Western Digital Corporation (NASDAQ:WDC) Goldman Sachs Communacopia + Technology Conference September 10, 2024 10:30 AM ET
Company Participants
Amitesh Bajad – IR
Robert Soderbery – EVP and General Manager, Flash Business
Conference Call Participants
Toshiya Hari – Goldman Sachs
Toshiya Hari
Okay, we’d like to get started. Good morning, everyone. Thank you for coming to day two of our conference, Communicopia and Tech Conference at Goldman. My name is Toshiya Hari, I cover the semiconductor space at the firm. Very happy, very honored to kick off day two with the team from Western Digital. We have Rob Soderbery, EVP and General Manager of the Flash Business.
We also have Amitesh Bajad from IR. Before we dive into questions, I’ll hand the mic over to Amitesh for the disclosures.
Amitesh Bajad
Thanks, Toshiya. I’ll quickly read the safe harbor disclosures. We will be making forward-looking statements in today’s discussion based on management’s current assumptions and expectations, including with respect to our product portfolio, business plans and performance, market trends and dynamics and future financial results. These forward-looking statements are subject to risks and uncertainties. Please refer to our recent financial report on Form 10-K and our other filings with the SEC for more information on the risks and uncertainties that could cause actual results to differ materially from expectations.
We will also be making references to non-GAAP financials and a reconciliation of our GAAP and non-GAAP results can be found on our website.
Back to you Toshiya.
Question-and-Answer Session
Q – Toshiya Hari
Great. Thanks, Amitesh. Thank you to both for attending. Really appreciate your time. So Rob, I wanted to start kind of high level. You joined the company about four years ago. It’s been a journey with COVID and the market ups and downs. What’s been the journey like so far? And as you look ahead, what are you and your team’s top goals and priorities?
Robert Soderbery
Well, I think it’s a good time to pause and reflect. When we — this management team was created four years ago around Western Digital, we had a pretty simple goal for the Flash business and the HDD business. The very first step and I think the very first inspiration there was that we did want to manage these businesses in a more focused way. I was part of that transition in that I was brought in to run the Flash business, colleague actually brought in to run HDD. And so we actually started the process to think of these as more unique businesses at that time.
My own goal for the Flash business really straightforward. Number one was to increase our exposure to a more diverse set of end markets and particularly enterprise SSD. In order to do that, we had to do a bunch of retooling under the hood of our engineering capabilities and our execution capabilities because, frankly, those had been aspirations of the company and they have not been able to fulfill.
And so that was the second was real like a really replatforming. And then the third was to evolve our capital models and our supply demand, our investment models to build a more thoughtful way of investing in the NAND business.
So I think at this point, on the eve of separation, we feel really good about those initial goals. We feel really good about our guidance from the May 22 analyst conference, even though it’s been a crazy journey, and we feel good about sort of really closing out the first chapter and looking ahead to all the exciting things that are in the market in front of us.
Toshiya Hari
Great. So Rob, as you noted, you guys are on track to complete the separation of the two businesses by year-end. What have the key learnings been as you’ve gone through the process? And apart from the potential dissynergies you guys talked about on the recent earnings call, how do you see the separation impacting your NAND operations, if at all?
Robert Soderbery
Well, we are excited about the progress we’re making a separation. Frankly, it’s mostly a back-office problem, setting up separate transactional rails and infrastructure rails for the company, ERP, Salesforce and so on. And so that works nearing completion. We’ll be going through, what they call, the sort of the soft spin of soft close, where we practice that separation process internally. That’s here in early October. So we’re feeling good about where we are in that separation process.
From a business standpoint, the business has been very focused in terms of having a Flash business unit and having focused on this market. But clearly, the separation does give us a chance to be more articulate about what we’re going to do with Flash as well as be more focused and 100% of the energy of the company behind the Flash business and the HDD business.
And then, of course, I think the big benefit of the separation is from an investor perspective, these assets really behave quite differently and they’re likely to have different shareholder returns policies, different growth curves, different volatilities. And so we think that from an investor perspective, they’re both interesting investable assets, really appealing to different types of investors.
Toshiya Hari
Got it. Shifting gears a little bit. On the near term, on your most recent call, you guided like-for-like pricing in the current quarter to be up in the low single digits on a sequential basis. That’s a bit of a deceleration relative to the past few quarters. What are your expectations?
What’s driving that? And what are your expectations from an industry supply-demand standpoint beyond the current quarter?
Robert Soderbery
Well, I guess it’s first important when you think about pricing is that very few of the markets are actually priced at the average price of the market. All these markets have a bit of display around that center point of the market. So when you’re looking at pricing, you’re looking at two effects, you’re looking at how is that median price moving and you’re looking how is your mix and mix maybe it changed by seasonality, by what specific end customers are doing.
So within our current pricing guidance, I think we’re reflecting the fact that in today’s market, there’s a general slowdown weakness in the transactional markets. Now that weakness is 60%, 70% higher volume than it was in — so we’re not taking about a return to COVID kind of weakness, but we’re talking about a little bit of end market softness causing conservatism in those transactional channels and that really affects things that are consumer facing.
So the consumer business, channel, some of the PC OEM. At the same time, the AI boom hit the SSD business squarely, we caught the big wave March, April of this year. And that continues to power forward, and that continues to be a ton of demand around eSSD and particularly moving to high capacity. So it’s really a tale of two markets here that blended sort of single digit up, reflecting kind of the mix of those markets and what’s happening inside them.
Toshiya Hari
Okay. Just as a quick follow-up, Rob. You didn’t mention mobile there. Is mobile kind of — would you bucket that with the transactional more consumer-facing applications? Or would that be sort of in the middle between a very strong enterprise SSD market versus the more trend…
Robert Soderbery
Mobile seems to be doing a bit better than PC. So I don’t know it’s booming, but before I talk more about AI, but our sense is that the mobile vendors have been a little bit more precise and going after the AI-enabled device opportunity and maybe consumers are seeing the more obvious audio and video and collaboration type of — social media type of applications on mobile being directly connected to AI benefits.
Toshiya Hari
Got it. Medium- to long-term NAND demand drivers, I wanted to sort of touch on, I think big growth expectations for the industry have generally come down over time, not just for you but for the industry overall. I think that’s due to a lot of large numbers and also the unit growth potential, if you will, across things like PCs and smartphones have come in a little bit. What are your most recent medium- to long-term big growth expectations, which applications are you most excited about? And I think at the 2022 Investor Day, you had guided to a 10% to 12% revenue CAGR target. Is that still the relevant appropriate range for you?
Robert Soderbery
Well, there’s a lot going on that question. So I’ll refer everyone to my new era of NAND webcast, which gives an hour answer since I probably don’t have time for that today. But essentially, the way to think about this is the sort of the last decade of NAND has really been driven by elasticity effects. Consumers don’t really need a terabytes in their mobile device. But if you give them that terabytes at the same price is 5, 12 gig, they’ll take it, right, on the come.
And so you’ve had elasticity effects really driving a lot of NAND demand. The problem the industry has had is a lot of that demand was somewhat empty calories. So you were chasing price, you’re driving volume, you were driving exabyte growth, but you were not necessarily driving revenue growth and definitely not driving gross margin growth. I think we’re shifting to a market which is going to be more organically driven. And so you’re going to see the NAND industry growth rate from an exabyte perspective, come down.
But that does not imply that you’re going to see revenue line come down. I think it implies you’re going to see the gross margin line in a much healthier place as you really fundamentally see organic drivers where we don’t have to use price. And clearly, what’s happening in AI right now is all organic. We’re not inducing anybody to buy devices. So we see a more organic driven what does that equate to — equates to a 20% type of growth rate instead of a 30% type of growth rate with AI being plus or minus 5% on that number.
Toshiya Hari
Okay. Great. And on that point, as it pertains to AI, as you noted, there’s been a very sharp increase in demand, especially for high capacity enterprise SSD. It’s very difficult from where we stand to sort of disaggregate that growth into AI versus restocking. What are you seeing in your business today? What are your thoughts there?
Robert Soderbery
Well, parts of that are — when you — we’ll step back and look, I don’t want to keep pitching my stuff, but we did go out and we talked about the AI data cycle kind of unpacking what AI is driving. And big picture, there’s three drivers — three AI drivers. The first driver is this high cap phenomena. That one is easy to track because essentially, right now, all of this high and ultra-high cap is going into AI and these are people — back in the day, we would have called this ETL, extract transform load in the database days.
These are people doing all of the preprocessing on their data sets, sectorization and so on. And if you’re building an AI data center, you need to spend a couple of hundred million dollars building a high-capacity, high-performance fast data lake for data preprocessing. So that’s our opportunity, number one, easily trackable by tracking ultra-high capacity units. So Gartner and so on, start reporting that.
Opportunity number two is in training and inference in the data center, that’s mostly driving lower capacity, higher performance, PCIe Gen 5, very demanding. We just launched our PCIe Gen 5 product in that space. And that is harder to disambiguate with core hyperscale consumption. And then opportunity number three is the edge. And inference at the edge with mobile first, we see client, auto and other end user application. So we’re working on that, trying to assemble an AI number — for right now, it’s 100% of that high cap and a significant amount of the Gen 5 stuff.
Toshiya Hari
Okay. And specific to some of the areas where perhaps you don’t have 100% visibility, what’s going on from an inventory restocking standpoint, again, specific to enterprise SSD. Do you feel like your customers are consuming inventory, building inventory? Is it sort of flattish? Any views there?
Robert Soderbery
Well, I think it’s very normal. In some cases, people were selling out of specific SKUs. High cap is definitely fully sold out. I think some of our peers have sold out. We’ve seen other segments where customers come to us and said that their suppliers are revectoring bits into AI.
And so that’s benefited us. We’ve reached number one in client’s market share and client PC market share in both the OEM and the channel. So it’s really the first time we’ve ever been number one in both halves of that market. And that’s because some of the bits have flowed out behind the scene. So definitely, essentially on allocation for AI, unlike these other markets.
Toshiya Hari
Great. There sort of a side question. We’re asking most of our companies at this year’s conference, how you’re using AI inside your company since AI is a big topic, whether it be R&D, whether it be design, whether it be day-to-day operations at Western Digital, do you guys actually leverage AI more so today than, say, two, three years ago? And if so, what are some of the benefits that you’ve been able to identify?
Robert Soderbery
We have a major program to leverage AI and conventional ML within the engineering organization. And my personal sense is that, that is a — it’s a very big, rich, deep vein that there’s tens, hundreds of millions of dollars of savings. There’s use cases really across the engineering pipeline. Think of a testing and qualification problem, right? You have 100,000 test vectors? Well, what’s the most efficient way to apply those 100,000 test vectors to test a problem, fundamentally is an AI and ML problem.
What order do you apply, what value to get out, which vectors — so anywhere in software development, software test, hardware development, sample reduction, tester reduction. So I think it’s super significant. I think it’s really a question of extraction time. And for us, we’re probably headed down a two-year path to take $100 million of productivity and deploy it in other places based on the benefits of AI and ML.
Toshiya Hari
Okay. It’s pretty significant.
Robert Soderbery
Very significant. Yes.
Toshiya Hari
Yes. Right. Cool. Shifting gears, I wanted to spend some time on your manufacturing setup and your technology road map. You guys have been involved in arguably one of the most successful and long-standing cross-border JVs and perhaps TMT. I think it’s close to 25 years. People always call us and ask us what we’re doing right. So yes, so what have been the sources of success? And from your standpoint, again, Rob, you’re still relatively fresh to this. What are the pros and cons of the JV setup in your view?
Robert Soderbery
Let’s start with what is the JV setup. So the JV setup has multiple parts. At an R&D level, our R&D activities for NAND are 100% co-mingled, robust dialogues, back and forth between the two companies, a 50-50 allocation of engineering, complete shared resources. We developed a common set of die, a common set of devices that we then take to market. At the manufacturing level, it’s similarly 50-50 co-mingled in that, if you go to Yokkaichi or Kitakami you don’t see a WD side and Kioxia side as one facility and there’s an economic JV construct on top of that facility, which splits the bits 50-50 coming out of the JV portions of the facility.
Everything beyond that, and that includes wafer packaging, eSSDs, UFS, all the system levels stuff, everything on that is completely separate and all the go-to-market activity is separate. So that’s the construct of that. And then from an investment perspective, we make our investment decisions 99% of the time in lockstep. So we’ll discuss our views on the market, we’ll determine the next allocation of capital into the JV and then both companies put that capital allocation into the JV.
I think that there’s a couple of different things that have made the JV really successful. The first is just the technical underpinnings. Both teams are extremely, extremely strong. Kioxia Toshiba, heritage of NAND, SanDisk, very much an influential early mover in the history of NAND. So you brought in two technical teams have a lot of respect and a lot of synergy.
And at this point, very co-mingled in terms of who does what. And so there’s a tremendous — just to the R&D aspect, a tremendous amount of synergy. And then the benefits of scale, 30%, 35% of the bits, large fabs in the world. Our — when I talked to Tim Archer, he tells me, I don’t know how you guys do it, but you buy less from me than anybody else buys from me relative to the amount of bits you build.
Toshiya Hari
It’s not good for them.
Robert Soderbery
So not good for them, but good for us. So that scale and that focus have resulted in systemically lower capital needs and a better return. I think really the proof point of that was in the downturn. We headed into the downturn. I think the talk track in the downturn was other NAND vendors were ahead of Western Digital, and we’re doing great.
We go into the downturn, it turns out those other NAND vendors that were ahead of Western Digital didn’t have the cost structure that was commensurate with what they’re representing and their numbers were horrific. Our numbers were not awesome, but on a relative basis, we showed much better performance through the downturn. And it’s because it was reduced capital intensity, better architecture and just a better portfolio. So big picture, look, the JV works. We’re super happy with it.
We think we’ll keep working. We’ve renewed through 2034 for the — or extended 2034 for new fabs in Kitakami and have a lot going on.
Toshiya Hari
Right. That’s a great segue into my next question. The market tends to focus quite a bit on layer count in 3D NAND. We’re often led to believe that the higher the layer account, the more advanced the technology. But at WD, if anything, you prioritize lateral scaling and logical scaling while also pursuing vertical scaling.
Talk a little bit about that dynamic and how you’ve managed to grow in line with the market despite, again, a lower capital intensity profile. I guess you sort of touched on that now, but if you can…
Robert Soderbery
Yes. So this new era of NAND topic essentially, as you increase the number of layers in your 3D NAND, you start to face diminishing returns, right? 50-storey hotel is a lot more expensive than 25-storey hotel. You no longer get incremental savings from adding layers. And that’s kind of where we are in the NAND space right now where if you look at an industry level, a new node can generate about 60% more bits than a previous node, but it only generates about 15% cost reduction. So lots of bit growth, relatively small cost reduction.
So it just reduces the incentive to be chasing layers. It’s still efficient, still more productive, still gives you more capital efficiencies. So you still need to invest in that direction, but it’s not like it’s been in the past where there was a huge return to invest in layer account. So let’s replace that? Well, I think the analogy here is in 2012, sort of the same thing happened in DRAM, right? DRAM technology innovation slowed, and what you saw was a proliferation of DRAM and focus into LPDDR for power, kind of DDR for capacity and now HBM for performance.
And we see a similar thing happening in the NAND market with this focus on mobile application — powerful mobile applications, right, capacity for big data applications and then performance for core inter thread, AI and compute applications. And that’s really where a lot of our innovation energy is going. And so it’s not even so much 3D scaling versus 2D scaling, it’s more about how fit are you for applications.
And that makes a big, big difference because — for example, I’ve delivered with BICS, I’ve delivered a 2-terabit die, that’s a 2-terabit QLC die. That’s all and wonderful, but if it’s not highly performing, it’s not very useful.
So we really focused with CBA technology of delivering performance on that 2-terabit QLC die that actually exceeds the TLC performance of the previous generation. So it’s a combination of the performance and the capacity and the other capabilities that really matter. So we think you should be looking for more of those differentiation stories. And like that device will be the building block of ultra-high cap enterprise SSDs.
Toshiya Hari
Okay. And Rob, you talked a little bit about this in your response, but can you speak to the significance of CBA technology. And when you compare and contrast BICS8 with BICS6, what are the performance gains? And to the extent you’ve made public comments, what should we expect beyond BICS8?
Robert Soderbery
Right. So BICS8 uses a technology called CBA, which essentially is wafer bonding. So previously, we would build the CMOS that does the logic and memory array on the same way for using the same devices or the same capital equipment. And the problem with that is that the memory array is — requires a different voltage temperature, environmental envelope than a modern CMOS process, always had to compromise your CMOS process.
Now what we’re doing is we built two wafers, a CMOS wafer and a memory array wafer, and then we bound those wafers together, we shave off the top and that becomes your device. So if you can pull off the bonding, that has a huge number of benefits.
The first benefit is you can optimize that CMOS wafer. So our BICS8 is a higher performance than competitive ninth generation NAND. We’re fighting up a level in terms of just pure performance, and that’s straight because we have access to better CMOS. The second thing you can do is you can think about upgrading the CMOS and the memory array separately. So if you want to have a low cap high-performance device versus a high-cap low performance device versus other configurations, you can upgrade those separately and develop a more interesting portfolio.
So think of BICS8. BICS8 is a great node. It ramps as fast as any or faster than any node we’ve ever ramped. We got through all the risks of building CBA and we’re just enjoying the benefits. Think of that as the foundation for a multigenerational road map where we can build on that.
It will be a long-life node. It descends to be a long-life node. So it’s a big, big leap forward to big derisk in the underlying business. It will be the heart of ultrahigh capacity on enterprise SSD as well as our next-gen QLC efforts.
Toshiya Hari
WD, I think, is in a very unique position given your presence in both NAND and HDD obviously, very soon, that might not be the case given the separation. But as of today, from your perspective, how do you envision the market share balance between enterprise SSD and near line HDD evolving in the data center going forward. Right now, it feels like both you guys on the NAND side and the HDD side are enjoying really strong growth, but how do you see that evolving going forward? And how would you characterize kind of the TCO differential today and how that sort of moves going forward as well?
Robert Soderbery
Well, so the — it’s a great time in both NAND and HDD. So on the HDD side, we’re sold out. We’ve got long lead times, mid- to high 20% CAGR from a capacity growth rate. So there’s no indications in the market that somehow there’s something funny going on between these two products. What’s new is that this AI data lake problem in the past might have been the kind of problem through HDD at.
But HDD doesn’t have the performance to go after the AI data lakes, and therefore, you’re seeing these high — HDD scale SSDs going after that AI data lake application. As of today, depending on who you ask and how you measure, there’s a 4x to 5x difference between per bit cost between HDD and enterprise SSD.
Maybe that comes down to 3x or 4x. So for conventional applications, think about your family photos on Facebook that you’re looking at once a year, None of those deep content applications are going to live anywhere but HDD, right? So that’s going to go forever, that works forever.
The question is what applications can afford 3x to 4x cost disadvantage because their TCO will result in a better equation. And there are definitely obviously, AI is one of those applications will be more of those applications. We are seeing interest from data center operators for space power and longer-term TCO benefits. So we see modest effects to the benefits of Flash over time, but it’s unlikely to be anything that’s measurable in terms of influencing that HDD topline growth rate, particularly given the capacity constraints on HDD.
Toshiya Hari
Got it. That’s helpful. In terms of your enterprise SSD competitive position at your Analyst Day a couple of years ago, I think at that time, you had about 8% market share. The goal was to double that to 16%, obviously, the market sort of went into a downturn post that. So all the design wins didn’t necessarily materialize. Where are you today in that journey? And how confident are you that you ultimately get to that sort of market share level given the pipeline you have today?
Robert Soderbery
Yes, it was a rough journey. We did all that fundamental work I talked about earlier in terms of retooling the technology platform. We arrived in — what was that the spring of ’22 there — actually a very competitive platform and essentially the eSSD market just vaporized in front of us due to this long-term kind of hyperscaler overrunning the system. So we had to pause. We almost went to zero.
We’ve now come roaring back, and we’re coming roaring back on the basis of three different platforms. First of all, the pre-downturn platform we built is still a good PCIe Gen 4 storage platform. And so that is ramping up, and that’s what you’ll see as we report the quarterly numbers and we talk about what we’re doing in the market, we report the quarter of the numbers. And we’re on track to get to about 10% of our bits. So it’s becoming material. It’s becoming a good part of the portfolio just based on that.
And then there are two new offers. The first offer is a 30- and 60-terabyte high cap AI PCIe Gen 4 device for that fast data lake application. That’s getting a lot of interest from the storage OEMs. We’re doing a lot of the AI data lake work as well as the hyperscalers. And then the second new platform is that PCIe Gen 5, a lower — higher-performance platform that’s got interest from major hyperscalers from China accounts from channel accounts. So those both have a lot of interest.
So when I stand back and look at that, I think that David and I have talked about like let’s get to a decent place for our own portfolio mix. And we think initially, that’s 10% of our bits moving to more of a fair share where we have the same share in eSSD as in the rest of the market, so I think 13%, 14%. So sort of the lower end of that range that I talked about. And I think that’s all very doable in the next couple of quarters.
Toshiya Hari
That’s great. Shifting gears, maybe a couple of questions on industry structure and how to think about supply-demand medium to long term. You’ve seen some consolidation in NAND over the past several years. You do still have five to 6 very capable, competent suppliers competing for market share. How do you at WD see the market structure evolving over the medium to long term? And do you believe you can still hit that 35% to 37% through-cycle gross margin target with the current setup? Or do you feel like there needs to be some consolidation?
Robert Soderbery
Well, we can clearly get to that 35% to 37% because there are now. However, that was not the goal, right? The goal was to get to 35% to 37% as a through cycle number, which requires in the up high cycle exceeding that, right, and having that balance out. I think — we think the market is radically more healthy today than it was three to four years ago. First of all, there has been material changes in the market in terms of the number of players.
But even more so, there’s been material changes in the number of players with disruptive aspiration in that we’ve been extremely vocal about what we see as our own capital strategy and how we see the market evolving.
I think competitors — but we’re a little more articulate in this market just because of where we sit and who we are. But I think if you listen carefully to what you hear in the marketplace, there’s a lot of alignment in terms of a focus on that organic demand. So benefits of focusing on organic demand or we don’t have to move price in order to spur that demand. And so now we’re talking about a market being driven by organic demand, a cost structure being driven by slower like cost reductions. The cost reduction has been the less and that profitability wedge increases going forward instead of decreasing growing forward.
So I think as we look at the marketplace, we’re pretty comfortable that, that’s a good forecast. Obviously, there’s a lot of interest in shifting capital to HBM as well as clean room space, which is in high demand. And so that’s also providing essentially a tailwind for NAND market. And then ultimately, all this takes. Those capital lead times are four to six quarters.
And so with the lessons of the market downturn kind of — in the very recent history, I think we’re all being quite conservative. And so there isn’t going to be some sort of snap back from a supply perspective, right? We’re going to — it’s fine at the industry if we chase demand. And it’s fine if supplies are a bit tight.
Toshiya Hari
I guess you sort of alluded to this in your response. But at the NAND event that you hosted a little while ago, you spoke to this market equilibrium rate of 18% to 20%, the WFE bookings and shipment numbers would suggest that your peers are onboard and sort of acting rationally. But is that sort of the perception you have?
Robert Soderbery
Yes, that 20% number, which is supported by that low double-digit WFE number. So I think we’re kind of creeping back into that low double digit versus WFE had crept up that $20 billion kind of number. So we have a really good model for exactly how much capital it takes for the industry to move and take the industry about $45 billion in capital to move from node to node. So you can kind of correlate the WFE investment to the amount of investment it takes to move from node to node to the bit to bit growth rate.
And I think there’s been this mindset that maybe we’re going to stay on this node-to-node arms race where every 18 months are going to a new node. The capital engines behind that are just not going to support that. So we’re going into this era in which there may be advances in technology, but there’s not going to be unlimited bits at that new fancy node.
So my view is the NAND industry has been a bit of a strange industry in that every 18 months, we delivered a new product. That new product was higher capacity, higher performance, lower power, came in a small package, smaller package had all these new things that it did. And the customer’s expectation is that they would be able to pay less for that new product than they would for the old product.
That’s very unusual for a technology market, and we’ve really reversed that where customers now understand that the high-performance, highly capable, state-of-the-art product is actually going to be in demand and not putting enough capital in to necessarily create a plentiful supply. And so behind the scenes, you’re seeing a rotation where, look, maybe a consumer doesn’t need those new bits, whereas an enterprise application or automotive application wants those bits. So the customers are sorting out behind the scenes who can afford to really to pay for what.
Toshiya Hari
Got it. I did want to pause here and see if we have any questions from the audience. If you do have a question, please raise your hand and we’ve have mics in the room. Okay. If not, one last question. I think, Rob, you spent a decent amount of time with investors and analysts maybe less so than David and Wissam. But based on what you hear, what you read, the questions that come your way, anything about the WD story about the NAND story, the markets you participate in, that we collectively underappreciate or overlook or misunderstand if you will?
Robert Soderbery
Yes. Well, I think if you stand back and you look at these two businesses, both have been through enormous transformations. If I go back four or five years ago, what was the setup for the hard drive business, we had massive excess capacity coming off of our long-term investment in client computing, right? The hard drive industry was scaled to produce hundreds of millions of client drives. All of those client drives went to Flash.
And so really, what’s been happening in HDD is, yes, there has been a technology progression for — and we’ve been fortunate enough to lead against our peers in the space from technology progression. But behind the scenes, we’ve consumed all of that capacity. We did — we were doing a good job of sort of consuming that organically. And in the downturn, we just restructured it out of the system. So we’re getting to this point now where the HDD business is set up in a fundamentally different way that you don’t have this overhang.
And now so to bet on HDD, you only have to believe one thing. You have to believe that the amount of data that we’re going to store in the world is going to go up over time. And if you believe that statement to be true, you can invest in HDD.
And I think you’re seeing that economically and financially in our numbers. You’re starting to see it in our competitors’ numbers. And there really is fundamentally nothing that’s going to change that trajectory going forward. So HDD has come out of this phase in a much better spot. And then similarly, Flash, to me, this downturn in flash in the NAND business was a cathartic downturn.
And I think people underestimate just how deeply it impacted the industry. There’s a lot of pride in this industry like our company is a company full of engineers and full of NAND technologists and full of innovators and there are people like that all over the industry. And all those people want a healthy industry and want a healthy go-forward model.
And I think it’s been a collective realization of, well, what can and can’t this industry support and what strategies do and do not work in this industry. So I view the Flash industry being on a fundamentally healthier track, and I see no signs of the issues that have plagued the industry in the past.
So the good fortune, good timing. We feel like we’re headed into this separation opportunity with two assets here that are unique. They’re different, probably appeal from a different investor base perspective, but both are set for performance set for excellence. So really pleased to be where we’re at.
Toshiya Hari
Awesome. Thank you, gentlemen, for joining us at the conference, and congrats on all the success.
Robert Soderbery
Yes. Thank you.
Toshiya Hari
Thank you.
Read the full article here