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

Good morning. Good morning. Thank you. Can you hear me? Thank you. I'm Tim Arcuri. I'm the semi -- semi equipment analyst here at UBS. Very happy to have Seagate with us on this next slot. We have Gianluca Romano, who is the CFO of Seagate. So we're going to talk about the company and I have a bunch of questions. So maybe you want to read the statement first, and then we can go into Q&A.

Yes. Thank you, Tim. Thank you for inviting us today. Well, first of all, 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.

Great. Awesome. So Gianluca, let's start by talking about end demand. Maybe you can sort of give us just your sense of where end demand is? And obviously, we've -- there's some digestion going on for nearline that's clear. You were very, very clear and your peer was clear about that last call. So can you just talk about sort of some of the dynamics in terms of demand for cloud and enterprise and some of the other end markets?

Yes. We discussed [ with our earnings ] release and a few meetings after the earnings release. There are different factors that are impacting demand in the short term. The first 1 actually is not really the U.S. cloud inventory. I think the major factor for us is actually the slowdown of the economy in China. That is in terms of revenue is where we get the majority of the short-term impact. And of course, this is related to the lockdowns and the COVID situation. But hopefully, it's starting to move in a better direction for the economy, but still very negatively impacting our business in the short term. The second part of still very important is actually the U.S. cloud inventory. And they have accumulated a bit of inventory, a little bit every quarter, but for a fairly long period of time, starting the COVID situation, many companies were buying a little bit more for their safety stock to avoid supply chain disruptions. And because COVID lasted basically 3 years and actually still it's not over yet. Companies have accumulated a lot of inventory. And right now, I think the fear of major supply chain disruption, especially on the storage space is not there anymore, and therefore, no big companies want to reduce their inventory. We are actually doing the same thing now, if you look at our own inventory. Before COVID, we were at about $1.1 billion. Right now, we are running at $1.6 billion. So even on our side, we have accumulated inventory to avoid supply chain disruptions. And now as our customers are doing, we will also reduce our own inventory. The third factor that is way smaller than the other 2, but is still material is consumer. And on the consumer side, of course, a high level of inflation and the war in Ukraine are, of course, contributing to a lower level of demand in the short term. We are still very, very confident in the long term of this business. There are so many new applications that right now are just starting, like artificial intelligence and machine learning and autonomous driving, Internet of Things and smart cities, smart [indiscernible]. All those new applications will generate an enormous quantity of data and where data will have to be stored? And as you know, 90% of the data storage is actually in the cloud is actually happening on hard disk. So very confident in the long term. But of course, we need to manage this short-term down cycle.

And just in terms of the shape of the recovery, can you kind of speak to that a little bit?

Of course, it's difficult to predict. I would say, in the current quarter, we are trending as we discussed at our earnings release. So we don't -- so far, we don't have differences there in terms of revenue. As you know, there are possibly a good sign from China in terms of how they manage COVID and lockdowns. So hopefully, that will be a positive to us. It will take a little bit of time. Those are things that doesn't really change in 1 week. It takes maybe a month, 2 months or 3 months. So a little bit of time. And then through our very low level of production that we have today, we are also trying to help our customers to reduce their inventory and try to go back to more normalized business model in the next few months.

Speaking of normalized, can you talk about gross margin? There's some atypical headwinds right now because of under-absorption and things like that. So can you sort of level set us in terms of gross margin? I think as I look at some of the Street numbers on gross margin, I feel like the -- what's being assumed is not sort of taking into account some of the headwinds that you have today, and it seems to me like gross margin ought to recover to a level that's higher than what the current Street numbers would show?

Yes. There is no reason why the gross margin should not go back to where it was before, once the revenue level becomes the same level. I would say, actually, through our restructuring and our reduction in force, we should have an even better cost structure at a certain point when revenue goes back into the $3.1 billion that we were basically a year ago, December quarter last year. So I would say, again, long-term but no reason to be pessimistic. I think we will go back to where we were before, even better. Now in the short term, as you said, we have a lot of underutilization charges. So we want to keep our production down until the inventory is clean. And of course, in the short term, that it is a big comp. It's probably where it could be bigger than what we were estimating at the beginning of the quarter, but it's all part of our decision and our strategy.

Right. Because if I look at December and I exclude the headwinds, you're sort of at a 26% gross margin ex those headwinds. So if I run kind of a normal drop-through on that, it seems like you don't have to get back to $3 billion to get to 30%, you could do 30% at something less than $3 billion, maybe $2.6 billion, something like that. So it seems like there is -- given all the cost actions that you've taken, you should be able to get to that 30% at quite a lower revenue number?

Yes. If you look at our results a year ago, we were about $3 billion, but we were well above the 30%. We were basically 32%. So again, I don't see why we should not be in the same situation or even a little bit better based on the new cost structure that we are going to have.

Great. Can we talk about the product road map for a second? CMR versus SMR? I still sense some confusion among investors that assume that you don't have SMR and they don't fully seem to appreciate that '22 TCMR product can be used as a '26 TSMR product. So -- and with some software changes on the customer side obviously. So can you kind of talk about the product road map a little bit?

Yes. There is a little bit of confusion. The technology, CMR technology with basically a software change, can be used as an SMR. So it's the same drive when you go into the final test to the software, an SMR drive. If you have a 20-terabyte CMR and the customer asked to have SMR version, you have to change the firmware. You have a 26-terabyte SMR. But the different, read and write, mainly the write is different. So it cannot be used for all the applications. But if a customer for certain applications prefer to have SMR because give more priority to capacity than random write compared to sequential write, you can't -- and it's just a software. We sell about 20%, 25% of our exabyte in SMR version today. We didn't talk about SMR too much until now because, for us, it was not a big deal. It's the same drive. You just change the software and you have an SMR. I think what is not maybe well understood is HAMR. Maybe people don't fully appreciate that. Also, HAMR can be used as fewer HAMR or as an HAMR drive. If you change the software and you have an SMR drive based on the MR technology, the 30-terabyte HAMR will be mid-30 terabyte in SMR version. So you need to compare SMR to SMR, you don't compare a base HAMR to an SMR. They are used for different applications. If you want an SMR version, now you're comparing a mid 30-terabyte SMR drive based on HAMR technology.

And I think you said that the portion of exabytes that are on SMR is probably in the 20% range today, roughly? Great. Can you just talk about HAMR a little more and sort of the investments you've made there and how that's going to help your growth margin over the longer term?

Yes. As you know, we developed HAMR for more than 10 years. A lot of R&D was actually going to HAMR. But at the same time, we had to continue to develop CMR. So we basically had 2 groups inside Seagate, 1 working on the CMR and CMR product, 1 working more on the HAMR technology. Of course, as you know, we have announced our 30-terabyte HAMR to be in the market beginning of summer, so in just a few months from now. And this is also part of our restructuring. We don't need to have 2 groups over anymore. CMR is basically at the end of its life. We probably have another product that's coming out in a few months, but then all the new products will be on HAMR. So it's a very difficult and long technology to develop. We are very happy that we are at this point, and we have product ready to come out. Again, a 30-terabyte HAMR beginning of the summer. If you want to use that HAMR into SMR version will be a mid 30-terabyte. So it's a big increase compared to the drives that are in the market right now. But what is real benefit of HAMR. It's basically the fact that HAMR will continue to grow capacity per drive based on aerial density. So you increase the capacity of the disk. Today, a 20-terabyte drive has 10 disks and 20 heads. If you go back in the past, the drive had 1 disk and 2 heads and then 2 disks, 3 disks, up to 10 disks. This is how CMR was growing capacity, adding bill of material inside the box. HAMR is very different. HAMR will stay on 10 disks and 20 heads, start at 30 terabytes. The second product will be a 36-terabyte and then 40-terabyte, still on 10 disks and 20 heads. This is a major improvement of this technology. It doesn't need to have more bill of material inside the box. But by the way, it's physically limited at a certain point, just increase the capacity of each of the disk.

Right. I mean the HAMR, there is pretty powerful ability to drive cost per terabyte down through aerial density alone without adding to the bill of materials is quite powerful. So in that vein, can we talk about sort of longer-term gross margin and your ability to drive gross margin and the cost curve of HDD versus SSD? If you look at SSD today versus HDD, the cost delta today with NAND is 8 to 9x. But even if it were cut to 3 to 4x, yes, that might be enough for customers to cut over to NAND. But HAMR, it seems like it's going to result in an improvement of the HDD cost curve versus NAND that is actually seeing some diminishing returns actually.

Absolutely. No, I think NAND already went through the technology improvement when they moved from [ planner ] to vertical, and that was a major cost reduction in NAND and starting at 32 layers and then 64 and 96, the improvement in the cost is diminishing. HAMR is just starting now. So we are at the beginning of this [ S-curve ] in terms of cost reduction. I also say what is a limited part of the market we really overlap on the legacy part, of course, like there is an overlap. So the low-capacity drive where the NAND can replace our disk assuming the cost is -- it's fairly similar. But on the high-capacity drives, if you go into a data center today, you already have hard disk in it. They are both used in the data center for different parts of the application. 90% of the data storage is on hard disk. It is cheap and it is what the hard disk has done for. And then you have a lot of NAND. But the NAND is done for that 10% of data storage that is used very frequently and is used to help for analytics and compute. So it's not -- they are not used in the same way. They have not been used in the same way since the beginning. And it was 90% a few years ago, it's 90% now. If you talk to cloud customers, they always tell you that 90% of the data storage will continue to be stored on our hard disk.

Right. So if you sort of take that and you project out to where the long-term margins can go, obviously, the customers have the power in the near term, just given the supply-demand dynamics. But given what's happening with HAMR, given what's happening with the cost curve, given the fact that all the client capacity basically has already been absorbed, you and WD are going to embark on a CapEx cycle if customers want more exabytes, then there's going to have to be more CapEx put in the ground for that once we absorb all these near-term issues. So the industry, it reminds me a bit of DRAM 5, 6, 7 years ago, and we can see what's happened to -- I mean, yes, obviously, they're going through a difficult time now too, but the margins there have been going up over time. So it seems to me that this is not a low 30% gross margin business over the longer term. There is no reason, would you agree that this business can't be a high 30s, 40% gross margin over the longer term?

Yes. A couple of years ago at our Analyst Day, we gave a financial model where the gross margin was between 30% and 33%, but we said this is before HAMR. So at next Analyst Day, we will have probably a different model. Of course, we need to ramp a good quantity of HAMR to impact the P&L. But of course, we believe it is accretive to our gross margin. And as I said, it's not only the force product. Actually, the most important is when you scale even higher, 36, 40 terabyte with the same bill of material as we discussed before. I think that will be really the time where we should get a good boost to our profitability.

Can we talk also part of that, the idea that you can push gross margin higher longer term? Is the concept of LTAs, and I'm not a huge believer in LTAs that they really hold. Whenever it's bad times for them, they want LTAs, whenever it's good times for them, they don't want to adhere to the LTA. But can you kind of talk about just the evolution of the LTAs and how much [indiscernible] that they have?

No, I think LTA is very good. It's good for us. It's good for customers, give some stability. I would say, we had and we still have LTAs on about 50% of the business. So it's not 50% of our revenue, it's 50% of the cloud. I hope 1 day, we will have LTAs on 50% or more of our total revenue. I think this is good, is helping us in how to manage our manufacturing. Right now, it's still not big enough. And I would say, customer, of course, now when they have too much inventory and they don't need the product, they are not maybe too happy to have an LTA in place if the volume is what they need. But our experience is they still buy what is in LTAs. Of course, they just push the problem a little bit further. So when you arrive at a time that LTA expires and you need to renegotiate the LTA, that is when they tell you for a certain period of time, we need less because if they [indiscernible]. So it's more a timing issue, but it's not [indiscernible] LTA. And actually, in the past, they were buying more than the LTA. So if there is an LTA for the certain volume, they buy the volume and they give you for an upside. And then when they do need the upside, they go to the LTA and when the LTA expire, they negotiate a new LTA, of course, considering the level of inventory base they have. So assuming they are able to manage their inventory and of course, COVID is a very difficult situation because I think they were managing their inventory. They just want it more. And now they don't want so much more anymore. So they need to readjust that. But longer term, LTA, I think, is good for us. It's good for customers, it's good for the industry to give a little bit of stabilization.

And so it's the first couple of weeks of December, I'm sure you're engaging with customers on LTAs for next year. What's the tenor of those discussions? Obviously, it's sort of max pain or the pendulum has swung to pretty much the max in favor of your customers. Right now, has that affected their appetite to engage in LTAs? And sort of how is the tenor of those discussions as you kind of head into next year?

No, I would not say they changed their mind on the reason why in longer term, the LTA is good for them too. So we still have LTAs. Of course, those LTAs for a period of time as a volume, that includes the fact that they need to reduce their inventory.

Right. So in terms of the duration of the LTAs, do you expect these to still be annual or like multiyear?

I'd say the majority is not annual. It's actually more on the 6 to 9 months. We have a couple of cases may be longer, but the majority are between 6 and 9 months.

And what portion is there a way to determine what portion of the LTAs, if we rewind a year or even 18 months, what portion of the LTAs were lit up to? Or was it just that the volume got pushed out, that's all?

Actually, the customer that had the LTAs they consume the product until the end. They were not consuming more than the LTAs, but they were still getting the volume that wasn't LTA. And then as I said before, at the end of the LTA, when we renegotiate the new LTA, the volume [indiscernible] went down because they have to use the inventory.

So do you think as -- have the LTAs actually created more volatility actually?

That's a good question. Again, I would say we are not in a normal time. This COVID situation is difficult to manage for everyone. As I said, even for Seagate. We created more inventory. No, because we were not planning in the right way because we wanted more inventory. And I think maybe our customers wanted more inventory for a while. And now we need to go through a period where we readjust. And hopefully, COVID will not create more disruption anywhere in the world, not only on the western part of the world, but everywhere in the world. And then the planning doesn't need to include that extra inventory for cause -- for safety stock. And I think that will be a new phase. But before we go into that new phase, we need to consume that additional inventory.

Got it. Can we talk about your capital strategy? It was notable with the recent debt exchange and the successful covenant renegotiation. I got some questions as to the interest rate on some of the new debt. But can you sort of walk through the capital strategy?

Yes, the capital strategy, longer term, we don't think it will change. We are -- we have always been, and we want to be very focused on shareholder return, and this included a good level of dividend and a certain level of share buyback. We were active in the market until September quarter. And then because of the down cycle, we said, okay, for maybe 2 quarters, we will go on pause in terms of share buyback. And we focused a little bit on our debt structure. Because of this fairly big reduction in the business, of course, we need to financially manage it. So we -- first of all, we renegotiated our covenants that were based on 4x in terms of leverage, and we renegotiated to 5x and we also reduced a little bit our debt through the exchange notes that we have just concluded last Monday. So -- and we also said, probably we will reduce our debt even more in the next 3 to 6 months. It's just part of the financial management of the downside.

Right. I think you replaced 4 tranches of 2029 and 2031s, all of which was at roughly 4% -- roughly, and you replaced it with 2032 debt at nearly 10%. So can you speak about like the changes in the debt service requirements?

Yes. So first of all, we reduced the principal amount by about $200 million [indiscernible] is on a lower debt. I think overall our interest rate moved from like 4.5% to 4.9% as a total company. Of course, we need to trade a little bit between the debt value and the interest, but of course, in this period of time is much higher. This new debt will basically temporarily increase our interest by about $10 million a quarter. But as I said before, we are going to buy back more or redeem new notes and reduce our debt. I would say probably when we arrive at the end of this fiscal year, that is June, we'll probably be at a similar level of interest. So we still reduce our debt to where the interest go back fairly close to where it was before.

Right. So there's a penalty in the near term, but you're going to continue to work down the debt throughout the year?

So you reduce the debt, the interest rate is a bit higher, but the interest dollar that you pay more or less are fairly well aligned.

Got it. And can you just maybe speak more broadly to capital return? And how committed you are to the dividend and sort of how you think about the liquidity and the free cash flow profile on the capital return in the near term and looking out?

Yes, as I said before, we are very focused on shareholder return. So this is something that is part of our strategy as a company. So we want to continue to do that. Of course, we want to protect our dividend. We said we think we will have a positive free cash flow every quarter even during this severe downside that I think is a good sign of our ability to create cash from this business even during a very difficult time in terms of revenue. We are reducing our inventory. That, of course, is also part of how we generate free cash flow in current quarter and next quarter, probably more in the next quarter. So again, it's something we need to manage in the down cycle. We don't think it'll be a long term. We think it will be a fairly short term. The big customers that have already discussed about their CapEx for 2023 were fairly optimistic, for sure an improvement compared to 2022. With a fairly big CapEx, we don't think they can wait only the September and December quarter to spend that level of CapEx. So we think if not March, at least June will be already a fairly high level of CapEx, and as you know, hard disk is a percentage of our CapEx.

And just in terms of the inventory, I mean, obviously, you don't need more inventory at this point, you reduce utilization in December, very low in October, November throughout this quarter. Do you expect utilization to come back much in Q1 or to work down the inventory, you're going to probably keep running underutilized?

Yes. As you said, mid-December is when you really talk to customers and see what is their forecast for the next 3 to 6 months. So we need to see what is the level of the inventory at this point and see when they want us to ramp back production because finally, it's based on what they want to commit to us in terms of volume, but also very important, as I said before, is the situation in China. That is a big part of our volume. China is important for China cloud, China enterprise OEM, China video and image application, consumer mission-critical, all the segments are fairly big in that part of the world. So when that part start to improve, we need to ramp up. And of course, when our major cloud customers in the U.S. have consumed their extra inventory, this is when we need to ramp. So right now, as you said, our production level is very low, and we want to keep it low until we have those good signs in terms of demand. This is, in the short term, creating underutilization charges that are very expensive and impacting the P&L actually fairly importantly.

But your -- you sound optimistic that China is going to come back maybe towards the end of Q1 into Q2?

Well, of course, we have a little bit of the same information. So it looks like we have -- there is now a support from the government to go back to a stronger economy. I think they probably want to have a stronger economy in 2023. 2022 is almost over. I don't think they want to influence the GDP of 2022 at this point. So probably they wait it to be longer and then when we enter in 2023, hopefully, they will relax some of the lockdown conditions and also support and stimulate economy, as they said, after the meeting -- the big political meeting they had in October.

Can we actually talk about CapEx for a moment? You said that you'll be below the target range of 4% to 6% of revenue. But really, you're being very careful to not cut CapEx related to HAMR. So can you talk about that process and maybe hold our hand a little bit? When you have such a critical technology transition, people get a little worried when you start to cut CapEx? Are you cutting the right CapEx? Are you potentially damaging the road map? Can you kind of talk about that?

Yes. Usually, in this period of time, you don't need CapEx because you are now producing more than the capacity you have in place. Actually, you produce much less. But because of HAMR, we still need to invest. We need to -- even if we spent CapEx that is fungible for HAMR since at least [indiscernible]. So it's already a couple of years that we are buying equipment that can be used for HAMR, but we need to be more. So we are still buying those equipment. We said we will be below our normal range of 4% to 6%. It will be a little bit lower than the 4%, but we still need some investment because we announced the product for early summer, and we need to ramp.

Got it. Also another question that I often get is 1 of the large hyperscalers, not just one, but a general trend lately has been to extend the useful lifetime of servers and network equipment and things like that. Do you see signs of real extension of the useful life of the drives? Or is this more of just an accounting adjustment on your customers' part?

I would say there is a business reason why they refresh the hard disk, it is more a capacity. So if you keep -- if you think what was the high capacity 5 years ago is probably out of the capacity you have right now. So it's not that the hard disk doesn't work after 5 years. It can work. Probably a big part of the volumes still work, technically work. But for that physical slot, you can replace a hard disk with a new hard disk with double the capacity. So it's the revenue that our customers generate from that physical slot can be doubled. That's why it's important for them to refresh because in the same physical slot that is a limitation today, they want to have more revenue and to have more revenue that we need to refresh and replace old hard disk with new hard disk. And this is why HAMR is so important because you go into a phase where you refresh and you have old hard disk drive, what today we say mid capacity, maybe 12, 14 terabyte and you can go to 30 terabyte or mid-40s if you use that in SMR. So the benefit for the customers is enormous, not only because of the cost per terabyte, but because of the revenue they generate from that physical slot.

Great. And I just -- I'm going to ask you the question that I get asked a lot around the Department of Commerce in investigation, and I know that you can't say too much about it. But is it -- does it change how you run the company at all? Or is it just -- is it just a legal matter? That's all.

It's a legal matter. It's not changing how we manage the company and our strategy. It's a legal litigation, so I cannot say much except that we think we have a very solid position. Our internal legal team, our external legal team, they are very comfortable that we have a solid situation. We need to go through the discussion with BIS and hopefully solve this matter as soon as possible.

Great. And then can we talk about free cash flow, I think you've committed to generating free cash flow every quarter throughout 2023. Can you kind of talk about the trajectory of free cash flow and maybe any other levers -- I mean, obviously, you're pulling the inventory lever, you're pulling other levers, you're cutting CapEx. So can you talk about that?

Yes. Probably. The current quarter is where we generate a lower free cash flow because the revenue is lower. And because -- even if we are starting to reduce our inventory, we are still not paying our account payable for things we have purchased in the September quarter. So very little bit of time before you realize a decrease in inventory and increase in cash flow. So probably this quarter will be the -- hopefully, will be the bottom in terms of free cash flow, and then you will see an improvement starting the moment.

Got it. And then we haven't talked about the VIA market yet. Can you just can you talk about VIA?

Very important market. So with an image application of smart cities, smart [indiscernible], extremely important. It will be everywhere in the world. Today, our major customers are actually in China. And this is 1 of the reason why China is so important to us, while they manage projects everywhere in the world. So it's not that it's a Chinese application, it's everyone in this world. But our achievement, our major customers are in China. So we think also that segment will improve and we'll go back to where it was before and even better. I think it is part of -- it's 1 of those applications that will continue to grow in the mid to long term.

And you don't see the risk for any export restriction there?

Well, you never know about restriction. Even, our restriction, of course, no, will be impacted as everyone else. In general, if you have a restriction at that time, the demand is moving to someone that has no restriction because [indiscernible] projects disappear. It's now that you start having a video camera in New York, you will still have it. So someone else will take care of that application and will become our new customer. So the important is all those applications are based on hard disk. So whoever is the customer, wherever it is in the world, we can serve the customer. Today, the customers are in China. But if that will change, for us, it is not really a major problem.

And just to kind of level set, China is roughly 30% of revenue, give or take?

Yes, we said -- yes, we said last year, when we were at $3.1 billion, China was about 1/3, so about $1 billion.

Great. And maybe just to wrap up, we're sort of running out of time. Can you -- so we talked about inventory, but can you talk about the risk of inventory obsolescence? Is there any -- I get some questions about the potential to have to take a write-down on inventory. That doesn't seem in the cards, but can you kind of talk about that?

Right now, we don't see that. Again, we are keeping our production very low. So that actually our inventory decline. I don't think we have part in our inventory that will become obsolete. I don't think.

Great. Well, we're out of time. So thank you, Gianluca.

Thank you, everyone.

Thank you again.

Thanks a lot.