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 risk and uncertainties. Please refer to our most recent financial report on Form 10 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 reference to non-GAAP financials, and a reconciliation of our GAAP to non-GAAP results can be found on our website. Thank you, Joe.
Great. Well, thank you, guys, and thanks for coming. I know this is your first time here as management of a pure-play NAND company, so I was joking up in your ears. You actually have as Western Digital, but now you're here representing SanDisk. Just maybe talk about the path to getting to this point, and it seems like your enthusiasm from here, what value this separation unlocks.
Yeah, so first of all, thanks for having us. Yeah, it's been quite a path from here through a strategic review of kind of what's the best thing to do with Western Digital and how to situate the company to announcing this in October of 2023, I think, to getting here pretty much on schedule as we kind of were shooting for, and now I have the whole process done, and to be here on stage running SanDisk. It's great to have the company back. It's an iconic brand. A lot of people have said very good things to us as we did the investor day and did the launch.
I thought over the last couple of years that this is actually the right outcome for Western Digital, and I'm very happy to be here and super happy to be very happy with Irving and the team that's going to run Western Digital. They're going to do fantastic. That's a great franchise. I enjoyed managing it for the five years that I did, and I think making a few changes to that, but I'm thrilled to be now part of SanDisk and to drive this company forward. I think we've got an incredible opportunity in front of us.
Well, I enjoy it too. We don't have that hard drive conversation getting in the way of the real stuff. It's all good. So maybe we could talk about the cycle a little bit and your big picture view. You've talked about maybe a smarter supply-side management going forward. There are two pure-play companies now that are public that can kind of lead the way to that behavior. I think investors are kind of like, "Well, we trust you guys. We trust Kioxia and you guys," but somebody else may disrupt the apple cart. Just how do you respond to that when you sort of think about the industry supply-side dynamic?
You know, Joe, I just thought deeply about this industry the last five years, being a participant in it, going through, we've kind of reached mid-cycle levels a couple of times and then turned around. What led to that? Obviously, went through a pretty traumatic downturn. And to me, this is just we're thinking about how to manage the business differently. I think a lot of the things people thought about the industry. I think the industry was really set up for a long time thinking about just growth, really drive growth in the industry. And that happened. I mean, we had one chart in our investor day. It was 20, I think 2005, where the industry shipped an exabyte of NAND, and now we're 970 something. So we've got the growth side figured out, and now it's a question. And also we have the technology side figured out.
I mean, NAND is one of the most productive semiconductor technologies out there. We've got 3D NAND that we can produce more NAND. We apply R&D. We can get more output. And it's really just about thinking about the business from a little different perspective about how do we now shift into more profitability and growth, not just growth at all costs. The way we talk about the business, the idea that it's very informative to me talking to people like our CTO that's been in the industry for several decades. And in the past, when the market was when you're in a downturn, the way out of it was to go to a new node. That was always the answer, go to a new node. Why? Because a new node gave you significant cost downs, and then that got you out of the downturn.
In the 3D era, that instinct led to just oversupply. So in many ways, you just accelerated the downturn by now just oversupplying the market. We have to think about it from the supply side. We have to think about how much supply we're putting in the market. The amount of cost downs we're getting out of the technology are just not as much as they were before. In the industry, it's just much, much, much bigger. So this growth at all costs is just not a strategy that just doesn't make sense anymore. So we're kind of articulating a different way to think about this business. It's still a fantastic business. You can't ship almost any modern technology product without NAND being a part of it. So the growth drivers are there. They're going to be there for a very, very long time.
Industry is constantly thinking of new use cases for how to use NAND. Now it's just thinking through that supply-demand dynamic in a little more sophisticated way from the supply side and the demand side and keep that balance, keep it more imbalanced.
I guess people can be skeptical about it, but it does seem like everybody has lowered utilization in an environment where gross margins are still not where you want to be, but healthy, still positive, still 20% plus. That seems different. I guess how do you the game theory aspect of that is always like, well, if everybody else is doing that, I'll just keep my fab full because it does have a big negative impact on your cost structure. But we lived through periods where stocks went up because their own factories were lit on fire. We shouldn't need that. We should have the discipline to not.
It's not the right approach.
So just how do you maintain that discipline when it must be tempting to keep your fab full when you're a relatively smaller player and everybody else is cutting utilization?
I think everybody knows how that story ends. I mean, that's what happened in the downturn. It was like just continue to pile on the supply. Like I said, the R&D is very productive in this industry. That's a good thing, unless you do too much of it. Look, we're not the only industry that goes through that. Like most industries, you can oversupply if you want to. I mean, people don't do that. That's what professional management is all about, is thinking about the industry and the structure of the industry in a way that leads to long-term profitability for everybody involved. And the amount of CapEx we're putting into this business, we need to think about that in a more structured way than we have in the past. So to me, this is just a maturation of the industry. It's nothing magical.
It's not like other people haven't gone through it. Again, I think the way the industry has evolved has been spectacular. I mean, it's just been spectacular growth. There's been, again, our technology from an R&D perspective, it's one of the most scalable technologies in the world from a semiconductor point of view. All of these are good things. It's just getting the supply-demand balance, thinking about that in a different way than we've thought about it in the past, where in the past you had this fully elastic, I could always put more supply out there and it would be absorbed. We went to the 3D era. The node transitions got more expensive. The CapEx went up. The productivity went up, and I think that old way of managing the business of just move the node forward just accelerated the oversupply.
The cost downs weren't enough to rescue the profitability of the business, and I think that's what, unfortunately, it took this historic downturn where the industry lost $30-$40 billion to maybe have that hammered home, and again, I think now you're seeing the industry, and well, let's just talk about our business. We're managing it in a different way. How do we think about this from the supply side and the demand side, keep these balanced, support pricing, be able to invest, and take advantage of this long-term growth in this very large market that we have, $65 billion TAM going to $100 billion before, let's call it 2029. That's a tremendous market to be in, and those are really, really good dynamics. You always want to be in growth markets. We got that part figured out.
We now just need to manage it a little bit differently, get the profitability where it needs to be, and it'll be, again, it paints a to me. It's an enormous opportunity.
And I guess, I mean, I've done this long enough that I'm sort of in the mode of I hope that you're right about all this, but I do worry that it doesn't play out. But at the same time, I have two years of almost minimal CapEx in the NAND space, and so I have very low organic supply, plus you're taking supply-side actions. We're seeing CapEx pick up just a little, and so maybe should I be alarmed about that? And it's interesting to see utilization coming down and CapEx going up a little at the same time. But can you just talk about the dynamics underlying that low CapEx and whether or not people do lower utilization? I mean, there's just not that much supply growth out there.
That's right. I mean, you can believe or not believe what I'm saying about how we manage the business, but to your point, if you just look at what's going on right in front of our eyes, we've got two years of historically low CapEx in the business. We all know it's a business that once you start applying CapEx, it takes a while for supply to come out the other side. And then we've got taking capacity offline, to your point. Look, this kind of rule of thumb, always run the business for lowest unit cost economics. People are pulling back on that, saying, "No, we need to support pricing." So that's happening at a point way, way earlier than it's happened at any other point in the cycle, if you will. So I think the near-term setup is very, very strong.
At the same time, you've got very strong demand drivers. Enterprise SSD continues, the cloud market, we all see the CapEx that's being spent there, continued CapEx being spent there. We've got PCs. We're optimistic as we go through 2025, that's going to get better, whether it's COVID refresh, the Windows 11 upgrade, AI PC, and then smartphones. We're seeing unit growth for the first time in a long time, again, and content growth on the back of AI. So we think the demand drivers are there. The supply side has clearly been constrained, if you will, over the last several years and continues to be by more immediate actions on a week-by-week basis. So we think that sets up for a very good period going forward.
Then I guess just bringing all that into the near term, I mean, March quarter stuff, you've said that. You guys have a view of kind of stabilization in Q2 and then improvement in second half. Can you talk about what gives you conviction in those two things, in the stabilization and then improvement?
Yeah, well, part of it is being in the market all the time, and then part of it is these bigger demand dynamics we just talked about. When you look at, we do a lot of very detailed work. Luis's team does a lot of very detailed work. We go out and we talk to all of our customers. What products are they going to launch? How are they thinking about their business? We build bottoms-up models, and then we kind of align that with what we see as the supply side of the business, and it's what gives us confidence that as we move through 2025, the market's going to get stronger.
Great, and I guess we'll get into the longer-term gross margin model, but how do you think about the prospect of kind of a supercycle? I mean, a lot of the earnings in memory companies in a 10-year window will come in an 18-month period. So it's really important to have those really outsized upturns. It seems like two years of low spending; we could say that, so when you guys talk about 35% long-term gross margin model, we've seen numbers with a five in front of it before for you guys. And do you think that that could happen again at some point down the road? I'm not asking for a when, but it could happen.
I think it absolutely can happen. I mean, look, we had a very strong recovery off the bottom of the downturn, which was we were down significantly, I think 55% peak to trough. And so we saw a very strong rebound from that. And then we kind of hit this turbulence over the last couple of quarters where I think realistically we probably thought the PC market was going to come back faster than it did. All these dynamics we talked about earlier, that didn't happen. It led to this kind of bit of oversupply we see in the market right now. But we think that's transitory. And again, we're pulling back on production to support ASP much earlier than we would have done in the past. So we think all those same trends are there, and that plays out over the next throughout 2025.
We're probably more bullish on 2026. And as we said, the supply side has not changed. There hasn't been a lot of CapEx spent. Now, there has been a little bit more CapEx just would be natural. It is still a growing market. There's going to be some CapEx spent in this market, but not in a way that we think oversupplies the market over that timeframe.
Okay, great. So I wonder if we could talk a little bit about SanDisk from a technology and competitive standpoint. Your gross margins and your competitors, sorry, and your partner, your JV partner, have been generally higher than the competitors over a fairly long period of time. And maybe dig in some of the reasons, maybe starting with foundational technology, BiCS 5 today, migrating to BiCS 8. Can you talk about the leadership potential of those technologies?
So look, the JV has been something that's been pretty spectacular. I think we're going on 25 years now of the JV. And I don't know how many JVs and technology are out there that have lasted this long, but it lasts, and it's still not just hanging on. It's thriving. And we're putting out some of the best technology in a long time and continue to lead the industry in fundamental NAND technology. And that is extremely important for building not only industry-leading technology from a performance point of view and all the other things that are important for the products we build being world-class NAND, but from a CapEx efficiency point of view. And it's always been a focus of the joint venture to focus on CapEx efficiency. The R&D teams really think about this.
How do I get that incremental petabyte of supply for the least amount of incremental CapEx and really make that part of the design process? And again, when you've been doing this for 25 years and we're able to invest as the largest player in the industry because us in Kioxia work hand in glove on this, when you look at the teams, they operate just as if they were one team. We put in 50% of the R&D, they put in 50% of the R&D. We're working on exactly the same roadmap. And when you run that model for 25 years across 17 generations of technology, what you get is what we showed at our investor day, which is a long period of time where we believe we're significantly more CapEx efficient than the industry.
And not only CapEx efficiency, but now here we have BiCS 8, where we're doing wafer bonding, build the CMOS independent of NAND stack, bond the two wafers together. That gives us a lot, gives us a very good product, very high performance. You're able to build the CMOS separately. You're not kind of making compromises where you have one wafer you're trying to do both. So it gives us a very good product, gives us a lot of optionality in the future of bonding different CMOS to different NAND stacks to have derivative technologies that can come out more quickly and with lower CapEx. So we feel very, very good. And we're just starting that journey. We're just starting moving into BiCS 8.
And that's probably where I'm most excited about the business of what we're going to be able to do with Ultra QLC that we announced, the performance of that node, and that's kind of all in front of us at this point.
Yeah, you said 10%, I think, by the end of the calendar year and then 30%-40%.
This fiscal year.
Yeah, this fiscal year. Yeah, sorry, sorry. And when I talked to Alper when we were doing the roadshow about does this give you leadership technology durability for ESSDs? 100%, absolutely, it does.
Yeah, I mean, I think we've got on our enterprise SSD story, we've kind of got a bunch of trains that arrived at the station at the same time, which is what you have to do. I mean, in a market that's complex, you can't just do one thing. You got to do three or four things right and get them all to align. And that's why it's a hard market to be in, and it's why it's a very attractive market. We've got BiCS 8, so you got to have a good node. If you don't have good NAND, building a great controller is not going to make up for it. So we've got really world-class node with QLC performance to build on top of. Now we've spent many years building a clean sheet controller for those high-density enterprise SSDs that we're just starting to ship.
That controller is going to take us many, many generations into the future. Everything we've learned from a systems perspective about how to build NAND systems, whether it's our kind of industry-leading client portfolio, what we're able to do in consumer and build controllers in that part of the business, now we've brought all of that expertise over the last three years of building this product. So you put those two together and you end up with a really advantageous position, and why we're able to announce something like Ultra QLC, where we see our performance just be not only ahead of the current generation of products, but even when we benchmark them against what others have announced as their next generation of product, we're still exceeding on performance and power and things that are very, very important for that market.
I mean, you've been underrepresented in enterprise SSD for a while, and I know it's had a nice rise in what's been a mixed environment right now. But are you going to be able to get to at least your fair share, maybe higher than your fair share a couple of years from now when all these technologies come together?
Yeah, absolutely. I mean, that's what we want is we want to have the optionality to mix when the getting is good in any market. We clearly have that in consumer. We clearly have that in client. We have that in gaming. We have that in mobile. And now as we get the portfolio where we need it to be, and that takes years to build new products, we believe we're heading into feel very strongly about we're heading into a period we'll be able to mix up into that business depending on what the economics are across the portfolio. Now we're putting this product in customers' hands literally in the next month or so. So we'll go through qualification cycles and all those kinds of things.
But that's what we want to do, have a great product that convinces the customers to actually go through the qualification process, which is not a simple process for an enterprise SSD, can take many, many, many months. And then once we go through that process, we'll have the ability to mix into that based on what the economics are in any given quarter.
It seems like that enterprise SSD market may have overshot on the upside a little bit. You were saying there's been more of a price correction there. What's your sense of the why of that? It seems like there's still pretty strong demand from cloud.
Yeah, I think I said a couple of months ago that we're in for a couple of choppy quarters, and it was because I was seeing this dynamic in the business where clearly you had a little bit of oversupply because of kind of came back from the downturn a little too rapidly expecting this PC market to come in and absorb some of that.
But that enterprise market, that was the one.
But the enterprise.
That's the one that.
Yeah, I mean, the enterprise market's been on fire, right? So you had a big spread in pricing across different markets. That tends to consolidate a little bit, and then the whole thing moves up again. So that's kind of what I think we expect to happen.
Okay, and then maybe a couple of other elements of the story that are underappreciated royalties. When I covered SanDisk pre-acquisition, there was a pretty large amount of royalty foundational patents that SanDisk, by being there in the early days of this, was collecting. And I know you're not quantifying it, but you did talk about it at the roadshow. It seems like it's material, or at the analyst day, I should say, it seems like it's material to the long-term earnings. Just how important is that royalty stream?
You want to take?
Yeah, so we haven't quantified it, Joe. I think the most important thing is it shows the strength of our R&D, right? And it's something that's where we started at the beginning of the company and something that we believe will continue to monetize going forward. And some of the new technologies could potentially lead into royalty streams as well.
Yeah, I think we really wanted, at the investor day, we wanted to show people the whole company. And we don't want people to forget. I got to ask a couple of times, what's going to surprise us about SanDisk when we were leading up to it? And it's pretty easy to answer is the innovation side of it. And it was really like this R&D team that, to Luis's point, this R&D team that defined this foundational technology for NAND that really is required to be in the market is still there. And they're thinking of new ideas. And we know how to run this play of royalties as well. And I think we got a couple of new ideas. We'll see how those play out. But it still is a part of the business that is nice to see come in every quarter.
It's not as significant as it used to be, but it's still there.
And then the consumer brand, the SanDisk brand, you've talked about it being close to a third of your business, really surprisingly high and durable margin structure around that. Can you talk about the strength of the brand and why it's so lucrative?
Yeah, I mean, I guess one thing I've been saying is there's only one company that can sell a SanDisk product. And we're back as SanDisk. And it's kind of an interesting business because we sell hundreds of millions of devices to individuals around the world every year. And then we sell $1 billion worth of product to single customers. So we have a huge dynamic range in this business. And clearly, the SanDisk brand, enormous respect for the people that built this brand over the last 20-25 years just have really built a powerhouse brand. We sell in every country where it's legal around the world. We sell on every e-tail platform. And not just sell, but on those platforms, we draw incremental share and profitability for the products where we play.
We have a very, very sophisticated team internally that's been running that play across the globe for decades now. So it is a business that gives us better through cycle profitability, and it dampens volatility in the business. Now that we're separate from Western Digital, it's just a bigger percentage of the share of the business. It worked there too. It'll work on a larger scale here. That's something we look forward to developing more as we go forward. We saw it over the last three years. We developed a brand in gaming. That became a very, very strong brand. I think we have incremental opportunity to continue to build on top of the brand equity that we have in the business.
Okay, great. So I want to ask a couple of questions on the financial model, and then I'll open it up to the room. I guess starting with the 35% gross margin, how did you arrive at that number? I've always sort of talked about a cycle that until the last few years, it was sort of different, bottomed at 15%, peaked at 55%. So you're right in the middle of that number. You've achieved that number two quarters ago. So it doesn't seem like that's a stretch number. Just how did you kind of land on 35% as the right number?
Yeah, we did a lot of modeling, Joe, good scenarios, tough scenarios, and we think this is a reasonable place to land. We think we can do better, right, and what we want to do is avoid the lows of the past. That is very, very important, and hopefully, we can still capture some of the highs, but yeah, we think that's a good, healthy model. What we want to do is create a sustainable business model, and that 35% is the foundation for that sustainable business model.
Great. And can you talk about your gross margin versus your competitor, your competitor JV partner with Kioxia? Because it's the only valuation comp that comes up a lot. Can you talk about the gross margins and operating margins versus Kioxia?
Yeah, the gross margins are pretty close, right? I mean, over time, we've been maybe a little bit ahead of them, right? If you look at many, many years together, if you look at the last few quarters, depending on how you account for inventory write-downs, you may be a little bit off, right? But the last quarter, we're very, very close. Our guide for the next quarter is very, very close. So I would say they are very comparable. And they are better than the industry, right? And one of the reasons they are better is because we spend less on CapEx, which means less on depreciation. And it's part of this lateral scaling that we have, which allows us to develop very performing nodes without needing to spend the CapEx that some of our peers need to spend.
On the OpEx side, yeah, they are a little bit better, right? The reason for that is structurally, they have two advantages. One is they're 50% larger, right? They get 60% of the output. We get 40%, so that's a little bit more than what we do. In addition to that, they're based in Japan. The yen has been moving from whatever, JPY 110-JPY 120 to JPY 150. They've been benefiting from that. As we said in Analyst Day, we're looking at efficiencies, and we'll keep on making progress. Don't expect a significant change from one day to another. We want to make sequential progress by making the right investments in the business so that we drive that growth, but at the same time, becoming more efficient in our spending.
Yeah, I wanted to ask about that because the OpEx number implied in your $10 billion long-term target was about 15% of revenues, right? That you had to get to 20% operating margins on 35% gross. That's actually kind of the same dollar amount that you're spending on OpEx today. So the message is that you can scale revenue, and OpEx doesn't have to necessarily scale with it. Obviously, that's true if it's just price. But how do I think about OpEx versus revenue?
Yeah, I mean, you're right, Joe. We're looking at efficiencies. We want to prioritize innovation because that's the lifeblood of this company. We want to continue to do that. But even within R&D, there are ways to do things more efficiently, right? So we're looking at spans, layers, where we have our talent, everything. And then we're looking very thoroughly at G&A and sales and marketing. We want to operate as efficiently as we can so that we can invest in those things that matter so that we can win with our customers while winning with our shareholders.
Great. And then the last element of the financial modeling was low free cash flow, sorry, long-term free cash flow margin is in the low teens.
Correct.
So it strikes me the only part of this that you fully control is the delta between operating margin and free cash flow margin. So can you talk a little bit about that? And obviously, those numbers are actually pretty good relative to some of your memory peers because your CapEx depreciation are a lot closer.
Yeah, I think the first change is actually talking about free cash flow. I mean, we want to focus on free cash flow because at the end of the day, that's the only thing that matters, right? The amount of dollars you can bring in to the bank, and therefore, you can do whatever is needed with that. So we are increasing the focus on free cash flow, and we think that based on that operating margin and then the CapEx that we need to sustain the business, we can get to that level, which we think creates a good return for our investors.
And if you get to the higher margin levels in a super cycle kind of environment, does that OpEx, or sorry, free cash flow depreciation delta change, or should we think about free cash flow margin lagging profits by a similar amount?
No, just I mean, the biggest contributor to free cash flow is going to be profit, right? So if we are in the 50, right, in the high end, if we're in an upcycle and gross margin is higher, we'd expect the free cash flow to flow through.
Yeah. All right. That makes a lot of sense. Let me see if we have questions from the audience. Thought we did, but then we don't. All right. Can you talk about consolidation? I think there was obviously a lot of discussion before these two events, the Kioxia IPO and your spinoff, about different elements of industry consolidation. How do you think about that? Do you think that we need consolidation for the industry profits to improve? And what's your openness to thinking about consolidation down the road?
First of all, we have a tremendous amount of conviction around the strategy we laid out at SanDisk. We think we have an incredible opportunity. We think this is a fantastic franchise. We think the market is a great market to be in. Like I said, it's got built-in growth drivers. So we've got a lot of conviction in the company and where it's headed. I think any industry that's a high fixed cost industry, consolidation always helps at some level if you can get it done. It's not a simple path. I think I've been pretty consistent since I joined Western Digital or soon thereafter that we were open to that. However, it's a complicated chessboard. I still think that that's true. There's always a lot of speculation in this industry, but we're going to focus on driving the company forward.
If those opportunities present themselves, we'll take them very seriously.
Okay, great. And then a couple of more green shoot technologies that came up around the analyst day: Matrix NAND, 3D XPoint kind of NAND, which is a storage class memory, as well as High Bandwidth Flash. Can you maybe talk about those? And just I know they're pretty far out. I'm not looking for a timeframe, but why those are worthy of your investment.
Yeah, I thought it was important to show them at our investor day because there are important things that we're working on. And I think it shows a number of things. One is I think they're both very high potential. They're also things that we're building. And I think that I think it's also a demonstration of what I said earlier is this R&D team that defined foundational NAND IP is still there and thinking about some of the most important questions in the industry today. So they're two very different technologies. The 3D XPoint memory really started six, seven, eight years ago out of our research team. So one of the things that was fascinating about Western Digital is we had a little lab inside the company where we could build wafers. I mean, it's very unique.
Kind of could solve this kind of lab to fab gap, which is how do we experiment with different technologies, different material science. And we also had this very interesting group of people that, like a lot of people from the hard drive business with material science background, and then put together with a whole bunch of semiconductor people. And out of that came some ideas for new material science and new ways to think about memory technology that started this iteration process that went on year after year after year to try and refine it, learn, keep refining, keep learning. That eventually led to a point where we needed a production line to continue the work of driving this to a product. And that's where we did the partnership with IMEC. That's been a great partnership.
They put a whole, I think, $750 million worth of tools behind that product, and we have a team over there that's building wafers every single day, week, month, and continuing this iteration process to continue to develop this memory technology, so we still got a little bit of ways to go, but we'll be building things we can put in customers' hands here relatively soon that we'll see how the technology works, and there's a lot of optimism behind that technology. I mean, anything that you've been continuing to do for seven plus years has some promise behind it. I think it's very easy to be skeptical on new memory technology, but again, you keep going in things that have promise. You continue to invest behind them, commensurate with what the risk is.
I think it's an interesting statement that a lot of people have looked at this technology from the inside and made the decisions they've made, including the U.S. government, where we have a contract with the U.S. Air Force to develop a derivative form of this technology for radiation. It has an interesting property in that it can be radiation hardened relatively in a relatively straightforward way, which solves a big problem for people in the defense industry. So we're going to continue working on that. And then high bandwidth flash is very, very different. This came out of some work, just I would say literally over the last 18 months, where Alper, who you mentioned earlier, CTO, kind of challenged the team, which is, "Hey, we've got this great scalable semiconductor technology. What can we do with it?" And that led to a bunch of ideas.
I think as Alper pointed out, NAND has always been designed from the point of view of scalability, cost efficiency. What if you thought about that differently and you designed it for bandwidth, maximum bandwidth, and then through some industry connections, we ended up talking to some of our biggest customers at the highest, highest level that we're looking at, "Hey, how are we going to build our data center two, three years from now," because we have to scale inference across a whole global footprint, and doing that the way the architecture works today is probably not the best way to do it, so we started working with those customers and doing a lot of deep simulation of what customers thought workloads were going to look like in two or three years, inference workloads.
And then we kept this, our R&D team kept taking that input and continuing to refine and refine and refine the design to the point where we think we have something here where we can solve a lot of the problems where people intuitively think, "Oh, flash couldn't work for AI because it's not fast enough," or whatever, but when you start to think about it, it's like, "Hey, for read-only applications, you have this huge density and this tremendous scalability, and you're reading everything in a pipelined way.
What if we redesign the NAND a little bit to massively parallelize all that?" And that's like an overly simplistic view, but that's essentially what's been happening over the last year of very, very deep work with customers on what they expect workloads to look like several years from now with this world-class NAND R&D team to see how we get these pieces to fit together. And that work continues. We made a decision going into the investor day to kind of show this to the world. We could have kept it private for a little while longer, but I think that this team and SanDisk has driven the industry to lots of standards and lots of foundational IP around NAND. And we think we have another very, very interesting opportunity to do this again.
Yeah, well, anything that reduces cost of HBM would be welcome. I'm afraid we actually have to wrap up. Sorry, we're at three seconds left, so some questions, maybe we can take them after. Thank you.
All right, thank you, Joe. We appreciate it. Thanks, everyone.