Thank you for standing by, and welcome to Micron Technology's fiscal third quarter 2022 post-earnings analyst call. After the speaker presentation, there will be a Q&A session. To ask a question, press star one on your touchtone telephone. I would now like to hand the call over to Farhan Ahmad, Vice President, Investor Relations. Please go ahead.
Thank you, Latif. Welcome to Micron Technology's fiscal third quarter 2022 sell-side analyst callback. 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. As a reminder, the matters we will be 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, specifically 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. Latif, you can now open the call for Q&A.
Yes, sir. Again, to ask a question, please press star one at this time. Our first question comes from Tom O'Malley of Barclays. Your line is open.
Hey, thanks for taking my question. I just wanted to ask on the long-term through cycle model. You guys laid out 30% from an operating margin perspective. I understand that there are some China headwinds that are impacting that I think are, you know, slightly more exacerbating this down cycle, but you're already in the guidance moving below that 30%, and it sounds like there may be some more weakening to come. Can you just talk to if you think that 30% is still viable now that you've seen these cuts, and just any of the moving parts as to how you think that can sustain through a cycle? Thank you.
Tom, I'm Mark, I'll start. I mean, it is a through cycle model, so it's not intended to be a floor, it's not intended to be any point in time. I think that, you know, I think the way to look at it is we're not gonna, you know, base any given quarter on where it stands relative to the long-term model because there are a lot of conditions that, you know, that impact obviously the short-term results.
You know, we have entered a period here where the market weakened considerably in a very short period of time, and we're just, you know, we've given a fourth quarter guide to reflect the best view we have at the moment, and we'll update through the quarter.
Gotcha. Appreciate it. My follow-up is just on the cost side, going into the quarter, you guys had talked about just limited cost reductions in both DRAM and NAND. Can you just give any color on just the reported quarter on how cost kinda shook out between the two buckets?
Well, we do expect for the full year for our cost to still outpace industry costs. In the third and the fourth quarter, improvements have been flattish. Little bit better on the front end because, you know, the leading node activities and those benefits. A little bit of inflation, but that inflation's more manifest on the back end, where, you know. The net is that we end up with basically flattish costs down in third and fourth quarter. Long term, you know, we're still planning to be in line with industry cost trends.
Tom, I'll just add in there that, you know, a lot of that has the reported COGS obviously have to do a lot of mix adjustments. When we have, you know, strong growth, you know, record quarter in SSDs, you know, we have stronger growth in some of these high-value solutions. Obviously, the COGS number is mix adjusted, but when we talk about outperforming the industry in terms of cost reduction for the year, we're talking about at the memory level where our, you know, technology leadership in 1α and 176-layer have allowed us to outpace the industry this year and perform well at the front end level.
Obviously, mix of it moving towards high-value solutions and as that, you know, happens over the last quarter and this next quarter, those are, you know, impacting the all-in cost, but obviously providing higher margin as well or higher price.
Thank you.
Thank you. Our next question comes from Joe Moore, Morgan Stanley. Your line is open.
Great. Thank you. I wonder if you could quantify at all the impact of the supply reduction that you guys are looking to do, the utilization impact of that. I guess, can you compare it to what you did in 2019? Do you think it's something similar to that, more significant than that, less significant than that? Just any color would be helpful.
Sure. All right, Joe. You know, we're still evaluating, you know, what we will finally do. I think what was important for us to give guidance and clarity to everyone on was that we are reducing WFE. You know, that will have an impact on supply bit growth for both DRAM and NAND from our prior levels next in 2023. You know, at this time not really ready to compare it. We're still working to finalize what our plans are.
As we mentioned, we're gonna have a combination of trying to manage CapEx, manage utilization, particularly for some of our legacy nodes, and take a look at, you know, maybe less cost-effective nodes, how we can optimize those, how we can optimize the manufacturing footprint, space as we transition to more advanced nodes. You know, we're gonna be ramping our 1β and 232-layer technologies. We think it's very important to make sure that we continue to maintain leadership and launch those technologies, so that we can, you know, have the product portfolio benefits from those as well as the cost benefits from those. Then the other piece, of course, is managing inventory as Mark was talking about on the call.
We'll kind of look at you know all those areas together to be able to determine what the right mix of supply bit growth you know CapEx utilization and inventory is, and cost reduction is for next year. We'll provide more color, I think, on the next call.
Great. Thank you very much.
Thank you. Our next question comes from Toshiya Hari of Goldman Sachs. Please go ahead.
Great. Thank you. I have one quick clarification then one quick question. On the clarification for the fiscal Q4 guide, based on your commentary, it seems like the sequential decline in revenue, that's mostly coming from bits and pricing is perhaps, you know, down low singles or something like that based on sort of the strategy that you're taking. Is that a fair assumption?
Yeah. If you do the math on things, you'll see that volume's roughly two-thirds sequential.
Got it. Okay.
So-
That's helpful. Then, as my question, this one's for you as well, probably, Mark. On OpEx, you know, you talked about some of the initiatives having a bigger impact in fiscal 2023. I guess last time you were kind of in the ZIP code from a revenue perspective, OpEx was kind of in the low $800 million. I realize there's been broad-based inflation, but how should we think about OpEx in the early part of fiscal 2023 as you guys flex down?
Yeah. You know, you can back into OpEx long term as a percent of sales through the model. I won't give you a number for OpEx in 2023 because, you know, we haven't done our plans. We'll give guidance on that at a future call. I think the comments today were meant to reflect that, you know, it is an environment that's changed very quickly and the demand environment's weakened, and we've responded quickly across all major areas of spend. We are, you know, the OpEx number is down from what it was in our original forecast here. That gives you a sense that we're already dealing with it.
Our expectation is that as the market outlook becomes clear, we will, you know, adjust our OpEx spend appropriately. Now, you know, the company's in great position of leadership across technology products and manufacturing. This is not a, you know, take a sledgehammer to, you know, OpEx. It's a scalpel, you know, scalpel exercise, and we work to get OpEx down and for an appropriate level for the conditions of the business.
Mark, sorry, just to be clear, the sequential increase in OpEx this quarter, that's just timing of quals, as you guys noted on the call.
Yeah. We had in the third quarter, the OpEx was low. It was actually about $100 million off of the guidance. Two major drivers. One, product and technology spend, as you talked about. That was actually not due to sort of fewer programs. It was due to actually very good performance and fewer wafers required for technology qualification. That was the largest part of the variance. The next was just, as you can imagine, our forecast came way down, so the incentive comp was adjusted. Since we're this far in the year, it was a sizable you know, accrual adjustment. Those two things were drivers. In the fourth quarter, of course, those benefits you know, become headwinds for the fourth quarter.
We're kind of back up to this level, you know, just over $1 billion, and sort of makes it look like more of a sequential decline than in reality it is. We'll just continue to work it in a disciplined way from here.
Great. Thank you.
Thank you. Our next question comes from CJ Muse of Evercore. Your line is open.
Yeah, I wanted to just follow up on prior question around cost downs and whether you're contemplating any material kind of change to utilization rates and how we should think about that impacting cost downs, you know, in the coming quarters.
Hi, CJ. So we are evaluating utilization. I think as I said, we are evaluating utilization. I would say that, you know, we're really, you know, with having technology leadership, you know, we definitely believe that, you know, the inventory at the leading edge that we're generating is very cost effective relative to the rest of the industry, relative to even what will be available for us in future quarters. So, you know, we're really, you know, on the utilization front, focused on thinking about optimizing older nodes, legacy nodes.
You know, of course, keeping in mind the legacy demand that we have for certain customers as well. But it's not as if I think we're at a position where on the leading edge we're gonna be making, you know, significant action or drastic action. I, as I mentioned before, when previously when I think Joe or Tom had asked about this. You know, we're still planning on ramping our 1β and 232-layer technologies later this year. We think it's very important that we maintain the technology cadence and the, you know, the learning and the, you know, drive the adoption of those technologies. That'll happen in fiscal year 2023 as well.
Very helpful. I guess as a just quick follow-up, are there any targets you have for inventory that you can share with us and in what kind of time frame?
It's CJ, the only target that, you know, we can provide is the previous level of inventory target. I guess maybe you'd call it. You know, I think we'd like to keep it below 150 days. We do think it'll go up in the fourth quarter on DIO. I think I gave a number of a couple of weeks. We'll just have to see where it goes from there. That's a function of, you know, market conditions, as you know.
Thanks so much.
Thank you. Our next question comes from Srini Pajjuri of SMBC Nikko. Your line is open.
Thank you. A couple from me as well. First on the PC and smartphone weakness. You know, you talked about inventory correction. Just curious, I mean, do you think it's a one-quarter event? Meaning, do you think the inventories, you know, customers will normalize, you know, after this quarter? Or do you see, you know, further, you know, potential for kind of a correction going forward if assuming that, you know, demand kind of stabilizes here? Then I have a follow-up.
Typically, we see that, in any given segment when customers start to correct inventory, it is a couple of quarters worth of a process. It takes several months to get down to their own internal target levels of inventory. That, of course, assumes a predictable macroeconomic environment and demand trends in their own segment. Obviously, you know, there is concern about the trajectory of the macroeconomic environment, so that can further create another vector of uncertainty that has to be dealt with. Typically, it takes a couple of quarters for the inventory to get normalized.
We're talking both PC and smartphones, Sumit, that started this quarter, so we're talking two quarters basically from current quarter.
Yeah. I mean, that's the reason we said that, you know, we see this playing out sort of in the second half of calendar 2022 as the biggest portion of the impact of the lowered sell-through rates in the end markets in those segments as well as the inventory correction.
Okay. Got it. On your own balance sheet inventory, Mark, you know, you talked about $150 is kind of, you know, where, you know, you'll start getting uncomfortable. If I look at your historic inventory, I think it peaked out around 130s, you know, mid-130s. I'm just curious, what changed, you know, in the last, I guess, you know, year or two that, you know, you are more comfortable even at $150, I guess? Why has that level gone higher? The other side of the question is that, you know, you did say that, you know, you want kind of, you know, you're going to maintain pricing discipline and also use your inventory to kind of, you know, supply into the market next year.
I guess at some point, the inventory will come down as you lower your production, you know, growth. What is the minimum level that you're comfortable? I mean, I'm looking at, I guess, you know, a few quarters ago, you're in the low 90s. I'm just curious, you know, what is the minimum level that you're comfortable? When do you have to ramp up, you know, production again? At what level? Thank you.
Yeah. There was no news today on the call, actually. It was more a validation of ranges that have been set before. I did comment on, you know, there are some headwinds in the sense of on DIO around more complex processing and, you know, that the cycle time in the fab would be longer and more WIP. There's more complex modules, and there's the associated components with those and assembly times and so forth. I think all that is a headwind, but at the same time, you know, the team's doing a lot of things to offset those. I was validating that, you know, the number that we've provided before of about 150 is a maximum level.
Normal levels have also been talked about before, around 100 days. The number that you cited, low 90s, is, you know, a level that we would view as, you know, good lean inventory levels for our business.
Got it. Thank you.
I'll just add one comment there, just Srini's comment. These things always are the trend in the industry, and it doesn't matter whether you look at the logic in this logic side or you look at memory. The trend is longer and longer processing time and cycle time in the fabs, right? That's more fab WIP. For us specifically then, you add in the fact that we're transforming more of our business and, in general, more industry demand moving towards high-value solutions like SSDs, like data center DRAM modules. These things have longer cycle time in the back end, as well as more component inventory required to be held for the back end as well.
When you think about all of those things, all of these, whether you're talking about the lean inventory range or the max comfortable inventory range, those are things that, you know, generally across this industry, as technology becomes more complex and products become more complex, are gonna continue to keep going up, you know, over time. The other thing just to keep in mind, and we talked about this at our Investor Day, is one of the trends in the industry over the last several years now, again, speaking about memory, that's helpful to allow us to hold more inventory is the cost declines are lower now than they were a decade ago, from technology transition.
You know, that's another one of the areas where we're able to hold more inventory because, you know, we have very good cost-effective inventory, especially being at the leading edge now of the industry, and we expect that inventory to continue to be cost-effective through this next period while we're working down our inventories. Very helpful. Thank you.
Thank you. Our next question comes from Steven Fox of Fox Advisors. Please go ahead.
Thanks. Good afternoon. Two questions. First, on the gross margin guidance for the quarter, I think you're guiding down about 500 basis points roughly quarter-over-quarter. Can you sort of break that out between, I assume a bulk of that is just sheer volume, but anything you can provide color on how much is related to mix, pricing, et cetera? I had a follow-up.
Yeah. I won't go into the details, but the sequential decline third to fourth is just driven by market trends or, yeah, dynamics in the market. You know, the second quarter to third quarter was driven by a higher and higher mix in NAND, that decline.
Got it.
Third quarter, fourth quarter.
In terms of.
Go ahead.
Third quarter from second quarter. Yeah, I got that. Just in terms of-
No. I would just. I'm sorry, just one thing to add. You know, beyond the fourth quarter, you know, as we've talked about, as these inventories are elevated and the market works through this period, we would expect that to weigh on gross margins. You know, I just wanna make sure that that's clear.
That's helpful. Just on the long-term agreements, I mean, I understand, you know, this is unusual times, but, I mean, you just spent the last several quarters putting some long-term agreements into place, and they seem, as soon as things got tough, to unraveled a bit. I guess it sounded like you were kind of describing some benefits that are still coming through from the long-term agreements, even as demand slows here. I'm just not clear on what those are. So, can you maybe talk to the, I guess, that comment and the question I just put out there? Thanks.
Sure. In terms of the long-term agreements, you know, we have been working on these for several years. Used to be just over 10% of our revenue under long-term agreements five years ago, and it's now 75% of our revenue under long-term agreements. Substantially all of our large customers are under long-term agreements in terms of purchases. I have mentioned this in the past, that these agreements were never structured to be take or pay agreements, that our customers would buy the volume come hell or high water.
It's more like a very strong, you know, good level of, planning between our customers and us to ensure that we are driving alignment between their demand, our supply, allocation of bits to different customers, different segments, and the mechanism to drive shared dynamics in different customers and different segments. Our goal with these agreements is really to, you know, use these as a mechanism to, drive the portfolio mix to where we want it to be. We have mentioned in the past that our goal is to keep our supply bit share consistent over time in both DRAM and NAND, and strive for a bigger and bigger portion of the industry profit pool through, moves in our product portfolio and mix.
That's where we have had tremendous success, not just in improving our product portfolio, but also using these agreements to drive our product portfolio mix into a place which gives us better profitability and improved stability. One example is that in the auto, industrial, networking, graphics type of segments, the margin stability through the cycle is much better compared to other segments of the market. Our goal has been to drive higher share in those segments, and these long-term agreements become critical over time in driving that share. Now, that does not mean that, you know, we don't work with customers to have them purchase inventory that is consistent with these long-term agreements. Of course, we do.
I think to be fair to our customers as well, when these kinds of significant macro events occur, it's not like, you know, that is a feasible outcome. I think we use these in a way that meets our overall company goals in a very effective manner, recognizing what the, you know, limitations are of these agreements and what they are meant to do and what they're not meant to do. Hopefully that helps.
Yeah, that's great color. Thank you so much.
Thank you. Our next question comes from Vijay Rakesh of Mizuho. Your line is open.
Yeah. Hi, guys. Just a quick question here. I know you mentioned inventories are high on PC handset, and I think, people are seeing that. But I was wondering if you could give us some way of quantifying that. What are the inventory levels on the PC handset and the, you know, on the cloud side, now and versus, you know, what are normal levels?
Yeah. I think in terms of PC and smartphones, and then, you know, you look at cloud. I think it's a generalized comment across the industry that as the industry semiconductor industry, and this goes beyond memory and storage. You know, as the industry went through a significant shortage in semiconductors, definitely that was one driver of customers wanting to have higher levels of inventory. Another was all of the geopolitical risks, in terms of impact to shipments. Then the last piece was shutdowns happening in different parts of the world due to COVID-related restrictions being put by individual local and federal governments around the world.
All of these conspired to cause customers to feel like they need to have better level of inventory and customers who couldn't get to those inventory levels, and one good example of that is automotive companies. You can see the, you know, significant impact on revenue that those segments and those companies have suffered as a result of not having adequate inventory. I think across several segments, you know, there has been higher level of inventory than what existed pre-COVID. You know, the question is, how will that trajectory of inventory in terms of weeks of sales change over time, especially as the industry gets to a phase, the whole semiconductor industry gets to a phase where fewer and fewer components are in short supply.
I think that part is, you know, different customers in different parts of the market will make their own determination on these things. Coming specifically to the PC and smartphone portion of the business, the volumes in each of these in terms of TAM for 2022 are down roughly 10% from the expectations for 2022 at the start of this calendar year. With that degradation in sell-through expectations, obviously the inventory levels consequently from a weeks of sales perspective appears even bigger, and there is a need for that adjustment.
I think those were some of the catalysts that created that impact on our Q4 and what we have projected for, you know, the rest of the calendar year, this year, in terms of when we expect the inventory correction to sort of play out in those segments.
Got it. It looks like you're talking more to a 10% discrepancy, but are you seeing a higher inventory level on the DRAM side or the NAND side? If you can give a little bit of color around that, and that's it. Thank you.
You know, just for clarity, right, that 10% was a number, it was the number of units that are lower in the TAM expectation for calendar 2022 for PC and smartphone global unit sales versus expectations from six months ago. The impact on our revenue, of course, is not just on the reduction in end demand sell-through, but also on top of that, the inventory reduction, right? I just wanted to clarify that the inventory reduction is on top of the reduction on the end market weakness.
Got it.
I just wanted to make sure that it's not just that 10%.
Yeah.
Yeah, 10% is for the full year. Second, the impact to us in the second half is more, as-
Yeah.
He said in the call.
Yeah, that's a very good, very important point that for us, most of the impact of that full year is going to occur in the second half of the calendar year.
Got it. Thank you.
Thank you. Our next question comes from Sidney Ho of Deutsche Bank. Your line is open.
Hey, thanks. In terms of the demand side of things, you talk a lot about PC and smartphones, but you also make a comment that enterprise OEMs are adjusting the memory and storage inventory because of shortages in memory components and macro. I want to see if you can give us a little more color. Any precedents that you can point to to see how long this correction could play out and maybe the magnitude of those kind of adjustments in the past? And I'll have a follow-up question. Thanks.
Yeah, I think the issue on that side is different than the issue on the PC and smartphone side. Whereas in PC and smartphones, the end demand trends are clearly weak, and the sell-through is clearly weak, and there has been a substantial degradation in expectations for this year's TAM from six months ago to now. By contrast, the server business and the trends of end demand, even in enterprise server on our OEM customers in that space, are healthy. Our customers continue to report a robust backlog and good server demand. I think the concern is that you know, some of them are obviously data center generally across cloud and enterprise and including small and medium businesses.
The builds of servers are not being able to be completed to the extent of the end demand because of shortages of certain parts that are still persisting, including, for example, network interface cards as one example, and then there are others. So that's one factor that's preventing them from completing builds and meeting end demand. The other factor is just the general cautiousness that is creeping in, driven by some of these macroeconomic factors. I just, you know, we are trying to just be extremely transparent and lay out what we see and what we hear very openly, because, you know, that caution is happening, causing them to reduce some of their inventories or wanting to reduce their inventories in spite of, at least today, what is clearly robust demand on the end customer side.
Okay. That's helpful. Maybe a follow-up question is, Mark, I want to elaborate on a comment earlier that you said beyond fiscal fourth quarter that the planned inventory build will weigh on gross margin. I thought it would actually help your gross margin because you keep your utilization higher than the current demand indicates. The other question I have, is there a framework that we can use to help understand where gross margin is gonna bottom, and when is any kind of inflection point we can think about? Obviously, pricing is a big factor here, but are there things that are within your control, whether it's technology transitions or maybe utilization starts to improve? Thanks.
There's no comment about the, you know, we're not guiding the quarter. Mine was just a market commentary that, you know, the market's gotten weak, and, you know, these conditions of the market will continue to weigh on margins. That was the only comment. It's not a guide, and I'm certainly not gonna comment on anything beyond that.
It was more a comment on inventory in the market, the customer inventory adjustments and that environment continuing rather than anything about our inventory.
Yes.
Okay. That's clear. All right. Thanks.
Thank you. Our next question comes from Rajvi Gill of Needham & Company. Your question, please.
Yes, thank you. This question is kind of a follow-up on the sequential decline. If you look at the August decline, it's 17% down sequentially. That's the biggest decline sequentially for August in kind of recent company history. You talked about, you know, an abrupt change in demand, a considerable change in demand. I just wanted to see if you could compare and contrast what you're seeing with this level of abruptness in demand versus, say, when, you know, revenue declined pretty significantly back in kind of February of 2019 or May of 2019 when we were in the middle of the trade war with China and there were sanctions against China.
Is it similar in terms of certain dynamics, in terms of a significant cut in inventory that you're seeing from your customers? Is it more concentrated? It seems to be more concentrated in mobile and PC versus, say, last time, but I'm just curious how you're thinking about that. Because we've had a few of these very abrupt macro impacts to demand.
Yeah. I mean, I think if we look at the environment, and I understand the desire to use a framework from past declines, but if you look at the 2019 environment, that was the time that we had been seeing some weakness in the demand trend on the backs of seven or eight quarters of extraordinarily high increases in DRAM pricing that had caused, you know, DRAM margins to become unsustainably high and the pricing to become unsustainably high due to significant shortage in the market. I think that environment was entirely different, and the motivations of customers, and the actions of the customers was also pretty different. You know, each environment brings with it its own sets of unique issues.
I think the way to think about the FQ4 number, yes, I mean, definitely, the environment worsened quickly late in our FQ3. You know, we have had primarily the consumer portions of the market, the PC, smartphone, and consumer segments, channel, et cetera, be the ones that are driving a lot of this weakness, especially in our FQ4. I will call out that within that context, you know, we spoke about in the Q&A earlier today, the impact of China has been pretty significant. The weakness in China is very pronounced. We spoke about the trajectory we were on, that we believe we were on a quarter ago, for FQ4 versus our current guide for China revenue.
Our China revenue is down 30% approximately from that trajectory to our latest forecast. That itself has, you know, impacted our consolidated revenue by approximately 10%. Very significant impact coming from China. Of course, you know, we mentioned some inventory correction here and there in the other segments that we mentioned in our prepared remarks within the price OEM forward. That's sort of the big picture that we just wanted to highlight. Now, like you said, every environment is different, and how this plays out, you know, I'm sure you'll be watching for trends when other companies report to try and triangulate, you know, where this is going.
Yeah, no, that's super helpful. Just to follow up on that, sizing up the China impact was important to understand that. Any kinda, you know, fresh data points or insights now that China has, you know, reversed the lockdowns, because this appears to be primarily, you know, a China lockdown situation combined with some of the other factors you talked about. If the China lockdowns reverse, are there any indications that that region is stabilizing or is it still to be determined?
I think the recent data points from the last month on, you know, smartphone sales in China still continue to be pretty weak. A lot of the consumer spending, you know, remains to be seen how it actually starts to improve towards electronics and those type of items versus, you know, pent-up demand for travel and things like that. It's not just about the overall consumer spending, it's also about the categories of those spending where consumers are choosing to shift their dollars. You know, a lot of inflation out there, just like everywhere else. I think there is a lot of that. Also just keep in mind that the China economy and its weakness, you know, was a lot more pronounced than most parts of the world, and did proliferate beyond just the consumer segment of the market, right?
The weakness in China was a broader economic weakness when the shutdown just shut down all economic activity, right? Much of the economic activity gets shut down. Even though everyone sort of intellectually knows and understands that that is happening, you know, this is probably one of the first few data points to understand the magnitude of some of these impacts.
Thank you.
Thank you. Our next question comes from Nik Todorov of Longbow Research. Your line is open.
Yeah, thanks for taking the question. Two questions on inventory, and then Sumit, you touched on one of them. When you speak to your customers in those conversations, where are they trying to land in terms of inventory on their shelves? Do you anticipate the inventory downstream to normalize to pre-COVID levels, or you think there's gonna be a new normal?
We think that the inventory will normalize to a level that is higher than where the pre-COVID levels were. I don't think enough distance has been put between the shortages and you know, the experiences and memories are still very raw of that time. I would be surprised if the inventories went all the way down to pre-COVID levels. There is still uncertainty around, for example, China's zero-COVID policies and what that means in case more COVID cases come up in different parts of China. Those kind of uncertainties will probably cause customers to have even after they rationalize the inventory, a higher level of stable base, at least in the near term to medium term, compared to pre-COVID levels at a little bit higher levels.
You know, how all of this transpires, I mean, this is just something we'll have to see. We're working very closely with customers. As things improve, you know, sell-through improve over time, you know, China may decide to stimulate the economy in a significant way and, you know, that can improve end demand. Those kind of trends then can quickly eat up the inventory once the sell-through improves, right? These are all very dynamic numbers. They have their individual weeks of sales targets, but then the absolute value of the inventory is heavily determined by the trajectory of the future sales expectations. That's really where, you know, a change in that trajectory, improvement in expectations of future sales can quickly cause things to improve.
Again, you know, there is the environment of the macroeconomic weakness that everyone's focused on, so it will take some time.
Yeah, makes sense. Second question is, can you talk about the inventory levels between DRAM and NAND, maybe both internally in the channel and customers? Are there any meaningful differences between the two?
I mean, our inventory levels are higher in NAND than in DRAM. I think if you look at the impact, in terms of the TAM on all of these changes in unit forecast for PC, smartphones, consumer markets, obviously that's being felt in both DRAM and NAND markets. Internally, our own inventory is lower in the DRAM segment.
Okay. Customers, is it fairly similar or is it, you think that there's more NAND again?
I can't see too many differences between DRAM and NAND. When customers go into this inventory reduction mode, they tend to reduce all semiconductor inventory of parts that they feel are readily available. When, you know, different parts, even beyond memory and storage, are not in shortage, across the semiconductor space, you know, they tend to go up and down sort of in a similar trajectory. Now, the inventory of products at customers may be at different levels. Different customers may have different levels of inventory of different semiconductor products. Generally, when they go into inventory reduction mode, it's pretty broad-based approach.
Got it. Thanks for the answers.
Sure.
Our next question comes from Timothy Arcuri of UBS. Please go ahead.
Hi, thanks. I just wanted to go back to this issue of LTAs. I guess I'm a little surprised that customers would be able to reduce demand this quickly, unless you're getting something out of it. I mean, I certainly understand that it provides you visibility and it helps you plan CapEx, but is there something that you're in a better position today versus if the same thing happened, you know, three or so years ago when you had less on LTA? Are you selling at a higher price maybe in August than you otherwise would have if you didn't have these LTAs? I'm just trying to.
You know, because people will say, "Well, what good are these LTAs if, you know, customers can, you know, stop buying to, you know, to this magnitude?" I had a second question too. Thanks.
Sure. Yeah. I mean, I think, you know, if you look at customers, in the past, you know, going back five years, there would be situations where customers would, not be very clear or specific as to how much they intend to purchase from each of the suppliers. The end result could, in many cases, be that the amount of bits that these suppliers are targeting to produce for any given customer is 10 or 120% of their customer's TAM. You know, when there are these LTAs that are used pretty pervasively in the industry now, it definitely helps to rationalize and bring supply and demand into a better balance from a planning perspective. The reason I say from a planning perspective is, of course, actuals can deviate from plans when these kind of macro events occur.
When they occur and, you know, these LTAs provide us with a very good basis to better understand how our customers are thinking, not just about the next quarter, but about, you know, multiple quarters, it just gives us a mechanism to have a dialogue, for longer visibility into, you know, how they expect things to proceed. As you can expect, you know, when we think about our own supply, we think about making changes to our CapEx plans, our supply plans. Any changes that we make have an impact several months out, right? I mean, these cycle times are pretty long. Consequently, you know, we need those, longer views from our customers in order to, understand exactly how best to create a better balance between supply and demand in the industry.
I understand your concern that, okay, if they're not going to buy based on the LTA volume, what good are these LTAs? You know, we have repeatedly said that these LTAs are never meant to be take-or-pay agreements. That was not the goal for which they were set up. There is no pricing in these agreements either. They're more of volume-based agreement, and the price is negotiated every quarter or every month, depending on, you know, what kind of customer arrangements there are.
You know, I think the way to think about it is it's a very useful mechanism to drive alignment between supply and demand over the medium term, to drive our portfolio direction, share gains, and things of that nature, and reduce you know an apparent level of customer and suppliers targeting 110-120% of a customer's TAM because the customer will have LTAs with pretty much all the suppliers. More difficult to do that in a normal environment.
Can I just follow up on that? In this case, if a PC and a smartphone customer is taking bits down by, I mean, it has to be at least 40%-50% sequentially. If they're doing that, they pay the exact same price that they would have if they took what they said they would. There's, you know, not even like a little bit of a price kicker you get if they, you know, bits are down 40%-50% sequentially for a customer?
Well, there are things that happen on the pricing front, which I can't get into. Those are all things that we discuss with customers based on, you know, what their needs are and how it deviates from the LTA and so on. What exactly do we do with these customers when those kind of situations occur, I'm not at liberty to discuss in this call. You know, clearly the intent is that if there are things that are happening that are outside the control of our own customers, you know, we work with them to come to a resolution that both sides can feel good about. That's what we do. The details of that and how it changes with price, et cetera, is not something I can address here.
Okay. Thank you.
Thank you. Ladies and gentlemen, this does conclude today's conference call. Thank you for participating. You may now disconnect.