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

Okay. With that, let's get started. Good morning, everybody. My name is Jim Schneider. Welcome to the Communacopia and Tech Conference. Jim Schneider, Senior Equity Analyst at Goldman Sachs. And it's my pleasure to welcome Seagate and CFO, Gianluca Romano to the stage with us today. Thank you for being here.

Thank you, Jim.

Now Gianluca, just start off with a very high-level industry question. The industry has done a great job of structurally improving itself over the last 18 months or so in terms of higher utilization, big recovery in terms of pricing and so on. So maybe just give us your high-level view of what are the key elements of what's driven that improvement in the industry situation from a supply demand perspective recently.

Okay. And before we start, let me remind everyone that I will be making forward-looking statements today, and you can learn more about the risk associated with those statements on our website. Yes, the industry has improved a lot, I would say, over the last more than 2 years. As you remember, during the COVID pandemic, many companies were moving data from on-prem into cloud. And that was, of course, bringing the cloud consumption fairly high during those couple of years. And then the COVID situation improved, but that also meant a little bit of reduction of the growth of the cloud. And so we enter into a little bit of a downcycle. And a little bit about 2 years ago, maybe 9 quarters ago, the cycle started to turn. At this point, we didn't have any more inventory in our WIP and we start to have a different discussion with our customers. We told them we need to have more visibility. We need to have more predictability on what we start. We cannot have a business where we start a lot of volume in terms of wafer ahead or media. And then we have those variability in demand. And because we started the production we need to go to the end, we need to push those volume to you. And of course, this is not helping on the financial health of this business. So we said we want to change it, and that will be beneficial to both of us. More visibility for us, better plan for you in terms of what you can receive in terms of storage. So we started the build-to-order. And the build-to-order means customers are giving us orders and then we start production. And of course, when we start production, we already have a certain volume, a certain price, a certain mix, certain time for delivery so everything is way more predictable. And we started that 9 quarters ago, on top of being more predictable, of course, we said no, now we need to have a much more fair pricing discussion. And so we're starting to increase the price of our products slowly but very consistently. So every time we were renegotiating a new build-to-order, customers that were buying the same product they saw our price going up and customers that were moving to our next product where we, of course, have a cost per terabyte reduction, they were seeing a little bit of that benefit going to them. But then if they're not moving again, price was going up. So that has helped us to improve not only our revenue but also very importantly, our profitability in that period of time will basically more than double our gross margin, and we have increased a lot our revenue. We said at the beginning of the calendar '25, every quarter, we'll see a higher revenue and a better profitability. And we have done it until now. And I'm sure we will do this quarter and next quarter aligned to that visibility that we already had at the beginning of the quarter. So this is what has changed is this ability to work with our customers in a different way. And I would say right now, customers are used to that. They understand how to optimize their TCO that is qualify next product, go to next product, they get more terabytes per unit, which is a major improvement for them, and we get more predictability and more profitability.

Great. And can you maybe give us a sense of what current exabyte capacity are you running at this point, where your utilization levels are, both for yourself and for the industry? And then how much runway do you still have to meet demand without increasing unit production?

I think in terms of units, no, we are fairly full. And I think the industry is fairly full. So I think we all sell the units that we produce. In terms of exabyte is a different story because you can mix up in capacity and generate more exabyte. This is why we invite our customers to move to the new product. If today, they buy 24 terabyte and they want more exabytes instead of buying 2 units of 24. Now they can buy 40 terabyte and fairly soon a 50 terabyte drive and achieving the same result without the need for us and the industry in general to have more units. So this is where customers get the best result. Lower units more exabyte. That is the best way to generate profitability for everyone in the storage business but more importantly, to give the best TCO to our customers.

Then if you talk about pricing for a second, if you think about where you sit with your customer discussions today versus a year ago or 6 months ago, what is their willingness today to accept incrementally higher prices? And when you think about pricing, do you think about it on a per exabyte basis or a per drive basis? And are they willing to pay more on one but not the other?

I will say there is no difference, I would say, possibly today, the gap between supply and demand in the cloud segment is even a little bit bigger than what it was 6 months ago. So it's a little bit more focus from our customers to get exabyte more than working on the pricing. So the methodology has not changed. We have used the same methodology for like 9 quarters, there's no reason to change. They know that it's better for everyone when they qualify the next product. They -- of course, they receive more exabytes from the same unit, we have a cost reduction when they move up in capacity, so good for everyone. When they don't move to the higher capacity drives, they need to pay a bit more for the old drive.

Let's say we kind of find ourselves in a situation down line at some point when the industry we discover has overshipped end demand and your customers are looking to cut orders to you, how does your build order strategy, protect you, if at all, in that kind of situation? And to what extent is there a cushion for you? Or does that create a cushion in the next potential down cycle?

Yes. I think the very important change compared to the past, is the visibility. So when we have a build to order is basically covering 3, 4 quarters. Now, this is the time that we need to produce a product. So we produce the volume that they need. And they have good visibility in those 3, 4 quarters because it's basically kind of independent from their final demand. Now those 3, 4 quarters depend on how many new data centers, they are complete, so they need to install hard disk, how many data centers they want to refresh and so replace all drive lower capacity with new hard disk drive higher capacity. When you go no longer, of course, that then depends on what is their view of demand. But I would say, for us, it's very important to start product that we know they will be sold, they will be sold at a certain time. They will be sold at a specific customer with the price that we have already discussed. There are no discussion at quarter end for volume or pricing or anything. That was the normal way to run the business a few years ago. You have a quarter end and you discuss some time shipments over the last few weeks, sometimes shipments for the next quarter and the pricing associated all that completely disappeared from this industry. So the visibility helps us. Now we don't have the need to push product to customers, they don't want. We just produce what they want and the condition that we have mutually agreed.

Okay. That's helpful. Maybe 1 more structural question before you, which is a question I get from investors all the time, it recurs every 18 months or so, which is, if you think about the level of price competition or cost per bit -- competition between solid-state drives and hard drives. What's your -- how do you summarize the delta in terms of either pricing per bit and total cost of ownership between the 2 technologies today and where do you expect that to go over the medium term?

The delta in cost is huge, it's like between maybe 6 and 8x. I don't think it's so important, honestly. The infrastructure of the cloud is built in a way that they buy both our disk and NAND but to different users. So the optimization of the storage is done with hard disk, but to run the application, you need NAND and DRAM and GPU and CPUs and many other components. So the data is moved from hard disk into an SSD. They run the application that they want to do and then they save it back into our disk. So the delta in pricing, I don't think is so important. We have seen different timing and different cost curve and different pricing for those 2 technologies and the cloud infrastructure has not changed. It's still 90% of the data is stored on our hard disk. 10% is actually what is in use, so is on an SSD. So I don't think it's important, but in any case, but it's an enormous difference in terms of cost per terabyte. I think HAMR actually will increase this gap. Now as I said before, the cost curve is very different in terms of timing. The NAND went from planar to vertical and then from 32 layers, 64 layers, you generate a lot of more gigabytes per wafer, so a lot of cost reduction, then you go to 96, 128. So the benefit of that technology start to decline. It's still giving cost reduction but start to decline in terms of impact. HAMR is just starting now, where we go from 24 terabyte PMR into a 30-terabyte HAMR into a 40-terabyte HAMR that we started to qualify last month. So the cost curve of HAMR just started. So the difference between the decrease in NAND cost and the decrease in hard disk cost, I think for -- at least for a certain number of years, is not going to reduce. It's actually going to increase.

Great. And so that kind of segues into the next topic, which I want to discuss, which is HAMR. That's the technology you've been very vocal about and which you've been through a fairly lengthy qualification with your lead CSP customer on recently. When you look forward to qualifying other customers on the HAMR technology, do you anticipate as many kind of stops and starts or as many kind of complications with those next customers as with the first one? And would you expect that to be -- over what time frame do you expect those to be completed?

No, I would say the force qualification took a little bit longer than what we were planning because it was the first time we were qualifying this technology. We had to find the right configuration for the cloud. We were actually fairly quick in qualifying HAMR in other segments. But when we had to qualify HAMR in the cloud, it took us maybe 6 to 9 months longer. So this was customer #1, then customer #2, 3 and 4, no problem, went very fast. Actually, we were -- so a little bit surprised by the time of the call of the last customers. So we qualified it a bit earlier than what was planned. And so we announced last week that now we have 4 cloud customers out of the big 7 that are fully qualified on HAMR, 3 of the major 4 U.S. customers and one which is outside U.S. And we also discussed at our earnings release that we started the qualification of 40 terabyte drive with 1 of those 4 customers in July. So everything is progressing right now, very smoothly, no problem.

And then when you think about the revenue ramp from those different customers, would you expect the other 3 customers to ramp to the same size or and/or as quickly as the first one.

I would say it's kind of independent from the technology. The ramp based on their need. Of course, when they are qualified on HAMR, they want to buy HAMR because it's bigger capacity per unit. So they have a better TCO. And so they will try to get more HAMR. HAMR also depend from our ramp. And of course, it takes us a little bit of time to ramp all that volume. As we said, we wanted to qualify first, get a build to order and produce the product that they want. But you will see a good improvement in HAMR production every quarter. And of course, the more customers we have more volume we can allocate to HAMR compared to the prior technology.

And then when you think about the cost competitiveness of HAMR on sort of a price per bit basis or cost per bit basis relative to advanced PMR technologies. Where does that stand today? And then what do you expect to -- how do you expect that to evolve over the next 18 months.

Yes. At our Investor Day end of May, we presented a slide where you can actually see the difference in cost per terabyte. And we show our last product in the old technology so in PMR technology and then the fourth product in HAMR which is a 30 terabyte and then the 40 terabyte and the 50 terabyte. So you can see in terms of cost per terabyte, the last PMR product and the 30 terabyte are fairly close. 30-terabyte HAMR is slightly better but not so much better. Of course, we go up in capacity from 24 to 30, so 6 terabyte more. But of course, HAMR has some components that we don't have in the old technology and some components that are a little bit more expensive, but it's already better. Then when you go from 30 terabyte to 40 terabyte, that's a big cost saving because now you have the same number of heads, the same number of disk, the same kind of components that you use, of course, some components continue to be a little bit more expensive, but you get this benefit of basically a very similar build of material and 10 more terabytes per unit. So that's a huge difference going from first generation HAMR 30 terabyte to second-generation HAMR 40 and then 50. So this brings back what I was saying before in terms of cost curve between the hard disk today and the NAND. The hard disk cost decline is happening right now. The NAND cost decline happened years ago when the technology moved from planar to vertical.

Yes. So more or less on parity at 30, but a very clear crossover by the time you get to 40. Yes. And maybe -- how do we think about HAMR's impact on your gross and/or operating margins? How -- what is the path to margin accretion on either front will look like?

It's a big help. It's already helping a little bit today. As you know, we will ramp much more volume. And as I said, the 30-terabyte HAMR doesn't have the same benefit of the 40 terabyte HAMR. So we will see much more in the near future when we start ramping the 40 terabyte HAMR. So I think this fiscal year is mainly the 30-terabyte ramp and the 40 terabyte call and then probably next fiscal year, you will see the benefit of the 40 terabyte ramp in high volume. Now we are also in a kind of unusual situation where the 30 terabyte and the 40 terabyte they are fairly close in time, mainly for what we discussed before, the 30 terabyte was a little bit delayed in terms of call, but the 40 was not. So they're very close. So today, there are both products that can be qualified. In the future, I think between 40 and 50, probably you will see between 18 and 24 months. So it's more the normal cadence for those products. So in that period of time, you grow from 40 to 50, 50 to 60. So there are huge improvement and increase in terms of terabytes per unit and that takes more time than what you have seen today between 30 and 40.

Yes. And then relative to your portfolio, VIA is a relatively small piece of that, but I think you started disclosing recently that the nearline mix within that is relatively heavy. Maybe talk about the growth of that subsegment and what it's looked like for you recently? And then would you expect that subsegment to kind of outgrow the broader VIA business over the next 1.5 years or so?

Yes. As you said, it's not a big part of our business today. The cloud is the one that is really taking the vast majority of the exabyte that we produce, but it's an important segment. And we have seen a change a little bit in their business model. In the past, they were basically providing the video system. If you have the need for a surveillance of the certain buildings, they were providing that with low capacity drives. So as the video camera and they are all connected to a fairly low capacity drive. But you couldn't save or store your data with those customers of us. So you have to move the data into your on-prem data center or to a public cloud. And a few quarters ago, they started to buy high capacity drives, not low capacity, high capacity drive. And so we were working with them, and they said that they now offer a kind of a simplified cloud. So if you have their system and you want to keep your data into a kind of simplified cloud, they can do it for you. So they offer this service. So they buy higher capacity drive to do their own little cloud. This is why we have basically separate VIA in between what is a cloud service and what is a client service that is still related to the video camera system.

Great. Also maybe I want to ask you about the legacy market. Again, not a huge piece of your mix. But clearly, that segment has been -- as mass capacity has increased, has been on a bit of a downward trajectory, if you think about the longer-term health of that segment or the growth of that segment, do you expect some stability for legacy? Or do you kind of keep -- expected to continue to decline on either an absolute or relative basis in the out years. And if there's stability, what applications would support the long-term presence in the market.

Yes. So -- and we discussed this at our last earning release. But we look at the business in 2 major segments, the data center that, of course, includes the public cloud, on-prem data centers or enterprise OEM, that part of the VIA data center business. And the edge. And the edge is more client consumer mission critical, the VIA client. This part, we probably do not expect to grow. And especially the client base is declining. So the low capacity drives, they are probably more and more replaced by the NAND because the NAND is used to run the application. They don't really store your data in your laptop. But you need a NAND to have your laptop performing well. But if you are an engineer, maybe an AI engineer and you generate many terabytes of data, you don't store in the laptop, you store in the cloud. So what is a different location where the storage is done. Now many years ago, when the hard disk business, the majority of that business was client. Everyone was storing data in their laptop or in their PC. Now everyone, especially people that create a lot of data store the data in the cloud. So it's not that it's a replacement of the storage. The storage does move to the cloud, and so hard disks moved to the cloud. But now more and more, if you want to use a high-performance laptop or desktop, you probably don't need to have an hard disk there. And you can use a NAND and just save your data or in your on-prem data center of your company or in the cloud. So I think the edge will probably continue to decline as we have seen in the past, and we have no different trend in different period of time. And cloud or the data center will continue to increase. However, we don't know about future application. So if there is an application where it requires a decentralized storage, then you could see hard disk and the edge starting to grow. So it depends from what will be or what kind of application we will need in the future in terms of storage.

Great. I'd be remiss if I had you and didn't ask you some financial questions, so let's do that. And 1 of the questions I get most frequently from investors is regarding gross margin. So maybe provide your perspective on how much room you have to expand gross margins over the next sort of 12 to 18 months from current levels? And what are the main drivers of that? Whether that be pricing, cost reductions, HAMR mix or otherwise.

Well, at the beginning of this calendar year, we said every quarter, we will see higher revenue, higher profitability, and now we are going exactly in that direction. We said -- we discussed before about pricing. So the pricing strategy is not changing. It is the same. So we expect similar results. And we will have a big help from the mix moving from -- the technology is less important but moving from 24 terabyte to 30 terabyte to 40 terabyte that is where we generate cost reduction and, therefore, opportunity for us to improve our profitability. We have not discussed about calendar '26 -- calendar year '26, yet. So we will do, I guess, fairly soon. But I don't see any reason to have a different trend. As I said before, demand, especially in the cloud space is very strong. Even stronger than what we were expecting a couple of quarters ago. So we are going in the right direction, and we don't see a reason for any change.

So stable pricing, better cost and better mix.

I'd say, a better mix but also drive better cost and the pricing strategy has not changed.

Great. One of the takeaways I had from your Analyst Day a little while ago was your OpEx efficiency that you're trying to drive. And between you and your main competitor, there's a 4 percentage point gap in your latest targeted financial models. So what would you attribute that difference? And could you maybe address the level of confidence you have in achieving the somewhat higher level that you've laid out?

Yes. We have reduced our OpEx in the last more than 2 years. And the main reason is, in the past, we had to develop a new technology and we had to develop product in the those technology because we are still having development in product from the PMR technology. So we had, let's say, 2 different things. Now we don't develop anything in the old technology anymore. So we don't have PMR development anymore. Everything is in HAMR. So we don't need 2 teams. We just need 1 team, focus on HAMR. We don't need to develop the technology really anymore. We have the technology so we just need to develop future products. That is not easy, but it's, of course, requiring a smaller pool of resources compared to what we were doing in the past. So we were able to reduce our resources. And we said at Analyst Day, now our target is to achieve about 10% of revenue in terms of OpEx, and we will see in the next several quarters where we are.

Okay. Very good. And then you've also made a lot of progress in paying down your debt. So has the $5 billion gross debt target shifted at all? And how do you think about your real estate target for debt levels in the next sort of 1 year or 2 years or longer?

Yes. We had a first major goal that was to achieve $5 billion. We were above $6 billion a couple of years ago. So we wanted to reduce that to $5 billion. I think $5 billion is very manageable for us in any part of the cycle if any cycle will happen or even in the future. I think we can go even a bit lower. I think a good way to think about our debt is probably between $4 billion and $5 billion. We don't have any maturity in the next couple of years. So we will see when is the right time. But probably between $4 billion and $5 billion is where we want to be.

Good news. Good flexibility -- ability to have that flexibility. Maybe just on free cash flow for a second. I mean the CapEx intensity of the business is pretty low at this point. You're on track to generate substantial free cash flow over the next couple of years. How do we think about your business from a free cash flow margin perspective? And is there any target you feel comfortable sort of attaining over the medium or even longer term?

Yes. Free cash flow will be strong. We have actually generated a good free cash flow during the 2 years of down cycles on fiscal '23 and fiscal '24, we generated enough free cash flow to completely pay our dividend and so supporting and protecting our dividend. Last fiscal year was a little bit of the opposite. We had to rebuild a little bit our working capital. Now we had to stretch that working capital in fiscal '23 and '24. We rebuilt it during fiscal '25. Last quarter of fiscal '25, it was June, we generated a very good free cash flow. And so I would say that part is done. I think now our free cash flow will trend very similar to our net income will be more aligned to that. So I expect this fiscal year to generate a very good free cash flow.

Great. And then maybe I'll just conclude on the final question regarding M&A. How do you think about your M&A posture going forward? Is that something that's on the table for you? Do you see any interesting adjacent product areas that could be complementary for Seagate over time? And sort of how do you think about that piece of the capital allocation strategy?

Well, I'll say we are mainly an hard disk company. We like this industry, especially right now. And the industry is very consolidated. So there is not much that is possible to do in this industry. We have done a little -- some small acquisitions in the supply chain. I don't exclude we could do something in the future, but nothing that we have a visibility or a specific target right now.

I think with that, we're almost out of time. But Gianluca, thanks so much for being here. We appreciate it.

Thank you very much.

Thank you.