. So it was part of the upside for the quarter. So we were a little bit worried about September and, and then December. But actually, you know, the, the September was fairly flat, maybe slightly up, but fairly flat. And December, I think, will be another, you know, good quarter. So, for sure, it has improved, since the down cycle. Unfortunately, March is, seasonally down also for VIA, as we were discussing before. But after that, I think, we'll, we'll restart growing. It's an important part of our business. I would say the majority of our customers are in China, but they manage all the smart city projects around the world, you know, including U.S., by the way. So it's not so important, you know, where the customers is.
For us, important is how many smart cities projects are approved and supported by the, you know, the local authorities, etc. So it's a good segment. Of course, we don't expect that to grow like the cloud, but we'll have sequential growth in the next few years.
Helpful. Okay. Turning to some CFO questions here. So your gross margin's hit 33.3%, which is highest gross margins in a decade. You have been talking about that for the last couple of years moving to this point. So we've gotten to the point where everyone's been bugging you about.
Okay.
Now everyone says we're gonna go from here, right? The price of success. Can you talk about we talked pricing. That's one factor in helping gross margins get better through calendar year 2025. What other levers can you pull? You have the HAMR mix. Anything else that we should be thinking about, and maybe an initial view of where you think things can go in a reasonable year?
What I would say is the short answer is we are going higher. You know, we have discussed about the pricing environment that is not changing from my point of view. So we will continue to implement the same strategy. We will have the positive impact of HAMR. We will have, in general, the positive impact of mix moving more and more into Mass Capacity out of Legacy. So all the costs, of course. So I would say putting all things together, I continue to see improvement sequentially quarter after quarter, mainly because of the mix and because I know there are every quarter some LTAs that expires and need to be renewed. And as I said before, that renewal usually comes with a little bit of price increase.
There is no reason for us to stop the strategy that has worked very well in the last, you know, five or six quarters. Until the supply-demand environment is different from the current situation, we will apply the same strategy.
Okay. In terms of capital returns, again, a period of time where people were questioning the dividend side and we've worked our way through that, and now people are asking on the buybacks. Can you talk about your framework on buybacks when you would see those increase? Obviously, it sounds like the operating model's in a really strong place. You now have HAMR qualified. The view is that structurally, you have some really good drivers at the calendar year 2025. What are you thinking about the buyback?
Yeah. In general, for sure, all the return, you know, we focus on two drivers. One is dividend. And, you know, in the up cycle, we tend to increase dividends every year. And in the down cycle, we tend to protect our dividend. So we have done that during the down cycle. And when we came out, you know, last October, we announced an increase in dividend. So this is our stable part of the shareholder return. And then there is the other part that is, of course, share buyback. We have done a lot of share buyback in the past, at a very good share price. Right now, we are focusing on the debt reduction. So we still have $5.7 billion of debt. We want to go to about five. We have a fairly good, fairly big maturity in January, about $500 million.
So we will repay that, and then we will do a little bit more in the following quarter. But once we have achieved that level of debt, we will, we will go and start again the share buyback.
It sounds like you've got a lot of great drivers heading into calendar year 2025. Thank you, as always, for being here, Gianluca, and good luck in this coming year.
Thank you very much. And thank you again for inviting me.
Yeah. Be well.
Thank you.