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

Good day, and welcome to Micron's Post-Earnings Analyst Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to your speaker, Mr. Farhan Ahmad, Head of Investor Relations.

Thank you, and welcome to Micron Technology's First Quarter 2023 sell-side call back. On the call with me today are Sumit Sadana, our Chief Business Officer; Manish Bhatia, our EVP of Global Operations; and Mark Murphy, our CFO. Today's call is being webcast from our Investor Relations site at investors.micron.com. As a reminder, the matters we are discussing today include forward-looking statements regarding market demand and supply, our expected results and other matters. 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, including our most recent Form 10-K and 10-Q for a discussion of the risks that may affect our future results. Although we believe that the expectations reflected in the forward-looking statements are reasonable, we cannot guarantee future results, levels of activity, performance or achievements. We are under no duty to update any of the forward-looking statements to conform these statements to actual results. We can now open the call.

Thank you. Today's first question will come from the line of Brian Chin with Stifel.

First, in the prepared remarks, there was a discussion of an insurance recovery in the February quarter. Is it fair to assume that's 100% fall through in the model? And then if I back that out of the guidance, that's the gross margins at around 5.5% plus or minus in fiscal 2Q?

That is correct.

Okay. Got it. Got it. That's helpful clarification. And maybe one more for clarification -- thanks Mark. I think one of the questions you said planning for lower utilization to persist through the balance of the fiscal year. And so are you thinking sort of a range for what that underutilization could look like, like a 10% to 20% below full? Or is that sort of you'll go -- you'll do what needs to be done as you move along through the year? Any sort of additional clarification on that?

Sure. And I could have been more specific, but it was a long answer as it was. I think on the back end, the utilization will improve in the second half of the year as volumes increase. On the front end, which is -- was the purpose of the announcement and of course, is the larger effect to your point, we're assuming 10% to 20% through the balance of the fiscal year.

Got it. Okay. That's helpful. And then kind of my follow-up -- and this, I think, also relates to the question you were asked. But when you think about the transition to DDR5 next year and it's tied in data centers, ties and client PC, the various platforms, products, et cetera, progress, it sounds like in mobile as well. But doesn't that run the risk of impairing some DDR4 inventory at some point here? And also kind of tied to that, when you get -- where do you think you can get inventory days back to by the end of the fiscal year? And what level does this need to be back to in order to regain pricing leverage in the industry?

Yes, I can take that. In terms of the DDR4 versus DDR5, the DDR5 ramp is underway in the client space. And somewhat behind that in the service space will really start ramping in earnest through calendar 2023. And the crossover with DDR4 doesn't occur in both segments, client and server from our estimation right now till mid-calendar 2024. So there is still substantial level of DDR4 runway in both segments for quite a while. And so we are not so concerned about that particular aspect, although the overall level of inventory is certainly something that we are focused on trying to improve. And so that's sort of the color that I just wanted to provide. And what was the other question -- sorry -- you had asked?

And the other part of that was just -- you talked about the days inventory peaking in fiscal 2Q, but where do you think you can get inventory days back to by fiscal year-end? I think prior guidance was around 150 days. And what level -- I know it's hard to answer it this way but what level does it need to get back to in order to regain pricing leverage?

On the pricing part, the -- obviously, as the inventory improves over time, the pricing environment will improve and the rate and pace of that improvement is heavily dependent on what happens to the supply in the industry. We have provided you a view of what we are doing with supply. We have taken a lot of aggressive actions in '23, reduced our CapEx, reduce our CapEx further on the WFE side in '24. And also reduced wafer starts and highlighted that really the improvement can be accelerated if the supply growth in the industry becomes negative in '23 calendar year in DRAM and flattish for NAND. So what will actually happen to our inventory is obviously a function of what happens to the overall industry supply-demand balance and not simply a function of what we do. But we continue to take appropriate actions through the year.

And Brian, on the days, as Sumit said, it's that far out. It's really a function of how the market is recovering, but we expect days to peak in second quarter. So we were at 214 in the second quarter. So we'll be above that. And then we expect to remain at elevated levels, but down in days in the third quarter and then down days in the fourth quarter, closer to 150 obviously we are today.

One moment for our next question. And that will come from the line of Joe Moore with Morgan Stanley.

I want to make sure I understand the mechanics of any kind of lower of cost to market inventory adjustment you may have to make. Is that a function of overall gross profitability? Or do you evaluate that on a product-by-product basis where if certain products go negative on gross margin that you have to take a charge there?

Yes. Joe, we look at a single pool of inventories. So -- yes, and that's disclosed thoroughly in our 10-K filing.

Got it. Okay. And then in terms of the underutilization, is there any thought of pulling that cost forward in time by taking a charge? Or that -- it sounds like it's just going to be kind of first in, first out on the inventory cost.

Joe, that charge will be a function of the inventory accounting on the lower of cost or net realizable value. Well, there are some period costs, Joe, I think I referred to that as well, but the majority is in inventories, but there are some period costs that dropped down in the period that they're incurred.

And just to be clear, for our underutilization, we've taken actions to reduce both DRAM and NAND by 20%. And those are happening across all of the notes that we have meaningful output. And we are kind of expecting to continue that until we see market conditions improve. So the numbers that we're giving you are just sort of the estimates we have right now based on these actions that we've taken.

One moment for our next question. That will come from the line of Steven Fox with Fox Advisors.

Just to follow up on that one. In terms of, if things got worse from here, is basically the actions from here mainly related to how you manage wafer starts because I would assume there's a point where you don't want to cut it too far into future growth plans? And then secondly, in terms of wild cards for the second half, you talked about some of the trends you're seeing positive in PCs and cell phones. I'm just curious, how are you sort of put a -- gaining those in terms of where you're most confident you're seeing improvements also the bottom and where you still have question marks on end markets.

I'll take the first part just in terms of the future growth plans for CapEx or wafer utilization plans. I think you've seen, Steven, that we remain flexible, right? After the -- at the last call, we had talked about some smaller underutilization actions in DRAM and NAND. And then as we saw conditions throughout F Q1 get worse, we took action and we kind of made the public statement informed you back in November. That we were going to go to 20% utilization, and we said then that we're reducing our CapEx view. So I think right now, based on the trajectory of recovery that Mark and Sanjay outlined, this is our view. But I think we remain flexible to both -- adjust CapEx or adjust wafer start and utilization depending on the trajectory of the improvement in inventories and recovery in our own inventory and as well as demand outlook that we see. Certainly trying to make sure we prioritize maintaining our technology learning for these new nodes, particularly 1 beta and 232 layer, which are really dynamite nodes.

And in terms of end markets, different end markets will behave differently when it comes to the recovery process. And we definitely stay very focused on each segment of the market, each customer, each geography, they all have their own individual dynamics and we are appropriately planning for inventory as well as our overall projections for growth and support of that growth based on the individual dynamics at each customer, in each segment, in each geography.

One moment for our next question. That will come from the line of Vijay Rakesh with Mizuho Group.

Just a couple of quick questions. On the -- as you look out to 2023, I'm just wondering if you look at your key markets, PC, handset and server, what you are thinking -- or what you're estimating, I guess, in terms of unit growth and content growth, if you could parse that out?

Sure, yes. So in terms of some of these end markets, we also mentioned some of this in our prepared remarks but on the PC side, after a pretty sharp pullback in calendar '22, we're estimating in calendar '23, there's another low to mid-single-digit percentage reduction in PC units in calendar '23, that will take the total amount of PC units sold close to where it was in 2019 before the big run-up happened due to COVID, in those PC units. So full replacement back pretty much to those levels. On the smartphone side, again, pretty sharp pullbacks in calendar '22 in terms of units. Expecting a stabilization. So right now, we are modeling flattish to very slightly up unit growth in handsets in 2023, largely driven by strength in the second half of the calendar year especially as some expectations that China's recovery will start to regain its footing after it goes through some challenging time in the next few months as it recovers from initial phases of COVID through the reopening process. So that's what we have modeled on that front. In servers, we do feel that the growth will continue in content both in DRAM and NAND. Of course, even in PCs, we expect continued growth in DRAM and NAND content. PC, the penetration of NAND is mostly completed over 90% but content growth should continue. Smartphone content growth with 5G shift should continue. So more handset percentage as a percentage of the total in 5G in '23 versus '22. So that should help content. So our content mostly continue server unit growth as well, I mean these trends in '23 compared to historical trends will get passed on the server side, especially in terms of units based on the expectation of some level of macroeconomic impact on enterprise demand, both enterprise and cloud data center demand will get somewhat impacted, we think, in '23, but content growth will continue. I think as we look at memory and storage demand apart from all of these end market trends, there is obviously a pretty significant inventory component at the customers that comes into play. And the progress in improving inventories at customers is different by customer and different by segment. PCs and smartphones entered this inventory correction first and have made more progress compared to data center inventory, which is still at pretty high levels right now entering calendar 2023, which is why the DRAM and NAND demand from data center customers is heavily impacted in 2023 because of the level of customer inventory in that segment. So there are these end market trends and then there is the impact of inventory that ultimately, the interplay between them determines the demand that we see and the industry suppliers see on the DRAM and NAND side.

Got it. And just a quick question on the DDR5 side. I think you mentioned exiting 2023 at 30%. That ramp seems to be a little bit slower. Is that because of the price premium or to DDR4? Or is it being gated by some of the new product ramps?

Well, there should be -- there should be availability of DDR5 product, it's more driven by the rate and pace at which the DDR5 adoption in the industry happens based on the ramp of specific CPU platforms. The ramp is continuing on the client side, but really 2023 is an important year for the server platforms to really embrace and drive volume in critical mass or some of the newer platforms that are critical for DDR5 adoption in the data center. And so it's a matter of the rate and pace of that more than the availability from a supply perspective.

One moment for our next question. That will come from the line of Sidney Ho with Deutsche Bank.

Maybe just one more question on the underutilization charges in fiscal '23, $460 million. Just want to make sure I understand this, if utilization doesn't go up in the next 3 quarters, like Mark, you said 10% to 20%, does that mean that, call it, $230 million in the quarter will continue into fiscal Q1 given that there seems to be a lack of when inventories flush through to -- through the income statement. And if kind of the $230 million a quarter right now to think about or there's a different kind of linearity.

So we've -- we are going to 20% underutilization. Let me make that clear. Those underutilization charges, some of them end up in period, but a minority of them. The majority of them net of the cash cost -- I should say, the variable cost savings that we have. Those net costs, the majority of those are put into inventories, those inventories clear -- those higher cost inventory is clear in the second half, primarily for fiscal '23. Now some of those inventories -- because we're underutilized for the balance of the year, some of those charges for '23 COGS will end up clearing in fiscal '24. And as I said on the call, to the extent that volumes are better than we think, more of those higher cost inventories will be pulled into '23. And to the extent that volumes are end up being lower than we think, more of those higher cost inventories will be cleared in '24.

Okay. That's clear. My follow-up question is on free cash flow. I know for the first quarter, you have negative $1.5 billion, if my math is right. How are you thinking about free cash flow for the remainder of the year, Q2 and then maybe for the full year?

Well, what -- Sidney, can you repeat the question?

Yes, sorry. So as it relates to free cash flow for fiscal second quarter and the trajectory for the rest of the year.

Yes. So on free cash flow, we did indicate for first quarter while we ended up incurring $1.5 billion negative. Now in the September quarter, we also said at that time the second quarter free cash flow would be challenged and since we had that earnings call, the conditions have worsened, so incomes down versus what we originally thought, the working capital build is worse in the sense that inventories are higher than we thought from the second quarter. And then the receivables drawdown that we enjoyed in the first quarter will pass, and we'll be more flat on AR in the second quarter. And then CapEx, as we have mentioned before, remains elevated in the second half. Now it's down versus second quarter, but it's still around $2 billion. So we would expect second quarter to be more negative than the first quarter. And then just as we said before, we would expect free cash flow to improve in the second half, relative to the first half as a function of the business environment improving. The volumes increasing and then -- and the capital investment decreasing in the second half, and that's the profile for the year.

Yes. And Mark mentioned, CapEx will be elevated in second half. I think you meant second quarter.

I meant second quarter.

Second half, clearly, the CapEx is down, quite a bit.

I did mean second quarter. Yes.

So for the full year, we should be thinking about -- still thinking about negative free cash flow for the full fiscal year?

Full fiscal year it will be free cash flow negative.

One moment for our next question. That will come from the line of Pierre Ferragu with New Street Research.

Yes, I just want to make sure that we have the correct COGS assumptions for DRAM and NAND. So if we combine your assumptions on bit growth -- on DRAM bit growth down and then flattish with cost per bit down, say, low single digits for both DRAM and NAND, would it make sense to assume that DRAM COGS will be down, say, mid- to high single digits and low to mid-single digits for NAND.

Actually, because of these underutilization effects and some of those absorption -- or lower volumes and stuff on what we talked about on period costs and back end. DRAM will be down slightly and NAND will be up. Now if we were to strip all those effects out, then NAND and DRAM would be better, but in this year, it's going to be a challenge year from a cost standpoint. Now it's -- the cost downs resume in the fourth quarter, but it's challenged in 3 other quarters.

Okay. How should we think about 2024?

I haven't commented on '24.

One moment for our next question. That will come from the line of Tim Arcuri with UBS.

Mark, I had 2 sort of sets of questions. The first is for fiscal Q2, it sounds like the story versus 3 or so months ago where you thought that it would be pretty flat versus fiscal Q1. It sounds like bits are still up, maybe they're up a little less than you thought, but the real delta is more pricing than bits in fiscal Q2. So I wanted to confirm that. And on the fiscal Q3, it sounds like bits are going to be up just like you thought they would be before. But in fiscal Q3 revenue is up, but I think you had suggested that May quarter would see such a big snapback in bits that bits would be up year-over-year in both DRAM and NAND. So is that still the case? It sounds like probably not, but I just wanted to ask that. And then I had a follow-up there.

Yes, we're starting to do variances on old forecast and stuff, which makes it difficult. But Tim, to answer your question generally, I would take away that the profile is roughly the same that we had commented on before and that volumes would improve first quarter to second quarter and then further strengthen through the year. And then the market conditions are challenged in second quarter. And as there is slide at the end of the tunnel on inventories and supply-demand balance improving that those conditions would moderate in the back half of the year. So the profile stays about the same, but to your point, the volumes are a bit lower than we thought when we gave that guidance in September and the market conditions certainly are more challenging in the near-term period.

Got it. And then just last thing. So I know you're not giving CapEx for fiscal '24 and just saying that it's lower than what you thought it would -- that it was going to be prior to this. But the standing guidance is sort of you target like mid-30% of revenue. Is that still the way to think about it? Or because of these actions you're taking that fiscal '24 CapEx could be below that mid-30% of revenue target?

Boy, Tim, it's early to be giving comments on that. And we do have higher construction spend, which pressures CapEx up. However, we are absolutely committed to free cash flow generation in '24. So that's going to play a very important role in the level of CapEx spend we have, along with just the view on the market conditions and various other things, short and long term. So difficult to make the call now.

And as you're aware, the funding -- federal funding process is still working for the U.S. chips grants, which will impact our reported CapEx for the construction projects that we've targeted here in the U.S. will be kicking off some of it this year, but some of it is for sure in fiscal '24. So still early to comment on that, which is why we gave you some more color on the WFE part, which isn't really impacted by that in that time frame.

One moment for our next question. And that will come from the line of C.J. Muse with Evercore ISI.

I guess first, just a little modeling help for February. I guess based on the comments around mobility, but it's accelerating, can I interpret that as roughly -- bits for NAND growing perhaps roughly 2x as fast as DRAM? Is that kind of the right way to think about it?

Yes, I would not just think of NAND growing 2x faster than DRAM. I think it's just definitely our mobile revenue is growing sequentially from F Q1 to F Q2. But we're not providing further guidance on what part of that, which is really growing. But as you know, the overall revenue is down quarter-on-quarter.

Okay. I guess maybe a deeper question on the inventory side. I guess where do you think you'll see green shoots first in terms of things returning to normal. As that occurs, how do you see kind of the mix playing out for you guys?

If we look at the overall business and the different parts of the market, the automotive market is the most resilient, but obviously, it's a smaller part of the whole business. And industrial markets probably come next in terms of relative stability. Now the PC market and the smartphone markets were the ones that entered this downturn first, and those customers have made some progress on the inventory. So we do expect that those customers will likely end up improving their inventories to healthier place sooner than, for example, data center customers will do have high levels of inventory and have decided to start improving their inventory position aggressively relatively late compared to the other segments like PCs and smartphones, just based on the end market demand changes that have been occurring over the past several months. So we do expect that the PC market is likely to show some stabilization. But just keep in mind that while we are talking about customer inventory and we are looking at bits of inventory at our customers, our customers are also looking at DIO inventory and how many days of inventory they have. And it's obviously compounding their challenge because the forward-looking sales when it weakens due to macroeconomic environment and so much and individual segment trends, then the weakening end market trends does impact their view of what they can do on the inventory front, even if bit inventory reduces, DIO inventory still have more work to do. So those are the dynamics that are playing out. And we do expect that these server customers will take longer to get to a healthier place compared to the other segments.

One moment for our next question. That will come from the line of Chris Caso with Credit Suisse.

Yes. Just maybe a bit of a bigger picture question. And maybe you could give some comparisons to the last cycle and some of the prior cycles. And what you talk about is the supply-demand balance here is the worse than the last 13 years. Over the last 13 years the industry has structurally improved. So what do you think is the reason for the supply-demand balance that's occurred now? Is it solely market conditions do you have anything to do with structural conditions? And what does that mean for future cycles? In terms of how we read this for the bigger picture.

Tough to extrapolate based on 1 data point. Certainly, since the industry consolidated about a decade ago in DRAM, we have had relatively well-contained cycles and relatively healthy -- very healthy through-cycle profitability, certainly in DRAM across cycles since that point. In this particular cycle, definitely, there has been a demand shock and a set of exogenous factors that have come together in a way that is quite unusual. So you've seen 40-year highs in inflation, very rapid increases in interest rates, so macroeconomic environment, certainly, one for the history books in terms of multi-decade sort of challenges on that front. We have award in Europe that has not happened in many, many decades. We have COVID impacts and big shutdowns in China impacting demand in the second largest economy in the world. And then you have whole host of factors and individual companies and segments as a result of supply chain disruptions, et cetera. So there's a lot of moving parts that have hit the markets in these segments. So there has been definitely a pretty significant demand shock. And also included in that demand shock is the other side of all of the inventory buildup that happened when so much of the business went into a shortage, probably the biggest set of shortages in the semiconductor industry in a long time, probably ever. And then customers responded with building inventory in whatever areas they could get their hands on. So a lot of factors coming together created that demand shock. I think on the supply side, which is now where the burden is on the supply side to respond to that demand shock and cut supply and bring the industry into a healthier place. We have definitely taken pretty aggressive actions. We have got the wafer starts in really fast order. We have dramatically reduced CapEx in 2023. We have signaled very clearly that we are also reducing WFE CapEx in 2024 from 2023. And we have also responded to the changes in the market environment by lengthening the time between technology node ramps. We've taken a lot of aggressive action from our side and as the rest of the industry, hopefully does what it will individually decide to do. We are hopeful that the industry will become healthier as time goes by through calendar '23 and inventories will start to improve slowly from the peak that we're seeing in fiscal Q2.

One moment for our next question. That will come from the line of Krish Sankar with Cowen.

This is Eddy for Krish. You guys have impressive technological leadership, but your competitors have a bigger sale. So net-net, in terms of cost per bit, do you think you're at least at par with competitors? I'm asking because I'm wondering if there is a scenario where competitors become more aggressive on pricing, which will further elongate this down cycle.

So Krish (sic) [ Eddy ], I can just answer our competitive position. I had given some color on that at our Investor Day that we were certainly trailing in terms of cost competitiveness a few years ago. But by getting the technology leadership, even at our lower scale, we've become very competitive on our costs in both DRAM and in NAND. And I won't give specifics in terms of us versus others, but we're definitely very competitive now, thanks to the technology leadership that we have. And that -- having that technology leadership does give us flexibility in terms of how we manage our business and being able to -- even as we made the decision to -- and we just explained to delay the ramp of these new technologies, we expect they're still going to be very, very competitive when they launch when we ramp them as market conditions improve and they're going to -- the yields are coming up really nicely on both of them right now. So we feel pretty good about our underlying cost competitiveness and technology competitiveness here and into the future. And the scale that we have, we expect to maintain that share through the cycle as we're to be able to make sure that we can -- maintaining the ROI on all of our investments -- our R&D investments.

The one area that I think you know that there is an impact is related to foreign currencies, exchange rates versus the dollar. So definitely, the dollar has appreciated a lot versus the Korean won and has also appreciated a lot versus a Japanese yen. We do have some manufacturing in Japan, but not as much as some other competitors have -- for example, in NAND. So there is definitely some competitive impact on costs related to currencies, but we also feel like the dollar has already come off by 10% in recent weeks, on its peak. And hopefully did the future direction of the Fed slowing down and ultimately stopping. We're hoping that the dollar will start a weakening cycle that would be reversing some of those impacts on a competitive basis on cost that is due to the currency itself.

Got it. And if I can squeeze 1 more in. How should we think about the OpEx savings of $150 million? Are they related to production? And should we expect them to come back once utilization goes back to normal?

No, they're not related to production at all being COGS. It's related to R&D and SG&A activities. And clearly, the headcount actions are part of that, but very careful specific program-related actions, work with vendors on renegotiating on rate and scope of activities. Various other productivity projects and other items in SG&A as well.

As usual -- from Q4 of '23 to 2024, we'll have usual transition that happens in most companies where you know we have bonus coming back, salary increases, those sort of things, but other than that...

Thank you. Thank you all for participating in today's call. This concludes the program. You may now disconnect.