Mics on, yeah? Okay, great. Good afternoon, everyone. Great pleasure. Just to remind you, everyone, I'm Mark Newman from U.S. Bernstein's, U.S. IT hardware analyst. Great pleasure to welcome David Goeckeler, CEO of Sandisk.
Thank you, Mark.
also Luis Visoso, CFO.
Thank you.
Thanks very much for joining us today.
Well, thanks for having us. We're happy to be here.
Also, just before we get started, I wanna wish you a happy birthday.
Happy.
One year yesterday, I believe.
Yeah.
Yeah, that's right.
From the spin-off.
Yeah, yeah. It's a quick year.
Public listing of,
Yeah.
Republic listing.
Republic listing.
Of, Sandisk.
Yeah.
It's been.
It's been a lot of fun.
wild ride.
Yeah, it's been a lot of fun. A lot of fun.
Fantastic.
Hey, just wanna get the safe harbor, if you don't mind.
Absolutely. Go ahead.
We will be making forward-looking statements in today's discussion based on management's current assumptions and expectations, including with respect to our technology and product portfolio, our business plans and performance, market trends and opportunities, and our future financial results. These forward-looking statements are subject to risk and uncertainties. We assume no obligation to update these statements. Please refer to our annual report and Forms 10-K and/or 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 and non-GAAP financials results can be found on our website. Thank you.
Great. Thanks. Thanks so much. To kick off, I would just like to touch on this crazy pricing environment we've got currently in NAND flash. On the earnings call, the last earnings call, I asked you about Long-Term Agreements. Given how fast memory prices are rising, and given the shortage in supply in the market, there's some pros and cons to Long-Term Agreements. I remember on the call you talked about how Long-Term Agreements have transitioned from price negotiating tools to more like critical supply assurance mechanisms, I believe was how you put it on the call.
Just maybe if you could just talk a bit more about how you think about those pros and cons, because lots of investors are asking about: Are you signing up to LTAs, given that pricing is going up 50%, 60% a quarter? Is there gonna be any chance that you have any hard price caps that restricts your upside further down the line? Of course, there's also some benefit to having a long-term agreement, which is if you could have more stability longer term. You know, just think through those pros and cons. It's obviously a hot topic on investors' minds right now, given the current pricing environment.
Yeah. I mean, there's a lot to talk about here. Before we get into the specifics on kind of agreements and what we're trying to do, let's go up a level. Really what's happening here is I think the NAND market is going through some fundamental changes, right? Fundamental structural changes. It's catalyzed by, you know, something we started talking about, on... not this past earnings call, but the one before that. If you look at calendar year 2026, now you're gonna see data center is the largest market in NAND.
Mm-hmm.
You know, over two forecast cycles now, we've gone from believing that market was gonna grow mid-20s to mid-40s, to now mid to high 60s.
Mm.
in calendar year 2026. Clearly, the demand side is moving very rapidly, and that, those customers are becoming the largest consumers in the market. That's a big change.
Mm.
of a market that for the last 10, 15 years has been a device market predominantly. What that is catalyzing is thinking about this. You know, how do we get supply-demand aligned? How do we do planning? How do our customers do planning? What's the business practices about how we work with our customers? Kind of all that is changing very rapidly in this environment. It's not just that price is moving very quickly. I mean, price is just an artifact, and we have a market. A market has supply-demand. It's not either short or long or anything else, it's just a market, and that rationalizes every single day, and the way supply meets demand is through price.
Things are very dynamic right now, and that's kind of catalyzing people in the market to think differently about how we're gonna do business together. I would contend that is a extraordinarily positive development, because this has traditionally been a market where, on the supply side, we try to project what demand is going to be, right? We've been very transparent about what we're investing to. You know, we launched the company a year ago. We went on stage.
A year and one day.
A year and one day. like, a year and three weeks ago or something.
Mm.
we went on stage and said, "This is what we believe about growth in the market. This is what we're investing behind. You know, mid to high teens bit growth. We're putting billions of dollars behind that thesis. We're putting hundreds of millions of dollars of R&D behind that thesis, and then we're gonna see how it plays out," right? You know, we were bullish on the market a year ago.
Mm.
We got some pushback on that. Market has turned out quite... You know, we thought the market was gonna be undersupplied through the end of 2026. We got some pushback on that. We're kinda seeing how that is all being worked out right now. I think this kind of now turmoil in the market is causing, you know, quite frankly, especially on the data center player side, if you look at what they're going to need over the next several years in supply, right? What their demand is going to be. They're being more transparent with us on this 'cause we're qualifying new products for them. You know, we've been talking about new enterprise SSDs we've been building, we've been qualifying at new hyperscalers. They come to us with their demand forecasts, right?
This isn't, like, next quarter, you know, what's happening right now, but you look at the demand forecasts out in 2026, 2027, 2028, and there's some pretty impressive demand numbers.
Mm-hmm.
You know, our reaction to that is, "We're not gonna be able to supply what you're asking for with the way the market is working now," which basically, we just project what we think demand is going to be. We invest to that, and then really every quarter, the world's NAND supply is sold, and there's a clearing price every quarter.
Mm-hmm.
That's kind of the way the market's worked. I think what the realization is, that there's a new buyer, new use case, new demand driver, new economics of that buyer versus other buyers perhaps. It's kind of catalyzing this change for us to think about, how do we get this supply and demand equation better aligned, right? From our perspective, that is really positive because if we have a better view of demand, it's more likely we're gonna get it right as far as what we invest to. They give us a better view of their demand, it's more likely they're gonna get what they need than not get what they need.
Mm-hmm.
That's catalyzing this conversation, which is, okay, how do we rationalize this into an agreement that lasts longer than a year or lasts longer than a quarter for pricing?
Mm-hmm.
We're working through that, right? We think that's extremely positive as far as, you know, what we're trying to do in our business, is we're trying to obviously optimize the value of what we're producing every single day. We wanna dampen the volatility in this industry, right? We wanna stop going through these kind of very episodic, you know, we're oversupplied, we're undersupplied, all this kind of stuff, and kind of smooth that out a lot more because we think what we build is extremely valuable.
Mm-hmm.
If we can get more consistent in the market, we think that that's good for everybody. That's kind of what we're working through. Luis i s the one that's like on point on talking to everybody about these and putting together. I'll let him talk a little bit more about how we're thinking through them.
Yeah, Mark, success for us is a sustainable and attractive financial model, right? How do we make it over longer periods of time? It's not about maximizing every single dollar, but making sure that over time we're maxing that value. On price, to your question, right, we may have a portfolio of fixed pricing, variable pricings, that ensures that we can perform very well in an upmarket and very well in a downmarket, and outperform most times.
Mm-hmm.
That's what we're working to deliver. Think about a portfolio of outcomes with different customers, with different nodes that maximizes that value in any possible scenario.
Do you get any kind of, if your customer's giving you a projection of 60% growth, 50% growth over the next one year, two years, three years, or whatever time horizon, what kind of guarantee do you have? Do you have some kind of guarantee that they're gonna buy that much? Is there appetite for prepayment from some of your customers? Is there any kind of level of commitment, or is it more just based on their projections?
I'll say, I mean, first of all, I think everybody comes into this conversation, you know, from a very, very good place of saying, like, "We wanna work this out." Like, "We wanna get to a model, a business model, that works for everybody." Because we believe we build a very valuable product, they believe we build a very valuable product. Their companies are some of the most spectacular companies that have ever been created. Their business models are incredible. They don't wanna be short of components. Everybody comes into this from the right place, and it really is just trying to figure out, what is that model where we feel like the economics are attractive for us on a long-term basis, they feel like the economics are attractive for them on a long-term basis, and then we make sure that.
You know, things are gonna get tough at some point, right? There's gonna be quarters where things, you know, maybe are not what we thought they were gonna be. Our business is such, like, we have to build a fab, and when we turn the fab on, wafers come out, and they come out every single day, and there's more tomorrow than there was yesterday.
Mm-hmm
'cause we're investing for growth. What we're looking for for them is what Luis said, that predictable, sustainable, growing, financially attractive demand. It's how do we put a business construct around that, whether it's one year, two years, three years, up to five years, that we both have confidence we're gonna get what we need out of that? They're gonna get the, what they need, we're gonna get that predictable demand, growing demand, that we can run our business in a very, very financially attractive way. I think, you know, we're working through that right now.
Mm-hmm.
In many ways, we're kind of inventing a business model. Look, we're very aware of what's been tried in the past, what works-
Mm-hmm
... doesn't work.
Mm-hmm.
I think kind of the premise of your original question: Is this gonna be regrettable at some point? We're pretty convinced that's not gonna be the case, right? We're trying to maximize these two things, as Luis said. We need to get very attractive financials for our business, and then we need to sustain that.
Predictability and economics, basically.
Yeah.
The two things.
Predictability, consistency. Remember, the fab runs every day. You can't take a quarter off. You can't take two quarters off. That's where the cycle comes from, right? Customers, I mean, again, everybody comes to this from the good place. They're not necessarily trying to create issues for us.
Mm-hmm, mm-hmm.
They just don't run our business. We run our business. It's about understanding what each of us need out of this and putting a business practice around that works for everyone. Look, we're a deep participant in the market, right? We have very unique insights on what people are willing to pay for different products, what pricing is gonna be in the future. You know, I think you should trust that we're going to, like, set this up in a way where it's attractive for everybody involved.
Did you ever really expect that prices would be up this fast, though?
Look-
When we talked, six months, 12 months ago, of course, you were bullish. I was also bullish. However, it's things turned out to be, I think, a lot better than we all expected. Is that not true?
I would say it's happened very rapidly. I mean, in many ways, I think about our business as like this coiled spring, right? We have been investing R&D for 25 years to build some of the most sophisticated semiconductor technology in the world, 3D NAND, right? It's incredible technology.
Mm-hmm.
We have 25 years of accumulated experience of doing that. We've got 25+ years of spending billions of dollars on fabs and infrastructure. I mean, we have some of the most spectacular fabs in the world, in Japan with our partner. For a very long time, the value of that has not really been recognized. I mean, I know this very clearly because when I took this job at the separation of Western Digital, everybody told me, "Are you crazy? Why are you going into the NAND business? Nobody's ever made any money." I had the conviction that this was a really, really valuable franchise if we could get the business model right, if we could get the economics correct.
Mm.
with our customers.
Mm-hmm.
We were clearly providing something the world needed. It was clearly very difficult, it was clearly very capital intensive. I think that because of these changes in the market, I think this cannot be underestimated. The biggest thing I see right now is a lot of people with a lot of old game film on the way memory worked. "Oh, this is just a cycle. It's gonna come back. You know, I've been doing this for 30 years. Let me tell you what's gonna happen in the future." We look at that and go, "You know, maybe it's a luxury we haven't been in this industry for 30 years.
Yeah.
When I look at it, the buyer is different, the use case is different. You know, we're in one of the most spectacular deployments of AI. I mean, it's incredible what's going on around us.
Mm-hmm, mm-hmm.
I think just waiting for the past to come back is not a very good strategy. What we're really trying to stay very close to our customers, and they get it. Like, it's like, "No, look, we need to figure out a way that we can drive this business forward in a different way." This is a spectacular business and a spectacular industry, and if we get the business practices right, everybody's gonna get what they want. Everybody's gonna be a winner here. We're gonna kind of get out of this, "Hey, let's try and guess what demand is. Let's try to guess what supply is. Let's try and figure out-
Mm-hmm.
We have to invest years ahead of time, and once we invest, it's like a 10-year horizon. We need to get that consistent supply. You build the fab, you fill it with equipment, you do all the R&D, wafers come out every single day.
Mm-hmm.
Every single day, the wafers are coming out, right? You gotta sell them.
Mm-hmm.
I think that what's happening right now. Look, you see all the artifacts, the pricing, did you think it would happen all this fast? I think what you're seeing happen is a very big, very liquid market with, you know, some of the most spectacular technology companies on Earth are all transforming that market in a very short period of time to get what they need out of it in the future. Because in the future, it's more valuable than it is in the present.
Mm-hmm.
Nobody wants to be in a situation where I can't get what I need.
Mm-hmm.
we have to figure out a different way to run this business, and we're doing that in real time.
Is the calculus now changing, the gap between the demand and supply being seems like so wide right now? Is it now changing the calculus where companies such as Sandisk and your partner, Kioxia, considering to actually add capacity? So far there's not been.
No, no.
There's not been much. Actually, in the NAND industry, there's not much capacity additions, but perhaps besides YMTC in China adding a little bit, there's not, actually not much capacity in NAND flash coming online. Is that something you're looking at, given how strong demand is right now?
Again, let's go back to our Investor Day. We got up on stage and said, "We're gonna invest billions of dollars in this business, you know, hundreds of millions of dollars in R&D every year, some of the most brilliant technologists around, and we're gonna invest all that to grow high teens for as far as the eye can see." It's a pretty big commitment. we are spending a lot of money.
Mm-hmm.
Again, up until like July or August, people had conviction that, like, in the December quarter, pricing was gonna be down.
Right.
Right? That was the prevailing wisdom. I had people calling me up and saying, "Oh, you're not gonna make your December numbers because pricing will be down in December. The market's gonna be oversupplied.
Mm-hmm.
It's a little quick, like, a couple months later and say, "Are you going to, like, go in for billions more?" I mean, I think what we need to do is we need to get this conviction around what is the long-term demand, and the way we get that conviction is have the conversations we talked about earlier.
Mm-hmm.
We continue to invest for that bit growth. We're very comfortable with that.
Mm-hmm.
By the way, I, you know, I'll challenge your terms a little bit. I mean, markets don't really have shortages. They just have balance.
Mm-hmm.
You can lower the price of any product, and demand will go up.
Right.
That doesn't mean it's an economic thing you want to invest behind.
Right.
we need to understand what is the sustained growth rate of attractive business?
Mm-hmm. To maximize profits.
Yeah, to maximize the value of the franchise we own. That's our job, right? We have a responsibility to do that.
Makes sense. On data center, growth, can we talk more specifically about Sandisk mix?
Sure.
Because, you know, where are you today? I know data center is increasing as a percentage of your mix. Can you talk about quantitatively or qualitatively, like, where you are, where you expect to get to in terms of your exposure to data center?
Yeah. I'll talk about a little bit. Luis has a lot of detailed comments on that as well. Look, I kind of see three, like, big arcs of horizon, of Sandisk, right? You know, Sandisk, I wasn't around at the time. I have an enormous amount of respect for the people that were. I mean, Sandisk was a tremendous consumer franchise, IP franchise. It really built its name on that. You know, I think you can look back at the period of time that Sandisk was a part of Western Digital was the rise of the client business, you know. In retrospect, it makes a lot of sense to me. I mean, NAND was replacing the hard drive in the PC.
Mm-hmm.
We exited. Sandisk now sits today as, like, 25% share in the edge business, and I think the third arc going forward is that arc of building out the enterprise business and the data center business.
Mm-hmm.
We've basically been at that since I came in, right? Reorganized the company and put some different people in place. You know, we're very optimistic where we are with that product portfolio. We've been transparent about that. You know, two quarters ago, I think we were low, mid-20s sequential growth in data center. Last quarter was 64% sequential growth in data center. We see that pace increasing throughout the rest of the year. We feel really good. And in many ways, it's the strength of that portfolio that, again, is part of what's catalyzing these conversations on the more long-term business agreements. People see that enterprise SSD portfolio is very valuable.
Mm-hmm.
Right? I think you can assume that going forward, we're very confident of our position in that market. We're always going to have a balanced portfolio, you know, it's good to have a very strong market to grow into as a business owner.
Yeah, no, I think you covered it, David. I think the only thing I would add is just as data center is, the market is exploding, we're bringing amazing technologies to the market, right? With BiCS8, and we have our Stargate products. Everything is happening at exactly the right time, and that gives us momentum.
Actually, on your Stargate comment, can you talk a bit about this Stargate, the what? It's 128 TB Stargate QLC ESSD, I believe it is, now in qualification. Do you think that unlocks new incremental TAM as it potentially replaces hard disk drives with the QLC?
You want to take that? I was with you right until you said the hard disk drive part, right?
Yes.
look, there's two. You know.
Gotta ask, Dave.
Oh, I know. The, you know, there's two big products that are driving the data center business right now on enterprise SSD. It's that compute-focused drive.
Right
... an 8 TB, a 16 TB PCIe Gen 5, very fast interface.
Mm-hmm. Mm-hmm.
That product, we have a great product in that part of the market we introduced a couple of years ago.
Mm-hmm
... that's driving all this growth we talked about. We have a new product coming out, which is more the storage side of it. 64 to 128, 256, 512, even terabyte is on the roadmap.
Mm-hmm.
That product is what we had codenamed Stargate. That is a complete, clean-sheet ASIC build for that. We have lots of runway. You know, we introduced at that 128 level is what we're qualifying.
Mm-hmm
... We'll start rolling that forward.
Mm-hmm.
There's a lot of pieces that make up that, right? It's BiCS8 to start with, which is phenomenal QLC performance. Even on BiCS8, 2TB die, right? You know, the form factor is not that big. You gotta have a high-density die. It's the ASIC and all the work that's been done over the last three to four years to build that. Again, as Luis said, all those pieces are now coming together in a world-class product, and the market is responding very. Our customers are responding very well to that. That product has not started shipping for revenue yet. All this growth we've talked about has been without that product. When that kicks in, we expect that will, of course, drive better growth for us.
Mm-hmm.
Now, is that product replacing HDDs, is that what's driving the growth? I'm not a big believer in that thesis.
Right.
Right?
That's more That would go more into the KV cache, the separate storage tray that NVIDIA is talking about?
That would more be the compute one. This is just it's a storage class, like the data lake you're going to create.
Okay
of storage.
The storage-
Yeah
the storage equipment.
Right.
Okay, got it. Got it. Also, you talking about BiCS8. BiCS8 has been very successful so far. there's been some recent commentary, like, chatter, market chatter about you accelerating BiCS10.
Mm
technology to 2026 to meet hyperscaler needs. Anything you can talk about on BiCS10?
I think you called that right as chatter. Like, I mean, we have a BiCS8 plan. We've been very consistent with it. You know, we expect to exit this fiscal year with it, the predominant node in our portfolio. BiCS10 is a great node, don't get me wrong.
Mm-hmm.
We're not accelerating BiCS10.
BiCS10.
to meet demand. We can meet the demand and the bit growth we need with the CapEx plans and the BiCS8 plans that we have.
Mm-hmm.
As you can imagine, a very sophisticated fab plan behind month by month, what is the nodal mix gonna be? There's a lot of work that goes into making sure that's set up right.
Mm-hmm.
It's, you know, not something you can change overnight. Look, BiCS10's a great node, don't get me wrong.
Mm-hmm.
We're always working on multiple nodes. That's one of the great things in NAND. I think this is why one of the reasons the NAND market is so spectacular. We just have Moore's Law, if you want to call it that, is alive and well. There's a lot of R&D productivity that we can put into this market, and that's a big part of NAND now becoming a big part of the AI architecture. It's the most scalable semiconductor technology. It's inevitable it was gonna become part of that architecture. As models get bigger, caches get larger, content lengths get bigger, you have to bring the most scalable semiconductor technology. That's not. There's no negative comment on DRAM or HBM. That is, like, brilliant technology.
Mm-hmm.
At some point, you need the scale, and that becomes a big tailwind for NAND. I think for the last, maybe up until two quarters ago, everybody was asking me, "Does AI really impact NAND?" You know, I think that question is kind of being put to bed now.
Right. Right.
The answer to that is, like, definitely yes.
I think, Jensen Huang put that to bed.
Well.
Yes.
We completely agree that, like, Yes.
Switching gears a little bit, how do you view the competitive threat from the Chinese? YMTC, in China, there's been some, again, market chatter, although I think this time it is actually backed up, this market chatter, to some extent. There was an announcement from the Pentagon saying that YMTC and CXMT have been removed from the list of entities that are working with the Chinese military.
Mm.
That announcement was then rescinded, I understand, so slightly confused about what's going on there. I'm, I doubt you want to talk too much about that specifically.
Mm.
Just, you know, how do you think about the Chinese, that competitive threat and how that may change going forward?
I mean, a couple of comments. I mean, anybody that's in this business is very capable, right? We take all of our peers very, very seriously. When we talk about our market numbers of supply, demand, they're completely in all of our numbers, right? You know, we have that factored in. You know, again, this is why we say things like, "We're very confident with mid to high teens bit growth," 'cause we're factoring all this stuff you're talking about goes into our thinking and our equations.
Mm-hmm.
You know, the thing... I, you're right. I mean, I certainly can't speak for the U.S. government. I take it at their word. You know, that was a DOD list. It was retracted right away. It wasn't the, it wasn't the stuff out of commerce or export controls.
Mm
... and all that, I'm sure that'll get clarified over, you know, going forward.
Mm-hmm. Got it. Can we talk a little bit about HBF, High Bandwidth Flash? You recently formed this partnership with SK hynix on HBF. First of all, any progress you can talk about on HBF? I also wanted to ask, how is your partnership with SK hynix, like how do you actually collaborate with them?
You know, again, you started with happy birthday. If I look back on what... you know, we made a lot of progress during our first year, and one of the things I may be the happiest with is the progress we've made on HBF. A year ago, you know, this week or whatever, we got on stage, and we introduced this whole concept of High Bandwidth Flash. Let's just be kind about it, we got some people that were a little bit skeptical...
Mm
about what we were saying, right?
Yep.
What are these guys talking about?" Or, "That, you know, is this real?" No, we were very serious about it because this came out of some work of our engineering team. Again, you know, I think one of the things about Sandisk's heritage is our R&D prowess. Lot of, you know, company really known for innovation, and You know, some of our engineers started thinking a number of years ago when they... I mean, they saw AI. They're very close and, you know, they kinda have this conviction: "Hey, we've got very scalable technology here.
Mm.
How do we make it, how do we apply it to AI? They started thinking about these questions, and if you're a NAND designer, you've kind of spent the last 25 years of, like, being told constantly, "Give me more density. Density, density. I want more density. Give me more density." You know, I mean, okay, 2D, 3D, like, memory holes closer together, like, all kinds of crazy stuff to always deliver more density. If you actually start asking those same people, "What if you start thinking about higher bandwidth instead of just more density? What if you start thinking about how could you increase the endurance as opposed to just more density?" You know, people started thinking about those questions: "Oh, well, we actually can come up with some ideas for how to solve those.
Out of that comes this idea High Bandwidth Flash. we believe we can take the flash technology we have, again, the BiCS8 technology we have, going to BiCS10, going to BiCS11, and all the whole roadmap, and we think we can change that by building a new die that is much more suitable for AI.
Mm-hmm.
Out of that cause we believed, I think our R&D team believed, that, you know, at some point, this AI technology is gonna need the density of NAND. NAND, you know, anytime something is ten times-
Mm-hmm
... something, like, that's an area worth spending time on.
Mm
'cause that's a major difference.
Absolutely.
We made a lot of progress. You know, going into the Investor Day, we talked a lot about, do we wanna talk about this? We decided we should talk about it. We're asking people to invest in the company. We're spending time on this. We think we're very optimistic about it, so we talked about it. You know, there was a little bit of skepticism at the beginning. One of the companies that wasn't so skeptical was SK hynix. They called us up and said, "Hey, can we work together on this? 'Cause we think it's a good idea." That was very important, 'cause if you're gonna make a market, it's hard to make a market by yourself.
Right.
It's better if you make a market, you have an ecosystem. We're working with them on the specification of how the HBF system would work. What are the interfaces to it? Those kind. We're not, like, collaborating on the die. We're not collaborating on the controller. We're doing all that stuff ourselves. They're doing all their piece. We're collaborating how HBF would fit into a system. You know, again, looking back over our first year, I am just extremely happy about where we sit here a year from now, and nobody is asking skeptical questions about HBF. They're talking about how HBF is the future, like, you know, bringing. You know, models have gotten bigger, caches have gotten bigger, contents lengths have gotten bigger, number of tokens have gotten bigger.
Like, all this kind of stuff has kinda went in the direction we thought it was gonna go. Kinda came to the point where at CES, somebody very important in the AI world got up and said: "Hey, storage needs to be rearchitected for AI." We're like: "Yeah, we agree with that.
Mm-hmm.
We're ahead of that a little bit.
Mm-hmm.
I think it's where we've landed after one year is, I think, just fantastic. You know, the relationship with our partners at SK on this technology is great, right? They're working on their NAND stuff, but we're collaborating on the system, which helps everybody adopt it. You know, we're working with customers on how they could use it in both devices and the cloud, because again, this is not like we're building a plug-compatible piece of technology for something else. You know, the system is gonna have to change to accommodate our technology. You know, to do that, you have to know exactly what the use case is gonna be, how is it gonna scale, all these kinds of things, and that's quite difficult.
Mm-hmm.
We're going through that process, and we're optimistic that, you know, that's going to lead to something very interesting, and we'll have more to say about that when we get more details around that. In the meantime, We're building the NAND die. We expect to have that towards the end of this year.
Mm-hmm.
Then we're working on the controller and the system to put it together. We, you know, maybe a year from now, we'll have that. We can put in customers' hands.
Mm
where they can start using the system, kinda understanding how it would work with their infrastructure, so.
How's customer interest so far?
Um-
Do you have already preliminary discussions with customers about this, or is it more working with the GPU makers at this stage?
No, we've been working with customers for quite some time on this technology and this idea. You know, it's a long process. It's like pure innovation, right? It's pure, like, you're building something new.
Mm-hmm.
It takes a lot of collaboration. You know, we're experts in building storage. They're experts in building data centers or building devices, and it's the collaboration of those two expertises, putting them together on the what's possible to build here, that allows us to build.
Mm
a better system, and we're in that process.
Presumably, you still have to work with the GPU makers, though, don't you, for this?
Well, Yeah, we've gotta work with the people that are building the infrastructure.
Right. Right. What does success look like here for HBF? I think you talked before about it's really focused more on inference, not training.
Yeah.
Is that right?
It's inference. It's inference. It's not, you know. Look, I mean, nobody would sit up here and say: We're gonna replace HBM in training. That would be, like, a crazy thing to say. Like, that's like brilliant technology.
The speed requirement.
Inference is very different. The model is already built. You know, it's a very predictive read. It's a big model. You've gotta pump it into a CPU somewhere or a GPU, you can run inference against it. It's a very different kind of problem. Also inference is, you know, where you really have to scale, right? You build the model, you propagate the model, you drive inference everywhere, whether it's on your device.
Mm-hmm. Mm-hmm.
... whether it's on your phone, your laptop, or in the cloud. If it's in the cloud, you're gonna have to scale inference around the globe at incredible levels. I think that's where, you know, when you start to talk about technology that's 10 times more dense, that has a scalability roadmap in front of it, once you hop on that NAND roadmap, now you still have those nodal transitions in front of you that is providing you-
Mm
... significantly more supply every time we turn the crank-
Right, right.
... on the node. That's a great place to be if you need a lot of storage, and it turns out AI needs a lot of storage.
It does. Certainly sounds very exciting on paper, so we're just watching it closely, hoping to.
Stay tuned.
...hoping to hear more.
We're excited about it.
Yeah, yeah. I guess, my follow-up question was around, we talked about how strong demand is. Actually, you made the comment, David, about this demand versus supply. In the end, shipments, there's a clearing price. In the end.
It's a market.
It's a market.
It's a big market.
So-
Like, we don't set the price.
So-
The market sets the price.
My question really is on the, on the demand side, and what are you seeing happening there? There are some customers that are not gonna be able to get enough. I'm hearing that from some of the OEMs. HP just reported yesterday, they talked about how they can't get enough memory, that they're lumping DRAM and NAND together. Are you seeing that in the market? Are you seeing that some of the customers that essentially, maybe they're not as important as hyperscalers, are not sure, or they can't afford? Are you starting to see some kind of demand destruction because of this high pricing?
Everybody's important, so let's start with that.
Of course. Of course.
Everybody is important, but I'll let Luis-
Of course.
He's very close to them.
We treat every customer very importantly. Some of them are more willing to talk to us about their longer-term needs. They are willing to make commitments longer term, and some others would prefer to have this quarterly process, where we negotiate price every single quarter. In that process, you're not gonna have all the availability, right?
Mm-hmm.
In a market that is tighter, you know, you're gonna give preference to those customers for with which you have a longer-term agreement.
I see.
That's what we're doing, but every day, we're making sure we're allocating the bits to the right places, and if customers are willing to come to us and talk about longer-term pricing, I mean, we're happy to do that.
Mm-hmm.
Happy to do that.
Any other long-term opportunities we haven't talked about that you're excited about for Sandisk going forwards, that perhaps we haven't talked about or Wall Street's not aware of currently?
Look, I mean, I think our consumer franchise is in fantastic shape. It's an area we're investing in the brand, and I think there's, I think that that's just a tremendous franchise. People tend not to focus on it as much because the other ones move faster, but we have incredible reach. We have incredible brand power. You know, we're now treating that more like a real consumer business.
Mm-hmm.
I'm very optimistic about what's gonna happen there over the next couple of years. I think our technology roadmap is extremely strong. I mean, I think BiCS8 is a extremely strong technology, and it stems from our relationship with Kioxia. Again, we invest together in R&D. I think people tend to focus on the JV as, like, a manufacturing-
Mm-hmm.
which it is, but it's really that R&D piece where we're able to invest the number of engineers as if we have the highest share in the industry. It's that, you know, sustained 25 years of working together that allows us just to bring incredibly powerful technology at the lowest incremental cost per bit to get that incremental technology. It puts us in a cost leadership position. We just extended the JV for another five years.
Right
... in Yokkaichi. I think that was, you know, a huge step forward for us. You know, again, it just highlights a little bit of that, this difference on supply-demand. You know, we're putting money behind where are we gonna get supply from 2030 to 2034.
Mm-hmm.
People are very focused on where they're going to get supply for next quarter, right?
Right.
We need to align those timeframes a little better.
Right.
That's what the business practice piece is all about. As Luis said, the enterprise SSD portfolio is coming together as a world-class portfolio. It couldn't be at a better time. It absolutely couldn't be a better time.
Mm.
Right? The pull is there. It's hard to push a new product into the market. It's a lot easier if people are pulling you along. That's in a very, very good spot. You know, I think the team inside of Sandisk is just completely on fire as an independent company. I mean, it is just a very, very exciting place to be. We're having a lot of fun.
Mm-hmm.
You know, maybe we have the audacity to believe that a lot of that old tape that people keep playing in their mind about waiting for. You know, when is the next cycle gonna come? Like, we're very focused in making sure this is a sustainable model, and we kinda get some of that out of the system. We change the way the industry works, or we change our piece of it. I can't change anybody else.
Right.
We can change our piece of it so everybody wins, and I think as we do that, this is just a spectacular franchise.
Mm-hmm.
It's kind of why I'm here and why I chose to come to this business a year ago, because. That's why
Right
... you know, six months before that, agreed to join me. We just think this is an unbelievable opportunity, and we're having a lot of fun doing it, and we think there's an incredible amount of value creation in our future.
I think also on your comment about, you can't control what others are doing, but I think, you can influence via some kind of leadership in the industry in terms of telegraphing how you're managing supply-
Mm
... how you're-
Mm
... thinking about things, or how you're thinking about economics, then that can actually impact the industry.
Mm.
It can actually impact, but you can't guarantee that they're gonna-.
Yeah
... do the same thing. You can't guarantee that they're.
Look, I think.
add more supply,
I think we have, I wouldn't go quite as far as what you were saying, but, look, I think we have an interesting position as a standalone NAND company. 'Cause we can talk about how we're gonna manage the NAND business, and we can talk about how we think the best way to run this business is. I think we try to be very transparent about that, right? I mean, I watch all of the, you know, a lot of the stuff in the media and all that, it's like, oh, my gosh, you know, like, there's not enough or whatever.
Mm-hmm.
It's like, look, we've been saying for a year, this is what we're investing to.
Mm-hmm.
Nobody, like, you gotta ask for more a little earlier. I think people are listening. That's good. Like, again, everybody is very important. We want everybody to get everything they need to... I mean, our customers have spectacular businesses, and, you know, we're envious of what they're able to do. We think we're providing an incredible technology, and that technology is not easy, and that technology is extremely IP-intensive, and it's extremely capital-intensive. There's not a lot of companies in the world where you have both of those.
Mm-hmm.
You usually have one or the other. We have both. What this is about is getting the right return on that business for what we're investing.
Mm-hmm.
Again, I think everybody wins, and I think it's an incredibly dynamic market. I don't think, you know, I'll just say it one more time, if your frame of reference is, "Oh, this is just another upcycle, and ride the price curve, and all this," I don't think that's what's happening here. What's happening here is a very big, very important technology for the world is undergoing a fundamentally structural transition, and it's gonna come out of it in a very different place.
Long-term new memory paradigm, isn't it, David? New memory. Didn't someone write that 10 years ago? I think. I wanted just to see if there's any questions from the audience. Hands up if you have any questions. Everyone is just so convinced, I think.
Good
... by the argument today. Okay, well, I think we'll wrap up.
Thank you.
Thank you.
All right.
Thanks very much.
Mark, thanks to you .
I really appreciate it.
I really appreciate it.
Thanks very much.
Thank you, man. Thanks, appreciate it.