Welcom to the Barclays Tech Conference. I'm Tom O'Malley, U.S. Semiconductors and Semiconductor Equipment Analyst. Lucky enough to have Seagate here, Gianluca Romano, CFO. Thank you so much for being here.
Thank you, Tom.
It feels like it's been a blur going into this correction and then coming out of it, and I think that most companies that are here today would say that, you know, they're near the bottom, if not at the bottom, and improving. So I want to start out broadly and just ask about your expectations for the Nearline recovery versus the last time we spoke on earnings. What are you seeing in terms of customer inventories and order patterns? Okay, bang.
Okay. Before we start, let me remind everyone that I will be making forward-looking statements today, and you can learn more about the risks associated with those statements on our website. Yeah, so I think now we have a similar view. I would say at our last earnings release, and now even more recently, we have talked about the bottom that we possibly achieved in September quarter. So we are now coming up from the lower part of the down cycle. We have seen improvement in different parts of the business. In last quarter, when NAND pricing was really starting to improve eventually, the majority of the benefit was coming from video image application. This quarter, we see a better recovery from the cloud. And not only U.S. cloud, I would say U.S. cloud and China cloud.
So in both now major regions, that part of the business finally start to recover. And so we, we recently, discussed about our view for the quarter. We see, the revenue coming as we guided at the midpoint of our guidance. We are also being, fairly successful with our pricing increase, as we recently discussed. And, you know, we have a little bit more volume coming out in the next quarters, so we have a possibly lower amortization charges in the current quarter. So profitability could be a bit better than what we guided, at our last earnings release. So good news, of course, coming out from, from the bottom of the cycle, so still not a huge revenue, but at least with some sequential improvement.
And we also discussed about future quarters, where we see continuous improvement on the top line, despite the March quarter will have some seasonality impact on the legacy part of the business and possibly on video image application. But we still see Nearline both cloud and enterprise OEM to more than offset that seasonality. So we expect another growth the next quarter and even either.
Just wanted to dive a little bit into the comments on Nearline you're seeing from U.S. cloud guys and others. When you look at that business, is the vast majority of Nearline today at these lower levels than U.S. cloud guys? Or can you just give me a feel for that mix of business? Because I know historically they've been the bigger customer, but at these lower levels, I would just like to understand.
Yeah, every quarter is a bit different, but I'll say before going into the down cycle, cloud, non-cloud, inside the Nearline space, were very similar. So, because most of the time is now very close to 50/50, but then some quarters is 65/45, or even a bit higher. But depends on the quarter, has always been fairly similar. Then the down cycle are different dynamics. Now, enterprise OEM actually corrected earlier than cloud, but then also started to recover earlier than cloud. Now cloud is recovered. So I'll say we still be in that range, you know, 50%- 55%, or maybe 60% of cloud. But in future, you will see cloud being higher percentage. So, in the longer term, you know, when you look three to five years, we think cloud growth will be higher than enterprise OEM.
Enterprise OEM will still be a good growth, but cloud will be higher. The other segment that will continue to grow, possibly similarly to the enterprise OEM, is video image. You have those three sub-segments growing and the rest of the legacy business that actually will continue to possibly decline year after year.
Yeah. Breaking that down just a little bit further, could you talk about areas that you're seeing a more accelerated recovery? Because I know you kind of laid out those moving parts with those businesses before, but, you know, where are you a little bit quicker than you would have originally anticipated?
In the short term is cloud. So now we have good recovery. Now we were expecting, of course, a correction after the inventory utilization, but it may be coming a little bit earlier, but can be bigger than what we were expecting, you know, a few months ago. Again, it's not a big jump in revenue, but we have guided, but it's a good starting point for the upside. And as I said, we expect more in the next few quarters. And as I said, it's not only U.S. cloud, even if U.S. cloud probably 70% of the cloud business, but even in China, now we are seeing a good sign of recovery.
You mentioned earlier that you were having success with some of the pricing discussions that you had with customers. Me going to a cloud guy with any sort of pricing discussion seems very scary. So how can you have those conversations and walk away with a better result?
Well, I would say, first of all, they, not they have an interest in having a good supply chain, and they understand that right now, when you look at, you know, through the bottom of the cycle, the hard disk industry, you know, has not been in a good shape in terms of profitability. I would say Seagate, in particular, was at least generating good free cash flow, but the rest of the industry was not only being in a net loss position, but also consuming high level of cash. So that is not a something that can be sustained for a long period of time and could put their own business to a certain point at risk. So starting from that situation, of course, they, you know, they are more, they understand better why, you know, the industry is going a certain direction.
Second, you know, the industry is very consolidated, especially when you go into the high-capacity product, where there are only two suppliers. So, you know, we have the same, I think we have the same focus on improving profitability and improving cash flow because we have to, and because we have to support our own supply chain. We don't want that something will happen to our supply chain and then create a problem to us first and to our customers later. So, I think, as you said, it's never easy to negotiate an increasing price, especially when you are still not through the bottom of the down cycle and not start recovering, but it's still only the recovery. So it's not easy, but, I would say not only with the cloud, but everywhere in our business, we are increasing pricing.
No, we don't have crazy increasing. No, we do slip a few points every quarter, but very consistently, since already two or three quarters, so it's not something we just started.
Cool. So I think there's a debate right now in the industry as to what the traditional server market should grow at. And I think you've seen, like other end markets, a large ramp up of the qualified period of digestion. And on top of that, you've had the origination or at least the advent of AI, which seemed to steal some wallet share as it came to market this year. When you look at your business, I would assume that there's an underlying expectation for what the traditional server market grows. Can you give me your view for 2024?
Yeah, and I think a bigger discussion on the impact of AI, you know, in the short term and long term. But I would say, well, our view is similar to what you can see from other, you know, analysts, you know, that are working on, on the projection. So it will grow. It will probably grow a little bit lower than what we have seen in the past, I think is overall, it's a bit of a timing issue because AI is consuming a big part of the CapEx, uh, from our customers, you know, to create the infrastructure of AI, but doesn't need today the storage. The storage is already there. What they need to create is an infrastructure with all the GPU and CPU and other, other components that they don't have today. So AI cannot be used without that infrastructure.
So they need to create the infrastructure. When the infrastructure is ready, you will see the move to storage, because AI, first of all, AI works much better when you can access a lot of data. So people, businesses, governments that want to use AI will tend to delete much less than what they were deleting before, because they want to have a large data of data to use AI. Secondly, more importantly for the impact of storage is AI will access the data and create something new, but something new is new data. For example, if you are in manufacturing and, you know, you have a new factory, you need a new manufacturing flow, you ask AI to do it for you, and AI will generate a new manufacturing flow, and that is data that will be saved in the cloud, in cloud disk.
Then you do it again. Every time you have a in your manufacturing flow, or you have a new factory, or you have a new equipment, you ask, you use AI to optimize your business, and that is new data going to the cloud. This is still not happening, and it possibly could be four quarters. You know, today, the use of AI is very limited. Now, I don't know how many people here are really using AI. It's very limited, but in few quarters, you know, it will be normal practice for people and for businesses to use AI.
You talked about 1 million HAMR unit shipments in the first half of 2024. Is that guide dependent on a server market that is improving, or do you think that happens regardless of the market?
No, that million unit, no, it's not dependent from the server. No, we are qualifying the product. We are through the end of the qualification with one main customer, and that volume will probably go mainly to this one customer, because it is the one that is qualified at that time. But starting next quarter, we qualified many other customers. So now we have requests for qualification from all the major cloud customers, enterprise OEM. We also have requests from internet application customers. So the interest into that technology and that product is getting bigger and bigger. So in the first six months, we will sell about 1 million units. After that, when we're going calendar 2023, calendar 2024, that volume will continue to increase and will be more diversified to a bigger number of customers.
When you look at the qualification process for a cloud customer, how much of the learnings can a customer take from the guy that got qualified first? Because you imagine that it's just the time horizon of running a product. And if you see that work somewhere else and you, you've used storage arrays in your data center in the past, is it possible for the second to accelerate that qualification process, or is it going to be the same time frame, no matter what?
Well, I would say, first of all, they will start the qualification with a high confidence, because they know if the product is working, is qualified by one of their peers. And in their that is huge scale and fairly complex architecture, so it's not an easy, difficult qualification. This is why we are working now with this partner. So I would say, if you want, that we follow, we'll start already with, you know, a different level of confidence. And then, now, usually qualification takes, you know, between four and six months. So that is the time that we need to work with many customers. So it will be a good effort for us because we try to qualify, you know, three, four, five customers at the same time, is not easy. But it will take the time that is needed.
The important is through our calendar 2024, we want to have all those big customers, including our enterprise OEM, including some of the video image application qualified. So the volume will continue to grow, the number of customers will continue to grow, and then can be diversified. We should have also a, you know, a good impact to our pricing.
Do you think that any of the sales of those HAMR drives in the first half will be cannibalistic to the recovery in your broad unit line? For example, your main customers... Why would they go out and buy more drives now? Aren't they incentivized to wait until they have this new product?
Well, there is not enough volume for the customers to wait. Even so, even the current customer, which is qualifying HAMR, is still buying CMR. So the benefit of HAMR is you can use HAMR in the same data center where you have CMR. So they can mix, they can mix vendors, technology. So they can currently buy a 30 TB HAMR, and a 24 TB CMR, and use in the same data center. So depending on the volume that they need, they buy a little bit of both. And in future, I think, you know, the majority of what we will sell will be HAMR-based, but for sure, you will have a transition between current technology and HAMR because of the volumes are big.
I mean, 1 million at this point, the cycle is quite significant, but when you get back to normalized run rate, slightly less so. But do you have a feeling for when you would expect crossover from a HAMR product perspective? You're saying 1 million in the first half. You're saying higher than that in the second half, so clearly not in 2024. But how quickly can that accelerate? 2025, 2026, or are you still thinking out the decade?
Oh, no, we think we will have a lot of products on the HAMR technology operating calendar 2025. So you will see sequential improvement. Now, exactly, crossover, depending if in exercise, the crossover will happen, much earlier than in unit, of course. But again, the focus now is qualifying customers so that they can buy HAMR, and then we need to have the time to ramp, because the ramp is not a short cycle time. When we start to get demand and orders, then we need to ramp. So everything needs a little bit of time to happen, but we are going in the direction we were expecting, a little bit faster than what we were expecting. So confidence is high, and so far, so good.
I want to shift gears to... You specifically called out China as weak in the quarter and kind of pointed to some uneven recovery there. Is there a specific area in China that you would think that is causing the weakness that out? And then you're saying uneven recovery. Does that mean up and down, or does that mean kind of flat now?
Well, I would say every segment is a bit different. So now we see China now starting to recover, so we think this will continue to recover in the future quarters. When you look at video and image application, is more seasonal. So December is high volume. Because we are raising pricing, we think, you know, customer has, have anticipated a little bit of demand in June and September, so that helped those two quarters to be particularly good. They're still buying a lot in December, but probably lower than what we saw in September. So that normal seasonality this year is, is not happening, not the same, mainly because of, but, but of course, they're still buying good volume.
You know, after the seasonality in low seasonality in March, they will come back to high volume starting June and then they compare to the calendar year. Then when you go into, let's say, the legacy part, that is still very depressed. So the consumer side and even the initial, because it's fairly the part that they're still waiting to recover. On top of the other segments grow, because, you know, they are growing, but they're not even close to the levels they were before the down cycle. So we expect that to continue to grow very strongly. But even the legacy, no, but usually decline, we think right now it's so depressed, but maybe it will rebound a little bit before following the normal rate of decline.
Yeah, I wanted to ask you on that, because traditionally in technologies that are winding down, even when you see a big setback in demand, you don't often see that come back and recover because people see the opportunity to move on. What gives you the confidence that, like, your other segments that have, I would say, growth trajectories in the future, you know, people should get back to the normal order rates? But what gives you the confidence that a legacy business that you've already talked about kind of winding down, step up before it continues to decline?
The main reason that also know the technology that is, the placing hard disk in the legacy part, has also the same cycle, so they're also very depressed. So I don't think there's been a change from hard disk to NAND in consumer and mission-critical, maybe in client, yes, but not too much in consumer and mission-critical. So if demand comes back up, it's not that they have changed the design to a NAND design yet. So I think what they will do in the future, because that is the part where, you know, NAND has the benefit of the speed and the time. So that part will continue to move to NAND, but possibly you could, you could see a little bit of rebound compared the very low level of demand we have today before getting to the normal trend.
In the past, you know, our revenue decline for legacy was about 15%, between 16% and 20% per year. Now, it's a small part of our business. It's only 20% of our revenue, so not a huge impact. And it's also now the part of the business with a lower, lowest gross margin. So with the transition we know we are going through, but it's not anything that will really affect the overall growth of the company.
I wanted to ask you about competition, some of the financials. You had your competitor here, and they talked about last quarter, half of their exabytes being on the 26 TB platform in the Nearline business. Can you just... I think that there's a narrative that before HAMR, if you are at higher capacity drive or what you claim to be a higher capacity drive, you will soak up the demand prior to technology transition. Do you think, one, that there is a lead in their product portfolio prior to your transition to HAMR? And do you think that you're seeing any shift in share right now?
Yeah, no, you can look at the market share last quarter, but there's no shift in share. No, despite, you know, what they are saying on the 26 TB, we have, I think we still have the leadership, even on CMR, even before HAMR. We launched our 24 TB CMR before competition. That drive, if it's sold in the SMR version, can be 28 TB. So I would say these two companies are fairly well aligned on the current roadmap. The difference is we have a major advantage on HAMR, and we are starting to sell HAMR next quarter. So it's not that we need to wait another year to take benefit of that better technology and better product. We will start next quarter, and we will go through the calendar 2024 and later.
So talking about the current roadmap, I don't see Seagate has a disadvantage because we actually decided to be developed that 24 TB CMR or 26 TB SMR. To be sure, we didn't have a period of time of disadvantage, to be at least at parity, and then taking the benefit of it.
All right. I want to switch to the gross margin side. So recovery got pushed out a little last quarter. Could you walk through the pieces related to underutilization right now? I know you've talked recently about how it could be a bit lower as things... Why is that happening? And then you also expect the charges to persist into the first half of the calendar year. Even if it's a little bit better, talk about your view on where underutilization goes.
Underutilization, now, now we expect it to be lower because now we start to ramp it a bit higher volume than what we were thinking. So the recovery is happening, and it's happening slightly higher than what we were expecting. Still not, as I said before, not a huge jump in revenue, but a start to be a good improvement. So that will decrease our underutilization charge. Now, we are still far from our capacity level.
So we will have underutilization charges for a few more quarters, but the amount is decreasing as we increase the production. The capacity is about 25% lower than what we had at the peak, but we can rebuild that capacity with a little bit of time. So the reduction in capacity is coming from taking equipment offline and, of course, reducing the number of people in manufacturing. So we had to lay off about 10,000 people through those very painful seven consecutive quarters of decline. So once we arrive at full capacity, and as I said before, we still have time, once we arrive there, we need to do two things. We need to put those equipment and maintain them, and we need to start adding a certain number of people. We will not need to add the same number of people we had before.
We can achieve that level of capacity with a lower number of people. So, but we still need to go out and hire 5,000-6,000 people. So it's big number. So between, you know, hiring people and putting that equipment in the manufacturing line, it could take, you know, five to six months.
So that seems like something that's going to happen over the longer term. You're talking about underutilization being a bit better. Is there any headwinds to the gross margin that you can think about over the next couple of quarters that would be worse than the offset of that? Is there anything from a pricing perspective, costs associated with HAMR, from other charges that you're seeing elsewhere? Can you just walk me through? Because there are a lot of moving pieces.
Yeah, no, pricing, I think, our strategy is fairly clear, right? We always adjust our strategy depending what happens in the market, I know it's a, it's a very competitive landscape, so we're always monitoring what's happening in the industry. But our point of view, profitability is still very low, so we still need to work with customer in improving our pricing, so we will continue to do it. So next quarter, we expect pricing, you know, a few points better and possibly even the following quarter. In terms of cost, cost should be better because underutilization starts to decline. The volume of HAMR is still fairly small compared to the total volume that we move, so the impact with P&L is limited. But we said, that what is going to be accretive to gross margin.
More volume of HAMR you will see in our result and better margin you should expect.
I want to switch gears over to the Non-HDD portion of the business. It's kind of guided to stay at these levels through the end of 2023. Do you have a view in enterprise going into 2024? I would say others at the conference have kind of talked about invisibility and probably bias more. What's your view with this?
So our Non-HDD has two main products: system solution, but it is based on hard disk, because inside the system, now we have mainly hard disk, SSD. We also have, now a fairly small business on SSD. I would say, you know, SSD will follow the trend, of the overall SSD, both in consumer and enterprise SSD. The systems, we think long term, this could be a business that, you know, can grow or finally depend how our customer will want to manage their architecture. If they want to manage their architecture, like a big cloud is doing today, they will use, they buy hard disk, and they buy all the other components, and they create their system inside the data center. If they want to simplify, they can buy a system that is basically maybe ready for usage.
Now, if they think they can get a better result with their architecture, they will continue to buy, as they do today. If they think system is an easier solution and finally, you know, similar results, maybe they will move more into the system. But it's basically serving the same, the same need in a different way. So it's only depending from how customers want to manage.
We're running out of time here, so I had one last broader one that I wanted to touch on. So you've continuously noted in the past the importance of being iterative. That question abated slightly as... But I look at even in a critical revenue environment and even in a pretty efficient working capital environment, in the back half of 2024, I do see cash levels still getting down to $low hundreds of millions, at least in my model. Maybe it's different from yours, but what additional actions can you take from here that help you protect cash? And which would you use first?
Yes, we are always very focused on the cash balance and generating positive cash flow. So even though last quarter was hopefully the bottom of our down cycle, we were producing a positive cash flow, and we also had to pay a lot of the restructuring costs in the same quarter. So I think now we have a good track record on generating free cash flow. As you said, we have to use a little bit of working capital in the last several quarters to generate also with that cycle. So when we go into the up cycle, maybe we use a little bit of cash to change a bit our working capital. But in general, now the business will generate strong free cash flow now through the calendar 2024, and then even much better to the calendar 2025.
Thank you for being here. I really appreciate it.
Thank you.
Have a great week.