Micron Technology, Inc. (MU) Earnings Call Transcript & Summary

Good afternoon. My name is Latif, and I will be your conference facilitator today. At this time, I would like to welcome everyone to Micron's Post-Earnings Analyst Conference Call. [Operator Instructions] It's now my pleasure to turn the floor over to your host, Farhan Ahmad, Vice President of Investor Relations. You may begin.

Thank you, and welcome to Micron Technologies Fiscal Fourth Quarter 2021 Sell-side Analyst Call back. On the call with me today are Sumit Sadana, Executive Vice President and Chief Business Officer of Micron; Manish Bhatia, EVP of Global Life Operations; and Dave Zinsner, Chief Financial Officer. Today's call will be approximately 45 minutes in length. As a reminder, the matters we will be discussing today include forward-looking statements. These forward-looking statements are subject to risks and uncertainties that may cause actual results to differ materially from the statements made today. We refer you to the documents we filed with the SEC, specifically our most recent Form 10-K and 10-Q for a discussion of risks that may affect our future results. Although we believe that the expectation reflected in the forward-looking statements are reasonable, we cannot guarantee future results, level of activity, performance or achievements. We are under no duty to update any of the forward-looking statements after today's date to conform these statements to actual results. Latif, you may now open the call for Q&A.

[Operator Instructions] Our first question comes from the line of Srini Pajjuri of SMBC Nikko Securities.

A couple of questions. First, on the mobile in the quarter that you reported, I think revenues were down mid-single digits. Given the strong demand commentary and the seasonal strength, I was surprised to see that mobile was down sequentially. So if you can talk about that. And then on the outlook for the November quarter for bits going down little bit. I understand the PC market is going through an inventory correction, but PC is a relatively small portion of the overall DRAM market. And it looks like the cloud market is pretty strong and mobile is strong. So I'm just trying to reconcile that comment with your demand commentary in other segments. If you can maybe talk about how the other segments are faring sequentially as we go from August to the November quarter in terms of bits, I think that would be helpful.

Sorry, what was that second question again? I didn't get pass the second question.

Yes. The second question, your guidance for bits sequentially declining. I think you attributed that to the PC market. I believe the PC market is about 10% of the overall DRAM bits. So I'm surprised to see that PC is having such a big impact to move the entire, I mean, bits down sequentially. So if you can talk about what's your assumption for cloud bits sequentially and also the mobile bits sequentially, I think that would be helpful.

Okay. All right. Understood. Yes, I can address that. So the first question about mobile. So I just wanted to mention that part of what we see in terms of the specific business unit performance is also driven by how we decide to allocate bits for shipment into various customers and to various segments. So it's not only the dynamics of the segment itself that's at play, but also our own allocation strategy internally that's at play. So one important thing I just wanted to call your attention to is our company inventory at the end of FY Q4 was below our target level, which means we practically shipped everything that we could ship that was not nailed down and then some. So we have unsustainably low level of inventory right now, not enough inventory to service adequately customer demand and manage the mix change. So we have to improve our own internal inventory position. And that is one of the reasons why we had a decline in sequential bits in mobile, not related to end market demand. End market demand is strong. And also, I'd like to mention that we did have very strong shipments in our storage business. If you look at our storage business results, our NAND shipments were very strong there. So overall, I would say that our inventories are at very low levels and our mobile end market demand, to your first question, remains very robust. So now to your second question about why is F Q1 bit shipments forecasted to decline modestly. PCs are more than just 10% of the DRAM market, and they are definitely higher level, significantly higher mix for us than that 10% number you mentioned. So it is an impact to that portion of the market. And when -- you can appreciate when some of these purchase slowdown occurs in any one of these segments, the overall goal for some of these customers is to make those corrections very quick so that they can move on to more normal ordering patterns in future quarters. And so typically, most of these things end up being sharp and short when they have those kind of collections. So that's one aspect of it. And the other aspect is, again, we have very low inventory. So that's another thing that's impacting our sequential bit shipments. And the last aspect, as we have already mentioned in our prepared remarks, we are seeing some constraints in our own supply chain to be able to meet all of the demand that is placed on us by our customers. And those are things we have done a really good job over the last many quarters of avoiding some of those constraints, whereas the rest of the industry had to struggle with some of those constraints. But we are having some impact, nevertheless. So that's also showing up in some of our ships.

Okay. And if I can maybe follow up on the same topic. So you're guiding for bit growth to resume, I think, in the second half of the fiscal year. And you said that your fiscal year bit supply growth will be somewhat below the industry growth. But I think just looking at my model, I have to make pretty aggressive assumptions in the second half even to get to 10% bit growth. And I know you're guiding for like mid-teens, maybe somewhat below mid-teens. Just curious, what will get you to that aggressive second half ramp in terms of volume, is it just 1-alpha? Are you thinking about building inventory in the first half? I'm just curious as to how you get to that level.

Yes. I think...

I can handle it. I think that your question, Srini, was mainly about DRAM, but I can answer both for DRAM and NAND. And the inventory levels are very lean for both of them right now. And so as Sumit was saying in the previous question, it's just difficult for us to be able to adjust the mix when they're -- when our customers have shortages on components, they want to -- they change where they want the mix of supply to be and we have to adjust that. And with lower levels of inventory, we're not able to do that in the near term. We think that the component shortages will get better for everyone, for customers and for us as we go through the next several quarters. And that gradual improvement will release more demand across all segments. And then we'll be able to align our limited supply better with the supply -- with the demand requirements of the broader market as we go through the year. In terms of what the key drivers are. The key drivers for both DRAM and NAND are the leading technologies, 1-alpha for DRAM and 176 layer for NAND. Both of those are ramping very well. Sanjay mentioned some statistics about how well the yield is going, how well the shipments are going on the prior call. And so those really are still going to be the workhorses for us through the year, and we do still expect a record year in terms of revenue for the year.

Our next question comes from Ambrish Srivastava of BMO Capital Markets.

This is Jitendra for Ambrish. Dave, your office guide for fiscal first quarter is up around 3% on a lower sequential revenue guide. Are there any onetime costs that we should be thinking of from the beginning of the fiscal year? And how should we think about OpEx trajectory for the rest of the fiscal year?

I'm totally sorry, I totally missed the first part. Is the whole question around OpEx?

Yes.

Okay. Well, let me try to -- I'll talk about OpEx and then let me know if I answered your question. So OpEx will be up sequentially this quarter. Some of that -- we had a relatively light year last year for fiscal '21 for pre-qual expenses, and those are starting to come back as we work on the next node. So that's driving some of it. And we also have salary increases in the first quarter, which will impact the first quarter and actually have a bigger impact in the second quarter. We're also making investments in R&D and in SG&A to bring -- to invest on the product side and the process technology side and make more investments in each of those areas, which is likely to increase our headcount, and we feel it's prudent. Nevertheless, we think OpEx as a percent of revenue will be at a fairly low level, barely low as it compares to not only our memory peers but also our -- the overall semiconductor business. So we feel like we're still prudent to investment on the CapEx side, even though we're growing CapEx at 15%. And as Sanjay mentioned, also, we expect a record year in terms of revenue. So that will help fund obviously the OpEx increase. After that, I mean, hard to say beyond fiscal '22 what we think on the OpEx side, but I think that this OpEx as a percent of revenue roughly in the low double-digit to low teens is roughly where we'd like to operate.

Great. As a quick follow-up. We just wanted to check if there is any impact from the latest power restrictions in China. That has any impact on the manufacturing for Micron?

So yes, we're tracking that closely, and we do have some component supply that is coming from regions that are impacted -- and the -- that is -- as we mentioned, we do have some areas of our component supply chain as well as some of our silicon supply chain that is starting to have some impact on our ability to ship products this quarter, and that's baked into our guidance there. So we're tracking that very closely and making sure that we can recover and meet demand for our customers. But that is one of the things that is creating some impact for us in our supply chain this quarter.

Our next question comes from Mehdi Hosseini of SIG.

Two follow-ups. David, I want to go back to your bit shipment. In the prepared remarks, you talked about the shipment growth resuming in the second half of fiscal year '22. You also highlighted the decline in the November quarter. But you left out February. What should we assume for February? Would it be flat before the growth resumes? Or would February be also down?

Yes, I mean we're shying away from that. We don't -- we gave a lot of detail, obviously. We're not trying to guide every quarter. I guess, I would just say that typically, the favorite quarter is seasonally weaker than the first fiscal quarter. That said, given that the first quarter is a bit weaker in terms of shipments, it's likely that this might be a more muted environment from a seasonality perspective in the February quarter. And so that's -- I think that probably is as much as I can give you on the modeling front.

Okay. And then you mentioned something about free cash flow seasonally weaker in Q1 fiscal year '22. Did you mean to decline, not actually having cash burn, correct?

Yes. I mean, again, we don't necessarily forecast out free cash flow every quarter. But given that CapEx is more weighted towards the first half, it's likely to have a more meaningful impact on free cash flow. So I think it's going to be relatively muted. Hard to say whether it's positive or negative, we'll have to kind of see how things go from a business perspective and how work to capital plays out. It certainly will be weaker than the fourth fiscal quarter from a cash perspective. But having said that, we are expecting strong free cash flow for the year.

Okay. I just want to make sure you even imply cash flow. It just it's going to be a positive free cash flow, it's just that maybe compared to a record in Q4, maybe it's just down sequentially.

Yes, I think it's going to be meaningfully down sequentially.

Our next question comes from Tom O'Malley of Barclays.

Mine on the OpEx side as well. Can you talk about how contingent that 15% raise on R&D? Is some of that bit growth coming in the back half? And I guess another way to say that is, let's just say you get to the May quarter, bits are supposed to come back on whether that's supply or demand getting a bit better, and you're not seeing it come on as fast. Are you still expecting to spend that much? Is that a fixed number or would you dial that back if things were a bit weaker?

Of course, we will manage OpEx based on business conditions, of course, yes. We are anticipating that like I said that we see record revenue next year. So we're optimistic that we'll be making that investment.

And I guess, kind of in conjunction with that and to a question earlier as well. In that May quarter where you're starting to see some of that bit growth, is that more from your expectation that some supply growth, some supply comes back? Or is that really seeing some material demand drivers improve? Can you just talk about the mix there versus what you're expecting, which is more of a contribution to that bit growth in the back half?

Yes. In terms of the back half, a couple of things will happen. Number one, we feel that our internal inventory position, which is below our target levels right now would have improved. We think some of the component shortages that are taking place, both for us as well as our customers who don't have adequate match sets in certain areas of their business, those component shortages would improve as well. And we also feel that the continued growth in several of the segments that we have already highlighted are going to create a very good demand environment. We expect good demand in the data center side with a lot of upgrades happening to newer processor platforms with higher content of DRAM and NAND. We also expect more than 50% unit volume growth in 5G shipments next year and continued very robust growth significantly above average industry levels of growth in both automotive and industrial to continue. So a lot of strong demand trends that are underpinning it and coming together with some of the improvements on the supply side that are creating some shortages right now, we feel will help drive the overall business trends.

Our next question comes from Sidney Ho of Deutsche Bank.

I have a question on gross margin. You talked about the mix changes being the predominant factor impacting gross margin in the November quarter. Can you be a little more specific about what is driving this mix? I thought less PC should be better for ASP, maybe there's just high mobile client SSD. Also on the second question here as well. For fiscal '22 cost down, you talked about DRAM to be better than the long-term trend and NAND to be in line. Does that comment exclude the COVID costs you are assuming for the next 2 quarters? And is mix change a big impact on a full year basis, when we think about the full year?

Okay. I didn't quite get the second question, but okay.

I'll repeat that.

No, no, that's all right. Manish picked it up, and I can add color. I think you're -- so, I think you're thinking too much on the gross margin guidance as it relates to mix. All I'm saying is cost and price are not a big determinant of the gross margins. We have to create a range because there's a lot of different outcomes, the mix of NAND versus DRAM, the mix, as you can point out, of the different business units, the mix within the product categories. We can't call it within a 200 basis point range. So we gave you a 200 basis point range. It might be at the high end, it might be at the low end. It somewhat depends on how we get, as Manish already talked about, what bits come out of the line and how we do in terms of shipping those to customers. That's -- it's no more complicated than that.

Okay. So the second part of the question...

On cost, right? So you had a couple of questions in there, Sidney. One was the clarification on the front end. The comment we made about cost down trajectory and then the second part was around mix and how the mix would impact, I think. So what we said is that the front-end cost, meaning the memory level cost before we get to assembly and test driven largely on the back of 1-alpha and 176 being really, really great leadership nodes for us and ramping well with yields and production output and everything else, those are going to deliver really good cost declines and DRAM maybe even above the long-term range that we've talked about. And in NAND, kind of in line with the range. So we feel good about those. But when it comes to the full product cost, we have some of the costs to mitigate COVID in all of our assembly factories, materials cost increases that we're seeing across the supply chain and then yes, mix will increase. DDR5, LP5, data center SSDs, these are all going to continue to grow as a part of our mix as we move to more high-value solutions, and those impact our end kind of pure cost reduction capability. But as Dave mentioned, when you're -- the portfolio is moving, then you're going to -- the mix will also help you on the pricing side. And so that's why the margin side is still -- we have a pretty healthy guide for the margins here, at least the one that we gave for FQ1.

Next question comes from Karl Ackerman of Cowen & Company.

I wanted to understand perhaps a little bit better what the right mix of work in process is for your inventory. You mentioned how you're trying to improve your finished goods inventory, that's had a very record low. It sounds like you're building that up this quarter. But are you able to expand your packaging and/or assembly and test suppliers to free up some of the bottlenecks on inventory?

So I'll take that one, Karl. So broadly, there's a couple of elements of work in process. Obviously, what's in the fab is the largest part of our work in process. And that part is as you implement longer and long -- newer technology nodes, longer process flows, so that is one of the headwinds for us. for the whole industry really in terms of monotonically increasing portion of our inventory because fab cycle time is getting longer every single node. In terms of the constraints in the back end, one of the things that's a challenge is really the lead times in terms of how you're able to get components and get assembly capacity across our network. We've done pretty well with being able to meet customer demand across our both captive back-end as well as our noncaptive back-end. I wouldn't say that that's really driving so much the -- an increase in WIP. It's more that when we work with more subcontractors and use their volumes, the cycle times are longer than they are in our own factories. And so that does maybe elevate a little bit our back-end inventory. But our -- I think as we mentioned, our finished goods inventories are at the leanest level they've been in many, many years. So we're able to -- whatever we're able to produce, we're able to match to demand and move -- and generate revenue.

Got it, that's very helpful. I'm not sure if I was limited to one question, if I could just sneak one more in. I want to ask about automotive. Has your ability to service automotive demand improved or stayed the same over the last 90 days? I ask because last quarter, you indicated demand was far outstripping your ability to supply. I'm wondering if OEM production shutdowns have allowed you to better meet existing demand and perhaps broaden your design wins as you think about fiscal '22.

Yes. Our design wins continue to come in at a very strong pace. We have, as you know, #1 market share in our business, in the memory space, in automotive and industrial by a wide margin compared to our next competitor. And so we have had tremendous momentum and design wins, and we have some really great technologies and products we have introduced and are also moving our customers to newer nodes of technology. So really a lot of positive momentum in that space. I think the challenges of supplying to the level of demand that we have in automotive, and I'll put the industrial segment in that same bucket because they use very similar products and similar nodes of technology. Both of those are continuing to have demand levels that significantly outstrip our ability to supply to those. We are working very hard. We have made improvements internally, but we are still short of where we would need to be, we would want to be. Some of those shortages relate to internal production capacity. Others relate to some of the components that we need in our supply chain to be able to ship. So there are shortages of both kinds that are impacting that particular segment. And the demand supply balance is such that it will take us a while before we will be able to catch up. And I think you have already seen that. Broadly speaking, in the semiconductor industry level, it's also the segment that is the most challenged in terms of getting to that supply-demand balance. It will probably be the longest in that segment where supply-demand balance will be achieved in the industry, probably not till the end of next calendar year, maybe into 2023.

Our next question comes from Chris Caso of Raymond James.

I just say my question is, I know that it's the goal to...

Chris, sorry, you're not coming in clear. Can you try to get closer to the mic, if you're on a speaker phone, maybe just pick up the handset?

Sorry about that. Yes, is this better?

Yes.

Okay, sorry, headset problem. So I said I know the goal over time is to modulate the CapEx to stay in line with industry bit demand. Have you taken any steps to do that? And I guess the question is, has what you've seen in PC been enough to convention to make any changes. And I know you've given us a CapEx number now, but I know it's even in the planning process for a while? Or is this transitory enough such that it really hasn't had an effect on what your capacity planning has been, again, not just on a 1 quarter basis, but over a couple of quarters?

Yes. I mean growth has been stronger both in DRAM and for calendar '21 versus what we expected, for sure. And so that obviously is an influence on what we think in terms of the growth. And then, of course, you're right, the PC market is softer, but we do view it as transitory or an air pocket and so forth. We do view the long-term bit growth rates of both DRAM and NAND to be very similar to what we have felt the long-term growth rate would be, which is mid- to high teens in DRAM and about 30% in NAND over time. And so as we obviously look at this every month really with Manish to determine what makes sense. And we haven't really changed our view from that perspective. That being said, if we do start to have a feeling that bit growth assumptions are too high, we, of course, would immediately adjust CapEx spend to react to that. But so far, based on our view, we haven't.

And I think this time, we went in the script, Chris, giving you a little more color on some of the aspects that drove our CapEx guide that are not really going to contribute to bit growth in the near term, right? We've got the investments in EUV that are going to benefit us a few years down the road. We've got investments in other R&D for other products. We've got assembly test, which won't add to bit growth, but will help with our supply chain resiliency. So these are some of the examples of things that are in our CapEx guide that are not tied directly to our view of the next 12 to 18 months' worth of market conditions. And then I think we are being pretty disciplined even within that. We explained that our DRAM CapEx for transitions, we've done a good job with 1-alpha so far, and we're going to leverage that with over the next year and our DRAM CapEx will be lower year-on-year. So I think we're managing prudently with sort of the near-term market dynamics as well as keeping in mind the long-term strategic technologies we need to invest in for the future.

Got it. And if I can just have a quick follow-up. And recognizing that the CapEx is not necessarily apples-to-apples because of some of those things that don't affect bit demand. I mean is the net of it that the CapEx spend that we see this year is in line with what your view of industry bit demand is as we look at the full year?

Yes, I think as we said that we think our bit supply growth for calendar year '22 will be in line with the long-term CAGR, and that should be -- in fact, the whole industry has this challenge of having shipped -- depleted inventories over the last 12 months, maybe even longer. And so in order to be able to have the industry growth rates catch back up, you're going to need to -- you're going to see that it's going to be difficult for the shipment growth rate to be above the long-term CAGR. That's why we think that in CY '22, there'll be healthy supply/demand balance. We'll be shipping new supply -- the whole industry will be shipping from new fab supply not from inventory.

I think the other thing is that -- and we'll see this year-to-year that while we believe that capital intensity over time is kind of mid-30s as a percent of revenue. It will vary year-to-year somewhat as a function of the revenue number, but also a function of what node we're investing in and what that delivers in terms of bit growth. And so those things will obviously influence how we invest our capital on a year-to-year basis on the nodes.

Yes. And on that point, 1-alpha is a good capital intensity node, right? So DRAM CapEx is going down because it's a mature node now, and we are investing on it. And 176, our transition is especially expensive transition for us because we are not just going up higher layer count, we are also changing the notes, the stack of the NAND, and so it is more expensive. So NAND CapEx is also going up because of that because more of the wafers we are converting next year.

Your next question comes from Vijay Rakesh of Mizuho.

Dave, just a question on the -- I know you talked about inventory on the gains on the supply constraints. But just wondering what -- how we are seeing inventory levels on the cloud hyperscale guys and on the mobile handset side, too? If you could give some quantification as to where those were versus normal levels? And I have a follow-up.

Yes, in terms of customer inventory, by and large, we think that the inventory levels are in decent shape. I think on the cloud side and on the enterprise side, the inventory levels are in much better shape than they were, for example, a few years ago when everyone was all concerned about how that inventory would impact the market environment, let's say, in 2018-2019 time frame. So this time around, that definitely that inventory is in much better shape. And also, the data center companies are looking ahead to a fairly strong year of CapEx investments in server capacity and deploying new server platforms, both with newer processors as well as DDR5 capabilities. And so it's going to be an important investment year for our customers there. So I think on the mobile side, some customers have little bit elevated level of inventory. Others are in relatively good shape. Of course, there are market share shifts that keep happening in the mobile space. So it is not unusual for certain customers to have more inventory. But this time around, we feel maybe there is some element of geopolitical risk that's causing perhaps some of those inventories to be somewhat elevated and also a lot of challenges that most companies have faced over the last 18 months. Certainly, since COVID began is likely causing -- and all of these events that keep occurring every few days or weeks about some new supply chain risk that's coming up in the industry is also probably causing companies to have a strategic decision to carry a little bit higher inventory for a period of time perhaps. But certainly, things like geopolitical-related aspects don't have any end in sight. So we don't know what the outlook of that is in terms of its impact on normalization of those inventories. But overall, I would say some customers are on the mobile side with okay inventory. Others have a little bit higher inventory.

Got it. And just one other question. I know you mentioned maybe a couple of months here quarters to resolve this whole constraint issue. Is that more a commentary on the PC side in terms of the component constraint side? Or is that more a commentary on that the supply -- that OEMs will start to pull on demand again on all segments like data center and mobile and PCs you go into -- as you go through some adjustments here.

Manish, you want to take that or you want me to?

I was thinking -- I think the question was more about PC market, right?

Yes. Not just PC and mobile and data center, yes. sorry.

But the inventory of the...

Yes. About the components. You're talk about component charts in the end. I mean I think for the PC side, definitely, we talked about it as a 1- to 2-month -- sorry, 1 to 2 quarters. This quarter, next quarter should be getting resolved and seeing the end market demand start to come through more fully. Across the other markets, we don't see as much of constraints. From the end market side, PC is definitely the one that's there. And we think that the end market demand for all segments, whether it's mobile, whether it's data center, obviously, automotive which has been constrained quite a bit. The end market demand is very, very strong. And as Sumit said earlier, automotive will be one like -- that may take longer to be able to be resolved.

Our next question comes from John Pitzer of Credit Suisse.

Dave, I'm just kind of curious on the R&D guide you gave for the full fiscal year. Is the implication there that you're going to be outside of the normal range, and that's why you're calling it out? And specifically, help me understand what this R&D is going to? Is this really a qualification of new products?

No. I mean, I think -- I mean the reason to call it out was because 15% growth is actually higher relative to what we've normally seen in terms of growth in our OpEx. And in fiscal '21, we're basically flat with the prior year. So just wanted to make sure that investors were modeling that level appropriately. I mean I think that was more of the intention there. It's going through a whole bunch of things. I mean, we are hiring engineers for the product side, no doubt about it. We do have a heavier load of just prequalification expenses where the R&D groups are utilizing the wafer lines for their run as opposed to direction runs. So that obviously if -- and that activity kind of goes up and down quarter-to-quarter and year-to-year, and it's going to be a pretty heavy year in terms of activity around qualification expenses. So that's another aspect of what we're trying to highlight.

I think we had some -- just to start, we have some significant increases in investment happening in high-bandwidth memory and CXL, which is a completely new area that we have started investing in, continued investment in data center SSDs. So it's a set of significant ongoing investments on the product side because as we have mentioned, we are constantly trying to improve the product mix. And some of these newer technologies that we are going -- that we are working on, obviously, both on the DRAM and NAND side, the wafer costs keep increasing. So that constantly puts upward pressure on R&D and, of course, that's why it's not easy to -- for newer players in this business because it's a business of scale and it requires that level of scale to be able to make these large investments.

EUV R&D technology investment is also something that we have talked about as a program where we are making it a bigger program.

Got it. And then, Dave, maybe just trying to ask my question on the public call a little bit differently. I mean to the extent that what we do is kind of look at historical patterns and try a correlation, this cycle or this upturn looks a lot different than a normal upturn. I'm kind of curious, to the extent that the magnitude and duration is at least taking a pause here, are you arguing that that's completely attributable to some of the supply constraint issues that your customers are having? Or can we actually dare say that maybe your customers are thinking about memory procurement a little bit differently? And perhaps I know you guys have been striving to make this business less volatile and more predictable. Is that sort of what we're seeing? Or is it too early to make that call?

I mean from a...

What's causing the air packet, it is definitely the component side of PC space, there's no doubt about that. We have -- our embedded business unit is now at record levels. That is a more stable business within the portfolio. So I think that does help in terms of alleviating some of the pressure that past cycles have had on gross margins, for sure. I think we really view the next quarter, first Q1 plus Q2 is more just a function of the PC customers needed to cleanse out the results. They've got fairly strong demand at the end customer level and just can't build units without all the components. And as soon as they do, we have seen sustainably strong demand for memory across all of the various markets. And I think that once we get past that period where the PC market gets back into a better place, we'll see a resumption of our strong demand as well.

I think the other aspect to just keep in mind is, there is nothing normal about the time that we are in, right? I mean this is an unprecedented time in our industry and certainly in the memory and storage business as well, in the technology business and anything that uses semiconductors, it's an unprecedented time. And shortages of the type that we are seeing across a whole range of aspects of supply chain have never been experienced in the industry in the many decades that all of us have been working in this industry. So I would say that for us to expect that will be a "normal" upturn is sort of maybe not a realistic expectation. So really what all of these things that we have described in our discussion today with you are meant to describe an environment where a lot of the demand is stretching out because largely, it's difficult to meet the demand in the time frame in which it has been placed on not just us, but the rest of the industry. And that is going to enable us to look ahead and feel a level of confidence that we will be at record levels of revenue in fiscal '22. And that's all the reasons I've done. We think the supply will be pretty constrained. And so it's going to be a very profitable year as well. So sort of I just wanted to mention that. So Manish, you want to...

I'm going to say the same thing at those points.

And then lastly, Dave, very quickly, there's been a lot of questions around the gross margin guide. To the extent that the revenue guide is being driven by PC weakness, do we just conclude that PCs are a much lower-margin business for you? And if PC start to come back, does that become a margin headwind?

I mean, PCs are not a low-margin business for us. In fact, the PC segment is very solidly profitable and compares very well with -- in other segments that we serve. And so I would say that I would not think of PCs as being either a headwind or anything else like that. But I feel like I should just reemphasize what I mentioned earlier, which is it's certainly the PC segment that is part of the issue that we have. The other issue is, I just want to reiterate our below target inventory level and the aggregate memory and storage -- on that memory and storage side, as well as our own constraints that we are seeing on the supply of certain ICEs as well as the constraints that our customers are seeing across multiple end markets in terms of getting all of the semiconductors that they need to get matched sets of semiconductors for their products. So it's -- I wish I could say that it is only one thing, but all of these constraints are causing us to feel like we'll have a stretched out cycle.

Our last question comes from C.J. Muse of Evercore.

You hinted at mix into '22 throughout this call, but I was hoping you could isolate your commentary just on mix for calendar '22. And what should we be thinking about pluses and minus wise for you guys? Can you hit on DDR5, data center SSD? Anything else that we should be considering as it relates to gross margins and revenues in calendar '22? .

Yes, I can address some of that. So on the product side, of course, we will continue to ramp our newest products with 176 layer NAND and 1-alpha DRAM. So both of those technologies have good cost structures. So we are very eagerly looking forward to continued ramp of those products. And between the different product categories and segments, we do expect that the component mix of our NAND business is going to decline. The components typically have the lowest prices and we feel that the high-value solutions that we have in NAND will post a pretty robust improvement in fiscal '22. So we should end up with more SSDs in the mix and also more data center SSDs in the mix in fiscal '22. So we are certainly looking forward to that. And then, of course, we expect to have more DDR5 shipments in the mix, which is a new set of products that the industry is just on the cusp of adopting. Of course, the DDR5 volumes as a percent of our total is still going to be pretty small. It will take a few quarters for it to ramp to some appreciable levels. But fiscal '23 will be a very strong year for DDR5 shipments, and we obviously, expect DDR5 ASPs to be higher because those modules are more expensive and more complex than DDR4 modules. So those are going to be additional opportunities for us to continue to drive our mix. And longer term, we will have improvements in mix coming from high-bandwidth memory and CXL. So if I look out for the next 2, 3, 4 years. Those kind of products will start to become more impactful. And so we have what we believe is a multiyear journey in which our portfolio mix should continue to improve, continue to become better and continue to enable us to create more differentiation and higher gross margin opportunities for us as a company. And part of the R&D investment that Dave spoke about, that 15% growth, is meant to drive some of those investments that enable that differentiation for us to be able to drive higher gross margins in the future.

Thank you. Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.