Please welcome Chief Executive Officer, David Goeckeler.
All right, thank you, folks. Look, I really want to thank everybody for joining us here in person and online. This is an important day for us. We've been working very hard to get here, and we have a lot to talk about today. We're really looking forward to it, and we really appreciate all of you spending time with us. We'll have plenty of time for Q&A at the end as well. I'll go through the setup for the day in just a minute. Before I get started, I do want to make some introductory remarks. I think that before we get into the content of the day, I think I'm going to comment briefly on the news you all saw about Elliott Investment Management taking a stake in Western Digital.
I'm gonna actually read a statement to make sure we get this on the record, about what we think about this, and then we'll move on to the content for the day. A little more than two years ago, I joined Western Digital to create value. I strongly believe the company was well-positioned to capitalize on key forces shaping the technology landscape that is a critical part of all of our individual lives and all businesses. I have an even more conviction of this now, which we'll discuss in more detail today. While considerable energy went into rebuilding and refocusing the company, almost from the beginning, I also explored broader options to reposition the company to better capture the value created in the attractive markets where we operate, and continue to do so.
This is with the full support of the board, in consultation with advisors, and at multiple times included external interested parties. We maintain an open dialogue with all shareholders, and we genuinely value their input. On the topic that Elliott raised, the potential to unlock value by separating our two businesses, the board and management take this idea seriously, and we welcome shareholder input on the proposal. As with all of our shareholders, we look forward to discussing Elliott's proposal and constructive input with them directly. Having said that, today is not about discussing the pros and cons of various potential corporate actions or financial structures. This session is about the business we have today and how we're working to drive growth and shareholder value through effective execution of our strategy.
Over 65,000 people come to work here every day excited to drive that strategy, and we're going to talk to you about what we as a management team are doing to capitalize on the many opportunities before us today, even as we continue to analyze other structural options for the business. All right. That'll be in the, of course, in the transcript. Just wanted to make a comment about that because I know you're all thinking about it, and it's an important development. All right. Let me set up the day for you. What I'm gonna do is give you an overview of our business, kind of why I'm as excited about this opportunity in front of us all around data. Kind of what I see is the fundamental architectural underpinnings for what's driving that opportunity.
We'll talk about the segments that we operate in. I'll give some high-level thoughts on each of our businesses. Well, before I get to that, I'll go through the organization and kind of all the changes we made in the last couple of years, give you some context around that, and introduce the broader leadership team. I'll give some high-level views on each of our two businesses, flash and HDD. Then from there, we'll have Rob come out. He will talk about his business in flash, go through the strategy there, what we see the opportunity as, what we're managing to. He'll end that with giving you a model for how we think about what we can drive that business to over the next several years.
Ashley will come out and do the same thing for the HDD business. He'll go through that, how we think about that. We'll show you a model for how we're thinking about that business and what we're driving to. Siva will come out, and he will talk about the fundamental technology underpinnings of both businesses. HDD business, how we think about how we drive innovation there. I think all of you saw yesterday. It was a fun day for us. It was an enormous amount of innovation that we launched, things we've been working on for literally a decade, all coming together in products that can be launched into the most sophisticated and demanding data centers in the world. Siva will talk about how we see that in the long-term roadmap, in that technology.
He'll go through the same thing in flash, and really focus on how we see the fundamental differentiation that we have, especially with our JV partner, Kioxia, what allows us to drive a very capital efficient business, and really explain why that is. Then Wissam will come out, and he will bring it all together at the corporate level and kind of give a review of kind of where we're at and the results we've driven, and then talk about how we think about the corporate model going forward over the next three-five years. That gives you a little bit of setup. There's a lot to get through. We'll have plenty of time at the end for Q&A. With that, let's jump in. All right. There's a lot of
You know, let's talk a little bit about our markets. There's a lot of talk in the world about data. There's a lot of data in the world. There's a lot of way to talk about data, how much data is being created, the exponential growth of data. What I like to think about, and certainly those things are certainly all true. You know, in many ways, there's no lack of data in the world, and nobody expects there to be a lack of data in the world. What I like to think about is what is the fundamental underpinning of why that is the case.
This has a lot to do with why I joined Western Digital a couple of years ago, is I have a fundamental belief in this architecture that the world has settled on, and it's essentially an architectural underpinning that everybody else in the world is—it's a platform, if you will, that the rest of the world is building on. I think there's some very interesting points to it, that have fundamentally changed how we think about technology from, let's say, a decade ago. That is a very simple model of we have this world of ever-increasing intelligent devices, 6.9 billion devices out there. Rob had that data yesterday. Those are powered by a cloud. The cloud is something that is being driven by the most powerful technology companies in the world.
These two things are connected by high-speed networks. That architecture allows a fundamental rethinking of how applications are built, how software works in the world than the way things used to be. It basically drives this exponential growth as each of these three elements of the architecture evolve independently. Devices continue to get more powerful through the integration, sensors, cameras, the general semiconductor process driving things forward to make everything we use more intelligent, more powerful with an ability to create more and more data. I think video is a good example of that. I mean, everybody can record 4K video now themselves on their device. There's just an enormous amount of data being created through these devices. They're just gonna get more and more powerful.
They're connected to a cloud which is ever more powerful. Some of the most important and powerful technology companies in the world are investing heavily in that cloud, and in many ways just democratizing access to the most sophisticated software in the world. Every company in the world doesn't necessarily need to build it all themselves. If you want the most sophisticated AI technology in the world, it's there for you in the cloud, and there is somebody behind that driving that. Every six months, there's just more and more features for everybody to consume. Of course, networks just get faster and faster. Obviously, a lot of conversation around 5G, Wi-Fi 6, just moving on.
As the ability to connect these two things together increases, the experience gets better, opens up more avenues for innovation, more data is created and around we just keep going. It's really is a very, very powerful architecture, and I think it's just gonna continue to drive things forward. This world we live in of generating more and more data is just gonna continue to increase. The real question is, how do we harness that? How do we capture that data? And then how do we do something with it? That's what we're gonna talk about here today, how we think about that. I think Western Digital is very well-positioned to capitalize on this architecture. We can play from the intelligent device to the edge of the network, all the way to the heart of the cloud.
Everywhere you're gonna store data across this architecture, we have the ability to play in those markets. A lot of what we've been doing over the last couple of years is really fine-tuning our portfolio and really upping our execution to make sure we can play. You know, we're very intentional about where across that architecture we wanna play, where is the most opportunity to capture value in getting our portfolio structured in a way that we can get the best result possible.
You saw some of this yesterday with this new innovation that's coming out, you know, that a lot of that is most of that is simply because of the changes we've made in the last couple of years and the ability now to bring this, you know, what is an amazing amount of innovation on one day from this company. It has to be the most innovation that's ever been announced on a single day. We have ability to access the market all the way from the cloud titans. Very clearly, if you're gonna build a cloud in the world today, you're going to use hard drives for about 90% of your storage. So anybody that's gonna build a cloud is gonna be our customer, in that there's a lot of option value in that. We'll talk about that.
To OEMs, company's heritage as a client company means we have a deep relationship with every OEM out there, very large channel distribution business, and then on top of that, an at-scale consumer business that allows us to access every single consumer in the world. We believe that the company is well-positioned. It's a question of how do we get the portfolio tuned around that, and what is the strategy to go after all of those markets. Again, throughout the day, you'll hear more about that, and again, the fundamental technology underpinnings that we believe drive differentiation for the company. All right. If you look at our end markets, at the beginning of this fiscal year, we believe we simplified the way we look at the business.
We just think about these three segments that we operate in, the cloud, client, and then consumer. If you look at the long arc of the company, you see what probably should not be too surprising is we're moving more and more to a cloud company, a cloud-driven company. The consumer business, I think, is we'll talk about that. It's a little bit of the secret sauce of the company. I mean, that will be there for a very long time. In client, there's a lot of very good opportunity in client, but quite simply, there's just a tremendous amount of storage growth in the cloud. If you look at us over what we're gonna talk about here, this forecast period, you can see by the time we get to FY 2025, where nearly half of our business is driven by the cloud.
What I'm gonna do is just walk through these three segments, how we think about them at a high level, and then Ashley and Rob will go into a little more detail. Each segment plays out in either more or less intensity, depending on if it's the HDD or the flash business. All right, let's start with the cloud. Insatiable need for capacity. I think the amount of data being created. We just talked about this amount of data being created. The reality is our big customers would like to store more data than they can. It's not a demand side problem. It's a question of can we economically store it. We're constantly working on this question with them, how do we drive a TCO proposition that allows them to store more data.
It's not a question, what I said earlier, this architecture we're all building on and every company out there is leveraging to innovate, just drives an enormous amount of data. The question is how do we store it? How do we capture it? How do we harness it? This also mean that the workloads are growing in complexity. This is an important point we'll double-click on a little bit more. It's not really a one-size-fits-all kind of market. The cloud market is segmenting into different capacity points, different requirements, different pieces of equipment that we need for our customers to continue to build out this critical infrastructure. Just to get grounded in the numbers real quick. It's a big TAM, $385 billion. Storage is 29% of it.
You can see the HDD TAM is about 70% in the cloud. The flash TAM, 31% in the cloud. Are driven by the cloud. If we roll this forward a couple years, calendar year 2026, you still have the 29% storage, but you have another $50 billion-$60 billion of TAM that's been created. Also what's interesting, of course, is the percent of the TAM in each of our franchises is increasing in the cloud. You have a growing TAM, and you have a growing segment of that TAM that's addressable by our portfolio.
You know, it's not surprising that when you look at how you add this up across the whole portfolio and how we're innovating, that this becomes a larger percentage of our business and a big focus for our innovation portfolio. All right, we're well positioned in this. Now, this has been an interesting conversation for a number of years. How do these two technologies play together? Look, there's no doubt in the client market we'll talk about this. Flash and HDD are substitute technologies. The decline of HDD in the client is obviously undeniable. All of the laptops that we all have here today have client SSDs in them. As a company, we played that transition, I think, quite well to emerge with one of the largest client SSD franchises.
However, in the cloud, these are complementary technologies, and they will be complementary for a very long time. We'll talk about that. You know, two things can be true, they're both going to grow. You can see here that NAND or flash in the cloud, 37% year-over-year growth, obviously a very attractive market, something that we've been very focused on. You know, this is why we've been talking about for the last year plus, just been very transparent about getting an NVMe-based enterprise SSD qualified at these customers is extremely important to capture that growth we just talked about. We've been transparent about that process. We've had success with that process over the last year.
Again, the focus that was brought by the way we organize and said, like, the flash team's gonna focus on flash, the HDD team is gonna focus on HDD. We brought that focus about 18 months ago now, and it's working. You're seeing it show up in the products we launched yesterday and the focus teams around, like getting. It's not just ROI-based investment, all the things that general managers do, it's also a lot of focused execution. We've seen that. We're now one of three NVMe SSD suppliers supporting the hyperscalers. By the way, in this presentation, you'll see kind of hyperscalers and cloud titans used interchangeably. Different people have different terms for that. Very good growth there.
We've pivoted the portfolio in the last year and a half and actually delivered what is a very, very tall order from an execution point of view, is to build one of these products that will be accepted by a cloud provider. I'll show you a chart in a second about why the qualification process is so stringent and very difficult to get through. Of course, HDD, again, the backbone of the cloud, what carries a large percentage of the storage will continue to grow, continued exabyte growth there, 30% or more year-over-year exabyte growth into the future.
I think what you saw yesterday if you were watching was kind of a, you know, multi-year product strategy that is now showing up in the way the portfolio is back to a leading position, where we're leveraging what Western Digital has always been known for, which is leading areal density in the ability to deliver not only a 22 TB drive, but a 26 TB drive at the same time. We've been talking about SMR for a number of quarters now. It's being widely adopted by the biggest customers, and the ability to make such an important step forward, I think speaks well to our ability to be very competitive in this market. If you look at this data. I thought I would share this data. This is a single customer.
This is what it looks like when you get qualified for an enterprise SSD. I think what this says is, this is why the qualification process is so hard. Because once you get through the qualification process, you start deploying at scale, like literally the next week. This chart is actually very convenient because the qualification here happened in the first week of January, so we have a full year of deployment. You see how you take a, you know, what was an $800 million relationship with a single customer to a $1.4+ billion relationship with that same customer, and how fast it grows once you get qualified on an enterprise SSD. You know, there's a very good reason why everybody asks about this technology all the time.
Rob will get into a little bit more detail here. I know there's a question like, how much market share can you gain? Where's the ending position in all this? That's a fairly complicated question because we have a broad portfolio and a lot of optionality in our portfolio, and we don't just think about one particular segment is what we're playing in. This does show you why we work so hard to get through these qualifications, and how important the cloud customers are to us, and vice versa. We will exit this fiscal year with four relationships of $1 billion or more with cloud vendors around the world. Very big market for us and a very big position.
It's pretty clear as the cloud grows, we grow. I think that's where we've been trying to position the company. You know, clearly with our HDD franchise, that is a cloud business. There's a long tail of a client business, but all the action is in the cloud. Then now getting the portfolio in a position where we can participate in the mainstream parts of enterprise SSD growth in the cloud is a good place to be. I'll also say there's a lot of optionality that comes out of this as well. When you have a relationship this big with a customer, there's lots of stuff they ask you to do, right? Addressing their archival storage opportunity. Again, our big customers wanna store a lot more data. They have a lot of data. They tell us their customers wanna store more data.
The question is it economical? Do we have the ability to do that? How do we think about that equation? You know, there may be different access times for that data. There may be different ways how often we wanna use it. So there's this constant dialogue that's always going, like, "How can we use your technology to solve a larger part of our storage needs? You're our biggest storage provider. You obviously have a long history of innovation as a company. How do we work together?" I think this is a big model in the, I think as we all know this, in the hyperscale cloud titan world, this co-development model of how do we work together to innovate is a very, very important model. The same thing applies to a wider range of SSDs and HDDs.
As I said, these are not just one product SKU. This is things are splitting into multiple segments and different requirements from provider to provider to provider. It's not just a single product. Each company builds their cloud to different requirements. They have different ways they use the technology and different levels of intensity, different densities, different speeds they want. This is very, very clearly segmenting into quite a complicated market. One thing I'll say, these big vendors also have relatively large consumer franchises, and there's a lot of option value there that goes on about having a front row seat as innovation happens, whether it's around home automation, tablets, gaming, VR headsets. There's a lot of follow-on business that we develop out of that.
We're not gonna go into that in great detail today, but there's, you know, roughly three-quarters of a billion dollars of client business that's simply attached to this business, and that's grown quite a bit over the last couple of years. Not only is the cloud business grows as we grow, but it leads to follow-on opportunity across the entire portfolio. All right. Let's talk about the client market. Again, I would say this is big-time heritage of the company. The company has played the transition of the client HDD to the client SSD, I think, very well over the last several years. It has ended up with a very compelling client SSD portfolio. I'll start up here on the right. You know, the PC is the anchor point of the hybrid enterprise.
I think two years ago, before the pandemic, the number of PCs everybody's expected in the world was quite a bit lower than are actually in the world today. I don't think it's going back to pre-pandemic levels. There's always gonna be some ups and downs given the new point we're at. If you had a chance to see what we talked about yesterday, very interesting discussion between us, Intel, and Microsoft about just the enormous amount of innovation now that's gonna go back into the PC given it's the anchor point of this hybrid world for everybody. There's a whole bunch of innovation now that point of the architecture is gonna continue to drive evolution, and storage is a big part of that.
Again, take a look at what that interview yesterday. I think it was quite interesting to talk about how everybody in the ecosystem is thinking about this device now and how we have to just make it a lot better to really enable the hybrid model that we all lived through during the pandemic. I don't think anybody would say it was a perfect model. It got us through, but there's enormous amount of innovation that can continue to go back into these devices now. As the largest supplier of client drives, and you may ask me, "Well, how are you claiming to be the largest supplier?" When you count the consumer part of this, which is a lot of drives sold that way and through distribution, we are the largest supplier of client drives.
It's a very good part of the portfolio and a lot of work that's gone in over the years. Gaming, I'm gonna do a little deep dive on gaming at WD_BLACK, but this is just I think an example in the flash market of this kind of evergreen market where there's new applications that come around every it just seems like all the time there's a new application, new devices. What I said earlier, this architecture that we're building is these intelligent devices connected to the cloud, and the experience between those two things now fast enough that you can build on top of that just means more and more interesting devices, and gaming is a big part of that. How do we think about gaming?
Flash has come to gaming now, and we'll talk about our ability to play that market. Of course, the automotive, mobile, embedded. Let's just say there's a lot of opportunity in the client market. It gets somewhat fragmented, but where we maintain what we wanna do is, first of all, have the absolute best technology that we can play at any market. Especially in the mobile market, that's very, very important. Like, you have to have the absolute best technology. It's kinda like what I talked about in the hyperscale market of getting qualified.
If you can get qualified at the very, very top, mobile providers, that is a very difficult task and to stay there, and that's something that we have done for quite some time, and we will continue to do, but we will engage in that market in a selective way based on our ability to drive the best profitability across the company. Let's look at gaming real quick. I think what's interesting about this market is we have two ways to play this market, and I'll talk about consumer in a second. We have the ability to basically play in the console itself.
Because this is such a premium market where people really value a premium experience, over the last three years we've been able to create a brand, WD_BLACK, and again, the WD_BLACK is not just one product, it's a whole range of products across gaming consoles, from everything from the casual gamer to the very, very extreme gamer, that gives them the absolute best experience possible, and the product is built in a way that actually enables that best experience at a very detailed level. Then we're able to drive that product back through our consumer reach and go direct to the consumer.
Play both the OEM side of this and the consumer side of it and really use our consumer franchise and our ability to build brand and kind of what I'll show you is a relatively big machine here of how we operate a consumer market to create incremental value for us. This WD_BLACK piece here is some of the most profitable products in the company. This really gives us a way to drive more volume in very attractive markets and do it in a way that allows us to drive premium experience, premium value proposition for our customers, which means premium value for us. I mean, that's where any kind of value capture starts is with innovation.
You have to bring differentiated innovation to the market, and that allows us to bring differentiated value to all of us here that are investing in the company. All right, let's talk about that consumer market for a minute. One of the things you see around us here is we're doing some more branding to try and pull all of this together under the Western Digital umbrella. We have very recognizable brands, SanDisk especially is a very recognizable brand in the entire technology space. But in flash, it's a brand where we drive differentiated share and differentiated value out of those channels. We've just brought SanDisk Professional for kind of the high-end creator.
The WD brand, which is a broad brand based on the colors, WD green, red, purple for different segments of the market. That strategy has worked very well across HDD and flash. Then, of course, WD Black, which I just talked about, which is a brand we've created in the last three years. Obviously, wrapping these all under the Western Digital umbrella. You know, this is—I think this creates option value in the future as new segments emerge that we think are significant, where we can bring innovation that allows us to create additional brands under this umbrella. Over 300 million products sold per year through this channel, so this is an at-scale business. 45%, you know, the one segment in this you say is really on fire is this external SSD segment.
I think something you'll hear from Rob, like, the flash business is becoming an SSD business. That's just the reality. You see that everywhere. We just talked about it in the cloud, enterprise SSDs, in the client SSDs, in gaming, SSDs that go into your gaming devices, and external SSDs are the largest-growing segment in the consumer business. Flash is a big-time SSD business. The other thing I'll say about that is, a lot of times people ask me, "You know, Dave, can flash really continue to grow 30% a year, 35% a year from a demand perspective?" In calendar year 2021, if you look at...
Now there's obviously a lot of hard drive products in here, but if you look at the flash content across, you know, that scale of devices, it was up 34% year-over-year in the consumer market. You know, whatever angle we look at the market from, we continue to see this 30% kind of number growth, including across, obviously, a very broad spectrum of consumer devices. I'll just highlight our reach here. We do operate everywhere we legally can in the world, 300 million products sold, 351,000 storefronts, and 123,000 partners. Really, I think very good market reach. There is a really top-flight team behind this that manages it on.
Literally, if you're managing e-tail on a day-to-day basis at the scale we operate at, you're obviously on every platform in the world, all around the world. You're managing promotion on an hour-by-hour basis. The ability to have this infrastructure in place, and then when we see incremental opportunity like WD_BLACK emerging segments, the ability to not only create the brand, drive the innovation, which is a differentiated experience for our customers, but then have the reach to actually bring this to market around the world and capture value in that, I think, is a unique position that we have and something that we work to take advantage of. All right. As I said, Rob and Ashley are gonna go into this in more detail about how it applies to their particular business, and the strategy to leverage that.
Let me talk a little bit about our transformation as a company since I've come here a couple of years ago. We've been on quite a transformation journey. It's working. We'll talk about the results, but I'll give you a little bit of background into how I've thought about that, what is the strategy, what we're trying to do here, and why we're organized the way we are. This is really in one simple chart, kind of how I think about the company and the way we're organized and what we're trying to do. It was very clear to me when I came here that there was an enormous amount of tremendous assets in this company. The amount of innovation we're able to drive is incredible. Company. Siva Sivaram will talk about this. The ability to do very long-term R&D.
Not just R&D, but bring that innovation to market in a very disciplined way that drives value for our customers. The value of our patent portfolio is very strong. Our ability to continue to innovate is very strong. Things like what you saw, OptiNAND yesterday, UltraSMR, things that make very big impact on the business, that have been going on a long time. But we needed to execute better, right? We have two businesses. You can't just think about them the same every day. Just gotta think about how we're gonna organize the company to actually capture all this value that we're driving. That really started with over here on the left, like, product and operational focus is critical if you're gonna drive a innovation-based company. If you're gonna drive a product-based company, you have to have focus.
It's just as simple as that. Nothing happens without intense focus on a very specific issue and market, and I thought that was absolutely the first step we needed to make. At the end of the day, we're a product company. Like, we live and die based on the quality of our products and the quality of our innovation. Like, to really get that focus and make sure we're making all the right decisions was to go to business units, bring in general managers that know how to do this at scale, know how to make ROI-based investments, get the portfolio tuned right, and then drive execution through focus. I think, quite frankly, everything you saw yesterday, maybe the biggest day in the history of this company as far as launching innovation across the entire portfolio.
Like, we went to bed Sunday night, and the state-of-the-art in a hard drive for the cloud was 20 TB. By the time we went to bed on Monday night, the state-of-the-art for a hard drive in the cloud was 26 TB. That didn't just take one day for that. That was the culmination of Siva will talk about this, over decades of innovation. To actually get it to the finish line took an enormous amount of focus, and that's what Ashley brought to that organization. Same thing with Rob. I mean, we announced everything from new enterprise SSDs to WD_BLACK, to a new SanDisk Professional series of products. Just enormous amount of innovation across the portfolio driven by this ability to focus.
Same thing applies on operations and supply chain. Different supply chains, different operations for each business. I started on products. This is. We're just getting going on this part of it. I'll talk a little bit about this order of battle. Again, very specific focus there to drive the best possible outcome and drive innovation, drive cost in our portfolio. In the middle was very clear to me, where we do have you know we can operate as one company. We don't need to split these two businesses to operate like this. We could, but you know, very clearly, you know, if we have these two franchises, we wanna have one go-to-market team.
You know, that relationship I showed with that cloud titan that's over $1 billion, you don't wanna have, like, three or four people in there talking to them or two people about different parts of the portfolio. You wanna have one person driving that relationship. We have a very good team that does that. What I just talked about in consumer of driving this worldwide presence, like, you don't wanna do that twice. You wanna do that once across the whole portfolio. It was very clear to me that that was a place where we had a lot of strength. The leader of that team is the person that was here when I started, Gerry Cagle. He does a very good job. The team is world-class at what they do. Also the core technology evolution. Siva Sivaram will talk about this.
Like, I think you'll see that when you think about core technology and core innovation, like, having a larger team that does this across multiple domains is very important because you get this cross-pollination of different people with different levels of expertise coming together to drive innovation, that's where innovation happens. You know, the company has a long heritage of that, but one thing that's very impressive about this company coming in as somebody with an engineering background, the number of people we have in this company that are physicists, material scientists, experts in magnetics, like PhDs everywhere, semiconductor expertise at the highest level to build NAND flash memory, one of the most complicated disciplines out there.
We have that full dynamic range of expertise, the ability to bring those people together, drive innovation, and when you get the fundamental innovation right, hand it off to the business unit, and they productize it. Maybe one way to think about that, if so, you know, a business unit GM thinks about their roadmap for like the next three years, and then somebody else is thinking about how do I feed fundamental innovation into that to build better and better products. There's an overlap there, but also kind of an area of focus. You'll see when Siva comes up here, like, he's thinking decades into the future about what our technology roadmap is, not just what's gonna happen in three years, but what's gonna happen five years from now, 10 years from now, because these disciplines are extremely difficult. They're physical sciences.
They're not things that just happen in a year or two, and we have to think very long term. That's enterprise-wide capabilities. There were several parts of the organization we just needed to enhance our capabilities and organizations. That kinda lays out the order of battle for how we rebuilt the leadership team. You can see I joined in first quarter of 2020. Lynne Cox, who heads marketing and communications for me, joined the same time. If you're gonna take a new job, it helps to be able to communicate well with a large group of people, and it's like a lot of the stuff you see around here, a very professional team has been built to up our game significantly. You can see here, like the first thing to do is bring in general managers.
Like I said, product focus is very important in a product company, so you wanna get your products right. Ashley and Rob came in just a couple of quarters later, and like I said, everything you saw yesterday was a result of these guys coming in here, getting their teams formed, bringing the focus, like tremendous amount of good people in the organization. Don't get me wrong. Incredible engineering talent, a lot of good ideas, but bring that focus to actually bring things to market and, you know, relatively 18 months later, you all have the biggest launch in the history of the company.
Don joined corporate strategy, brought in Christine to rebuild our people organization, and then just within the last quarter or so, Wissam joined us in finance, and then Irving Tan joined us in global operations, very senior executive, that is actually based in Singapore. I thought, you know, we have 45,000 people in Asia in all of our manufacturing facilities there. It's a pretty good idea to have somebody in Asia that understands the region and is very, very sophisticated in how to think about operations. As you can see, you know, you don't show up on the first day, change 70% of the leadership of the company.
You have to go about it in a very disciplined and structured way, and that's exactly what we've done every quarter or two, bringing in another senior leader to continue to build out what is now a very, very sophisticated and leadership team to drive this company at the scale we need to drive. There's some folks here, long-standing leadership. I talked about Jerry. You're gonna hear from Siva. Siva is one of the best memory technology experts in the world. I think that will be obvious when you hear him talk. Very, very good at what he does. Then Michael Ray, our GC, longtime GC, just does an excellent job. One of the things I will say is that Siva and Michael have very deep expertise on the JV relationship. The JV relationship is extremely beneficial to us.
It has a long history of 23 years. Something doesn't last for 23 years unless it is very well structured and very beneficial to both sides. Understanding that relationship and that JV, both from a relationship point of view, which is obviously extremely important if you're gonna work with somebody for 23 years, and also from a structural point of view because it is a relatively complicated JV that was very well structured that drive both companies to continue to work together on an ongoing basis to both of our great success, quite frankly. I think you'll see that from Siva.
The fundamental technology that we drive and have been driving as a company for 23 years, we will show you it is demonstrably the best technology and the most capital-efficient NAND flash technology in the world. We also always look at the board as well. We have a tremendous board. We talked about adding some expertise. Obviously, I thought adding somebody that has expertise in Japan, just like we just hired a head of operations in Asia. Miyuki is a very senior executive that has been based in Japan, obviously from Japan, and also has very good knowledge of the entire Asia region as well. Again, we are a semiconductor company, and the fabs that we have in Yokkaichi.
Certainly Yokkaichi is one of the largest fab complexes in the world. Actually, I haven't been there yet. I've been here over two years and haven't been able to travel to Japan, but I look forward to doing that in the next month or so. The ability to have somebody like Tom that has semiconductor fab experience being the CEO of GlobalFoundries brings a lot of expertise to the board as well. A lot of focus on the leadership team. Very intentional about who was brought in, people that have scale and have done this at very large scale, as well as a set of expertise, the mix of people that have been here a long time, people that bring new perspectives, people that bring different skills.
I think we've landed in a spot that is very, very. I'm very happy with the team that we've built. You know, like I said, it's just Wissam, I'll give Wissam an enormous amount of credit. The guy showed up at the end of January and said, "Hey, we're gonna have an investor day. Time to get ready." He'll be on stage here later today to bring all of this together, of how we're thinking about the whole company model. Just done a fantastic job. All right. I'm not gonna dwell on this. We haven't just been waiting till everybody shows up until we start performing better. I think we have been performing better.
Nobody is ready to put up the mission accomplished banner yet and say that we're done, but it was very important to me that we continue to drive better results while we're driving this transformation of the company, and I think we have. Wissam will go through this in a more disciplined manner. This is just the timeframe, quite frankly, from when I showed up to if we meet the midpoint of our guidance through the end of the year, of kind of how the numbers have changed over that time. I think in a lot of dimensions, we have been driving better results, revenue growth, but not just revenue growth, margin expansion, you know, leverage expansion of operating margin growing faster. OpEx as a % of revenue is actually decreasing, and we're getting more out of it.
Again, this is what a focused organization brings, and this is what when you bring in experts that know how to do ROI-based analysis in portfolio management, they get more out of the same amount of money. That is a fundamental tenet of why we went to a BU model, which is bring the focus, bring the discipline, bring senior executives that understand how to manage P&Ls, and you get more out of what you're spending, and again, you get to a day like we saw yesterday. Okay. ESG, extremely important. Not an area, you know, we can't. We've been very focused here as well. We are a big manufacturer, so staying focused on this is something we have every commitment to do. I'm just gonna highlight a couple of things here. Again, we are a big manufacturer. We have big manufacturing plants.
How we think about our impact on the environment is extremely important. It's not only important to us, it's important to our customers. They ask us about this all the time. It's a very important strategic investment for us. One thing I'm very proud of in the last since I've been here, team has driven not just setting targets for us on reducing operational emissions, but science-based targets, right? Science-based targets, then we're gonna hold ourselves accountable to that over the next decade. Social progress, you know, we voluntarily publish the EEO-1 data, I think which is more and more companies are doing. Human rights impact assessments throughout our company and through our values, value chain as well.
I'll highlight, you know, our factories in Asia, really a lot of focus by the local leadership as far as driving excellence and to be recognized by the World Economic Forum as a sustainability lighthouse. Our factory in Malaysia, the first factory in Asia that was able to get that designation by the World Economic Forum. Really proud of that. Of course, you know, the way we conduct ourselves is extremely important to me and everybody in the company. To be recognized by Ethisphere for the fourth year in a row as one of the world's most ethical companies is something we have a very large commitment to. Okay.
I'm gonna make a few comments on the HDD and flash business, but I'm gonna leave most of this to Ashley and Rob. I'll make a few comments from my perspective. Let's start with the HDD business. This may be the most obvious thing that's said all day today that HDD is moving from a client to a cloud business. Everybody knows this. Why are you saying it? I'm saying it because of this second thing here. Like, one thing to realize, is that the HDD we're building today is very different than the HDD that was built for the client. It's very different than the HDD that was even built for other data centers.
The requirements of a cloud titan hyperscaler on latency, reliability, the temperature that they run that thing in, you'd be surprised how. Well, maybe you wouldn't be surprised how hot the data centers are they run in. Essentially, that HDD has to be pounded every day over and over for years. The building that device is a moving target. You know, what I talked about earlier, like the technology that you saw yesterday and things, you know, we've been working on those technologies, some of them, for over a decade, and the goal you're trying to meet is a moving target, right? The requirements just keep getting more stringent. You're doing all this fundamental research.
How do I get these two things to land and actually have a product that I can sell to these customers, and they can bet their data centers on? I mean, we take this very seriously. The world's largest data center operators, most sophisticated companies in the world, bet their data centers on our technology, something we take extremely seriously. Siva will talk a little bit about how we think about that. When you see technology in this space, just realize it's a long innovation cycle. The end line is moving all the time, and so it's not surprising you see technology talked about, but actually bringing it to market is something very, very different.
The other thing I'll say about this is there are new product segments emerging, you know, archival storage, customers talking to us about how do we take this fundamental technology and structure it in different ways to solve different problems. Where have we been focused in this, and where do I think Western Digital has sustainable leadership? Areal density. We can put 2.2 TB of data on a platter. That is better than anybody else in the industry. It's something we've invested in. If you talk to the teams, the engineering teams, they've invested in this for decades. You know, that's why, again, we're able to launch a 22 TB drive, CMR drive, so quickly, after we launched a 20 TB drive. Our strategy has been around EPMR, OptiNAND, SMR. I've been talking about this for a number of quarters.
What you're seeing is a number of technologies that have been commercialized over the last two years, EPMR starting in our 16 and 18s, and then OptiNAND in our 20. SMR's been under development for a long time, but OptiNAND allows us to build something called UltraSMR, an extra 20% when you're able to use basically the ability to put all the software in the drive and the system that we're able to do with OptiNAND. That technology is working. Siva will talk about how far that technology is gonna take us, then what are the next several arcs of technology that are gonna drive this industry forward. We think that gives us a very strong total cost of ownership position. The name of the game, any product company is driven by innovation.
If you bring innovation to the market, you can. Everybody benefits from that. Customer benefits, they get lower costs of ownership. We benefit, we build profitable products. You have to drive innovation, and I think we're gonna continue to stay very, very focused on this, how do we drive consistent, better total cost of ownership. We've talked a lot about the HDD business. Talk about it more. Ashley will talk about it. The market is changing. I mean, that's just the reality. Ashley will show some data about where we've been in this period of absorbing the client decline as the cloud rises. There's a lot of invested infrastructure, whether it's media, whether it's heads. There's been an over-a-decade-long decline.
When I looked at the data, I think it was 14 years of decline of that market, freeing up capacity for us to build capacity enterprise drives. We're kind of towards the end or at the end of that period, so it's not surprising that new business, the business model is also evolving as that changes. You're seeing more long-term agreements. Again, if we continue to innovate, if you continue to see things like we saw yesterday, and I think, you know, between Ashley and Siva, they'll give you a lot of confidence that we can continue to drive innovation in this space. We will have the opportunity to do more value-based pricing with our customers. We continue to drive the cost of ownership down. We will be able to, we will have a value-based conversation with our customers.
I think the industry will be more capital disciplined. It's striking to me, quite frankly, since I've been here, every quarter, I get asked the same question: Dave, what's the supply-demand balance in flash? I never get asked that question about HDD, because there's always been supply available. Nobody thinks about the supply side of it. There's lots of demand, but, like, how much are you investing for future growth in HDD? I think the HDD business is gonna start to be talked about in that way going forward. Like, the 30% exabyte growth at the scale we're talking about is incredible growth. I think Ashley will show you some data. The amount of growth that's out there is incredible, and we are going to have to invest for that.
I think how we make those decisions is gonna be very important. I think the industry is gonna be very, very disciplined in the way they think about that equation. Just a little bit of data on long-term agreements to reinforce this. When I talked about this earlier, I talked about the entire portfolio. What I did here is just focus it down to enterprise capacity, right? We're not gonna have long-term agreements on client drives. Again, there's still a lot of client drives in the world. It's important market, but the real action is in the cloud. If you look at the growth of this over the last three years, we continue to see our customers, like, receptive to a longer-term conversation around this supply-demand question.
I will admit to you, I was a little surprised when I showed up a little over two years ago and fully understood that the HDD business basically transacted on a quarter-to-quarter basis. It was like you know, it's like beginning of the quarter, we talked about, "Okay, how many drives are you gonna buy this quarter? How many are we gonna sell?" On a product that takes 40 weeks to build or something like that's a very interesting business model. Because again, it was like, "Hey, you have some-- you want some drives? We have some." But now it's a very different conversation, which is, "Hey, what are you gonna build through the rest of the year?
Like, let's talk two quarters, three quarters, even four quarters down the road. I think this phenomenon will just continue because everybody needs to do better planning in a world where we have to continue to invest in heads and media. All right, let's talk about flash just quickly. I don't wanna overstay my welcome. There's lots of other good stuff to come. Flash, kind of going more from a mobile to cloud business. Rob will talk more about this. You know, if you look back, mobile is still a huge consumer of flash, don't get me wrong. But the amount of enterprise SSD growth and the scale that's out there, the business is becoming a SSD-based business. We think the industry has been pretty capital, a lot of discipline over the last couple of years.
Nobody's going nuts. There's been incremental consolidation, so we think it's better structured. The technology maturity, I think is. We still have a long runway. Siva will talk about that, but it's getting harder. Each step gets harder, to continue to get the gains. Although we still see, you know, as Siva says, I'll steal one of his lines, Moore's Law is alive and well in flash. So we still have the ability to innovate and drive, die size improvements, more layers. He'll talk about all kinds of other techniques that we're working on that will come into play for years in the future to drive us to hundreds and hundreds of layers. But it's, you know, we have the advantage of a 23-year relationship with a partner where we do this at scale.
We invest as much or more than anybody in the industry in our technology roadmap because we do it together with Kioxia, and together we are one of the largest providers in the world. Where is our differentiation? There's one thing I want you to walk away from this whole session with. If you wanna know about flash, you need to think about the charge trap cell. Like, Siva will talk about exactly what that is, about why it's important, and you know, one of the interesting things in the industry is there's all this discussion of layers. Like, it's a constant thing, like more layers is good, like anything else. But it's actually not the case. Like, fewer layers are better.
Siva will explain why that's the case and why having the single best charge trap cell in the industry, and again, he will show you data where that is demonstrable, is really important. As I said, the joint venture scale, not just on R&D that leads to this great charge trap cell that leads to BiCS5, BiCS6 and going forward, but manufacturing scale. Then again, the consumer franchise is a gem. It allows us to consume more of our flash across the entire spectrum of flash, and it's a scale business where we can drive brand value in what is believed to be a commodity market.
I can tell you, people will pay, you know, there is value in innovation in the consumer market when you deliver high-quality, premium-branded products. There is value to be created there. Portfolio evolution, obviously, we've talked about enterprise SSD success, product success, big progress in the last year or so. The BiCS roadmap, and again, this idea, and Rob will go into this more, we're driving optionality. We have a lot of access to a lot of different customers in a lot of different markets. We have very good fundamental technology.
How do we drive optionality across a large market, large number of costs, large number of customers, large number of routes to market, different kinds of customers, to continually mix the product so we can continue to drive gross margin to the best possible place? Okay, I'm gonna wrap up here. That's what we've talked about. I think we have the technology franchises, the innovation expertise, like the ability to do fundamental innovation and then bring it to market to the benefit of our customers, and the partnerships to drive long-term growth as the world's data storage company. As I said, as the cloud grows, we grow, and it creates option value with these, you know, very, very sophisticated customers that have lots of franchises.
It's interesting, very interesting to me that when I got here, this breadth of market access. I work in other technology companies where you work very, very hard to have good relationships with the cloud vendors. I come to Western Digital, and every cloud vendor is our customer. Not just a customer, but a very, very deep customer because we provide the fundamental infrastructure for how they build their entire cloud. We have access to all of those customers, the biggest customers in the world, where we can sell billions of dollars worth of our products to them. But at the same time, we also have access to every consumer in the world and the ability to sell hundreds of millions of individual products through a consumer channel.
I think now we put the right leadership team in place that has the experience in operating at scale businesses to get the most out of what we have here, which I think a very enviable position as a company, in a part of the world, part of the technology world that is extremely important, which is data. A lot of data in the world. The question is: how do we harness it? How do we allow people to use that data then create value from it? We think that's a very important part of the world of increasing importance, and it's an area where we're gonna stay very, very focused on delivering the absolute best value proposition to our customers. All right.
What I'm gonna do now is I'm gonna hand it off to Rob, and he's gonna go through details on how he thinks about the flash portfolio and the flash strategy. Thank you. Rob?
Please welcome Executive Vice President and General Manager, Flash Business, Rob Soderbery.
Dave.
Good morning. It was just nicely two years ago that David called me and said, "Would you like to take a look at this flash business?" I have to admit, I wasn't instantly swayed by his pitch. After looking at the business and looking at the assets that were in the business, I became increasingly convinced that, as well as I think many of you are, the flash business was undervalued with respect to the core capabilities, the technology, the market access, the brands, and I fundamentally didn't understand why. I got intrigued. Make a long story short, 18 months ago, I joined and have been putting in place practices to really drive the business since then.
That started with people. As Dave talked about, really reassembling a flash business unit, sort of pulling back the people that have been distributed around the company, and not just within my organization, but collaborative across the entire organization. The second big area of focus was technology. Previously, the organization had been serving many markets, but it's serving those markets with individual technology stacks. It's pretty inefficient 'cause you end up replicating a lot of underlying infrastructure. So converging our technology platforms, so we could build multiple products on the same technology platform. You know, a third area we immediately started working on was portfolio. We do have wonderful access to many different markets and many different segments. In each of those markets, there's actually many swim lanes or different product lanes.
It's a pretty ripe area for portfolio optimization. Looking at both the revenue growth opportunity, but also the cost reduction and value engineering and margin generation, opportunity. Really digging into the portfolio. Then the fourth area of opportunity has been in just the world of our business practices. LTAs, pricing, learning how to value price. I come from Cisco, a long heritage of understanding the value of our products and delivering those to market, and so sort of bringing some of those traditional processes into the company. As well as some more just block and tackle things like how we manage our inventory, and particularly how we manage inventory through the cycle, counter-cyclical investments. Really working to rebuild some fundamental business capabilities.
My basic thesis was, you know, the raw underpinnings of the business, the engineers, the technology stack, our core capabilities in ASIC and firmware and testing and all that underlying capability was actually in quite good shape, and we could reassemble business focus here, get a lot better result. I think that's proving out. Finally, we spent time on two things. One was strategy, understand what the unifying strategy would be for the group, and then to set a very specific financial goal for the business, which we ultimately set at an increase of our through-cycle gross margins by 5%-7%. I saw plenty of opportunities, plenty of what was at that time missed opportunities, to deliver that kind of number.
That's on top of the overall cyclicality of the market. We can't control the cyclicality, but we can lift both, you know, trough and peak and the overall, through cycle gross margin. That's been our objective. First, I wanna frame a fundamental transition, though, in the flash market. Flash has been a mobile-dominated world, that consuming the majority of the bits, and that was a wonderful transition. But we're seeing a major shift to flash becoming an SSD-dominated space. That's very important because there's more technology, there's more opportunities for differentiation, there's more different form factors and capabilities. This is a big shift. It's a shift caused by three actually sort of separate, interrelated, but quite separate trends.
We've been talking about cloud and the tremendous capacity growth in cloud that's causing a dramatic increase in the demand for SSD. We now have SSD and HDD playing in their own swim lanes in cloud, but both of those are showing vigorous growth. In the client world, this notion of the desktop PC has become basically the edge of the enterprise, and it's anchoring our employees into the enterprise. That no-compromise experience, what we all look for, right? We want, you know, more power, more capability, more everything on that device. We want it smaller, we want it thinner, we want longer battery life. We want it rapid turn on. We want a really no-compromise experience. That's driven not just unit growth in client, but technology growth and reinvigoration of client differentiation. Then consumer.
Everybody's a content creator or content consumer. We're generating an unbelievable mass of content that's coming off devices. All those devices are opportunities for us, and it's going to, you know, it's ultimately going back to the edge and getting viewed. The consumer world, you're seeing this also a tremendous explosion. Within that, there's the last really remaining part of HDD transition, which is almost 140 exabytes of hard drive that's still sitting out there on consumer desktop and portable devices. That's a remaining opportunity. That's a really big number relative even just the size of our flash business today. Those three trends are really converging to create a shift to an SSD business.
I believe that's fundamentally good for us, because Western Digital, as a company, brings together three core technologies, all of which are extremely useful in that transition to SSD. It starts with NAND leadership, and NAND leadership in both design and manufacturing scale. Delivering the right performance, the right attributes, NAND attributes, the right foundation for us to build on. The second, and what I spend even more of our WD OpEx on, is the platform and systems capability. Just like at Cisco, we maintain a, you know, a deep, robust ASIC, a family of ASICs, and full system stacks optimized for each one of these markets we play in. That's about that increased technology differentiation. There's a rich heritage of systems innovation capability and demonstrated technology differentiating advantages.
I'll talk a bit about some of those. What David has been talking about, which is go-to-market strength. The traditional OEMs, the hyperscalers, the channels. I mean, it's very routine that a customer will look me in the eye and say, "Hey, look, all things being equal, you are my preferred partner. So even if you don't have the product right now, even if you don't have the capability right now, I'd like you to be in a position where you could be number one on my list." That turns into commercial terms as well, where it's not always a race for the bottom. It's about how we can deliver a whole set of capabilities to that customer. We're uniquely positioned in that transition, and our strategy reflects that.
While we see the market as a whole going to 60% SSDs by the end of our three-year forecast cycle here, the WD Flash business, we're essentially transitioning to an SSD-centric business with 80% SSD volume. It's driven by in those three markets each of those markets really unique advantages. The preferred partner of the largest hyperscalers with a compelling eSSD roadmap. In the consumer business, the transition, the remaining transition of HDD to SSD and growth of content. In the client SSD, the continued dependence of all of us on the devices that are in front of us today in terms of driving productivity and connectivity in the enterprise. This is a big transition for the WD Flash business.
I think we're well equipped for it, and it's fundamentally core to our strategy. We're gonna talk a lot about what that means specifically for eSSD and the cloud. The good news from my perspective as the GM is that, you know, SSD technologies are common. We have a single architecture, and so we'll get scale within Western Digital across all of these markets, and have optionality as we think about how we wanna serve each one of them. I'm gonna speak really briefly on Flash because Siva is the world's expert in this technology, and I want you to hear directly from him how we think about it and the relative advantages. The bottom line is we feel that we are very, very comfortable in our roadmap.
We're very comfortable in our roadmap because that charge trap cell that David was talking about provides two things. It provides better performance, which is gonna be well-matched to the needs of both client and enterprise SSDs, and particularly in the transition to QLC. It provides better efficiency, better capital efficiency, more die per wafer, density, better utilization of our factories, all of those benefits. We're very comfortable with where our NAND roadmap is, and we're ignoring the noise in the marketplace. At a high level, where are we? We are today right in the middle of that fifth generation BiCS, the BiCS5 transition. We were fortunate yesterday to announce that both our client and enterprise SSD drives are rolling into qualification and production respectively.
That will drive the BiCS5 transition through to completion in this calendar year. BiCS5 itself was a node built on the notion of capital efficiency. You probably saw quite a bit lower capital investment required in this cycle, as well as fast scaling. Very similar from an architecture construct, which allowed us to quite quickly scale the SSDs we're building on top of that. BiCS6 is right on BiCS5 heels, bringing increased density but also focused on delivering QLC and performant QLC. NAND performance has not been a strong differentiator. In the TLC world, we've been able to saturate the interfaces relatively efficiently. As we go to the QLC world, though, the performance is more marginal, and there's more differentiation and more opportunity for differentiation in order to saturate that interface.
We think BiCS6 QLC is really gonna shine in that dimension. Again, we'll have more data. We'll be ramping our consumer portfolio this year, and then the SSD portfolio will ramp in 2023. Then in a few minutes, you're gonna get a sneak preview of a project code-named BiCS+, which represents the next generation of BiCS technology. I did mention earlier this notion of consolidating our architecture. When we looked at the underpinnings as we moved to an SSD-centric world, it was clear there was a lot of assets here that were really duplicative. The company have done quite a bit of M&A, of course, bringing together HGST and Western Digital and SanDisk, and had a lot of sort of parallel activities. We were able to go in and consolidate those activities.
That allowed us to redeploy 30% of our product development OpEx to be customer-facing in terms of proliferating different products in different swim lanes and investing in customer realizable benefit, and then consolidating around a single technology stack. That technology stack is a very, very valuable technology stack. You know, we've talked, I think it's no surprise that the DIY market for, say, for eSSDs has been very fraught with peril for folks that have entered, tried to enter that market, 'cause it's a very complex technology. It starts in the core with NAND IP. The NAND IP doesn't stop, you know, on the wafer or on the hardware. The NAND IP includes the firmware, includes error correction, includes the algorithms.
NAND is a very strange quantum mechanical device, and it actually requires the software in order to make that NAND perform. That software itself changes from generation to generation to generation. Having 20 years of NAND IP in that software layer is extremely important to seamlessly make those technology transitions. But then it keeps going up the stack. We have a large ASIC portfolio, systems manufacturing, firmware validation, strong IP. We deploy all that IP in a consistent platform across those focused markets. It's not just a technology platform, it's a manufacturing platform. In our manufacturing facilities in Malaysia, in our facility, we have the ability to put wafers in on that fourth floor, and SSDs out on the bottom floor.
There's a tremendous optionality in terms of the kind of products we build, so we're not doing, you know, die sort and packaging and then, you know, shipping to a warehouse and then figuring out what to build. This is one single process. A lot of optionality, a lot of flexibility, and a lot of capability going into building increasingly demanding mechanical devices that are put into increasingly demanding hyperscaler and client qualifications. Another big asset. Now I wanna turn to the market segments and put a little bit more depth on the introductory comments in our market segments. First, where do we start? This is our position today. 37% share in consumer. We think that's a great position for us to be in.
We're very comfortable with that, and really focused at, you know, maximizing the product categories we play in, ensuring that we can play in the premium categories in all of the swim lanes, you know, all, in all of the GOs, in all of the channels, all over the world. This business has been something that we have been reinvesting in. Part of that 30% of OpEx reallocation I talked about was going to efforts here. We can drive high profitability, a unique branded differentiation out of the consumer business. It's very important. Consumer business also is an outlet for bits for us, and so it allows us to control our own destiny with respect to a certain portion of our NAND consumption. The client business is a great success story. We're at 20% share in client.
Part of that is based on those long-standing relationships. There was a desire by the traditional OEMs for us to partner with them, and part of that comes from execution. We're comfortable at that 20% share number in client. We think that's about the right number for us as a business. Cloud, we're at 8% share. The first thing you might say is, "Well, 8% sounds kinda small. Mmm, are you succeeding or not?" The good news is it's the correct 8%. It's 8%, and it's focused on NVMe, and it's focused on our top three hyperscalers. Strategically, it represents a very, very strong foothold in this hyperscale market. We went through a lot of learning process in the big four generation in insertion.
After that learning process, after that insertion, the products have performed extremely well, very high reliability, very high performance, have stood up best-in-class versus the other providers in the industry. It was a real credibility-building experience, a knowledge-building experience for our team, but a credibility-building experience for the hyperscalers. I'll talk in a second about where that number's going. Let's go ahead and do those in reverse order, and we'll jump into cloud. The first thing to understand is eSSD for the cloud is not a single product. The cloud itself is evolving, and you're seeing a fragmentation and increased complexity, so architectures are changing. Individual cloud providers are basically building bespoke architectures that are actually quite different from each other, and that's just driving an overall change in the demand landscape.
We're also seeing optimization for different use cases, compute and storage, and even optimization for workloads, right? The world's largest, you know, email workloads or file-serving workloads or streaming media workloads, you know, these are workloads that are big enough to be industries, right? They're product categories all by themselves. So all of this is causing a lot more complexity in the eSSD world, becoming a lot more demanding, providing a lot more room for differentiation. At the same time, capacity and speed are just as important as ever. Double-clicking on that, what do you see inside the cloud? Well, you'll see a variety of different workloads and use cases, and actually optimized stacks are being built for each of those workloads and use cases. Below that, compute SSDs, storage SSDs, boot SSDs. You know, compute SSDs, typically smaller capacities, higher performance.
They're on the PCIe NVMe transition. Storage SSDs, larger capacity, typically lagging on interface transitions, but with new capabilities to support virtualization of that storage, so we can efficiently partition them up. Then boot SSDs, those are being hit by read/write every day, so those are high endurance. We have different, you know, different product categories emerging, and even in these are differing between the different hyperscale vendors. A lot of complexity here. We think in general that's good, 'cause while there's five or six folks in the NAND industry, and three or four of those are capable enough to deliver in this marketplace, nobody can actually serve this entire marketplace.
We're having to start to focus on which aspects of the market we can serve, and that will drive more differentiation and more opportunity for a more balanced business relationship in each of these sub-markets. A big part of our value proposition with the hyperscalers is long-term alignment. It's the combination of how we transact with LTAs. It's how we think about roadmap. It's how we respond to unique requests. It's a brand and a value proposition we've built over time by developing a long and trusted relationship. Of course, key to that is the roadmap. BiCS4, the current generation, those are the high-volume products today, leading to BiCS5, leading to BiCS6. In each generation, we're expanding the scope of the roadmap.
Another area where I made a change coming in, and I think one of the root causes of some of our early eSSD struggles, was the team got a bit overambitious in terms of trying to serve too much of the market at the same time. With BiCS4, really honed in on serving the storage swim lane within eSSD. That's the highest capacity, so the best consumers of NAND, and it's the fastest-growing and the fastest-growing part of the compute landscape. Over there on the right, probably the most important figure on the page, which is the share goal. Today we're at 8%, we're setting a 16% share goal, you know, in a market that's growing at 35%-40%. You really have to grow the business in order to hit that share goal.
We think long term, something in the 16%-20% is about the right number for us. We wanna be in a position where we're overweight in hyperscale, and that share goal is the overall market, including enterprise and hyperscale. We wanna be overweight in hyperscale, but we also wanna maintain the ability to mix and match across the portfolio and drive the best overall mix for our business and not be overly dependent. That 16% goal, it's a good goal. We probably will plateau out in that 16%-20% range. Then at that point, we're really out of bits. We've allocated all our bits to the highest and best use of those bits, and we'll have a balanced portfolio.
Big news from yesterday was the launch of this device here, which is the SN650. This is our fifth-generation BiCS enterprise SSD. I pulled this one off the truck as they were loading them on and shipping them off to our first hyperscaler qualification. That qualification is underway. In fact, I'm pleased to report that our top three hyperscalers have been eagerly awaiting this device. We've gotten requests to really actually pull it in, which has been super challenging with today's supply chain environment, and we've gotten commitments to accelerate qualification. Everything we've seen so far, we expect a relatively smooth qualification, and that will lead to first revenues in the September quarter and good contribution this year from this device.
This is gonna be our high runner, data center product for the next 18 months to two years. A really important device in the portfolio, and I'm pleased at how smooth the technology development effort has gone, how smooth the work has gone so far with the cloud titans, the hyperscalers, and I think it bodes well. It's a 16 TB device, so twice the capacity. Improved TCO, so better power performance. E1.L packaging, so you can get denser form factors. Technology-wise, there's a lot under the hood. I mentioned earlier, increased virtualization of the storage layer in the cloud. That requires quality of service and performance guarantees. You know, if you're sharing a workload with somebody else on the same drive, you don't want your app to slow down when theirs starts using that drive.
A lot of technology going into making this work. A big delivery for the team and something that we think really de-risks that 16% share goal for sure. Now quarter to quarter, I know you're gonna be watching the data center number. That number's gonna go up and down based on how we allocate our bits, about who's consuming what product, about the tactical decisions. But we see, I think, a good glide slope up to that share position. Shifting gears here to client. If you think about that 8%-16% goal, you'd say, "Well, that seems pretty bold to double your share goal." Well, it's basically what what's happened in client, where we doubled our share from 10% to 20%.
We did that primarily by taking share away from the market leader in client, and we did it in a timeframe in which no other competitor essentially was able to move the needle in terms of their position. We did it really on the basis of two advantages. The first was technology leadership. In the client, we have innovated in technology, which is DRAMless client architectures. We're a full year ahead of the market in DRAMless, almost two years against certain competitors. Essentially that was a technology innovation. We're able to rearchitect the client drive to use the existing DRAM in your device as opposed to shipping more DRAM with the client device. You know, why is that important?
Well, it's important because, you know, we're trying to build a device that looks like this, and there isn't a lot of room for extra DRAM on this device. I have, in case you wanna look at the big one, I've got the big one here. Right? That one's pretty tight as well. Huge benefits for us in this kind of innovation, lower cost, lower power, more density. At the same time, the strong relationship with the PC OEMs, we have worked to interlock qualification cycles, interlock our portfolios, make sure we're delivering to every swim lane, premium mainstream, as well as the entry swim lane. Really strong success. It's clearly benefited from that leap in client volume from, the, you know, 280, 285, up to 350 in the pandemic era.
We think that that trend of hybrid is here to stay, and we think that it's just starting now in the transition of that device in front of you to become the singular device which with you engage with the enterprise. We're all used to this, but of course, you know, enterprise have been going through transitions, not with, you know, high-end knowledge workers, but with call centers, with all kinds of employees having to move to modern client platforms, and they're not done. Every enterprise is hybrid. Every enterprise realizes how important that compute environment is. Employees realize how it is. They complain about that compute environment. It's a dynamic world in which we need to drive, you know, retention and engagement with our employees.
There's still a lot of pressure behind the scenes in order to ensure the healthiness, particularly of the commercial PC cycle. I just showed you the SN740 device. We've completely refreshed the eSSD with fifth-generation BiCS. The SN740 is the equivalent product for client. This is our highest volume, mainstream, right down the middle of the fairway, client device, 50% better performance, 50%, almost 45% better battery life, built-in hardware security, you know, 30-millimeter package. A really, really attractive device delivered at the right time for our client partners.
We did have Microsoft and Intel on stage with us yesterday, talking about how we've worked with them on all of the features that you expect in the hybrid workplace, from power to battery life to turn on time to standby time, and all the things that we're doing together to make this an awesome client device. This is qualified in our lead OEM, and we'll have substantial revenues in the June quarter on this product. That's wonderful. It puts us really through the BiCS five transition from a product standpoint. We've really, you know, can clear the decks and focus all of our product energies now on BiCS six.
In addition to the PC client, we serve the broad set of clients out there across mobile, automotive, IoT, new kinds of endpoint devices. Our strategy here is actually quite simple. We're not trying to overserve these markets. We're looking to serve the premium aspect these markets and use them as part of our gross margin mix-up strategy. eMMC continues to be a very healthy business with good gross margins, UFS and UFS standards transitions across mobile, automotive, and IoT. We are pleased to serve the premium mobile phone market, so the high-end mobile phone market, really differentiated and focused on performance. Automotive has been lots of stats about the growth rate of the automotive sector that's driving both strong growth and sustained through-cycle margins, reducing that overall flash volatility.
IoT, and especially emerging consumer devices like the VR headset, where we had a very early and leadership position, and have worked with both the standalone innovators in the space, but also increasingly with the cloud providers who are doing so much of the consumer innovation, a big part of the portfolio in this area. Okay, let me touch on the consumer, which is actually the funnest part of the portfolio. It's a content world, and there's this content explosion. You know, we've all got our new iPhones, right? 4K video, as much as you want on demand. In professional world, they're moving to 12K video. Takes 15 minutes to fill up our largest and fastest memory card with a 12K camera. So it's just staggering. Imagine yourself putting that.
a new card in your camera every 15 minutes. You know, 500 hours of video uploaded to YouTube, 200 million people in creative and culture industries. You know, everybody's a content creator, everybody's a content consumer, and all that content is getting stored. It's getting stored not once, it's getting stored multiple times. At capture, it's getting stored when it's viewed, it's getting stored locally, it's getting in the cloud. This is all driving an amazing amount of consumption on the consumer side. David shared this number here, 6.9 billion devices. Amazing number of devices. 2.4 billion device shipments across gaming, camera, smartphones, and notebooks. Now to put this number in perspective, if I exclude our embedded, and we're actually embedded into all four of these segments, and I just look at our standalone after...
You know, branded delivery, we're delivering well over 300 million units. Our attach rate to things in the world is approaching 15%. That's just a pretty amazing thing to be attached to 15% of, you know, all of the devices in the world. That number continues to grow. The demands on those devices continue to grow. They're moving from being kind of these throwaway consumer peripherals to really important parts of consumer electronics. The consumer is supported by the brands. That SanDisk brand really appealing to the content, the creative, the value of your memories. The SanDisk Professional brand for the most demanding users. Launched that new brand last year. It's been a huge success.
1.7 EB shipped in our first year just with SanDisk Professional to a variety of markets. The WD brand providing for our internal drives, the color portfolio, as well as our NAS products. WD_BLACK. Huge success with WD_BLACK. You know, we're looking to drive that to be a billion-dollar business, driving a premium experience for gamers, at premium gross margins. You know, under the surface, the product teams are paddling furiously to keep up with the speed and scale requirements here. Just a few of the products we've launched over the last six months. My Passport and SanDisk SSDs up to 4 TB now. Our USB cards doing 400 MB/s. Our UHS-I, the world's fastest, SD card at 200 MB/s.
This is what's required to keep up with the devices and keep up with the consumers. I do wanna double-click on a significant opportunity in here for flash, which is the transition of the consumer market from HDD to SSD. Overall, it's 140 EB out there in consumer hard drives. Almost 100 EB of that is in the 2.5-inch sector. We think most of that is gonna go to SSD over the next couple of years, so it's a big opportunity. It's essentially the size of our current flash business. So very significant opportunity, and we're really well-positioned to capture it. We have the right products. We have the right brand. We have the relationship with the retailer.
If he wants to stock a few more SSDs and fewer HDDs, we have the ability to do that. We have the ability to take any capacity that we free up here and redeploy it into, you know, really limitless, you know, enterprise capacity demand. This is the very tail end of this transition, but it's a significant opportunity for the flash business that we'll be pursuing. Okay, let me take this through to closure here. I know it's been a lot of content. I talked about the world-class and what we're doing in terms of retooling the organization, the portfolio foundation, and then just ultimately, I think, hopefully you can see we're thinking about value creation, right?
That 5%-7% goal requires thinking about value creation in every aspect of the business, mix, capital strategy, through-cycle, business practices, portfolio. Where does that 5%-7% through-cycle gross margin, where does that come from? It is really four big levers. Of course, we need NAND cost leadership. We need to drive capital efficiency and keep on our long-term 15% cost reduction target. If we don't do that, we'll, you know, we'll have bigger problems. Key to that is then maximizing the use of the underlying bits because there's a numerator and denominator there, and the more volume we can ship out of the same investment, the better. The big lever is portfolio mix. The detailed slide on that, mixing up and across segments, investing for the right lineup.
Improving through-cycle economics. LTAs, better serving markets have better through-cycle GMs, counter-cycle investments, really working to mute that cycle, you know, lower or higher lows out of the trough. Operational improvements. There's plenty of room for scale, for cost reductions. We consolidate really on a single SSD platform and architecture underlying that supply chain optimization, inventory management optimization. We have a lot of different levers in the portfolio to drive against that five-seven number. Just double-clicking on the portfolio itself. We think about our portfolio in four buckets. That's premium. That's where we have a unique technical or brand advantage that we believe is very sustainable. Differentiated, we're delivering a product that right now in the marketplace is able to command a premium, mainstream, and then the commodity markets.
The goal is pretty simple. We wanna shift our mix up. We do that by serving the more premium product segment in swim lanes within a given segment, as well as mixing between segments. This will be the bulk of that gross margin increase in addition to sort of the taking care of business aspects of the other dimensions of the plan. We are well underway. Think of this as, you know, you're two years into a three-year plan in terms of delivering this, so this is what we've been working on, and we're well down the product development path. I do wanna touch on capital strategy because I think that's been a very misunderstood part of the strategy.
First of all, just to set some context, looking at the CapEx numbers of all the different vendors and trying to figure out what's going on is actually a very difficult challenge. It's difficult because there's notable CapEx variation. It's difficult because different vendors have different CapEx strategies. Some have lots of capital, and they're willing to deploy that in order to get to nodes faster, and they're willing to sacrifice maybe performance for cost advantages. You have a lot going on under the hood. We have been going through a period of super CapEx efficiency relative to our long-term averages, primarily 'cause BiCS5 was a very CapEx-efficient node, and so that's made our CapEx look below our expected rate.
As we regress to the expected capital investment cycle here, our strategy is very simple. The 14%-16% market share goal, we have a capital efficiency advantage, better cell design, economies of scale, ability to use all the bits. The simple strategy is leveraging that advantage, keeping that forward 13%-16% market share goal in mind, and then tactically making the right decisions at every point in time to make sure that we're investing capital in the right transitions that generate the best economics and letting the share and the capital investment ratios float to achieve that goal.
Wissam Jabre will translate that into, what it looks like at a company capital portfolio level, but we're very comfortable with our capital strategy and the affordability of that strategy. Finally, to wrap up here, we'll talk about those targets. On the revenue side, 10%-12% CAGR really positioning us to grow with the market, and flash continues to be on, you know, firing all cylinders from a market demand perspective.
On the gross margin side, I've talked about that 5%-7% improvement. That's a through cycle number. I think about that 29%-30% where the model showed in that, you know, 18%-20%, you know, range. How do we get the through cycle? 35%-37% is our objective there. Look, the NAND business right now is more of a top of cycle, sitting at that, you know, in that mid-upper teens level. If I look at this business right now, I see plenty of opportunities where we could have been in a 40%+ gross margin range. I see upside from here.
Obviously, you know, the underlying cycle itself will come and go, but 5%-7% over that through cycle estimate is where we're putting the stake in the ground. To wrap up here, I am super excited about this business. You know, the team is fired up. I feel great about our leadership team and about the assets we have in the flash portfolio, and we hope to be back in the not-too-distant future sharing more results and more opportunities with you. With that, I'd like to introduce my colleague, Ashley, to talk about the hard drive business.
Please welcome Executive Vice President and General Manager, Western Digital HDD Business, Ashley Gorakhpurwalla.
Hello. Good morning. How's everyone doing? I'm between you and the dynamic Siva Sivaram. Let me walk you through how we think about the hard drive business today. We'll sort of take the same flow of thinking through the market, how we're approaching that internally, why we think, you know, I'll foreshadow a little bit why we think our leadership and innovation matters. Then we'll talk a little bit about how we think that projects over a longer term. Okay? One cultural aspect I really appreciate about Western Digital is that we constantly strive to have an external perspective in order to check our internal thinking.
I've included a couple of external perspectives on the HDD business that I'll call out and share with you along the way that I think help illustrate our strategy. The first of which is sort of a layup. As we think through what David set up, which is our business across the portfolio, is quite broad actually today still, and we have a multi-billion dollar marketplace in endpoint drives, in data center drives for traditional businesses that are transforming themselves digitally. Really our go-forward strategy today that I'd like to focus on is that cloud segment, where the fundamental element of storage remains the hard drive, will remain the hard drive going forward. It's really key to unlocking sort of the cloud economics as we go forward. What does being the king of big data translate to?
Well, if you look at the visual, if you're in the back of the room, I apologize, you may not be able to actually see the client line from an exabyte standpoint. It's there, it's important. We'll be very competitive in client. We enjoy a great deal of relationships with our OEM partners and cherish those going forward. But really what we take as a great deal of pressure and importance for our business is being able to meet the cloud challenge. As we go forward, just for perspective, this year, the hard drive marketplace will deliver somewhere in the 1.5 ZB range, perhaps. And that's sort of the most it's ever been. Fast-forward a few years, maybe five years from now, we'll have to deliver that year over year.
We're entering a multi-zettabyte scaling business going forward, and it'll come fundamentally on the backs of innovation and operational excellence in the hard drive industry. What does that translate to? Well, the other side of the challenge is the opportunity. I won't belabor the point, but you could grab this, the history part of this chart, and you could drag it another five-six years back, and you'd have fundamentally the same situation, a decline in CAGR in terms of revenue, a dramatic drop, say, from 2013 of 600 million units in the industry to today, say roughly 250 million units. Some of the underlying dynamics of that, an incredible rise in per unit value, of course.
Also we've been absorbing and playing an efficiency game relative to the manufacturing capacity in the industry that we've had. It's gone through a very dynamic two years, so I've been here for a year and a half of that. Now we're entering a phase in which the marketplace, and I'll show you some data on this, through some of the scaling factors, has absorbed essentially that decline, and we begin an era of secular growth on the backs of meeting this cloud challenge. As we go forward, I'm really going to focus on what Western Digital's strategy is to meet this cloud challenge going forward.
David alluded a little bit to this, and Siva will bring it home, but if we are thinking about the hard drive as a single vector of innovation, perhaps just capacity, I don't think that will really change the balance of storage, capacity, density, and TCO. That balance is fundamental and super important for a cloud customer or any customer. We're seeing the emergence, especially from our deep engagements with our cloud customers, our hyperscalers, is that there are many vectors of innovation and many vectors of requirements that are changing over this period of time. Some of them drawn by scale, some of them environmentals, workloads, but fundamentally, it's just growth. We have entered an era of data creation, and the value of that data is accelerating going forward.
Our ability, for instance, at scale, to support our own customer sustainability and circularity needs is very important. A large manufacturer, we think through that from manufacturing to operation to retirement. That's very important through the supply chain of many of our customers, important to their employees, important to ours. Performance, access time, and reliability are all now new vectors of innovation within the hard drive. As we go forward, we'll talk about innovation across many vectors as opposed to maybe one magic silver bullet that helps meet the cloud challenge going forward. We'll do that sort of in a couple of ways. We'll talk about how the market has changed through the framework of our engagement with customers. We'll talk about innovation now in a fashion of tiering.
Not a singular monolithic drive that solves all the problems of the cloud and the cloud growth, but one that now is starting to tier into different workloads, capacities, trade-offs against performance. We're proud of our legacy of predictable capacity leadership going forward. We've demonstrated quite a bit of that just yesterday, which was a lot of fun to reveal what the team's been working on for a while sort of holding back a great deal of technology that we're working on that will come out later. That is a roadmap that allows our cloud customers to think about their own growth in the terms of being able to meet all the growth curves their customer needs going forward. Then, of course, we wanna be excellent at what we do across these operations.
David did a great job of setting up the sort of natural dynamics that are changing within our marketplace. Again, we enjoy a great deal of strategic relationships with our OEM partners today and will. That was based on, especially with the dynamics of compute and now the displacement of hard drives within those compute devices, that we're essentially not a device-led company in terms of hard drives anymore, we're cloud-led. Standardization was the way we operated, form, fit, function. A 1 TB drive is a 1 TB drive. Actually, it would be bad to break that compatibility. If you have that commoditization with a declining unit count, but an infrastructure of manufacturing capacity was built for many more units, you have the definition of commoditization.
Then now replacement cycle by flash, where form factor matters more, and that spec is really the fundamental reason the industrial design has changed, and now the rise of flash and compute. It's really moved, though, to a very diversified customer engagement for us. In addition to the OEMs, in addition to our retail partners, in addition to our distributor partners, now we have these deep direct engagements with our cloud partners. As David alluded to as well, this has brought up a very different level of engagement, where instead of form, fit, function, we've begun to talk about all these different vectors of innovation. Now we have differentiated solutions. Why?
Because we now have customers who have the scale and, more importantly, the sophistication to accept differentiated solutions into their infrastructure, monetize them, those differences in specialization, essentially the opposite of commoditization. As we go forward, the other really important element that we take very seriously is that the hard drive is essentially the best proxy in infrastructure for data creation and acceleration of data value. Networking, compute, offload, all very important, but our trade-offs with performance, latency, customer SLAs, the best proxy for data growth, the hard drive in the cloud. This is core to the offering of our customers. No longer just a spec, no longer a commodity, but key. That leads to, of course, value.
We can already see not only the long-term agreements, not only the deep engagements, we're talking about multi-year engagements relative to co-engineering, but now having the conversation relative to pricing towards that value and that innovation that we're putting into our roadmaps. What doesn't change is the focus on quality. How that quality is perceived is quite a bit different going forward. Unfortunately, if your hard drive today in your desktop or your notebook, if you have one, fails, you're quite unhappy, but only you. If you kinda go forward to zettabyte scale, a reliability issue at zettabyte scale takes down the world's services. There's a very, very different environment we're operating in going forward.
For a second external perspective, because quality is sort of in the eye of the customer in most cases, I'll leverage an annualized failure rate report that Backblaze, a service provider, provides publicly to all of you, to us, and to others. If I orient you just to the data, lower is always better if you're talking about annualized failure rates in a data center. This is an installed base that's been in service for quite a long time. Long, in fact, that they still track HGST separate from Western Digital. These systems have been in place for a very long time. What I would tell you is that if the two lines that are labeled are below essentially competitive benchmarks, that's the right place to be.
This matters to our customers, it matters to us. What's even more interesting is now some customers would like to engage with us slightly above the device level, so system level. Perhaps you've heard the terms JBOD, JBOF, so just a bunch of drives. The way to disaggregate storage. We have quite a bit of IP and technology in delivering platforms of storage to customers as well. We would see on average, a 62% decrease in failure rates when delivered through our IP due to thermal management, very sophisticated operational vibration technology. We can meet customers where they need to be met relative to how they wanna engage with us. Innovation across data tiers. Our innovation goes quite a while back. I just mentioned HGST and the integration.
If you track all the way back the origins and the roots of the WD hard drive organization, you trace it back to the invention of the hard drive, which go back to that timeframe, and you look at the volume it took to deliver early hard drives to today, in that same volume, we've added over a billion times value. Not a bad investment. A billion times value over that period from invention to today. Our intent is to, of course, drive that sort of ROI going forward for the next 50 years in hard drives. Another really seminal point on this is that can show you the inflection of how hard drive technology works, is the helium drive. The ability to seal the drives with helium changed.
For instance, we deliver or consume half the power per terabyte of previous air-filled drives, and really started another wave of capacity increases. What you'll find and what I'll talk about is the innovation that we deliver today is really based around meeting that cloud challenge. We're gonna be able to deliver this innovation sort of as a roadmap for different tiers in our data center customers. Today, dual actuator, this is the concept of having two methods to read and write in one drive, essentially. You can get, especially for sequential reads, almost 2X the performance. This is what I would consider in the TAM. This is an existing part of the data tier in a few customers. It sits essentially below the performance requirements for an eSSD, above sort of a conventional drive.
We have really where the bulk of data will be stored for the rest of our lives in big capacity drives, such as CMR drives, conventional magnetic recording. I'll talk a little bit about the difference between that and an UltraSMR drive, which has a little bit different write characteristics and is more suitable for large capacity, maybe not as low latency applications. Finally, I won't say too much about the specifics of our archive drives and storage units that we're building and co-engineering with a few customers. As a matter of fact, these might not become publicly available as they're specialized and differentiated products for certain large customers. We see with our roadmap and our development, we're right around the corner, essentially from delivering 50 TB capacity units.
Let's talk about that capacity leadership, and really, I want to impart upon you that this matters. Then I want Siva to back me up with just how we did it and why it's so important. We have a very novel way of thinking about extending what is perpendicular magnetic recording by introducing for the industry a novel solution around energy-assisted. We've commercialized now energy-assisted in our 16-, 18-, 20-, 22-, 26 TB drives. By the time we start shipping our 22 TB drive next month, we'll have already put about 20 million energy-assisted high-capacity drives in data centers of our customers. This is novel in the industry, but it is not. It is a well-proven, now commercialized, solution to driving the next wave of capacity increases.
EPMR, we think of as in generations, and EPMR 2, essentially, the design is done, and we'll be commercializing that as we go forward and driving more conventional magnetic recording capacities, and being able to rely on our architecture as we go forward. OptiNAND, we've talked about this before. We introduced it with our 20 TB drives. What's just the important takeaway here is that we've changed the control plane and architecture of our hard drives going forward. It's not a one-off, one-time development, but opening up a roadmap for innovation going forward. Again, unique within the industry. Improved capacity, we're able to now open up, our capacity relative to a few dimensions. For instance, extremely precise positioning of our heads allows us to have much more density of tracks, as an example.
I'll talk about UltraSMR and its impact in a second. Improved performance. We've essentially removed one of the biggest obstacles in hard drives and software development forever, which was a software developer had to choose, do I want write cache enable, a very fast performance of my drive? Or do I want write cache disable, a very safe way to operate my drive and never lose data in my storage stack? With OptiNAND, we're able to actually offer the performance of write cache enable with the safety of write cache disable, and the user doesn't have to do anything. Fundamentally, it changed again with ArmorCache, the ability for our drives to increase in innovation. Drive resilience.
I talked about that level of quality, and with OptiNAND, we're able to actually protect 50x more of the user's data as it's in-flight within the drive in the instruction set. This will continue. Hopefully, next time I talk to you, I'll have a new roadmap of just the things we've built on the OptiNAND architecture going forward. I'd like to tell you about it today, but it's still under development. UltraSMR. One more part of layering on the innovation. I've talked about recording physics, so EPMR energy assist that we've introduced to the industry. I've talked about the architecture and the control plane, OptiNAND. What if I put them together? Well, you can see the roadmap now. In 2015, I could have stood up here if I was part of this company and said, "Look, we've introduced SMR.
We can bring you, as a software developer, the ability to use a new Zoned Namespaces and immediately get 10% capacity. In a compute device, that 10% capacity probably wouldn't have paid off. You probably wouldn't have noticed. It met the spec. You weren't going to fill your drive. 10% at the scale of millions of drives in one data center or under one cloud. It's a huge impact on TCO. Now that we have OptiNAND, we've been able to build on the fundamentals of CMR and offer UltraSMR. Something others can't do. This now doubles that capacity. What's gonna be really important here is this is an enduring innovation.
I don't care if it's EPMR 1, EPMR 2, HAMR, heat-dot, ordered granular media in the future. SMR is an enduring capacity gain over the recording physics, so 20% increase. This is really the reason why David sort of said it in a throwaway way, but I wanna really emphasize it, is that you do have to do work to incorporate a host-managed SMR drive. You have to change a bit of your instruction set. That can be a big job or a small job depending on your software stack. Four of our five largest customers have now committed to us or are underway on moving into UltraSMR qualification. What does that mean? It means we've turned SMR into sort of an eight-year overnight success. For eight years, we've had the capabilities.
We've now added to that with our innovation going forward in an enduring way. Now I suspect over the next three years, we'll probably triple the adoption of SMR in the cloud. I think if you fast-forward five years, you know, we could safely project that we may ship at least at Western Digital, more SMR bits than we do conventional magnetic recording bits. Huge innovation. How does it play out for a customer? As well, if you give me the luxury of a quarter, next quarter, I can start to ship these 26 TB drives. Very, very big jump, leap forward from what was state-of-the-art last week. How does this play out for a customer?
Well, when we have the luxury of having these conversations and these roadmaps, they were not a surprise to our customers. They have these samples. They've been qualifying them. They're thinking through already their adoption plans. It plays out in TCO almost immediately. We did one example internally as a project, and this is basically taking racks full of infrastructure, common infrastructure that you would see in our customers, and keeping essentially the capacity, the total capacity the same, but building this with 20, 22, 24, and 26 TB drives instead as a study. You can imagine that if capacity and TCO have a very linear relationship that we're able to have our customers immediately start to see the benefits.
The benefits are the combination of exabyte essentially per compute node is very dense, and so less compute nodes, less rack infrastructure, less cooling, less HVAC, less OpEx. You start to build this very robust TCO model going forward. It extends the sustainability relative to carbon footprint and greenhouse emissions, being able to have less as you densify your storage. I would assert that for data creation, data value creation, and acceleration, the fundamental element that's gonna scale with it is the hard drive. How do our two franchises, you just heard from Rob, how do our two franchises that are leading this cloud expansion, how do they work together? Is there displacement? Is there conflict, or is it complementary?
Well, largely on the back of economics, at least, there are other reasons to tier your data center, workload, and customer SLAs. But on the back of economics going forward and projecting out, again, forecasts forward, the economics don't fundamentally change how our customers view how they tier their workloads. We anticipate there'll be small percentage shift towards SSD over this period of time, and you see that sort of in the different projections relative to bit growth. Layer it all together now because it's really a roadmap of innovation. It's not a single technology. It's not about just recording physics, not just a recording format, not just about architecture, but the combination of those.
We will be able to deliver not only next month a 22 TB to lead the industry in capacity, but over time, we see this technology being able to take us forward into greater than 32 TB per drive, and in the coming years. A couple of notes around operational excellence before I give the mic over to Siva Sivaram to talk about innovation again. How do we think about changing the operational capabilities within Western Digital for meeting the hard drive and cloud challenge? Well, first of all, platform commonality is important to us, so mix shift will help us here. In a sense, today, we have already reduced our client platforms by 25%, and we can actually cover the same set of customers, the same market reach with a much more simplified portfolio. It drives greater efficiency.
Our EPMR novel energy assist recording physics actually allows us quite a bit of efficiency going forward. Now we can take a universal design for recording physics and streamline all of our capacities, essentially from 14 TB to 26 TB today, on a similar design. That allows us then. How does that play out in streamlining us? When I think through my fab, I can reduce the number of different wafers I have to start in my fab by about 66% over this period. Even more importantly, our one important design that streamlines across those capacity points become doubles in utilization. It's a really important factor through our fab process. Supply chain leverage. A return to data growth, acceleration, driving secular growth within the hard drive industry is reflected all the way back through our supply chain as well.
The long-term engagements and agreements that we have with our own customers become essentially a ripple effect through our supply chain. Today, we're deeply engaged with our strategic suppliers on Western Digital's hard drive 2027 architecture. What does that allow us to do? Well, when we can think through growth over this period of time, we can engage on simplifying even at the level of, for instance, electronics and architecture. We'll be able to take that down again by about two-thirds, about 66% in terms of the platforms to cover what is gonna be a multi-zettabyte business going forward. It also allows us to reflect our long-term agreements back into our supply base. More predictability. We can drive with them the right level of capital discipline to meet our goals going forward.
This really starts to change the nature of how I think the hard drive industry and supply chain have been engaging in a declining phase. Automation analytics play into our capability of building essentially end-to-end smart factories. I'm sure one day we'll get Irving to come up and discuss this with you, but just a couple of things that I'm personally excited about is if you think about our capability in our manufacturing network to employ automation across the lines, robotic transport, storage automation. It really plays out in things like things that I can get excited about, OEE, you know, things like overall equipment effectiveness, product cycle times, manual touches go down, and we have a really significant, over this period of time, change in our direct labor productivity.
Also, for instance, we like to create data as well. We have a very large and vast amount of data in our factory systems, which is also helping us quite a bit. Our ability to use those analytics across our own inventory management systems is gonna unlock the way we think about cash usage in inventory going forward. It also, we can reflect that back to how we think about quality. For instance, we have a vast array of customer data coming back to us in telemetry, and in that way, we can use that data today to improve annual field return rates and even no-fault founds, which is very frustrating for the industry.
We see reductions of that, respectively, about 43% over this period of time in six, which are big numbers relative to the scale we operate with. Maybe more importantly, especially for customers, is we see an increase in customer sat and at the consumer level, kind of that net promoter score impact. One of the first changes we made in this VU structure is to bring the leadership and the resources and expertise within our head design, which has generated the kind of world's leading areal density capabilities and head and media, and putting that into a collaborative organization with our recording subsystems and mechanical expertise. That vertical integration is now reflected through our ability to now scale forward.
What's important for scaling forward with if you're gonna try to match the market growth of 30% in data creation is you should own the heads, the media, and the back-end test, and that's what we do. Our control over that also allows us to be very disciplined in how we think about modernizing it and expanding that to meet these needs. Finally, RIC thinking. R&D wise, for instance, another move we made was I built out an office of the CTO for hard drives to think much more long-term, to get us into a very long-term innovation thinking. We've already shifted just since I've been here, about 23% of our R&D OpEx to be focused on technology staging.
Thinking of that three-seven-year timeframe off of sustaining and what we think about as normal development. That's gonna help pay off in that innovation roadmap over this period of time. The other way to think about capital and heads and vertical integration is to just look at it at an industry level. If you look at it at an industry level, I could have pulled again this graph backwards, and it would have been a little bit more dramatic. It would have been a negative number in terms of expansion up until 2020. This is really playing out in two ways. The industry was absorbing. Heads are a good proxy for this growth. Obviously, there's a couple factors here.
One is, head usage per drive has expanded greatly, so the unit economics have gone up as capacity has increased, and the conversion is not actually one to one. A client head does not become a capacity enterprise or nearline head. It actually converts more like two to three to one, depending on where in the process you are. Either way you read it, this industry has to meet the cloud challenge in terms of growth, and it won't be off the backs of any absorption going forward. This is our opportunity, especially at Western Digital. We think about this as a very disciplined way to balance out the supply and demand, as David was talking about earlier, and to start that conversation, just as we have in Flash. How does that play out?
three-five-year financial targets, you'll see a market growth that we showed earlier of 5%, which represents sort of the mid-range of what we think our revenue growth over this period will model, and an expansion in profitability relative to the business as well. One way I think about this is, I mentioned it before, the two-year period we've just been through, it's been quite dynamic. We've, you know, gone from shutdowns on the supply side to changing our processes quite a bit and adapting basically just to keep our employees safe and healthy, all the way to now managing what is a highly inflationary ability to move equipment. Hard drives are quite heavy, so logistics, transportation costs have changed quite a bit in this two-year period.
We've had the return to office now sandwiched before we had the remote worker phenomena that changed the dynamics of our client business for a while. Right in the middle, we may have forgotten about it by now, but we had the crypto phase around proof of space and time. That drove an enormous but temporary spike in hard drive usage. On the demand side, been very, very dynamic relative to our own customers being able to manage the complexity within their supply chain outside of hard drives to meet their demand. As we go forward, if you'll allow me three to five years, I think we can start to normalize out that sort of volatility.
We're back to the fundamental thesis, which is the acceleration of data value and data creation can only be met by the final element of the hard drive. As the hard drive goes forward, we can no longer absorb the growth from a declining client business. We have to invest forward for innovation and for capacity, which now begins to balance out our marketplace. This really changes how we think about our profitability and value-based pricing, long-term agreements, all sort of the elements that underpin this next three-five years. Based on that, hopefully, you took away that we're quite excited about not only the progress that we've made in innovation, but how it's gonna play out going forward.
I think you'll be even more excited once I bring up Siva and he can talk to you a little bit about further about the technology. With that, I'd like to thank you and I'll circle back with questions in a minute. Let me bring out Dr. Siva Sivaram.
Please welcome President, Technology and Strategy, Siva Sivaram.
Wow. Now you can see why we all look up to Ashley. It's good to see a lot of friendly faces again face to face. Yesterday was a very exciting day. We got to see a lot of the technologies we've been working on for a long time culminate into products being shipped. The 26 TB, the big five enterprise SSDs, a refresh of the entire product portfolio. Underlying all that, as David was talking about, is our technology machinery. The technology development organization that I call spans from angstroms to IOPS. Physical sciences, magnetics, depositing extremely thin films, etching deep holes, transistors, devices, to on the other side, computational sciences. Complex ECC codes and algorithms, data structures, firmware. This is a breadth of technology that is being developed across in Western Digital. That machinery, where are we directing it to?
What objectives are we setting up for it? What strategies are we using to produce those products? That's what we're gonna talk about today. In developing that technology development strategy for delivering advanced products, those two pillars on the hard drive side, it's a very, very clearly articulated strategy that we have repeated several times. First and foremost, the basis of everything is areal density leadership. How much bits can I store on a platter? How do I make sure that the physical sciences, the recording systems are developed so that I can maximize the amount of bits that are stored? Go along with it is to reimagine the rest of the system so I can leverage that areal density leadership. Western Digital has always been the leader in areal density.
How do I take that areal density and convert it into a fully integrated hard drive that delivers consistent capacity leadership and TCO leadership? That's the single-minded focus of that side of the business. On the other side, on the flash side of the business, we've talked several times about this charge trap cell that we're talking about. I'll talk a lot more about it, but that power of the charge trap cell, that performance, that capability of that core of our device, how do I leverage that to produce the bit growth that Rob's asking for? The 30%+ growth in bits that we're asking for generation after generation. How do I get that with the lowest capital investment, highest capital efficiency? Capital efficiency, you'll see, is at the basis of the technology development in flash.
These two, leveraging the areal density to form an integrated drive and utilizing our very powerful charge trap cell and producing a capital efficient bit growth. That's the basis of the strategy I'll talk about. Underneath this is a research engine. That research engine, as I was talking about, is a vast collection of extraordinarily talented individuals with in-depth knowledge of magnetics. Almost 50+ years of expertise in magnetics. Four decades of expertise in solid-state memories. Put both of them together, 14,000 active patents. One of the most powerful patent portfolios in the industry. People pay us money to utilize our technologies. That forms the core, and that core when we direct it, you can see the advancements that are coming. These hits are coming one after the other. For instance, OptiNAND that Ashley was talking about. Taking NAND expertise and applying it to hard drives.
UltraSMR, which we'll spend a lot of time today talking about, and its analog on the flash side, on the solid-state side, Zoned Namespaces. Later on, we will talk about other new products that are evolving at this intersection of magnetics and solid state storage, new exciting products that we'll be talking about later in the year. Let's spend some time on the hard drive itself. Areal density, as we talked about, is a relatively simple metric to measure. How much can you store on a single disk? You imagine it is, here's a spinning disk. Individual bits have to be made very small through smaller and smaller grain sizes. Along the circumference, you make that very small, and then increase the number of tracks on the radial direction. The product of those two is what we are trying to produce.
2004, 2005, that PMR revolution started happening. Perpendicular magnetic recording. That had been in development for a decade ahead of it. About 2018, 2019, we first introduced ePMR, energy-assisted PMR. Our ability to write smaller bits through the addition of external energy input into that bit. ePMR, as a technology, is going to live through multiple generations, as Ashley was talking about. The first generation is out. We have already shipped millions and millions of drives. Over 200 EB have been shipped on this technology. It is not one that just happened overnight, it was a development that happened over a decade that got acceptance, proven, and delivered to the customer. ePMR is gonna continue to deliver hits, and Ashley actually showed that ePMR's successive generations of it are going to come in.
Of course, subsequent to that, you'll have HAMR. Subsequent to that you'll have ordered granular media that uses all of this. For the next 10, 12 years, we have development already happening for every one of these future transitions that we are looking for. I wanna make sure you see that these developments take time, take our focused effort to prove the reliability, demonstrate to our customers and show that there is multi-generational leadership in these. This slide, Ashley showed it as, "Hey, look, when we are innovating for the cloud, we do it with respect to making sure there is capacity available, there is cost of ownership available, there is sustainability, there is the storage density is high." I look at it from what this disk is going to be used in the data center. This is not your father's client drive.
Where in a nice air-conditioned room, you have a desktop PC that probably gets filled to 30%-40% of its capacity, and probably gets as a duty cycle of 10%-15%. These drives in a modern data center are stored at 60 degrees C. That's 130 degrees Fahrenheit. Continuously written over and over again. Daily drive writes per day of about a quarter or a half. Every two or three days, you write the entire drive fully. You write it over and over and over again for five years. I need to make sure I don't have failures anywhere in that field. We have moved the goalpost so much farther. If I were just making client drives, we could have been introducing technologies very, very often.
To prove and make sure that we deliver drives to the data center means that I have to prove it to our customers generationally. SMR was introduced in 2015, 2016. Only this year that knee of the curve is taking off. Multiple cloud customers taking off and starting to use SMR. We needed to prove that to them. I just wanna make sure you realize that we are making hard drives that are quite different than what we were making in the client days. This is the reason when we talk about technologies that are delivered, we do it over generationally. Taking that areal leadership, areal density leadership, making a fully reimagined hard drive as a product, what does that mean? We all know about the helium seal. The moment we introduced the helium seal, the viscosity of the medium inside went down.
We reduced disk flutter, we could stack more disk in. Automatically, we saw that quantum improvement in overall capacity we could deliver. That was done almost a decade ago now. We are still continuing to use the same Al-Mg media that gives us infrastructural advantage that I don't need to change the infrastructure. I continue to thin down that Al-Mg. Now I can add more disk. Yesterday you saw our 10-disk version coming out, and we continue to scale that. Recording format. Conventional magnetic recording, CMR, which is still the bulk of the workload. We introduced SMR. SMR, shingled magnetic recording, and I'm gonna talk about it a little bit more. Now UltraSMR. These give that from original 7%-8% to now almost 20% boost in capacity. Micro positioning.
Mechanically, the way I increase the number of tracks so that I have positional accuracy on the head, which means the shoulder, elbow, wrist motions, the high inertia to very low inertia, so I can get very precise positioning. Our innovation to go get that triple-stage actuation. Recording heads, we talked about PMR to going to EPMR, energy-enhanced PMR. Ability to access one disk, one head at a time to multiple access. Most importantly, we talked about OptiNAND. Conventional SoC with a lot of DRAM, which is very expensive, to now coming back and using NAND in it. NAND, which is a thousand times cheaper per bit than DRAM, and I can give a higher ability to store metadata in that drive. Put them all together, you now have a new hard drive. Helium sealed UltraSMR with thin Al-Mg substrates with energy-enhanced PMR, OptiNAND, and triple-stage actuation.
You can see it is not just, "Hey, what is your recording system? PMR, EPMR, HAMR." It is all of these systems put together and our expertise in every one of these facets, mechanical, electrical, signal processing. That's what makes the product possible. What is UltraSMR? What is it that we have done magically, suddenly we can give you 20% higher density from CMR? When we are introducing 14 and 16 TB drives, we could get with SMR a 10% increase. You could get a 14 TB to going to a 16 or a 16 going to an 18. It's interesting. It's a good step forward. Now we are talking taking a 22 and making it into a 26 TB drive purely through signal processing. What did we do?
SMR, for those of you who are not living and breathing hard drives all day long, is the recording track wider than the reading track. I'm able to read a thinner track than I am writing a track. I can shingle it. Like a roof shingle, I can go overlap them. That's the simple idea behind SMR. Now you start to process the data differently. You divide it into larger blocks and write them sequentially. That whole thing was called the SMR system. Now what we have done is add to it TDMR, two-dimensional magnetic recording. Then in case there's a scratch or a defect on the disk, divide up that large block into multiple sectors and divide the sectors all around the track. Add soft sector ECC.
Now you are able to guarantee that I can lose all the fat and make sure you get the maximum density out of it. Put all these together, I have this phenomenon called UltraSMR. UltraSMR is what took us from 22 TB to 26 TB, the world's first 26 TB drive that we are beginning to sample next month. UltraSMR is a word you'll remember over and over again, because this is, there's one thing you take out of the hard drive pitch today, it is how we got to UltraSMR. That's our esteemed colleague, Ashley, showing off his 22 and 26 TB babies. We are going to be shipping 22 TB next month and sampling 26 TB subsequently thereafter. All right, I'll leave Ashley's face there now and move it on to flash. What's happening in flash?
What exciting things are in store in flash technology? Before I go too far, let me talk about our partnership with our long-term partner, Kioxia. It was originally Toshiba, to Toshiba Memory, to now Kioxia, with whom for 23 years, since 1999, we have been working together in developing 15 advanced nodes of flash. The Yokkaichi Mega Fab is the largest flash-producing fab in the world. Together, we produce over 30% of the world's flash. That's all nice. What have you done for me lately?
In December of last year, we crossed over from BiCS4 to BiCS5, the fastest ramp in our history, from an 80%+ utilization of the entire fab in BiCS4 two quarters earlier, by end of the year, we shifted over to majority of our bits coming out of BiCS5, which is what led to the product that Rob was introducing, the new BiCS5 enterprise SSDs. Of course, we are primed to ramp BiCS6 right behind it. BiCS6 will ramp towards the very end of this year and end of this calendar year and into high volume next year. Y7, the new fab that we just announced, is coming into production in September of this year. Of course, we have this innovatively titled BiCS+, which is a placeholder for the next generation technology we'll be talking about today.
You can see that the hits keep coming in the JV. The JV is alive, working very, very closely together. The leadership teams all the way to the engineering level, very close for collaboration to make this successful. How does this work? How does the JV work? Technology is developed jointly between the two teams, whether it is equipment development, process development, integration, designing individual dies so that we can share each other. The technology and design are done together. This goes into the seven fabs currently running together for us, six in Yokkaichi and one in Kitakami. These seven fabs produce half a million-plus wafers a month, half a million wafers plus month out of here, and that output goes jointly to the JVs. There are three JVs, originally started in 2004 and extended all the way to 2034. These are...
The output of the fabs are fed into the JVs, and the output is shared 50/50 between Western Digital and Kioxia. There is a 20% output that goes directly from the JV fabs to Kioxia. That's the general framework of how the JV operates. Out of this come the technologies. In this 20 years of creating this is a conceptual drawing of BiCS6, the one that we are ramping. You can see this extremely complex structure that we are developing. You can see the CMOS underneath the array. On the far right side, you see the vertical bits, those memory holes. The planes are the word lines. On this, you see the staircases, these big Mayan pyramids, where you have to get up the stairs to get to connect to each layer. On top are the bit lines.
It's not something that you create overnight. This is a long-standing development in material science, in device, in technology development of various kinds to integrate a technology like BiCS6 together. This progression of every 18 months a new node from the 48-layer BiCS2 to 96, 64 layers to 96 layers to 112 layers to 162 layers, and then now to the next generation, 200+ layers that is coming up. Every 18 months, we do develop these technologies. However, when we develop these technologies, it is not everything is just in time. You develop core technology concepts ahead of time and have them staged so that when it is time, you integrate them together to be that product at the right cost point, at the right bit growth rate that you're looking for.
For instance, we put a lot of focus on lateral scaling. Lateral scaling means you wanna make sure those memory holes are closely together, smallest pitch. You remove as much of the overhead that they use as possible so that you can have as much of it producing only the bits and not just overheads. So that produces the lateral scaling, and then you multiply that by the vertical scaling. The more the lateral scaling, the less vertical scaling you need to get to the same bit growth. Vertical scaling is the number of layers you add. You make each layer thinner and thinner as you go along, so you don't have to etch deeper and deeper. So the combination of lateral and vertical scaling gives you the overall density. Of course, in order to achieve that, architecturally, you are scaling it.
Architecturally, you are trying to figure out, "Okay, can I etch all of them in one go? Can I etch them in multiple steps? Can I bond wafers? Can I double bond wafers?" These kinds of ideas are coming in, so you develop and stage them. Last but not least, logical scaling. You want to make sure that you are able to store two, three, four, five bits for each cell. You put them all together, you are able to deliver the most capital efficient, at that point, bit growth rate that you're looking for. Now, we talked about all of these, but all memory technologies in the world are developed cell outwards. You start with the cell. The stronger the cell, the stronger the core of the cell, all these work out better. There we talk about the charge trap cell.
The charge trap cell, for those of you who are not familiar living in semiconductor land all day long, is the world's highest volume shipped primitive semiconductor device ever made by mankind. It started shipping in volume maybe a decade ago. Now, by far, it's orders of magnitude more than transistors, resistors, capacitors. Put them all together, this charge trap cell is the highest volume shipping primitive cell there is. That cell is extraordinarily complex. That 3D that you're seeing is multiple materials and multiple layers, multiple electrical structures, charge storage, blocking layers, transistors, all into that. That cell, which we have developed over a long time, is the world's best. How do I say that it's the world's best so confidently? It's because every year we have this conference called International Solid-State Circuits Conference, sponsored by the IEEE.
We all self-report as to how well we are doing. This year's ISSCC conference, our competitors take that same charge trap cell on an identical performing device and told us that, "Hey, we can get 40 MB/s performance out of it." Of course, we reported in the same conference, not knowing what they were gonna show, 60 MB/s. Why that's possible? Because that cell has the best distributions power that I can now take that high-performing cell and leverage it to do others. What can I do with it? If I have a higher performance, clearly Rob's products have higher performance. That's easy. I can trade the performance for latency. I can trade the performance for endurance. I can trade the performance for retention. That's next set of it.
What you can do most is use that performance to figure out, I don't need quite as many planes to get the end per-product performance. I don't need to spend additional capital in trying to make a device. The biggest advantage of that charge trap cell performance is that our capital efficiency gets to be very good. I'll show you why. This is the industry's capital spending requirement over the last decade. You can see for every successive node, it is getting very, very expensive. It's that the technology is getting more complicated. Almost $1.5 billion per 1% growth. You're looking for 30% growth. That's $45 billion that needs to be spent every year by the industry. This is getting to be very, very complicated, very, very expensive, very quickly.
In this situation, clearly, you get a huge advantage if your own capital intensity is lower. What that charge trap cell allows me to do is the following. This is data reported by Yole, just came out less than three months ago. This shows over the last three years, where were you in December 2019, to where will you be at the end of December 2022 over the three-year period. For however much bits you had to however much your bits you are going to have, how much capital did you spend? CapEx per incremental petabyte. You can see when you are building greenfield, when you are not utilizing your existing capital, you are not reusing your capital equipment.
If you are making leaps, you are skipping generations, I'm in this generation, I'm gonna skip the next generation, I'm going to the third generation, your CapEx goes high. This is a clear indication of how our own expertise and ability to use our own equipment, our own reuse, and leveraging of that central cell allows us. This is three years, so it's in the 18-month cycle, this is multiple generations in it. You can probably plot this graph over the last 10 years, and you'll probably see the same result. We have always been the most capital efficient in our production of NAND. How does that transpire? We talked about the lateral scaling, the vertical scaling, the logical scaling. Here is again data from the Solid-State Circuits Conference.
To get to the same bit growth rate, we used 162 layers. The competitors used 176 layers. Our die size for that critical most important die for us, the Terabit X4 QLC die. All of you are gonna talk to me in the Q&A, I'm pretty sure, about QLC. This is a QLC die. You can see our die size is smaller with lower number of layers than anyone else. The most important number you should look at is that how much bits can it produce on a single wafer. On that QLC die, 100 TB per wafer. 100 TB per wafer. Barely five, six years ago, we will be producing one, 2 TB per wafer. We are producing 100 TB on that single wafer through the scaling across all of those factors.
Put that all into a flash memory roadmap for us. On the y-axis is the terabyte per wafer that I just talked about. You can see over the next 10+ years. We have a good roadmap that technology is being developed and staged to go all the way up to 500+ layers. We started with 3D NAND, two tiers. We started to scale the cell pitch. We started putting more staggers between the overhead lines. We probably will introduce wafer bonding, we'll do multi-tiering, multi-wafer bonding. These technologies are already in the works, being in place so that we can introduce them at the right node for every node that we'll be developing, that then Rob can convert them into products.
Which brings us to today's next node that is coming up, tentatively called BiCS+. Kioxia and us will come up with a more imaginative name than BiCS for this. This is engineered for the data center workloads. You can see, again, one of my baby pictures I love to show off. 55% bit growth rate per wafer when we go from BiCS6 to this node. You can see if I had an 18-month cadence, you can see that 30%+ growth annually out of this. 60% increase in I/O speed. The program bandwidth I just talked about, which we hold the world record at 60 MB/s on that cell, we are pushing that up even further, another 15%.
By the time this is developed and ramping, this would be a quantum leap in the performance of NAND. The big takeaways out of all of this, on both sides of the house, on both the hard drive and in flash. First, in hard drive, areal density leadership. Areal density leadership forms the basis of our strength, and then we reimagine every aspect of the drive so that we can now talk about shipping 22 TB CMR and sampling 26 TB SMRs, UltraSMRs. On flash, industry-leading roadmap with the best capital efficiency. There is more to come at the intersection of magnetics and solid-state drives for the future on both sides and new products. We will continue to drive this development engine to produce new technologies for Rob and Ashley to take into products. With that, let me call our CFO, Wissam, onto the stage.
Please welcome Chief Financial Officer, Wissam Jabre.
Thank you so much. Hi, everyone. Thanks again for joining us today. You heard from David this morning, the data opportunity and his vision for the company. You also heard from Rob and Ashley about their strategies for their businesses, as well as the financial and operational improvements that their business teams are putting in place. You just heard from Siva about the great innovations that our advanced technology teams are focused on. I'm here now to bring it all together and talk about the business model and our capital allocation strategy that are both designed to create long-term value for our shareholders. As I start, I wanna focus, maybe spend a few minutes on a few key elements of how we deliver shareholder value.
We have sustaining and growing both revenue and profitability in the large markets and growing markets that we serve. In parallel, we focus on cash flow generation, and the way we do that is by continuing to be disciplined about financial management and capital deployment. You heard us all talk about disciplined capital deployment, disciplined investment, but also ROI-based investments. We utilize these approaches because this is how not only do we grow profitability in a sustainable manner, but we also focus on growing our cash flow. The reason we wanna focus on continuing to grow that cash flow generation, it's to enable us to continue to delever, reduce leverage, delever the balance sheet, and get us in a position where we can restart capital returns to our shareholders.
Before I get into that and get to the model, let me just share with you some of the improvements which we saw over the last couple of years. From FY 2020 to FY 2022, we've pretty much saw our revenue grow by approximately 6% on a compounded annual growth rate. What this translates by business is flash grew by approximately around 12%, and the hard drive business grew by 1%. The interesting thing in the hard drive business is during that time period, we saw a good inflection in where the revenue generated from cloud now more than offsets the declining client and the return to growth in that business.
In fact, when we look at year-on-year, the business is planned to grow from $8.2 billion- $9.1 billion. Over the same time period, we saw phenomenal expansion in profitability. We ended FY 2020 at 26.9% gross margin. We're projected to end FY 2022 at 32.9% gross margin. This is a six percentage point expansion in a couple of years. In parallel, we saw a faster improvement in the operating margin, which is growing from 9.1% to 16.8%. This is really a great demonstration of the operating leverage in our model and the capability of expanding profitability at a much faster pace as we continue to expand gross margin.
With the new structure in place, the business units, and bringing the focus on operational and financial execution, we saw also the operating expenses as a percent of revenue decline over the last couple of years from around 18% to where we'll exit this year at around 16%. This is really a great achievement because when you look at the OpEx itself, it hasn't changed much. Within the composition of our operating expenses, we saw SG&A decline over the last couple of years, which is a trend in the right direction. We also saw R&D increase by around 3% over the last couple of years.
Now, that increase in R&D was really very, very focused because what we saw is really two things: increase in investment, but also rotating the investment towards a higher growth portfolio. You can see from what we've announced yesterday some of the great innovations that came out of that. As we grow our profitability and revenue, we also see our cash flow generation improve. In the last 12 months, we've generated an EBITDA of around $4 billion. The EBITDA margin also has expanded by almost seven points over the last couple of years. All in all, you can see how with the expanding revenue, increased profitability, our operating leverage is really working.
This is where, you know, when you look at non-GAAP EPS, in the last year, it's expected to grow for FY 2022 by 80%. Over the last couple of years, it went from almost $3- $8+. In parallel, free cash flow generation also continues to accelerate with our last 12 months at around $1.6 billion with a 40% growth. All in all, this last couple of years, actually last three years, we've seen good improvement top line. We've seen a four-year high for our profitability from a gross margin and operating margin perspective. We continue to see acceleration in not only the earnings power, but also the cash flow generation of the business. Okay, let me switch gears and talk about our plans going forward.
The markets we compete in are not only large, but also growing, which puts us in a really great place to capitalize on this opportunity. Our hard drive market is expected to grow over the next 3 years by 5% on a compound annual growth rate. Our flash market is expected to grow by around 11% on a compound annual growth rate. Over the same time period, we continue to evolve the company into a cloud-centric company. You can see that from FY 2019 to FY 2022, obviously on the hard drive side, our three-quarters of our revenue is now generated from cloud. On the flash side, we have a sizable part of the revenue that is generated from the cloud.
More importantly, when we look forward into 2025, we're expecting to have approximately half of our revenue generated from the cloud. That basically over the last or over sort of the snapshots from FY 2019 to 2022 to 2025, we see that cloud revenue grow from around 35% to 40, 42% to 49%. The value in that is, one, obviously this is a more stable market for us, but also more profitable. What this basically enables us to do is to continue to capitalize on the operating leverage in the business and accelerate our profitability and cash generation in a slightly more predictable way. On increasing profitability, you heard this morning from both Rob Soderbery and Ashley Gorakhpurwalla about their plans to with respect to operational improvements as well as financial, basically gross margin expansion.
I'll just pick on a couple of points here in the flash business. Just as Rob and Siva have pointed out, our NAND costs, we're pretty much the leaders on, in terms of the, on the NAND cost side. We continue to focus on that because that's a key element for us to continue the, not only the cost reduction, but the gross margin expansion.
In parallel, the business is focused on driving a better portfolio mix towards higher value products over the time horizon. On the hard drive side, as you heard from Ashley, there's key focus on gross margin expansion with many initiatives, starting from the top line into the cost reduction aspect, into the supply chain leverage to enable not only the cost reduction and the productivity in the sort of our manufacturing machine to help us drive costs down, but also focus on value-based pricing to also help on the top line. All these initiatives are focused towards driving gross margin expansion and operating margin expansion, and that's pretty much a testament of the, again, the operating leverage in our model. In parallel, there's a strong focus on optimizing cash flow.
Here, you know, there's the typical elements of cash flow. One is optimizing the working capital with increased focus on inventory management, as well as supply chain optimization and all the various elements of the conversion cycle. Basically reducing that conversion cycle to accelerate cash flow generation, but also in parallel, the disciplined capital investment. Disciplined capital investment really means that on the flash side, we wanna continue to be capital efficient and basically maintain that capital efficiency, meaning generating the lowest cost bit by putting the least amount of capital. On the hard drive side, we wanna continue to focus on aligning supply and demand. We obviously wanna continue to grow our business, but we also are very, very mindful not to overbuild capacity.
All of this is happening through really an ROI-centric lens. Why do we focus really on driving optimizing our cash flow is, again, to continue to reduce that leverage. In fact, over the last couple of quarters, not only did we continue to pay down debt, in fact, the last couple of quarters is almost $1.3 billion-$1.4 billion of debt pay down. We also amended our credit agreements to get us in a position where we extended our maturity. We do have a very low-weighted average interest rate that helps us also spend less of our hard-earned cash on servicing the debt and more on investing in the business and reducing the debt. This is our three- to five-year financial targets. You saw the targets from the flash business in Rob's presentation.
The goal is to grow at 10%-12% CAGR, and that 35%-37% through cycle gross margin. When you look at the gross margin range, and the reason why we feel confident that this is a margin that we can, we can achieve, one, we've pretty much shown that we pretty much can get close to those numbers. The reason we think these are really good targets for us, it is, as Rob described it's 5-7 percentage points expansion from where we've seen the business operate over the last few years. On the hard drive, we're targeting a revenue growth of 4%-6% on a compound annual growth rate, and you saw the growth in the industry being around 5%.
On the gross margin, the non-GAAP gross margin, we're targeting 31%-34% margin, which compared to the last 12 months at the midpoint, is almost 2-3 percentage points better. If you compare it to the last few years, it's almost more than 400 basis points better. Again, the reason we feel confident that these targets are good for us, one, we're almost at the 30%+ in the hard drive business, but also we have a long list of initiatives that the team is focused on to continue to drive the efficiencies as well as gross margin expansion. What does this translate to for the company? For the company, this translates to a revenue growth targets of 7%-9% CAGR.
Basically, when you aggregate the two businesses, it gets us to that range. Then on the non-GAAP gross margin, we're targeting a 33%-36% compound annual growth rate, which when you look at it, we're almost at the bottom end of that range, but we still have room to grow. Also, when you look at it over the last few years, we're pretty much almost around again 3-4 percentage points higher at the midpoint. From a non-GAAP operating margin, what this translates to is 17%-22% range. Basically, it implies around fourteen to sixteen percent of operating expenses as a percentage of revenue.
On the CapEx side, this is the cash CapEx, which is really the portion of the CapEx that basically shows up on our cash flow statement. We're targeting 8%-10% of revenue. Within this, if we sort of decompose that between hard drive and flash, it basically translates to approximately 4%-6% in the hard drive business and the remaining in the flash business. That's really where the three- to five-year financial targets are for us. Let me switch gears and talk a little bit more about the balance sheet. What we've done over the last couple of years as well, where we see our capital allocation framework shaping up. In the last three years, we continued to delever our balance sheet.
We paid down approximately $3.4 billion-$3.5 billion of debt, which really got us to a point in December of 2021 to transition to investment grade. That's really great because it enabled us obviously it gives us a better lower cost for our debt. Over the same time period, we continued to reduce that leverage ratio from a peak of around 5.7 to around 1.8 or so in the most recent quarter. The one point I wanna make here is these are the leverage ratios based on our reported EBITDA. Obviously, in our credit agreements, we have some add backs and. Before I switch to that, there's a couple of points I wanna make here.
One is we're committed to maintaining that investment grade, and also we're very focused on maintaining a lower leverage ratio. What this helped us do, all this debt pay down, it really helped us get to a solid capital structure, where from a liquidity perspective, we have around $4.76 billion between the $2.5 billion of cash and cash equivalents and the $2.25 billion in revolving credit facility. As of the end of fiscal Q3, we had $7.25 billion of debt with a leverage ratio that's based on the credit agreements of around 1.4x . What you see also from the capital structure is that our debt maturity now is 85% of it won't mature before 2026 or later.
That's pretty much gives us a lot of flexibility to operate, and we have ample liquidity to continue to invest and grow the business, going forward. Let me talk a little bit about our capital allocation framework. Here, we really have three priorities. One is to reinvest in the company. This is where we would continue to focus on our R&D investment, our disciplined capital deployment to maintain that growth in the business. The second priority for us is to continue to reduce leverage. The way we think of that is really twofold. One, we're targeting a leverage ratio to be between one and 3.5x .
Now, you could argue that this is a bit of a wider range, but the one point I would like to make here is obviously we still have a certain aspect of cyclicality in our business. At the trough of the cycle, we'll probably expect to be on the higher side in terms of leverage ratio, and at the peaks, we'll expect to be on the lower side. What that translates to in terms of the target gross debt, we really want to target a gross debt level that gives us buffer capacity at the trough of the cycle, and also, we're committed to maintaining our investment grade level. What does that mean from a gross debt perspective? When we look at the.
If I sort of look back at the last cycle, that translates roughly into a $6 billion-$6.5 billion of gross debt. As we go through that, continuing to delever the balance sheet, when we get to the right point, we will be restarting our capital returns to shareholders. Let me in closing just basically reiterate. Obviously, we're competing in great markets. They're large, they're growing. We have revenue growth that's led by really secular demand. We have a compelling product portfolio that you heard about this morning from the management team and also saw some of the innovations that we've announced yesterday.
We continue to focus on profitability and cash flow generation. Our profitability is really driven by solid execution, continuing to be innovative in our technology, and also evolving the portfolio in a way that provides us higher and higher value. We have a disciplined capital management that's pretty much at the heart of everything we do, and it's really designed to create long-term shareholder value. That wraps up my prepared remarks. What I'll do next, I'll ask the management team to join me on the stage, and I'll also ask Peter to help us maybe coordinate our question and answer session. As we're ready, we'll ask you to raise your hands when you're ready to ask a question.
All right. Wait for everybody to get up here. Thanks for indulging us going through that. We had a lot to say. There's been a lot of change in the last couple of years, so really appreciate it and we look forward to questions. I suspect there might be some. Okay, Peter, please.
All right, great. Thank you. Sidney Ho with Deutsche Bank.
Thanks.
Thanks for doing all the presentations. I have questions on the value-based pricing. David, you talked about the hard drives business potentially moving to value-based pricing.
Mm-hmm.
How have your conversations with customers been going on that subject, and how are you thinking about how it actually will be implemented? Is it kind of like cost-plus type of arrangement? Any goalposts that you can share with us will be great.
I don't think about it as cost plus. I think we understand. I'll let Ashley come in on this as well, but you know, we have a lot of insight into the value we deliver with each successive generation of drives. We know how to model that. We work very closely with our customers on that. I really think this starts with the next generation of drives, right? The 20 and 22 beyond that. Again, look, value-based pricing starts with innovation. It's not about repricing something that's already out there. It's about we deliver a better TCO proposition through innovation.
I think one of the things you saw today from the team is there's a lot of innovation that's been in the works that is now coming to actually being shipped in the products. That gives us a chance to change that conversation, which we are with our customers. Ashley, you wanna?
That was pretty eloquent. I'll stand on that. Yeah.
Okay. Maybe just a quick follow-up. On the gross margin side, in the past, your predecessors talk about the NAND Flash gross margin on the long-term basis needs to be more than 40% to make sense to put in the investment. Is that still the threshold you're thinking? How does that jive with the 35%-37% that through cycle margin target that Rob talked about? How does that equation change now that there's more entrants, more suppliers in this market? Thanks.
I think maybe I'll start and then Wissam and/or Siva may also have some commentary. I think we're very comfortable with the economic engine that we've defined up here. It's a better economic engine than we have been operating for the last few years, will produce more economic returns, and it's self-consistent. As I think about the investment levels as that, we're you know, rolling up to Wissam for him to aggregate, you know, those hang together with the number set we've provided. Siva, maybe you
Yeah.
Historical context.
You can see the combination of the gross margin that Rob's generating with the need for the investment and the overall bit growth rate that we are trying to target and the market share that we are targeting. They're all consistent in the sense that 35%-37% that Rob's talking about is the right kind of place for us to be in. Clearly, the cycles will go up and down, but he's talking about a through cycle number that we feel comfortable.
Yeah. Maybe I'll add just a couple of comments, Sidney. You know, I mean, obviously, the business is not there today, and it is economically sound. It is a cash flow generating business that's growing over time, and it's, and Rob and the team and Siva are able to reinvest in innovation that that continues to fuel that growth.
Thank you. It's Jim Suva here from Citigroup.
Mm-hmm.
Thank you so much. It's been a long time since we've gotten together in person. My question is, Dave, you brought in a lot of new management, a lot of changes. From us as the outsiders, we really don't see what's going on inside the company except for the presentation today, which was great. My question is now that you have two different business leaders, was the thought to get the two business lines to work together better or simply to have them more financially responsible or a combination of both? The reason I ask is, from the outsider standpoint, investors are a little bit disappointed. A lot of your benchmarks today have been based upon the year 2020, which was the trough year, so easy comps. You know, we've had a contamination issue and all these different things.
Can you let us know in your mind about running it or doing two different businesses, the main milestones or the thought you had behind doing all that?
Thank you. Yeah. I think when I came in and really dug into the business, it's about finding where we need focus and where we need you know, where we can leverage commonality between the businesses? Clearly, if you're gonna drive an HDD business or a Flash business, as I said, you need focus. You need to make really smart ROI-based decisions on how you're investing in your roadmap. You need to have a strategy for how you're gonna capture value in the market. They're different businesses. They're sold to the same customer, but they're different businesses.
As I talked about, even from a technological point of view, in some markets, they're substitutes, and in some markets, they're complementary technologies. It became very clear to me, it's a little bit related to the question Sidney asked, like you can operate the business much more efficiently if you have better focus and you make better decisions about what you're gonna invest in and what kind of products you're gonna bring to market. I think that's what Ashley and Rob did. Like, that's why that was the first move, which is like, we've got to make better decisions. We've got to get our portfolio. We've got to get our products in a better position, and we got to get our portfolio. We have a lot of portfolio optionality. There's a lot of markets to play in.
We only have 14-16, in Flash, 14%-16% of the supply. Thinking through that equation, really complicated question, and then how do you make the R&D investments to, like, land a product two or three years from now to maximize value. That same thing in HDD, like I think one of the things hopefully you saw today, it's like it's not just like a crank-turning exercise where, oh, like we have an 18 and then we do a 20 and then we do a 22. It's a culmination of years and years of R&D, and how do you think about bringing that to market in a way that like the largest data center operators in the world are going to bet their business on your product.
That clearly required very focused individual thought, and that's why it's where we started with setting up a BU structure, bringing people in to think about it differently. Now, there's other parts of the business, for, as I said, clearly, I mean, like there's no sense in duplicating a go-to-market organization to talk to the same customer and the same person. It's the same buyer. In many, many cases, literally the exact same buyer. There's no need to like, there's absolutely no need to drive separation there. I mean, what you want is focus on commonality. You don't want to run two consumer franchises in the world. You want to run one consumer franchise in the world. Those became places where the way we are organized is fine, and we just continue to leverage that.
Then, you know, what you saw from Siva is there's an element. You know, we're in a business driven by innovation and driven by long-term, like really difficult, as he said, what did you say? From angstroms to-
IOPS.
Angstroms, IOPS, whatever. People with material science, physics, magnetics, so that's very long cycle time stuff and kind of keep that in a way where you get some. You know, there's a lot of people in there. They work with each other. You get a lot of innovation value from people talking to each other. New things happen. OptiNAND is a good example. Like, you just create this environment where innovation can come out of it, and then you drive that into products. That's a little bit how we thought about how, at least how I thought about it and how I thought about it again when I say the order of how we're gonna do things. You can't. Like, it became clear early on we need to do things differently. The question is, what are you gonna do first?
What are you gonna do second? What are you gonna do third? What are you gonna do fourth? What are you just gonna leave alone? I think that's a little bit what we talked about. You know, at some level, you'd like to say, "Well, I'd like to do everything first." But that's just, you know, in a company of 65,000 people where you still need to bring in $20 billion of revenue and increase your profitability, you gotta move along that path in a very, I think, disciplined and predictable way so that you get the results as you go. Peter.
Hi, it's Tim Arcuri at UBS. I had two questions. First of all, if in theory you were able to split the businesses, let's just play that out and say that if you were able to do it. I think when you bought SanDisk, I think you had about $500 million worth of cost savings. So my question is, in theory, if you were to split these businesses up, would you have to add that cost back? Number one. Number two, I had a question on gross margin targets, and I was a little surprised that your HDD margin target is 31%-34%. I mean, obviously you have a lot of leadership. You have, you know, 22 T, 26 T, and your competitor's already in that range.
I guess my question is, you know, why is the target gross margin for HDD not better than where your competitor already is, given that you have so much leadership? Thank you.
All right, the first one. Look, I think like answering theoretical questions about the way things could be is a little bit difficult. The one thing I will say is I've been in this company two years, a little over two years. The comment I made at the beginning was a very serious comment. Not only are we looking to execute better, we're very open to what are other structures that are gonna create more value. I will say, when you look at this company, you have to evaluate each individual option in a significant amount of detail to answer some of the questions you're asking. And there are a lot of external parties that are involved.
You know, what Siva talked about, the partnership we have is extraordinarily valuable, and you wanna be very careful you don't like put too much jitter in that. You have governments that are very interested in what we're doing. We probably have the largest cross-border JV that's out there. There's like layer upon layer of things you have to consider, and we're very open to considering all of those. You know, you gotta make sure you don't break a lot of stuff to get to a model that's better. Completely open to a model that's better, but we have to go through that process. Kind of each individual scenario needs to be evaluated in a significant amount of detail to get to some of the answers that you're asking.
It's really hard just to say, "Well, what if we just did this? Would that happen?" I think we can go through that in more detail some other time. I guess on the gross margin piece, again, I'll let Ashley comment. Like, one step at a time. We are in a very complex world right now with, I mean, this fiscal year, we will have $244 million in COVID costs. There's a lot of unwinding to do, supply chain shortages. There's a lot to for the world to work through. As we unwind that, we have a lot of confidence in our technology and the value we're bringing.
As we start to see that take hold and this world gets unwound a little bit more and some of the variability gets taken out of it, I think Ashley said it well. You know, if we'll come back and update the model as we get incremental confidence in that.
Hi. Joe Moore from Morgan Stanley.
Hey, Joe.
I guess the biggest question I get about you guys in sort of thinking about the stock and the value is the NAND business on kind of the next five years, and not just for you told us kind of what you can control. As you think about the industry returns on NAND, I feel like you've achieved a number fairly close to what your
Forecasting on kind of go back 15 years in NAND. I guess, how do you think about the next five years of industry returns relative to what you've seen historically? You know, how much of, you know, when you think about 35-37, is that thinking that the industry environment's comparable to what we've seen better or worse? Thanks.
Well, I think there's, that's a layer question, so there's multiple factors on both the supply side and the demand side that drive that. Overall, we view the industry as getting better because the demand drivers continue. You know, clearly, there's a lot of noise in the market right now, but the fundamental things I talked about, which is cloud growth, you know, the world in which we operate in becoming, you know, so dependent on the performance of the client and this tremendous consumer growth along with, you know, continued strength in mobility, all that I think portends well for, you know, not just the short to mid, but the long term on the demand side.
From a supply side, technology, you know, as technologies get more challenging and the end user market that we sell into gets more complex, there's more room for differentiation, portfolio, technology investments, you know, that 80% figure from an SSD perspective. You know, a more complex and more diverse world which in a heavier intensity in terms of the investment required tends to produce, you know, a different set of economic decisions. I think, you know, in our modeling, we're not expecting, we're not relying on a market tailwind on that 5%-7%, but we believe that the market is gonna help us over time in terms of driving our long-term gross margin targets for sure.
Peter.
Yes, thank you. It's Mehdi Hosseini from Susquehanna International Group. Two follow-ups. These are practical questions, not theoretical. Going back to your target three-five year, it seems to me that the biggest incremental driver is going to be share gain in enterprise SSD. I wanna hear what are the key nuggets or drivers that gives you the confidence that you can do that, and I have a follow-up.
Yeah. I think that starts with the fact that eSSDs are complex. The reason why eSSD is attractive market is it's proven very, very difficult for both vendors as well as do-it-yourself models to prevail and succeed. It results in relatively few people arriving and delivering, and that results in a better economics, and that's what makes the market attractive. In terms of what gives us our confidence, you know, I think our confidence is built in the fact that we've got the flywheel of intellectual property and capability running. It clearly took the Western Digital time to get that flywheel spun up, and this was a multi-generational three-, four-, five-year investment to really become relevant in the NVMe eSSD space. That investment's paid.
I think we're gonna see that in this transition right now, which is the BiCS5 transition, which, from a development and engineering perspective, kind of our side of the problem, has gone exceptionally smoothly and reflects more of an incremental advance as opposed to maybe the heavy lift of the first generation. At the same time, we know exactly what has to happen to go through a hyperscaler qualification process, and we're, you know, very steeped in the capabilities to make that happen. That's the momentum signal that feels like we're over the hump. In parallel to that, you know, I think that there, again, this eSSD problem continues to be hard, and it's getting more challenging based on the proliferation of the different architectures and different design points.
That's all better market and more direct point of momentum that I think we can identify.
Sure. Just wanna go back to the recent earnings conference call, revisit the HDD nearline exabyte assumption. Can you remind me what is your assumption for calendar year 2022?
Yeah, I don't think we gave a calendar year 2022 number.
I don't think we did.
What should we assume?
Well, we didn't guide calendar year 2022, Mehdi. We've really focused our discussion on the fiscal Q4, which is what we guided.
The reason I ask is you just highlighted the fact that you're gonna be shipping 20 TB and the 18 is high-volume manufacture, and I'm just trying to figure out how the mix shift is going to impact overall calendar year exabyte share.
I think we talked about the crossover, but we didn't really guide exactly when by sort of pinpointed to the quarter and so on. I mean, obviously, nearline is still growing. It's a good growth business for us. I would probably leave it at that.
Revisit next quarter.
Exactly.
We'll always revisit next quarter. Absolutely. Who's got. Over here. Hey.
Thank you. Aaron Rakers at Wells Fargo. I've got a couple questions. I wanna first build on Mehdi's question a little bit on the enterprise SSD side. We've heard you guys talk the last couple of quarters about, you know, multiple different design wins. You know, I think there was three cloud vendors that you talked about, some system vendors as well. Your aspiration of moving from 8% to 16% market share, is that a BiCS5 aspiration? Do you think that that'll happen over the next 12 months, or is that a longer term aspiration? I'm just kinda curious to how we gauge the trajectory of that ramp, and I have a quick follow-up.
I think that will extend into the BiCS6 era. You know, that aspiration is very difficult 'cause this market's growing at 40%. You really have to grow fast in order to just sort of keep up with the market, which actually puts pressure on our bit allocation. Obviously, today, all our bits are deployed. As we're feeding more relative bits into the enterprise SSD market, we have to pull those out of other markets. We're somewhat constrained in how quickly we can move that ramp. The market responds better to either those thoughtful rampings. I think we'll see substantial progress in BiCS5 because that gets us into really a less encumbered bit allocation problem.
In BiCS6, what happens is we actually expand the portfolio, and that opens up dramatically more TAM. Think of it as a two-step journey.
Yep. That, that's helpful. Then the second question is on the hard disk drive business to the gross margin discussion again at 31%-34%. You mentioned earlier $244 million of COVID-related costs flowing through the model this fiscal year. How do we think about your ability, when do you think you can be at that 31%-34%? Have you invoked any kind of pricing strategy changes to kind of offset some of the COVID costs? Because I think the hard drive vendors in general have not necessarily followed the system vendors on some price increases. Just curious how you've done that.
Sure. I think, you know, David said it well on just sort of when we get there, you know, one step at a time relative to, you know, as uncertainty unwinds, you know, we'll have to update the model relative to that. Relative to pricing, you know, we've publicly announced that we raised our pricing going into this quarter. That, as you can imagine, we have multiple mechanisms and channels. Distribution channel is probably the quickest to respond to those. You know, it'll go all the way out to longer-term agreements would take more time. We've begun the process of passing on some of the inflationary costs to our customers.
Hi, Wamsi Mohan, Bank of America. Thanks for the presentation.
Two questions. First, on through-cycle margins, the 5%-7% increase. When we think about the underlying levers to that, you know, you spoke about cost leadership down 15% in cost annually. That's not very different from what you've said in the past about mid-teens decrease in cost structure. The thing that stands out the most seems to be maybe product mix, and I just want to understand is that of that increase in through-cycle margins, if you were to attribute the four things that you highlighted, is product mix really the key underlying assumption that gets you up into that through-cycle margin range?
It's a combination of product mix and value engineering primarily at the product level. We see the NAND and cost performance are performing well and continuing to perform well, but not necessarily moving into that five-seven. Today, my product portfolio, I have gross margins that vary between 20% and 60%. It's pretty obvious that if we can blend up in that range, we can generate higher gross margins. Specifically, we are as we, you know, enter new markets, the number, job number one is to make sure that we hit the performance, that we hit the scale, that we get to that market quickly.
The more as we mature in markets, that opens up opportunities for cost reduction, and we see a 3%-5% cost reduction opportunity within our SSD portfolio, which then will help from that translating to gross margin opportunity. It's probably 60/40 portfolio mix versus cost structure.
Okay. Thank you for that. Dave, I know this is a little bit of a tougher one, but in your opening remarks, you said that you looked at various options when you took over the role and then evaluated sort of the various options that you considered. Any color at all that you can share on what those options were, what your thought process around those was? Was there any outside interest in the assets at that point to the degree you can comment? Thank you.
I mean, I guess what I'll say on that is when I came into the company, it became clear to me we need you know, there's execution changes we need to make to get more out of what we had. I think you can see that, you know, we've made significant changes, and hopefully you see we're getting more out of it. Like I said, we got more to go. We can still do more. In many ways, we've gotten to the starting line with the full team now over the last even just the last quarter of what we need to do. We'll, you know, we'll continue to layer that on and on.
I think in any business, you're always gonna look at what playing field are you on and the environment around you, and can you know, how can that be influenced to help you create value at the same time? Without going into any particular detail with any particular party, we've obviously been thinking about that very, very deeply with all kinds of different ideas. Like I said, there's lots of constituents that have to be satisfied and have a say, and kind of navigating through that. If we, you know, get conviction there's a better model, then that's what we'll move to. I would say it's an ongoing process. As I said, we continue to do that. We're doing both.
I think that in any business you wanna pursue as many parallel options as you can to create value. They all have different time horizons, some of which you can predict and some of them which, you know, there's other factors outside of your control as far as your ability to execute those. Certainly, you know, all options are active. Again, I'm very serious about what I said. We always welcome input on this, and we hope to have a very constructive and deep dialogue with our shareholders.
When new people show up interested in our company and have ideas and are willing to put capital to work, that's very constructive, and we look forward to a very robust dialogue on that. Yes, Peter.
Hi, it's Toshiya Hari from Goldman Sachs. Thanks so much for hosting the event. I had two questions as well. First one, I guess kind of as a follow-up to the prior question on how the business is set up and your thoughts going forward. David, obviously you've introduced quite a bit of change over the past two years plus. You seem to be really happy with the joint venture set up. You seem to be very happy with the BU set up. I guess going forward, any further change that you're looking to introduce or any parts of the business or how you run the business where you're kind of not satisfied with as of now?
I mean, look, you guys can all tell me areas I shouldn't be satisfied with, and I have a list of things I'm not satisfied with. I'm not gonna go through all those here, but I mean, clearly, we can improve across a number of dimensions. You can always improve. I think in any business you can always improve. You should relentlessly be looking for how we improve. I mean, obviously, the inventory question has been asked of us a lot, and I think we have opportunity there. Just to give one small example. I'm very happy that, you know, after two years, we've put together a team, a leadership team that I believe can get the most out of of the company we have.
I feel very good about that, and we've been able to attract some very high caliber leaders that have executed very large businesses. I think we're in a good spot. Like I said, we'll continue to improve every single day, and be relentless in driving improvement in the company. There's lots of areas where we can still continue to improve, but I think that's the case in any business.
Got it. As a quick follow-up, one for Rob in terms of how you guys are thinking about the competitive landscape in NAND and what's contemplated in your medium to long-term guidance in terms of gross margins. You know, you talked about managing your business and managing capital spending to sort of a 14%-16% market share number, which makes sense, and I think a lot of your competitors kind of make similar comments. You do have, you know, five or perhaps six players in the market still, depending on how you look at it. You know, Samsung perhaps, you know, kind of in the same place in terms of how they communicate their market share aspirations.
SK hynix may be a little bit more aggressive, and then you've got YMTC, which is emerging. As you throw out that medium to long-term gross margin number, how do you think about the competitive landscape there evolving? Thank you.
Well, I mean, it's a good position with Samsung having 1/3 and ourselves and Kioxia having a bit over 1/3. It's a good position to have 2/3 of the market be signaling rationality and consistency. I think the other factor that we think about is just the double down what Siva was talking about. While the technology roadmap is amazing, it's extremely challenging and extremely difficult, and there's a long distance between, you know, a twinkle in the eye about what might be possible and actually having, you know, products in high volume production that are at scale that can run for three-five years, that you can build, you know, premium phones and data centers and clouds on. You know, the problem continues to be challenging and robust.
I feel that from a market perspective, we've got big players in the market signaling a degree of discipline, and we also have, you know, a robust and challenging problem in front of us. That to me, you know, drives in a more neutral to positive view of the marketplace than we have seen.
Hi, it's Karl Ackerman. I had two questions. The first one is for Rob. Rob, what's the mix of LTAs today for cloud SSDs, and what initiatives do you have to expand the mix of LTAs in your enterprise SSD business that may help you achieve your 16%-20% market share over time?
It's in the LTA for the cloud vendors, like, these are massive scale operations that consume huge amounts of products, and they need multi-quarter visibility to that, and they need a supply chain that works to support them in an efficient way. 100% of the cloud is under some sort of LTA. When we talk about the innovations that we're thinking about, it's about changing the structure of those LTAs to further incentivize the right behavior. The right behavior for us is, you know, smoothing of the demand signal and the supply signal, as well as a degree of consistency in our ability to supply. That has both benefits and trade-offs on both sides.
The way I think about the LTAs, we have a basic volume-based commercial frameworks that are working well, and now we're layering on additional mechanisms to drive longer term business agreements.
Thanks for that. The other question is for Dr. Siva Sivaram and Ashley. You articulated that SMR took about seven years from introduction to inflection. You now also show that HAMR won't come out until circa 2026. I guess, what are some ways to improve the time to market for newer hard drive technologies, such as UltraSMR, that would arguably drive a significant uptick to your 30% HDD exabyte CAGR and revenue target through 2027? Thank you.
Sure. Thanks for the question. Obviously didn't communicate how well we think our roadmap sets us up for success going forward. You know, the innovations are, as we, I think, both said, across the spectrum in the hard drive space. They take time to develop. The reason they take time to develop is because a couple of things. One, it's an integrated system, so you know, we have to change things together. The second is we are dealing with fundamental physics. You know, adding capacity to a drive, to give you a long answer is we have to reduce the grain size that represents a bit. It becomes less stable. As you reduce the grain size, makes it harder to write, we add energy.
You've seen the now commercialization that Western Digital's brought forward of energy assist in an ePMR spectrum, soon in an Flux spectrum, soon in the HAMR spectrum, right? It's essentially a continuation of our energy assist commercialization. UltraSMR is an interesting adaptation in that it actually came very quickly after our OptiNAND introduction, so the actual latency, if you will, to invention was around the architectural change, which was done with great discipline and with ensuring that we could bring this at scale to our customers. Once OptiNAND is in place, then we've been able because we've had...
We have very, very sophisticated algorithms and software capability to bring this essentially from, let's say, Q4 till today and introduce a 26 TB, essentially, you know, a 30% gain in ADC in that time period. Perhaps not impressive enough, I think pretty impressive. Our customers certainly agree with that. You can see that in the adoption scale of that. I think the roadmap that we have going forward is one that, again, we wanna lean on scale and predictability, because our customers have to know that our roadmap can actually come to fruition when we say it is. We're committed to that energy assist roadmap.
Okay, I think we have time for two more questions.
Dom Kim, IT Research. I have a question on Siva. It's very interesting to see that you show multi-wafer bonding in 200+ layer flash roadmap. It seems all NAND suppliers are moving in that direction, which Chinese YMTC is doing today. How should we think about the move in terms of IP, cost, CapEx, and so on? Any color would be great. Thank you.
Yeah. Multi-wafer bonding, first wafer bonding, then multi-wafer bonding, has been part of our roadmap for several years. It's not just we came up with it now. We've been working on this for now well over five, six years. All wafer bondings are not the same. You do it at certain level of maturity, at a certain use case for the future. You're absolutely right. YMTC talked about multi-wafer bonding and introduced it in their 64 or 48 layer nodes very early on, which we thought, we took notice of, but, that was not part of our own evolution. In the end, as you said, we focus on how much capital and cost per bit that it leads to.
When we stage these technologies, we find which technologies are added together to form an integrated product, and this is one of those arrows in the quiver. We will introduce it at the right time. I'm not saying right now it is. We're not predicting to you when we'll be using it, but these technologies are kept ready for use in the future.
One more.
Hey. Hello, guys. This is Eddy Orabi for Krish Sankar from Cowen. What a fun start few months for Wissam. We got an activist and an analyst day. Congrats on product announcements. Pretty impressive. For HDDs, based on your engagement with customers, how fast do you expect 22T to ramp? And second, on UltraSMR, does this drive address some more niche applications compared to CMR? Thank you.
Sure, I'll take both. You know, we haven't actually begun shipping the UltraSMR, so I'll go backwards for a second. The UltraSMR drive, though, does represent such a gain in capacity and capability in TCO that it's not a surprise to our customers, obviously, that we've been working with them on the roadmap for some time. The excitement level relative to adopting that into their infrastructures has been very strong. It is new technology. It needs their qualification, and we partner with them to bring that qualification to bear. It typically can take, depending on the customer, from OEM to a cloud vendor. They want to sync it up with their infrastructure changes, their internal refreshes, and their resources.
This can be a quarter to three quarters, depends on their own timelines, and we'll support that for them. Relative to, you know, the mix and the ramp of our high cap product portfolio, we sort of have 18, 20, 22 in the CMR space. We have a very premium product in the 22 space, and you'll have the same type of timeframe of ramp. Over, you know, say our next fiscal year, we'll have all three products being qualified in our customers, and we'll be managing that mix for them. Sort of from the very mainstream 18 to the more premium 20 to the essentially, you know, unique in the industry 22.
Some customers will choose based on their own refreshes, their own infrastructure, to stay with three in their infrastructure, to move off of one, move all the way to the 22. We're working with our customers on those ramps and qualifications today.
Okay. Look, I really wanna thank everybody for indulging us. We really appreciate the opportunity to go through the business, talk about how excited we are about this, and we look forward to continuing the dialogue going forward. Thanks again for your time today, and we look forward to talking to you in more detail later today if we see you around here and also absolutely in a number of events we'll be at throughout the quarter. Thank you again.
Thanks, everyone.