So yeah, maybe I'll ask.
First of all, it's good to have the Form 10 out. Did you read the whole thing?
I read.
Parts of it. Okay, most of it. That's good. All right, good.
Yes, so our Form 10 was published last week for those of you who haven't had a chance to see it. In it, you'll see that SanDisk, the flash business, that will be spun out, will be capitalized with $2 billion of debt and $1 billion of cash, with also added liquidity of around $1.5 billion of revolving credit facility. And so that is sort of the plan for it to be capitalized on day one. Any cash above the $1 billion cash and cash equivalent balances will be dividended to the hard drive business to be used to reduce debt for the hard drive business.
In addition, there will be a 19.9% retained stake that WD Hard Drive will retain, and the intent of that will be to be monetized and exchanged for debt for additional de-leveraging within the first 12 months post-spin. The idea is to capitalize, let's say, within 12 months post-spin, both businesses will have stronger balance sheets and will be capitalized in many ways at a very high, sort of, in some cases, best in class. That will allow the whole system relative to today's WholeC o to have a lower level of debt.
Great. And let's sort of first talk about the HDD business. And your competitor put out a case saying that they're beginning to add capacity, but they're having some issues, and it's costing them revenue in the March quarter. So part of the thesis, at least on your side, has been that you're not going to add capacity, and yet they seem like they're sort of bringing some of those tools that have been mothballed back online. So how do you-I mean, I'm sure that you read the announcement. How do you sort of think about the capacity dynamic in the HDD business?
I mean, today we are operating in a tighter supply environment. We have a good supply-demand balance for our business. We continue to manage effectively through that. We don't see the need for us to add any capacity. We always, obviously, monitor where the demand is. We also have an approach of order-to-build with our customers that gives us good visibility of demand for the next few quarters. And with all these, we're still comfortable with where our capacity is. In fact, the build-to-order gives us good visibility also for the next few quarters, which is good for us to be able to manage our own supply chain, but also the supply chain of our partners with respect to components and other types of things that go into the hard drive. And so, as a result, we don't see the need for us to do anything different from a capacity perspective.
Great, and I think I asked you this, David, on the callback, but if the customers hear that you're not going to start to build a drive for them until they place an order and the lead times a year, they're just going to place orders, in theory. So how do you sort of think about the game theory that a customer might play in an environment where it's strictly build-to-order and you're really holding the line where you're saying, "I'm not going to expand capacity"?
I think we extended the lead time from 26 weeks to 52, maybe a couple of quarters ago, trying to get more visibility. I think our interests are aligned, quite frankly, with our customers, where we don't want to short the market, and they want to be able to grow their franchise. We want to participate in that growth. So it's just a matter of getting that supply-demand understood. By the way, we are adding capacity because we build new drives, new higher-capacity drives. We have a new generation of drives. We now have a 32 TB drive on the market. You couldn't buy a 32 TB drive until we release that. Customers are adopting SMR. That gives them all that additional capacity on a per-unit basis.
There are levers for us to continue to grow exabytes, which is really what we're trying to do, which is match our exabyte growth to the growth in the industry. The visibility from the customers is if we're only going a quarter at a time, it's very difficult to get that right on the high side and the low side. I actually think our interests are fairly aligned that these are big relationships. I think our customers know, and we know, that we're not just going to deal with each other for one cycle and then we won't see each other again. I don't think it's in anybody's best interest to kind of game the system, thinking that they're going to get ahead because there's going to be another side of that the next time around the block on this equation.
I think we're just trying to get an industry that is more predictable, and we can better utilize our capacity and deliver innovation to our customers so they can continue to grow their franchises. We clearly went through a traumatic downturn. We weren't the only ones that went through that. I don't think it's in anybody's best interest to repeat that. I don't view this as some kind of adversarial situation where we're trying to out-compete each other or something. I think we're just trying to say, "Look, if you can give us visibility to what your demand is," and I think the customers want to do that. Sometimes they have variability in that, and we work with them on that.
But I don't think this is something where we're trying to navigate it in a way that creates a situation where somebody feels like they're disadvantaged or overly advantaged and they need to do something that is not in the best long-term interest of the business. And the hard drive business is fairly concentrated on the demand side as well. So I think that the process is actually working quite well. And I think we are getting more visibility. And I think customers are working with us to give us more visibility so that we can do the right level of planning to support their business. And I think as we continue to get that more visibility, I think there's been a lot of questions on, "When are you going to add capacity? Are you going to add capacity?" I don't know what our peer said today.
I didn't listen to it. But I think that if we get conviction that there's demand a year from now, then we'll make the system support that demand. But we have to get that visibility to make that decision. Otherwise, we're the ones that are taking all the risk to increase demand that might happen in the future. And I think that everybody can understand that's not the best situation for any of us to be in.
Right. Just with respect to how far you can push HDD margins, I think I've had a conversation with you before where you said, "Look, you go out to your salespeople, and you say, 'We'll keep on raising price until the customer says that they don't want the drive anymore because it's too expensive.'" And you're not even close to that point from what you've said with that.
Yeah, let's look at it from a little bit different point of view. I think I've been very, very consistent that if you continue that the key to driving better profitability is to continue to innovate, right? That you need to continue to innovate. If you build a better product, you can go to your customers and share in the benefit of that value that's being created. This is essentially what we're doing. I mean, we have a very, very capable R&D team. That team has plotted out a hard drive capacity roadmap that, quite frankly, has worked spectacularly well. We are able to very predictably now. We've entered the 30s. We've been saying for a while we can get to 40 on the current roadmap. We're showing that we can do that. We've brought new innovations to the market like Ultra SMR.
Our customers have adopted those innovations, which gives them a lot of advantage, and as we continue to build a new product that has better total cost of ownership for our customers, we can share in that equation. It works for everybody, so that's part of it, where we're not just going back and saying, "Hey, we want different pricing for the same product." We're saying, "Hey, we're investing in all this innovation. And this innovation, if you adopt it, leads to better financial dynamics for both of us. And we're just participating more in that process." Now you get to the question of, "Well, what's the limit of that?" Because that's the next question that comes. Okay, where are you going? What's the limit of this, and I don't know if we know that yet because that gets to what is the marginal value of your product?
At what point are you destroying demand because your price as opposed to expanding the whole market? Storage is a business where you want to continue to expand the pie. You deliver better total cost of ownership as you go. And remember, even though our margins are growing up, the total cost of ownership of our products is going down. So everybody is winning in this. And I think it's figuring out, and this would be something for Wissam and Irving to focus on over the next couple of years.
I think that the value of the product that we're building is very high. It's a very, very complicated, very sophisticated piece of equipment that is providing an enormous amount of value to our customers. And we're very proud of that. And so the question is, where is the limit to that value? And we don't know yet. We're going to continue to work with our customers to find that point. But the key to all of it is to continue to innovate and continue to build better products and build a better value proposition to our customers. And then I think that's a much easier conversation. Everybody wins.
Great. Maybe just a last question that sort of transitions into the NAND business. But eSSD has been taking a lot of the incremental demand around AI. And I guess it's not really demand that would have gone to HDD. It's more just incremental demand that's going to NAND, not to HDD. So how do you think about HDD in the context of AI? Because obviously, anything that expands data, hot data becomes cold data. So it's ultimately good. But how do you kind of think about the leverage that the HDD business has to?
So on the highest level, I think that AI has again shown that the world has figured out another way to monetize data. Right now, I take all this data that I have, I put it through a large language model, I train the model, and now I can use that model in an inference phase for new use cases that didn't exist before. So the incremental value of storing data has gone up because I have another way to monetize it. In fact, I've been seeing stuff just in the last couple of weeks where people are talking about they don't have enough data to feed into their models to train it. So that tells me that, okay, people need to store more data, right? So we put out this piece of work called the AI Data Cycle that kind of demystifies how AI has an impact on storage.
You see this kind of process where it's bookended by hard disk drives. The vast majority of data in the world is stored on hard drives, right? The vast majority of data in the world will continue to be stored on hard drives. In between now, we've created this new use case for storage, which is large language model training and inference. I think that is what you're likely to see in the storage market is enterprise SSD is where you tend to see new use cases come up that drive demand for the product. It's easy to start to think that, well, that means that's the cannibalization of hard drives. That's not true. I mean, people didn't do large language model training five years ago.
And that use case came about, and it primarily benefited enterprise SSD for the actual training phase of the model. But once you get the model built and start using it, and now you start pouring out additional data, that's going to go back onto hard drives. You kind of have this virtuous cycle where hard drive is a place where you store the vast amounts of your data, where all the cloud providers store the largest, vast amount of their data. Now we have a new use case in AI for training and using models. We pull data out of those hard drives. We put it in a data lake that tends to be on enterprise SSDs. We then pull that data out of the data lake and feed it into a GPU complex for training. That's another enterprise SSD use case.
And then out of that, we build a model. We put it into an inference phase. Again, that uses maybe more client SSD capacity on a PC, maybe more capacity for NAND on your mobile device. And then eventually, you're going to create data out of those models, and you're going to store that back on HDDs. So HDDs is maybe a little harder to put your finger on because it's not a new use case. But when you're storing more data, the most likely case is that's going to go on enterprise SSDs. And when you come up with a new use case for storing data, that's more likely to be on enterprise SSDs. So if you just look at the new stuff, you tend to think, "Oh, enterprise SSDs may be benefiting at the expense of hard drives." But I don't think that's true. What you have is both of them growing, and more likely enterprise SSD just growing faster. That's a good situation.
Got it. Maybe that's a good transition on enterprise SSD into NAND. And you talked last quarter about enterprise SSD being more than 15% of NAND bits last quarter. Many of your other peers are quite a bit higher than that. Some are 50%, even 60%. Most of this, though, has to be made sort of on 2xx layer process. And you and the JV don't have that much 2xx. And what you have seems like you're using to make enterprise SSD. How do you feel now that NAND is going to be its own business? How do you feel about the mix of capacity that you and your JV partner have in the fab? It seems like there's a lot of 128-layer capacity that if you want to play in enterprise SSD, you're going to have to upgrade.
So a couple of comments on that. First of all, yeah, we have a traditionally client-focused portfolio. Very strong in client SSD, obviously very strong in consumer, strong in gaming, strong in mobile. And the pillar that we've been building out over the last several years is enterprise SSD. And we feel good now that we are building that out. And it's an ascendant part of the portfolio. And that's good. It's always good to have a growth factor, especially one that's one of the largest parts of the market. NAND is still 70% client business, 30% data center business. So we want to be thoughtful about how we manage the portfolio. But we want to have as much optionality as possible that we can mix our bits in the best possible way.
We talked about this a lot over the last three to four years that the way I think about the market is we want to have as much optionality as possible and then be as agile as possible because it's a market that changes very fast, literally week to week, quarter to quarter, a week to week, month to month, quarter to quarter of where the demand is, where the most profit is in the market, and mix in a way that gives you the best return for what's out there in the market. And enterprise SSD, we feel good about the progress we've made. We've still got a ways to go. We'll continue to make progress as we go forward. Now, as far as the nodal mix in the fab, the enterprise SSDs we're selling today are all under 200-layer nodal product. We don't have to have 200-layer.
For us, that would be BiCS 8. We don't have to have BiCS 8 for these products. When we talk about that 15% of bits, that's a mix of BiCS 5, BiCS 6, depending on if it's QLC, high performance, what SKU it is. We feel very good that we have BiCS 8 in front of us. BiCS 8 is a tremendous node. We'll build controllers for it, and those will be great products. We will evolve to that node. That's going to take CapEx to evolve to that node. That will be a demand-driven conversation of moving to that node when the demand is there to support the investments to bring that to market.
So you don't feel that you've fallen behind your peers?
No. I think I was hoping this was something we could put to bed in the downturn that the layers race. It's not about layers. Can we please put this behind us? It's not about layers. It's about performance. It's about all kinds of things. It's like more layers are not necessarily better. More layers means more CapEx. So we feel very good about where the portfolio is, our ability to sell into the markets with the node mix we have in the fab, and we feel very good about what the roadmap looks like. We got a tremendous node in front of us. We are starting to build BiCS 8. We are starting to ship some products on BiCS 8. So that's fantastic, but we don't feel like we have limitations on the portfolio based on what the node mix is.
I want to talk about cost because I've always been very impressed by your ability to bring cost down and still not spend that much money, and you have a sort of I think of it as more of a brownfield CapEx approach versus some of your peers have had more of a greenfield approach to CapEx, so can you talk about sort of how you approach, you and your partner approach CapEx a little differently than what some of your other NAND peers do?
This is a great strength of the JV and why the JV is so important. JV is important for a lot of reasons. It's a great construct. The company has been working together for over two decades. Several points about that. I mean, one is we have engineering teams been doing this a very long time, right? And that they've got a lot of experience in developing new nodes in the most efficient way possible. Number two, I think we really think a lot about CapEx efficiency when we design new nodes, right? You can think about how you design a new node in a lot of different ways. How many bits are you going to get? How much cost down are you going to get? How much CapEx are you going to spend? Those are dials you can set in the equation of building a new node.
The teams and the JV stays very focused on delivering incremental capacity at the lowest CapEx possible, right? Getting that right. This is why, again, I won't bring up the layers thing again because I'll just perpetuate it. Yeah, that's something we feel very good about. We feel like that we've been able to bring capacity, continue to grow our capacity at roughly a third less CapEx than the industry in general. That's quite a feat. That's something you would expect from one of the largest providers of NAND in the world. When Kioxia and I when Kioxia and us are together, we have enough scale to invest a lot in R&D. That shows up in great products.
And can you just talk about just the supply-demand environment right now in NAND? I feel like NAND has been benefiting in some ways by all the air being sucked out of the room by AI and HBM. And so if an integrated producer that makes both DRAM and NAND wants to spend money, they're going to spend it in DRAM before they spend in NAND. However, it does seem like there's some increase lately in NAND spending. And some of the suppliers are talking about NAND wafer fab equipment being up 2x next year. So the first really big year, now granted off of a low base, but being up quite a bit next year. So the question really is, how do you feel about the supply-demand dynamics in NAND? And you had a pretty easy pricing comp in September, and it's back up in December. But sort of as you go into 2025, how do you think about that?
Yeah. I think the NAND, as far as just the near-term stuff, we'll just go back to what we said on our earnings call. I don't think it's new news. I mean, the PC smartphone markets are kind of at a normalized industry level and just kind of going along, waiting for incremental demand to show up. Gaming is something we shipped a lot. We prepared for the holidays in the last season. It'll be weaker this quarter. Enterprise SSD is the strong part of the market, right? A lot of demand in enterprise SSD. So I think that'll play out quarter to quarter. I think the NAND business, like I said, is something that moves very quickly. It gets choppy. It wouldn't surprise me we see a couple of choppy quarters here.
But, big picture, to your question, we still see an environment as we go into 2025 where we see supply under demand, right? I think they've gotten closer over the last quarter or so, but we still see in general, more demand than supply for the year. We look at the CapEx spending, and there's been historically low CapEx spending in the market for quite a while now, which again is not surprising, coming off of the downturn we had, and so we think the setup is good for the year. I think that when we get to the end of the year and we add it all up, it'll be a good year in the NAND business. As far as when does CapEx spending come back? I think that depends on demand. I think this is, again, we talked a lot about this in the new era of NAND.
We're not going to roll the nodes forward just because we can, right? I think that was like a 2D era way to manage the business, like manage it from the supply side. The nodes are very efficient now. It's very easy to oversupply the market if you manage the business that way. I think you have to look at it from the demand side and say, "Where is the demand? Is demand really going to be there?" And if we see the demand, then we'll change the nodal mix in the fabs to meet that demand. Changing that nodal mix will take CapEx. And so I think we're going to see what demand looks like as we go throughout the year.
Let's actually talk about that. If I look at bit growth, it was up 30% in 2021, flat in 2022, up 20% in 2023, and flat this year. If I just smooth all that out, it's sort of, let's say, mid-teens%. Maybe it's 20%, but it's mid- to high-teens%, something like that.
I think that's a good number, yeah.
But it does sort of argue that you have easy comps this year because it's flat. So it does, I guess, go back to what you're talking about where you should get a better bit growth year next year, and let's see what happens to CapEx. Is that your?
Yeah. I mean, yeah. And in your numbers, you got to put inventory in all of that.
Of course, yeah.
Yeah. Because I mean, it's different for everybody, and there's a lot of inventory built up. But yeah, I think that we see a high-teens number is a good number. I mean, and I think that I think we're coming off of this idea that the market's going to grow 30% every year, and we need to invest to that. That's not realistic. It's a big market now. I think a high-teen number is a good number. We'll invest around that. And as we see that real demand showing up across the business, we have the levers to invest when we see the demand.
So David, you're going to, of course, run the NAND business. And I know you have very, very strong relationships with Kioxia and Japan. But the question is, how will you run the NAND business differently now that it's going to be standalone? I mean, I'm sure there's things that will be harder. I mean, otherwise, you wouldn't have put the businesses together in the first place. But maybe there's other things that will be easier that you wouldn't have been able to do if the businesses were combined. So how do you think about how you might run that business differently than it's been run?
So first of all, I'll say I'm really looking forward to running the business. I think they're both great businesses. And I'll also say I have an enormous amount of confidence in Wissam and Irving to run the hard drive business. I think that's going to be fantastic. I'm going to miss it, but I know it's in very, very good hands. The separation is not about running the business differently. It's never been about running the business differently. The separation is about getting credit for how well we're running the businesses. We feel the assets are undervalued. I've always thought from the beginning that we could run these businesses as best-in-class assets if they're together. This idea that because they're together, we can only do one thing. I've never bought into that.
And I think we've shown now over the last couple of years, these businesses are executing as best-in-class assets. So now we're going to go through the separation, and we think we're going to get the value out of what they deserve. I don't think we're not doing it because we think we want to run the businesses differently. So I think the big difference that's going to happen through this transaction is what Wissam started with, which is we're going to end up with much better balance sheets that are appropriate for the businesses we're in. And this is something we've both had a very big focus on. It's one of the things I put a focus on when I first came into the company, which was get a stronger balance sheet underneath us, focus on that.
This transaction is going to put both businesses in a very, very strong position to match the portfolios that we've built, the teams that have been established, the great products that we're building, and the profit we're able to generate off these businesses. I think I'm looking forward to that part of getting two very, very solid assets out there in the market.
And how do you think about it? I get calls all the time about what the NAND business might be worth, and I'm not going to ask you for a number. But you're not the only one who's floating a NAND business either, right? So there are other NAND businesses that are going to be floated publicly from other producers.
That's great.
So, how do you think? I guess I go back to your question, and this is just the question that I get asked a lot, is that unless you're going to run the businesses differently. I understand that each business is capitalized a bit better than it was before. But unless there's going to be something different about running these businesses, why should the market put a better value on the combined entity? And particularly in light of other investment investable assets coming to the market in the NAND space.
Why should the market value them differently?
Yeah. I mean.
It's kind of an interesting question given all the questions I've gotten over the last several years, but.
Well, I mean, there are many people that believe that the market is efficient and the market was looking at the company in sum of the parts way anyway, and that just by separating the businesses doesn't really create much value. That's not necessarily my view, but there is a view out there. And so I'm just asking that question in the context of there are other for all the pool of money that wants to invest in NAND assets, there are also other NAND assets to invest in during this time as well.
I think that we were in the current configuration, there's people that want to invest in the dynamics. Wissam can comment on this as well. The dynamics of each business are different. They're different businesses. Although they're both storage businesses, they do it a different way. One's semiconductor storage, one's hard drives. In some sense, you have a lot of the same customers. The dynamics of the business are different. The markets are different. The capital intensity is different. The cyclicality is different. The growth prospects are different. The size of the markets are different. And I think that my observation after doing this for the last five years is there's different sets of investors that have an appetite for either one. And when you put them together, you limit the size of the number of people that are interested in that combined asset.
It doesn't have anything to do with how the asset is being run. It's the fact that, hey, I want a business with lower cyclicality, better cash generation, lower capital intensity, proxy to the cloud. That's a hard drive business. Somebody else wants a business that is in a bigger market, maybe more opportunity. They know it's going to be more cyclical. They know it's going to require more capital. On the other hand, it's a bigger pond you're playing in. That's a very different kind of investment. And I think when you put the two of them together, you limit the number of people that want the combination.
Great. Thank you for the time. Appreciate it.
Thank you very much.
Thanks for having us.
Thank you.