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

All right. I think we're just about to start. My name is Mehdi Hosseini, senior analyst at Susquehanna International. It is with great pleasure to host the fireside chat with the team from Micron. This is actually truly a global fireside chat. We have Mark Murphy and Farhan from the U.S. and Manish Bhatia, who manages operation for Micron, dialing from Singapore. And the format is that we will go through a set of prepared questions and topics for 30 minutes, and then we'll open it up to the floor for any question. You can raise your hand or, alternatively, just e-mail me your question. So with that, again, I want to thank the team from Micron, especially Manish dialing. And perhaps, Mark, maybe we could start with you giving us an overview of the quarter that you just closed, and then we go from there.

Sure, Mehdi, and thank you all for joining us today. Before I start to answer your question, I'll do a safe harbor. We will be making forward-looking statements. These statements have risks and uncertainties associated with them. I do refer you to the risk factors disclosed in our public filings. So to your question, we are not providing a quantitative update on the quarter, which ends today, but I can provide you and the audience some color. Customer inventories are improving, though they're still elevated, with some markets like mobile in relatively better shape than others such as cloud. We continue to expect improving demand through the year, but pricing trends remain challenging. Consequently, we expect lower margins in our fiscal Q3 than we previously expected. This outlook may lead to material inventory write-downs, which could or would have an adverse impact on our second quarter margins and EPS. While we're encouraged by supply actions across the industry, it's clear that there is still significant supply-demand mismatch in the industry. So as we've said before, we expect our DRAM supply to contract in calendar '23 to help better align supply and demand. To supply anticipated volume growth in the near term, we have ample inventories and adequate production. While our cost of goods are being impacted by the effects of underutilization that we've talked about before, our broad-based spend reductions are generally coming down in line to ahead of plan. For example, on labor, head count is now projected to be down approximately 15% versus the 10% we discussed on our December earnings call. These labor and other reductions will take some time to work through. For example, on the labor, we talked about that occurring through the calendar year. But the important point is that we're seeing progress as planned. OpEx, for example, is coming down. And we expect to achieve the $850 million per quarter level by our fiscal fourth quarter, which is down from over $1 billion in the fourth quarter last fiscal year. We will continue to evaluate and navigate these challenging market conditions and adjust our operating spend and CapEx as appropriate. We remain confident in the long-term demand for memory and storage and in our ability to capitalize on these long-term trends.

Great. Thank you, Mark. Thanks for the update. And perhaps maybe you can remind us what is your expectation for DRAM bit shipment this year. And if this year is going to be one of the most severe downturns in DRAM, then in -- next year look really attractive. So update on the bit shipment for this year and what is your current assessment looking into next year for DRAM shipment.

So Mehdi, and thanks again. Maybe I'll take this one. On our last earnings call, we had mentioned that we expect that DRAM demand for this year will likely be in the low teens and NAND likely to be around 20%, both of these slightly below what we believe the long-term demand for DRAM and NAND as we're continuing to work through inventory in this calendar year. And we will likely update our industry growth forecast at the time of our F Q2 earnings, but that's what we had -- that's the guidance we had provided before. We also said that we believe that the industry -- the DRAM industry needs negative year-on-year bit growth in calendar year '23 in order to have -- to return to normalized inventories by the end of the year and flattish growth for NAND for that to -- for inventories to normalize. And there have been actions across the industry that are in that direction. But again, we'll provide a little bit more color on our next earnings call. For us, we had said that our bit production in calendar year '23 for DRAM as a result of the CapEx reductions and the utilization reductions that we've taken, that we will be negative year-on-year in DRAM and up slightly in NAND. As a reminder, we had announced approximately 20% utilization reduction in both our DRAM and NAND networks in the late part of last year. And we haven't really provided anything yet on calendar year '24. It's too early for that right now. We'll have to see how things play out.

Okay. Great. Just to summarize, your -- the bit shipment for -- DRAM bit shipment, low teen; NAND, 20%. But on the supply side, a decline in DRAM and a flattish on the NAND side, right?

Well, we said that the industry needs that. For us, what we said is that we would be negative on DRAM bit growth year-on-year and on NAND, maybe slightly up.

Okay. Got it. Okay. And the inventory situation, I think we started with the inventory correction middle of last year. And it seems like we need the entire 2023 to work that down. Two follow-up questions here. If you could elaborate or quantify DRAM inventory for Micron and industry or at the customer side, same for NAND. And I have a follow-up here.

Yes. Maybe, Mehdi, start with customer inventories. As I mentioned in my opening comments on the customer inventories, they're still elevated, I'd say, across the board. Now some markets like mobile are better. Other markets like cloud are worse. We do expect our demand to increase through the year as we continue to see the supplier or these customer inventory levels decline and improve. So that's still the case. And again, that's going to support what we see as volume growth through the year. Our inventory levels are still very high. And as we said on the last call, our second quarter here, we expected days of inventory to be over 200 days with NAND being worse than -- or higher, I should say, than DRAM. However, as I mentioned in my earlier comments, given the challenging pricing environment and assessing our current outlook, yes, we could have inventory write-downs that are material in the second quarter. And of course, that would have an impact on reported inventory and the associated metrics. We did mention on the last earnings call, looking out through the year, we did mention that we expect to end FY '23 with still elevated inventory levels in excess of 150 days versus what's more normal, down around 100 days, which we're certainly keeping -- working towards longer term.

Given the setup that you just laid out with inventory, then could fiscal Q3 revenue be down again? Or there's enough of a shipment increase that would offset the lower pricing?

Yes. I mean, the market is clearly trying to get it splitting. And while we expect volumes to improve through the year and actually pricing, we expect the declines to moderate through the year. But the pricing environment currently is weak. And our third quarter could be flat to down or we believe will be flat to down versus the second quarter as we see it today. But of course, we've got -- this is the end of the quarter. We'll report at the end of March, and we'll have a clearer picture and more commentary at that point.

Sure. I think, Mark, just since we're on the topic, I think one of the things that I always struggle with, inventory write-down. So your book value will come down, but you can actually make up for it as we go through an upturn. Your margins could also improve as the cycle turns. So can you just give us an idea how we should think about the near-term and the longer-term impact of inventory write-off or write-down?

Well, I mean, you've touched on it. It's going to be a timing difference. So we're essentially pulling forward or writing down things and pulling forward that cost into the current period. And so the sort of the basis of that inventory would be lower. And going forward, that margin would be better with that inventory having been -- a portion of it having been written down. So again, we'll talk in the earnings call about -- because we're going -- we'll need to go through our procedures at the quarter end on our inventory balances and assess those against the pricing environment in our latest forecasts. And then we will size a write-down if and as required and, at that time, talk about how that will flow through the income statement, play out through the forecasting.

And just one final follow-up here. As we look into the second half of the year, your inventory is coming down, your customers' inventory coming down. And hopefully, supply and demand are more in balance. Is it going to take a period for customers to feel good enough to start thinking about a long-term contract -- pricing contract? Or is it going to take like 2024 -- sometime in 2024 is when longer-term contracts would resurface? How is the inventory correction going to impact price negotiation dynamics?

Well, that's probably a question that's better handled in the earnings call when we have a better picture of the forecasts and a more complete picture of the market dynamics and are willing to comment on that. I mean, I think what we are seeing is the market working to get its footing. We see some positive data points and hope that those positive data points increase in frequency and are more sustained. But right now, there's no doubt that it's a challenging pricing environment still. However, we do still see volumes increasing through the year. We do still see and believe that customer inventories are improving and that eventually -- and supporting the volume growth, customers will be replenishing their inventory levels. And eventually, prices will firm up and market conditions will improve. But we'll comment more on the earnings call, if there's any change in that outlook.

Moving on to the next topic, CapEx. I think last time, you've guided to fiscal year '23 CapEx of $7 billion to $7.5 billion. Any update there?

No update, maybe just reinforcing that we're certainly -- it's a big lever for this business, an important part of our business, both managing the cash flow and investing for the future and the technology investment required. But we still have no update range between $7 billion and $7.5 billion. That's down 40% year-on-year. WFE within that is down over 50% year-over-year. And I think what you'll see in '23, and you see this play out into '24, is the mix of CapEx is kind of a bit heavier on construction, and that's related to just long-term capacity planning. And then also related to fiscal '24, as we mentioned on the last earnings call, we do expect our '24 to be significantly down versus what we had originally planned to align with the supply-demand environment. And we expect '24 WFE to be down from '23 levels. So again, we see an improving demand environment through the year. But as we've talked about, we have a substantial amount of inventories to work through. And then we have the underutilization, of course, that Manish mentioned that we will reutilize our capacity or bring capacity utilization back up. And then we have some nodes that are in great shape and can be efficiently ramped when we need to, and Manish can talk more about that later. But we're feeling good about our current CapEx projections. But of course, we'll modulate that based on market conditions and the outlook.

Just a quick follow-up here. When will you include the CHIPS Act impact?

Well, that process is underway for the application, as you've seen. We have some assumptions internally, but -- and may be in a position to make some comments on it at the earnings call about what we expect. But we would expect substantial grants in addition to the legislated ITC in order to support any investment in the U.S.

Sure.

And it's -- maybe I'll just add -- maybe I'll just jump in on that one and add that our plans for -- that we've announced for both the Boise, Idaho manufacturing site that will be co-located for R&D as well as the larger-scale manufacturing site in Syracuse, New York. Both of those are really to meet supply for the second half of the decade as we talked about. So we're still a ways out in terms of the WFE CapEx for those coming into our window, not likely to be in -- or not going to be in '24. So really, it's just the construction CapEx that might start coming in there. And as Mark mentioned, we're still pretty early in the CHIPS process to understand how grants and investment tax credits will be able to impact those as we go.

A couple of other semiconductor manufacturers that have commented on CHIPS Act, it seems like there is a lagging effect. The funds will come in 1 or 2 years later, and then they will readjust the depreciation schedule. So it could have an impact on the margin and -- depreciation on margin 1 or 2 years after the actual application is filed. Is that a reasonable assumption?

Mehdi, I mean...

Mark, if you can comment on the accounting again.

Yes. Mehdi, we've got a lot of -- there's a lot of variables there that we'll talk about at a future date when it's appropriate. The important thing is just to recognize that we're being capital disciplined. We are investing the amount that we need when we need it. And the -- as it relates to the U.S. investment, it requires the full support of that legislation in the form of substantial grants and a full and long ITC applied to the projects.

Got it. Okay. Let's switch gears and talk about process technology, and maybe Manish could help me here. What's the update with the DRAM technology as you transition from 1-alpha to 1-beta and to 1-gamma? And what's the update on the EUV insertion?

Sure. Sure. Thanks, Mehdi. So on DRAM, we -- 1-beta is looking like a very, very good node for us. And as a reminder, our 1-beta node does not utilize EUV on the back of 1-alpha. We're still -- we've been able to push the advanced immersion capabilities and multi-patterning beyond others in the industry, and we feel really good about that decision with regard to both what we were able to deliver in 1-alpha, industry leadership in terms of time to market. And it's been a terrific node for us. And then 1-beta now that we've -- we're making very, very good progress on and expected to have really good bit density improvement, about 35% improvement versus 1-alpha, which in itself had 40% improvement over 1z. So these 2 have been really, really good nodes. We're bringing up 1-beta across our portfolio. We'll have 1-beta. We'll have solutions from -- in the mobile space, in the compute space, LP5 and DDR5. And then we'll have it in -- across our graphics and high-bandwidth memory as well. So really, it's going to be a broad node, and that's what we've invested for this year was to make sure that we could bump the yield curve on 1-beta and -- which is going very well right now, and that we'll be able to deploy it across multiple products so that as demand recovers, we're going to be ready to ramp the node well across all end markets. And everything is looking like it's going to be, again, another time-to-market advantage for us as well as a very, very strong and competitive node. With regard to EUV insertion, as we said before, we're going to be inserting EUV into our road map on 1-gamma. And that is -- continues to be the plan. In fact, we have now received EUV tools not just in our development center in Boise but also in manufacturing site -- in our manufacturing site in Taiwan. So we're getting ready for that 1-gamma insertion. That 1-gamma, as we announced or explained on our last call, is now scheduled for 2025. But we're exercising the EUV tools now and making good progress still. We'll have multiple layers of EUV on 1-gamma, but would expect that we'll have even more penetration of EUV in subsequent nodes later, more number of layers in subsequent nodes. So we really feel great about our DRAM road map and some of the decisions we've made. 1-alpha was a great node. 1-beta, optimized for DDR5 insertion and as well as LP5, is going to be a really strong node. We're getting excellent feedback from our customers from the early samples we provided and think that we'll continue to lead the industry in the DDR5 ramp as we have been.

Manish, when you say insertion in '25, is that fiscal year or calendar year?

I'm not sure that we've given specifics. I think it will be sometime in calendar '25.

Got you. Okay. And then just one quick follow-up. What would be approximate mix of DRAM bits on 1-beta like exiting this year?

I don't think we've provided that yet, but it will be a very strong node for us this year as we continue to ramp up on DDR5 and LP5 and take advantage of the benefits, yes. And we may provide some more color a little bit here. But I can tell you that it's performing very well both from a sampling perspective and from a yield perspective.

Yes. I think one concern is, okay...

It will certainly be a function of the overall demand environment and how inventories work down and how we deploy our CapEx and also the overall utilization, and you know how that works.

Yes. So utilization rate improving into next year and then just the pricing improving, that will give you the boost in DRAM gross margin. And then in '25 is when your wafer density with a step down again when gamma or EUV is inserted.

I mean, that's the next time that we'll have a technology. But I think we'll continue to get benefits from 1-beta conversion all throughout '24. And again, 1-beta is going to be a really, really good node for us.

Yes. I just want to make sure, Mehdi, that's not lost on -- we'll work down the inventories. We'll benefit once utilization picks back up. And then as Manish said, we have 2 very good nodes that are ready to go. And when they go HBM, that's going to be favorable on mix and then the next.

Sure. Okay. And then quickly, same question on NAND, where are we in technology? And what's the next step?

Yes. So we mentioned that our 232-layer is making good progress and it is. Although given the inventory situation, we have slowed the ramp of 232, again, continuing up the yield curve very, very well and also deploying -- getting it ready to be deployed across multiple end markets so that the mobile, the SSD products will all be getting ready to be qualified. And then we'll bring it and increase the volumes on 232-layer as we see the inventory situation improve and we're able to take advantage -- take more advantage of that technology. Still for this year, we expect 176-layer to be the workhorse for us in terms of our NAND through '23.

Okay. And then moving on to the mix, and it is for both of you guys, Manish and Mark. Should I assume that on the DRAM and inventory write-down, it's mostly like a DDR4 and trailing edge technology nodes?

Yes. We're not going to comment on the mix. It's -- all I can say at this point is, given the pricing environment and our latest outlook and one that we'll refine here over the next several weeks, it looks like an inventory write-down is more likely. And it will be a portion of our inventories. And I'll leave it at that.

And how should I think about DDR5 and the price premium due to the die-size penalty? In the past, we always would think of the next DDR as having a premium over the previous technology. Now given the downturn, I don't think compares -- or should I still -- should we still compare DDR5 price to DDR4? Or the pricing dynamics are different given the downturn?

Well, I mean -- go ahead.

Go ahead.

No. You're right that the die size is larger, and so it is more expensive. So I would expect a price premium. I mean, I think we would -- we expect a price premium for that product. And in terms of margin, that will remain to be seen. But certainly, the costs are higher not just for Micron, for everyone, given the specifications of the DDR5 die. And we feel we're very well positioned with our DDR5 portfolio. Again, 1-beta is optimized really for that product.

And I would maybe just also add to Manish's comment. There's the cost element, which should hopefully drive higher pricing in the market. And as Manish said, we're very well positioned on the product line. But also, DDR4 inventories are higher. As the new CPU platforms add Sapphire Rapids and Genoa, they can only use DDR5, and the DDR5 inventory levels are thinner than DDR4. So that ramp of those new CPU platforms into a bit tighter market should help with that.

Sure. And by the way, DDR5 will also be adopted for the PC application, no?

Yes.

Yes. We actually already -- I mean, I think some of that already happened last year. And Micron, we talked about our strong share in the PC space last year. But definitely now, with these new CPUs, we're starting to see the server wave begin. And we think that will take some time, though. We think that it will be -- yes, it will be happening this year in '23, but the crossover for DDR5, DDR4 will be in calendar '24.

Okay. Now I'm going to switch gears and read some of the questions that are sent to me. So I apologize if some of the questions are repeat, but I'm just going to read them to you. Going back to the EUV commentary, did you say that you have actually received an EUV tool in Singapore facility?

No. I said Taiwan.

Taiwan, okay.

Yes, that's right. Singapore is our NAND center of excellence. So no, I said it was in Taiwan.

I just want to make sure. And then for -- what about Japan? Would Japan also receive the EUV tool later on?

No. We're still working on our overall road map. But right now, the plan is for our 1-gamma to be manufactured in Taiwan and with -- taking advantage of the strong EUV expertise that there is in Taiwan across the foundry ecosystem there. And that's been -- that's where we've been at so far.

Okay. Gross margin question. If you're writing down inventory, does that imply February gross margin will be negative?

Does it imply February quarter gross margin will be negative?

Yes.

We're not going to comment on margins until we report the results. But the -- it depends on the size of the write-down, of course. And the -- it would materially affect the results if it's material inventory write-down, which the current pricing environment and outlook indicates that there would be a material write-down. Yes.

Okay. Would you ask the team if they would take further cuts to the wafer starts?

So we will -- as I mentioned before, we announced that we have approximately 20% reduction in wafer starts or 20% utilization actions in both our DRAM and NAND network. And we'll continue to evaluate that level in line with both the demand trends that we see and the inventory position that we see ourselves having. And we'll be flexible to be able to manage that to optimize for those factors.

And you said early on that smartphone inventory is relatively better. Has China reopening and demand for smartphone in China anything to do with it? Or was that just a broad statement?

I'm sorry, repeat the question.

Sure. At the beginning, you highlighted that smartphone inventory or memory inventory used for smartphone is relatively in a better shape compared to the other end markets. And the question is whether China reopening has had an impact here.

Yes. It's certainly -- I didn't call that out specifically. And yes, that would be one of many factors, I think maybe one of the larger factors. It's just that the market has been under pressure for so long and the inventory levels have been worked down for going on close to a year now that it just had more time to work it down versus, let's say, data center, which those inventory levels started being adjusted really in the fall. So I think it's a combination of those things.

Sure. And then a follow-up question here. Mark, when you talk about inventory write-down, is there a specific inventory? Are these all finished goods? Or does that also include work in progress? Any color on the mix of the inventory that has been write down -- written down.

Yes. Again, we'll provide more details on the call, but I do want to point people to our disclosures that we do -- we look at inventory as a single pool. So we're not going to go into and don't account for this at the product level. And so that as we -- we'll give more details on the size and how the write-down will occur in effect on future periods. But again, we look at inventories as a single pool.

Okay. Contract pricing, I'm just going to read this to you. I don't know how to rephrase it. As investors, we continue to hear and read that quarterly negotiations close after the actual end date of the quarter. What can you do to bring this process into the quarter and to better manage inventory? Or is there anything you can do to change the dynamics of the contract pricing negotiation?

I think the negotiation of the customers are ongoing. And as I go back to the -- how we see the market playing out, there's excess inventories at the moment. Supply-demand balance is [ 4 ]. Through the year here, we're seeing customer inventories improve and expect the market to firm up and pricing effects to moderate and eventually improve. I think it's important that, again, as I mentioned, we've seen some positive data points. There haven't been enough and they -- and not sustained enough, which is why the pricing environment remains challenging. However, we are seeing some positive data points. And I think it just bears repeating some of our longer-term themes that there will be memory and storage growth. There are strong secular growth drivers to this business. And on automotive and industrial data center and all the recent news on AI, for example, while it's not a material effect in the near term because this high demand effect is the dominant feature in the business at the moment, but it's trends like that, that long term will really drive the business. And if you believe that AI is going to drive strong growth in compute, then you have to believe it's going to drive strong growth in memory to sort of match there. So I think that's certainly -- we're managing the business, of course, for the short term and improving the financial condition. But we're also sustaining investment in technology products, manufacturing to take advantage of what is a great market to be in the long term.

Mark, just a quick follow-up there in terms of end market demand. You mentioned smartphone relatively -- inventory within smartphone relatively better. Data center is the worst. What about PC market? What do you see in that end market?

I'd say less clear. You're talking specifically about inventory levels?

Inventory and demand from the PC end market.

Well, we think PC end markets in '23, units, PC units, will decline. However, we do believe that PC bit shipments will be up in both DRAM and NAND driven by content growth. And then given where pricing is, is there some elasticity that, that growth is a bit stronger? We'll have to see. But we do see bit growth, even though we do see PC units down.

Okay. But if PC market was the first one to weaken more than a year ago, do you see any signs of a stabilization in terms of just the demand trends? Or is it still very cloudy or not clear?

We'll give an update on the earnings call. I think since our view really hasn't changed since the earnings call, I think that maybe points to stabilization, in our view. And it's just that the inventory picture in PC isn't as sharp as the distinction between, let's say, mobile and data center. I think it may be more on the kind of mobile side as far as inventory channel health, but let us make sure we do our full analysis and give a complete assessment on the earnings call.

Shifting gears more of a longer term. As we look into like '24, should I assume that content growth and unit growth would help with a rebound? Or would the content growth accelerate? Is there any color that you can give us -- and I'm thinking about PC units and content, smartphone units and content, the same thing for like the data center or servers.

I think in -- well, in PC, we talked about, we think unit -- PC units will be down. Bit shipments will be up driven by content growth. In smartphone, we believe we'll have content growth as well. Underlying smartphone units, we had said on the earnings call, we expected CY '23 smartphone units to be up slightly. Now it might be trending a little bit off that. Again, that's something we're keeping an eye on and will have a better picture on the earnings call. But we feel good about content growth. And maybe to your question earlier, things seem to be stabilizing in both markets. And of course, we're very -- everyone, including ourselves, are excited about the data center trends long term. It's going to be growing faster than most segments of the market, except maybe with the exception of automotive growing even faster. And that's across DRAM and NAND. So very excited about things happening there and our technology and product positioning for data center customers.

But looking into '24, perhaps auto could provide you with the fastest content growth relative to like PC, smartphone, data center?

Auto continues to be just a very strong grower. It's going to -- it will increase to -- through this decade to become a more meaningful part of memory and storage. And fortunately, Micron has, over the decades, developed a very strong position in that space. And those technologies, those -- all the qualification requirements, manufacturing requirements for that space, maybe Manish can talk more about.

Okay. Let me just...

Go ahead, Mehdi. I guess one more on manufacturing, but not -- go ahead with your next question.

I was going to -- I got 2 more follow-up questions. But why don't you finish? Because we're reaching the 45-minute mark. But -- well, you finish, and then I will wrap up the call with the final 2 questions that I received.

No. I was just going to add to Mark's statement. We're really proud of what we've been able to establish in automotive, #1 market share and great relationships with a number of different OEMs as well as the supply base to those OEMs. And throughout the pandemic, we had -- I personally had multiple different company CEOs telling us how our supply chain resiliency, our quality, both were things they really valued. And of course, our product -- our broad product portfolio across DRAM, NAND and NOR for automotive is unmatched. So it's really an area that we think we're going to be able to take advantage of. As Mark mentioned, automotive is the fastest-growing segment throughout the rest of the decade.

Got it. And Mark, I remember at CES, we spent like an hour debating the pros and cons of cutting back the utilization rate versus inventory write-down. So the question that I'm getting is asking, why not just peel off the Band-Aid and further reduce the utilization rate instead of working -- writing off inventories?

I think I mentioned earlier that we are remaining flexible to evaluate our utilization levels and in line with -- or the demand that we see and the inventory, I mean, that's exactly the trade-off decisions that we make. And we will continue to evaluate that.

So Manish, is there like a back-of-envelope formula? Help us understand how the dynamics of lower utilization rate would work against versus inventory write-down. Is there a formula you can give us?

Yes. I mean, look, prices are still well above cash costs. And so as Manish mentioned, we -- there is -- we evaluate inventory levels, how we can efficiently manage the manufacturing network. And we're just pricing relative to costs across the enterprise and make the best judgment we can. And again, we're always evaluating it with a number of factors and the market conditions, and we'll update you and investors as we make material changes.

It's more complicated than just trying to simplify these 2 knobs of utilization rate versus inventory write-down, right?

Yes. For example, one other factor is just trying to understand how you efficiently transition fabs from one technology node to the next, right? We have these 2 great technology nodes that we've talked about that are industry-leading, 1-beta and 232-layer NAND. And of course, as you go out, trying to figure out how you efficiently transition while managing CapEx and supply bit growth, that's just another factor. So yes, I mean, the decisions do have many different considerations.

Got it. Okay. So we're 2 minutes or 3 minutes past the 45-minute mark. Manish, Mark, do you have any closing remarks before we wrap it up?

I would just say that Micron continues to perform well and adjust into these market conditions through our costs, our spend program reductions or our spend reductions, our modulating CapEx but continuing to invest in technology and product leadership. And we're in a great position for when the market strengthens to take advantage of all of the long-term trends that we see in automotive, AI and other areas that we talked about today. And thank you for your time.

Thank you, Mark, Manish and Farhan, appreciate it. And look forward to additional commentary in a couple of weeks when you have your earnings conference call. For those investors, if there is any follow-up question, you know where to find me. And I wish everyone a great day. Thank you, guys.

Thank you, Mehdi.

Thanks, Mehdi. Thanks for having us.

Bye-bye.