Perfect. So why don't we go ahead and get started? I'm Aaron Rakers. I'm the IT hardware and semiconductor analyst here at Wells Fargo. Thank you for joining us this morning at Wells Fargo's eighth annual TMT conference. Pleased to have the Western Digital team, Rob Soderbery, Executive Vice President of the Flash Business. I guess I'll call it SanDisk. We'll see where, you know, the spin is progressing, but we'll get into that. But before we get there, I think we've got a safe harbor.
Yep. I'll quickly read the safe harbor statement. We will be making forward-looking statements in today's discussion based on management's current assumptions and expectations, including with respect to the separation of our businesses, 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 most recent financial report on Form 10-K and 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 GAAP and non-GAAP results can be found on our website.
That was perfect. From memory, right? You did that all from the top of your head. That's perfect, so Rob, thanks for joining us.
Happy to be here, Aaron.
Yeah, I appreciate it. You know, there's a lot going on in your business. A lot of things going on at Western Digital. But I wanna start at a high level. Like, I know that you gave an update on kind of the longer-term mindset of the Flash industry. I think it was a couple months ago. You know, I'm gonna get to kind of like what maybe is thought about in terms of the guidance for this current quarter of the Flash business. But why don't we level set? Like, how do you see the Flash industry? What's the key drivers? We'll talk about supply and demand, you know, how Western Digital currently views the status of the NAND industry.
Yeah, absolutely. We start with that macro. I did a webcast called The New Era of NAND, which is still available on our site. That really went through like what happened in the downturn. I think to invest in a new standalone Flash company, you wanna understand that question 'cause you definitely don't want that to happen again.
Yep. Yep.
Fundamentally what we saw was essentially an overdrive of the WFE number. I saw a $20 billion WFE going into this business for a couple years, produced a level of production volume that was unsustainable, and that drove a downturn. I think coming out of that downturn, the industry's actually set up for a much, much different dynamic, and I won't repeat everything I said in that, but there's a lot of reasons to understand the incentives have fundamentally changed, and so you're now in a situation where you have market participants who are all essentially aligned in terms of their strategy, and that's setting up for, I think, a sustained period here of a different dynamic in the market.
and the market itself is of course based on all the different demand swim lanes: client, consumer, embedded, mobile, and of course the one that we're spending a lot of time talking about, which is enterprise SSD.
Yep.
Probably the biggest thing that's happened now that is a little different is we're seeing a bifurcation of that enterprise SSD kind of market dynamic from the rest of the market.
Yep.
You have your client, consumer, mobile markets going through a little bit of a mid-cycle pause here while the enterprise SSD AI tailwind is at, you know, full blast.
Yep.
And so that's changing. That's resulting in a bit of a different dynamic for us depending on your mix. But we think that's a couple of quarter phenomena and that the general thesis that is in place here is the correct thesis, which is, you know, clearly a more balanced supply-demand environment, very moderated capital, sustained demand drivers on both the data center and the edge from AI and consumer and the digitization of our lives.
Yep. Maybe before we get into unpacking some of those end market, you know, vectors and the puts and takes around that, maybe just, you know, update us, you know, for those that aren't familiar where we're at in the spin, the progression of separating the company. I know the Form 10 just recently came out.
Right.
Which I think many of us, you know, have gone through, but just remind us of the strategy where we're at there.
Right. So spin announced in November of last year. You know, the motivation of that is really for the folks in this room is to create 2 investable entities that had their own economics, their own capital return policies, their own balance sheets, and that you really could get granular on. All the work to date on the spin was sort of the, you know, the legs of the duck under the water. So, you know, two ERP systems, 2 Workday systems, two SOX compliant public company P&Ls. All that work's essentially done. So we had a trial close on October 1st. We had another close on November 1st. All systems are go from an internal perspective. So now it's simply, we need to kinda get the Form 10 out, get the financing done, get the investor day scheduled, and then get the disbursement date scheduled.
So, we're in the mechanical.
Yep.
So the final stage of the mechanical process there for the spin. So everybody feels good. You know, we have made most of the executive hires. We have two complete executive leadership teams. A little bit crowded, but we've got them, we've got them jammed into one company right now. And I think both companies are excited to launch into the market.
That's perfect. So, you know, and for those that aren't familiar, you know, one of the things that was unpacked is obviously the capital structure, the balance sheets, and the differential between, you know, what goes where in the companies. But one of the common questions I get in this structure and going all the way back to when SanDisk was standalone is just the mechanics of the JV.
Mm-hmm.
Right? And.
Right.
When we look at the capital structure, the balance sheet, you know, it's not consolidated.
Correct.
You know, so how do I think about the Western Digital piece of the asset side in the JV structure?
Correct.
Meaning the equipment really, right?
Right. Right. Yeah. So, so there's sort of 3 numbers that you wanna think about in the JV. There's a book value number, there's an acquisition cost, and then there's what are those assets really worth.
Yep.
Like the number between, you know, the book value is relatively low because we depreciate over five years. The acquisition value's very high, like $20+ billion has been thrown out. And of course, the actual value of the assets is in between those two numbers.
Yep.
It's, you know, we have the ability to use this capital equipment for a very long time. So the average life of a piece of gear is well over 10 years. We've got, like I think 25% of our capital equipment is 15 years of average age. And it's all fully productive. So we have a big fully depreciated asset in there, which is really driving our production. And we reuse all that gear every time we do a new node, which is based on that node. So I think in the investor day, we'll probably come out with a specific number for the size of that.
Yep.
but it's a material fraction of that original $20 billion acquisition cost.
That was very helpful. So you depreciate equipment on 5 years, but the duration of the useful life is well over 10.
Correct.
Okay.
Absolutely.
Okay. That's very interesting, and just the JV, again, understanding that it's 50.1% ownership.
Correct.
Kioxia, you.
Right.
Just the mechanics of how that works.
Yeah. So it makes your, if you're building a model for the company, it makes it a little tricky because the JV is off balance sheet. And so EBITDA recycles inside the JV. So if you just look at our financials, you can't do a correct EBITDA calculation because essentially what happens is we buy product from the JV. Part of that, what we purchased from the JV goes to fund capital equipment that we already buy.
Yep. Yep.
And so that essentially is EBITDA, but that EBITDA recycles. And so it just results in our cash CapEx being lower relative to what it would otherwise be. Or it's sort of the same as it would be otherwise, but that EBITDA would be recycling on our own balance sheet instead of recycling.
Yep.
On this off-balance-sheet entity.
Yep.
But, you know, so that makes understanding the financials one step removed. So we'll get out with that sort of asset value number. I think the other big thing about the JV though is just the JV has demonstrated a long-term ability to be more capital efficient. And that comes from like a lot of different factors. And the age of the gear is really an output of that.
Yep.
Right? It's about the engineering of the processes. It's about the fact it's the largest fab in the world. We're the largest manufacturer in the world. In our Yokkaichi facility, we have 7 fabs. Those fabs are connected via our track rail system, and you know, any wafer can go to any machine and be processed in any step, so you essentially have the largest possible pool of machines to run your fab processes across.
Yep.
All that is translated into very efficient capital usage. The business isn't nearly as scary as you might think. We're more capital efficient. We have an aspect of leasing. We've got EBITDA recycling inside the entity. You know, we think the numbers, you know, we'll get the model out at the analyst day and we think it'll be pretty interesting for folks.
That's perfect. And just so we level set too, like we do have some visibility into your JV partner now that, you know, they're progressing to an IPO.
Absolutely.
It's not a 50/50 split of the total capacity between the 2 companies, right? There's some capacity that resides outside the JV.
40/40/20.
40/40/20.
The JV is 80%.
Right.
Split into two chunks of 40/40, and then there's an extra 20 for KIC.
Gotcha. Okay.
It's not homogeneous capacity, but it's roughly how it works.
Okay. I'm gonna move away from all of that mechanic stuff.
Okay.
We'll anxiously wait for the analyst day. Going back to where maybe I started the discussion, when you thought about the guidance into this current quarter, we talked a little bit about, you know, there's client SSD, there's some churn in the consumer market, the mobile market, you know, on the smartphone side seems to be still relatively tepid, but what underpins, you know, the expectations for the Flash business as we kinda think about this current quarter?
Well, I mean, this quarter's been a, the whole question is how fast can you rotate out of these client bits and into enterprise bits.
Right.
So we are, I think we're on a good trend there. We're not yet at my aspirational goal for the business, but I'm, I'm pretty pleased with the rate at which our enterprise SSDs are growing. We have essentially 2 new products in that lineup. The first is a PCIe Gen 5 compute product. That's the product that's qualified in the NVIDIA GB200 NVL. When NVIDIA qualified it, they, we qualified us first, and they told us that's the fastest Gen 5 compute drive they've qualified. So that was a really good sign. That's getting a lot, and that, and that's a drive that essentially attaches to GPUs on essentially a one-for-one basis. The second swim lane is this data lake swim lane.
So in addition to your GPU complex, you have this very large data lake, and you're doing what we used to call in the database days, ETL or extract, transform, and load. But you're taking, you know, if you're gonna put an image into a GPU, you need to vectorize and embed that image. And all that work happens in this pre-process. That all happens in these fast data lakes. We have a 30- and 60-terabyte device that's going into that fast data lake. We're shipping 122-terabyte device into that data lake next year. And that's all new. Essentially, it's new demand on the market that did not exist in 2023. And that, that's quite, quite significant.
So both of those products are essentially shipping for the first time in volume this quarter, and that's driving sort of this next phase of that ramp. And then we'll have our 122 terabyte product. So that's all the eSSD tailwind, and the headline news in the portfolio.
Yep. That's perfect. Maybe before I go double-click on that enterprise side, can you? I know in the Form 10 there's some disclosures around cloud and consumer and client.
Right.
You know, breakdown. But how do I think about the mix of the Flash business from a client SSD? I guess I'll throw gaming coupled with that.
Maybe retail consumer solutions, mobile.
Enterprise. How do I think about that split as a stand?
Think, yeah. I mean, think about client as the base of our portfolio. That's the meat and potatoes of our portfolio. We're almost 30 points of share in there. So we're head to head with Samsung. You know, on a good day, we claim leadership. That's the base of the portfolio. Along with that comes gaming. We have great share position in gaming. And so that's the largest bit allocation.
Next is consumer. Consumer's a great business for us. I think it's gonna be more important in the new company because it's bigger and growing as a function of the total business. It's countercyclical. It's higher gross margin, and so the consumer's sort of the next layer of, of bits, about 20%. So think of that first layer as about 40%.
Consumer, another 20 on top of that. On top of that is Enterprise SSD. That'll be about 15% of our bits in the December quarter. And that's up, I think, over five quarters. It was 1, 2, 5, 7, 12, 15. So.
1, 2 , 5 ,7 , 12, 15. I'm trying to.
Sorry.
Fit the.
Notch sequence maybe.
I know. Perfect.
I'm not guaranteeing the next.
the next data point in that, but we'll be a little over 15%, and then everything else goes into mobile.
or component markets. So that's kinda the extra bits on top.
Where do you stand on client, the client SSD market? 'Cause that's been a tough space. I mean, we've seen a lot of discussion around, you know, look, PC shipments are relatively tepid, you know, 1% or 2 plus or minus quarter.
But, there is a narrative around AI PC and.
And that's starting to maybe materialize in the next year. Are we through inventory digestion there? When do we start to see manufacturing business?
No, we're not.
Okay.
Client recovered very strongly out of the downturn. Unfortunately, the PC OEMs had even higher aspirations. So they essentially just overshot the recovered level. So it's not that PC shipments are way down. It's that expectations weren't met. So they bought some inventory ahead of maybe an AI PC, a 2024 AI PC boom.
Which didn't clearly happen. So now they gotta chew that inventory down. That's done by the middle of 2025.
Okay. Okay.
Client volumes could go up 30% from here. Our selling volume could go up 30% per year. Selling is significantly depressed relative to what they're selling out.
Okay.
The client PC OEM shipment volume can be held constant.
We can grow our selling volume by 30% as they consume their inventory.
Just getting through that inventory level.
Just getting.
That depletion.
Just getting to inventory depletion.
WD's view is that we will see either be it Windows 10 plus AI PC.
Yeah. And then that, like the refresh cycle has not gone away. The way refresh cycles work is, you know, they essentially come back. The longer they take to happen, the stronger they happen. And so I think we're headed for a big bang, Windows AI PC future-proof refresh cycle here, in 2025. And that's what our big OEMs are telling us. So they're not panicking at all.
Okay. On the enterprise SSD side, you know, you mentioned the progression to 15%. Is the target 15% for fiscal 2025? Is that what was that number?
15% for December quarter, for this quarter.
For December quarter. Okay. Not.
And for the full- year should be.
15-20.
15- 20.
For 2025.
The idea is that that's gonna continue to move higher.
Yeah. There's no, I mean, if you look at eSSD's about 25% of the total market. So that's.
For my internal goal, I'm not putting this out yet.
Yep.
Externally, but my internal goal for my boss is I need to be equally represented in the eSSD market at that level. So this 15%, I think it was a good initial threshold. We've been driving hard. We wanna kinda stabilize that kind of number.
Mm-hmm.
And then we have big product announcements in 2025. They'll get qualified in 2025 and kinda later in 2025, we kinda take the next run up in the market.
You mentioned the NVIDIA qualification, which was very notable. And congrats on that. Where do we stand on the hyperscale cloud qualifications?
Hyperscale cloud, so the things that were under qualification are now qualified.
Yep.
And there are additional qualifications starting on hyperscaler cloud. So we're broadly engaged with both the hyperscalers, the traditional OEMs, as well as the emerging storage OEMs, really across that entire landscape.
Perfect. So I wanna go back to kinda the earlier question a little bit, is that, you know, you've got the demand side. Just, you know, remind us again of how Western Digital sees like the bit demand, growth algorithm for the total Flash industry.
Yeah.
Like when you're thinking about allocation of CapEx, you know, what is the underlying views of your mindset or the company's mindset of bit demand growth in aggregate? And you mentioned $20 billion in CapEx. Is $20 billion in CapEx now normalized at $15 billion of WFE? You know, just.
Yeah.
What would you?
We have a model of the entire industry. So I have every fab. I know every node. I know the capital equipment costs. We know these answers really well. Let's ignore the minus two plus 2 quarters 'cause the market's still wiggling.
Mm-hmm.
So it's still wiggling after the downturn. But if you think about on a sustained basis, pre-downturn, you saw exabyte growth at 30%.
Yep.
We see that exabyte growth number stabilizing post-downturn at 20%. Pre-downturn, you had capital peaking at about $20 billion WFE. We see that number coming down to $11-12 billion WFE, probably in a 2026 timeframe. So kinda going up from here towards that $11-12 billion in 2026.
Mm-hmm.
And then we'll call that the business-as-usual market. It's pretty clear that, you know, AI's gonna keep going here.
Yep.
And so there's acceleration upside from that 20%. So I don't think there's a lot of downside in that 20% number. And I think there's some upside from a supplier standpoint. You know, I think we went from a calm mindset to pre-downturn to, you know, don't shoot till you see the whites of their eyes mindset.
Yep.
After the downturn. So to the extent there is, you know, more AI drive and we think that that number goes up from 20% on a sustained basis, we can revisit our capital policy.
Yep. And let's say we start to see maybe that moves higher. How, remind me again how when that becomes bit output, like the duration of time?
It's essentially a full year.
It's a full year.
It's a full year before you.
Mm-hmm.
The industry is not gonna respond quickly to demand.
Right.
That, you know, demand will essentially drive price.
Okay. That's perfect. So, I try and juxtapose all of this 'cause everything you just said is really, in my opinion, structurally good.
Mm-hmm.
For the industry, right?
Right.
The setup looks good into 2025. We've got, you know, separation. You've got a Kioxia IPO. However, it seems like the pricing dynamics, the spot market seems to be still under pressure.
Absolutely.
I get this question all the time.
Right.
Have you seen spot pricing?
Why is that the right way to think about this or?
So that spot pricing number is a little bit funny. So it, the spot pricing is pretty thinly traded.
Yep.
It doesn't really reflect what's happening behind the scenes. What's happening behind the scenes is that the this bifurcation and Samsung actually did a nice job of explaining this. Slowdown in client weakness to global consumers has caused a surplus of bits in sort of the older nodes and the older generations. That's driving weakness. Enterprise SSD in general requires, you know, at a minimum, 1xx and beyond.
Yep.
AI for anything that's performant. And so that's driving kind of more tightness in the higher end of the market. So that phenomenon, at least in my tenure, has not happened before where you've had these sort of two different. And so you're seeing, look, if the whole market were going south, you'd, yeah, you'd be a disaster. If the whole market were.
Yep.
Serving AI would be on fire. Instead, we're in this little bit of a mixed terrain. But I think the big message is that, that lower-end bit phenomenon is like a couple of quarter phenomena. It's not systemic. Like we're, we're undershipping those markets right now because of the inventory position in those markets.
That will just mechanically work through the system.
So this AI inertia around enterprise SSDs, you know, there's a push to 120+ terabyte capacity points.
Right.
You get some maybe reprieve or, you know, finalization of inventory digestion in client. We'll put mobile to the side. I mean, couldn't we see a scenario where things get somewhat tight in the next year? No?
Yeah, absolutely. I think we, with continued supply discipline. And remember, even for people, for all of us, as we invest in new nodes, we're taking unit capacity out.
Yep.
So we're getting per unit growth. We're reducing, in this case, wafer starts in a variety of different ways. And every manufacturer is slightly different. But so you're even though you're putting $11 -12 billion of WFE in, you're effectively not getting growth commensurate with $11- 12 billion. You're getting a lower growth rate, which is better for us in terms of setting up that tighter market. So we think that second half setup is quite good.
Okay. That's perfect. One of the things that came up last night, not to put you on the spot, but there was an announcement by a system vendor,
Pure.
You know, Pure announced a hyperscale win for Flash. Obviously, you're in the Flash business. I'd love to, how do you see Flash relative to Nearline Hard Disk Drive? You got an interesting purview being.
Yeah.
Still at both companies. So how do you see that kinda?
Yeah.
That dynamic?
Pure essentially at this point is the combination of Veritas Software, which I spent my former years at, Cisco. So I know those guys.
Okay.
a great, great company, right? They've innovated for many years. I think the key thing to understand here is that the HDD market in the hyperscalers is growing at a 20% CAGR.
It's enormous. HDD is five times larger than Flash, and it's growing at 20%. It's a zettabyte growing at 20%. Flash is 200 exabytes growing at maybe 30%.
It takes a long time for something that's 20% as large, even growing at that rate to make it then.
You're talking about new HDD capacity, incremental growth of 200 exabytes, almost the size of the entire Flash market. Look, both statements can be correct. This transition can be wonderful for Pure.
If they're a three-and-a-half exabyte player, so I can't even calculate their share. It's, it's like 0.0.
That's a total.
0.0x%, right?
So Pure could double or triple, and nobody would notice.
It's great for Pure. At the same time, it's not changing the fundamental economics. You know, it would take hundreds of billions of dollars of Flash capital to backfill that zettabyte of HDD.
Nobody's gonna do that. I think the better way to think about it right now is that there's jump ball workloads. For example, the AI data center workload is a bit of a jump ball between HDD and Flash. Flash has kind of got that.
Because of performance constraints. That's consuming all of our capacity, and I think it will consume the bulk of the capacity.
I would put that HDD-Flash substitution as a bit of a peripheral effect for quite some time here.
Yeah. I appreciate that. And I think just to kinda think about that, I mean, correct me if I'm wrong, Western Digital's talked about like the cost curve down.
For Flash is.
Right.
Mid-teens?
Right.
You know, the parallel line is.
It just a long time, right?
So you're starting at 6x, mid-teens.
HDD's on 2%-3. Like you can.
Get Excel out, and you can figure that out, and it still takes a long time.
So, maybe I'll pause. If there is any questions in the audience, I'm happy to field them. But otherwise, I'll keep going. Raise your hand if you do. I wanna talk a little bit about one of the things that I often, for a long time, have had a lot of discussions around is technology progression, right? This BiCS 6- 8. Hey, you know, the WD Kioxia has 162 layers. Other guys have, pick your number, 220, whatever.
Right. Right.
Why is that wrong? Like, walk us through the technology.
Yeah, so why that's wrong is that, so if I got a 100-story hotel.
my neighbor's got a 100-story hotel. We have two different strategies, and we wanna double the size of our hotels. You have two strategies. You can either build a 200-story hotel or you build a second 100-story tower next door.
Right? So the problem is diminishing returns.
It's really expensive to build a 200-story hotel. So you just get slaughtered. And what's happened in the industry is you're getting diminishing returns to layers. It basically, they're costing you a lot more. It takes, you have to give all your money to Lam to buy cryo etchers to do them. And so you essentially have diminishing returns to layers. We've always have said that the correct layer goal is the smallest possible, lowest possible layers while still achieving cost, performance, power, and capacity. So for example, the BiCS 8, even though we're the lowest layer count in the market, we have the highest capacity die. So we have a 2-terabit Kioxia die. So why is it that my competitor has.
300.
300 layers, and I have a bigger die that's faster and has a better cost position?
The reason is that there's a lot more that goes into the engineering of the device. Essentially, you think about one strategy is buy a lot of capital equipment, relax your design rules, you can get layers fast. The other strategy is conserve capital, go for a tighter, more difficult to manufacture, but ultimately higher performance, higher density, lower cost device, so we've always had a more of a craftsman approach to migrations. And I think we've said that to the market. We faced a lot of skepticism.
Back from the market. I think BiCS 8 is finally definitively proving that that strategy works. So BiCS 8, which is a 2xx device, is faster than the 2yy devices of our competitors and is competitive with the 3xx devices of our competitors.
BiCS 8 is higher capacity. It's 2 terabit Kioxia. Our BiCS 8 Kioxia performance, so we're gonna be doing some press next week on, on Kioxia, but our BiCS 8 Kioxia performance is twice our BiCS 5 generation. So there was this thinking that, well, Kioxia is low performance. It's not gonna serve these swim lanes. It's not like, no, no, we're getting a 2x in performance because of all the things we've done in NAND innovation. So, you know, we think that, you know, BiCS 8 and beyond is awesome. You know, our NAND strategy is we're completely unapologetic about our NAND strategy at this point.
Right. Yeah. I understand. I mean, I think part of it is too the y-axis scale or the x, y.
Yeah. Yeah.
You get more.
We get more bits per area, and then so therefore we need fewer layers.
Means your etch holes are closer.
'Cause etch holes are closer. And it's a very, I mean, these are an amazing piece of technology where you're custom crafting this 3D cube of stuff, and you're trying to get the whole cube in the smallest.
Possible area.
So as we work towards the spin, you know, commencing, where do we, how do we think about the BiCS 8 transition? Where are we at today? Where do you think we'll be at a year from now? It seems like given the fact, and it's a CMOS, you know, bonded array architecture.
Right.
It's a different architecture than what you've had.
Right.
Just the relevance, importance of that transition.
Yeah. So first of all, think about 0 technology risk. It was one of the fastest to yield nodes we've ever done. It has a lot of desirable characteristics. There's essentially the CBA was a technical challenge, but it's not a capital challenge or a manufacturing challenge. So we, we're really good. So the, the gun has gone off. We're, we're in the race. We're shipping BiCS 8 now. We'll just continually migrate our portfolio to BiCS 8 really over probably a two, two-and-a-half-year time period here.
I think the device that people are really waiting for is the 122 terabyt, ultra-high-cap device.
But that's not the only device we'll do. We'll have our client devices. We'll have consumer devices. So it will, and then we'll modulate our BiCS 8 investment to achieve that 20% growth target.
Perfect.
We talked about.
Perfect. And that CBA architecture maintains that idea that you can keep a 15% mid-teens per annum kind of cost down?
It does. But I'd say at the industry, it used to be that you got that 15% out of what I'll call brute force.
Which is, you just got so much technology advantages node to node, you'll get that 15%. Now that 15% is much more subtle. It's continuous improvement. It's longer capital cycles. There's a lot of sort of more, you know, back-to-basics manufacturing in that number. So maybe half of that is, say, nodal transition, and half of that is just kinda operational efficiency.
So you're, you know, your mileage will vary in terms of different vendors, but this notion that you could just get arbitrary cost reduction forever, which turned out not even to be healthy for the industry. I think that is, that's gone.
Okay. That's helpful. The final couple questions in the few minutes I got left, I'm just gonna kinda rapid-fire a few things off. I'm not gonna ask you like longer-term gross margin. You know, we'll wait to see maybe analysts say, you know, update a long-term target. But the enterprise mix, does enterprise come into the portfolio at a higher structural gross margin for the company?
Yes and no, so enterprise in a normal market, compute will be the highest gross margins. Storage, the compute eSSDs, storage eSSDs will be kind of at or sort of average gross margins.
Okay.
Then client, and so on. Now, in a downturn, the Enterprise SSD is also cyclic. So you have to be a little bit careful about your Enterprise SSD mix.
And that's why, you know, some of our competitors that were heavy enterprise SSD had a pretty rough go of the downturn. So we like a balanced portfolio with, you know, it's with the benefits of portfolio economics helping us on a through-cycle basis.
Right. And that's prefaced against a through-cycle Flash gross margin target of 35%-37%.
Yeah. The May 2022 was 35-37. You know, I actually will never bite at the apple here when you come out to.
Investor Day.
Okay. What's in the final minute, what do you feel like I didn't ask that I should be asking you? Like on the Flash business. I'm sure, you know, there's a lot of different other vectors, but, you know, to me.
Well.
It's all about how we feel about the structural aspects of.
So something we haven't talked a lot about is consumer.
Yeah.
I love the SanDisk consumer business.
Okay.
I've wanted, since I got here, to invest more in consumer. We do 200 million units per year, branded consumer units, you know, single-package units to consumers. We've been doing that for 15-plus years. So think about millions of consumers, you know, individual consumer purchases. So that SanDisk brand is imprinted around the globe. And it doesn't matter like which Uber I'm in or which, you know, bellhop I'm talking to. Like when I tell them I'm from SanDisk, they know who I am. You know, we think, and that business for us has been very good on a through-cycle basis. It has been very good. Now it's been attached to an HDD consumer business that was shrinking.
So that's despite some headwinds. You know, we think that consumer business is a really nice part of the portfolio. In addition to being a great go-to-market vehicle, it's another key part of our through-cycle economics. It consumes lower-yielding bits for us. So it works in the fab economics. It's part of the SSD. It's a very important part of the overall SanDisk story above and beyond the 20% of bits it represents.
Yeah. And just to be clear on that point, like leveraging that business to kinda take on the leading edge of node transitions is kinda gets you up the learning curve.
Yeah. You can, you can take the leading edge. You can take the trailing edge. I can, you know, if I need to sell more ESSDs, I can shift that business to a, a different node. So it gives me a lot of flexibility in how I consume, and it's all captive, and I don't have to share, share it with anybody. So, you know, I have 20% of my bits are pre-sold every year. I don't have to worry about those.
Which is a great, great place to be.
Perfect. With Rob, why don't we wrap it up there, and I appreciate you taking the time.
All right. Looking forward to see everybody, soon.
That's our successful investor day.
Awesome. Thank you so much.
Okay.