Lam Research Corporation (LRCX) Q4 2023 Earnings Call
Lam Research Corporation (NASDAQ:LRCX) Q4 2023 Earnings Call Transcript July 26, 2023 5:00 PM ET
Company Participants
Tina Correia - VP, IR and Corporate Finance
Tim Archer - President and CEO
Doug Bettinger - EVP and CFO
Conference Call Participants
Harlan Sur - JPMorgan
CJ Muse - Evercore
Krish Sankar - TD Cowen
Tim Arcuri - UBS
Brian Chin - Stifel
Toshiya Hari - Goldman Sachs
Atif Malik - Citi
Joe Moore - Morgan Stanley
Vivek Arya - Bank of America
Stacy Rasgon - Bernstein Research
Blayne Curtis - Barclays
Vijay Rakesh - Mizuho
Operator
Good day, everyone, and welcome to the Lam Research Corporation's June 2023 Quarterly Financial Conference Call. All participants will be in a listen-only mode. [Operator Instructions] Please also note today's event is being recorded.
At this time, I'd like to turn the floor over to Tina Correia. Ma'am, please go ahead.
Tina Correia
Thank you, and good afternoon, everyone. Welcome to the Lam Research quarterly earnings conference call. With me today are Tim Archer, President and Chief Executive Officer, and Doug Bettinger, Executive Vice President and Chief Financial Officer. During today's call, we will share our overview on the business environment, and we'll review our financial results for the June 2023 quarter and our outlook for the September 2023 quarter. The press release detailing our financial results was distributed a little after 1 o'clock PM Pacific Time this afternoon. The release can also be found on the Investor Relations section of the company's website along with the presentation slides that accompany today's call.
Today's presentation and Q&A include forward-looking statements, that are subject to risks and uncertainties reflected in the risk factors disclosed in our SEC public filings. Please see accompanying slides in the presentation for additional information. Today's discussion of our financial results will be presented on a non-GAAP financial basis unless otherwise specified. A detailed reconciliation between GAAP and non-GAAP results can be found in the accompanying slides in the presentation. This call is scheduled to last until 3 o'clock PM Pacific Time. A replay of this will be made available later this afternoon on our website. And with that, I'll hand the call over to Tim.
Tim Archer
Thank you, Tina, and welcome to everyone joining the call today. Lam posted strong results for the June quarter with gross margin, operating margin and earnings per share well above the guided ranges. CSBG at 47% of revenues in the quarter, continues to stand-out as a key area of strength and stability for Lam in an otherwise weak wafer fabrication equipment spending environment.
Operationally, we achieved significant improvement in on-time delivery and critical backorder performance in the quarter and now see these metrics back at normalized pre-pandemic levels. Our focus on the resiliency of our global manufacturing and supply chain network is delivering greater predictability in the short term, but more importantly, positions us to scale more efficiently when WFE growth inevitably resumes.
Looking near-term, we see WFE spending in 2023 tracking to the mid $70 billion range with the upside coming from domestic China related spending as well as strong growth in high bandwidth memory or HBM related demand. By device segments, we expect overall memory WFE to be down in the mid-40% range compared to last year and non-memory segments to be down approximately 10%. We continue to see second half 2023 WFE tracking higher than first half.
While 2023 is a down year for WFE, the long-term growth dynamics for the semiconductor industry are strong. Emerging growth drivers such as generative AI are only in their initial stages of adoption and will be fundamental to driving increased investment in both memory and foundry logic fabs over the next several years.
Advanced AI servers have significantly higher leading edge logic memory and storage content versus traditional servers. And every incremental 1% penetration of AI servers and data centers is expected to drive $1 billion to $1.5 billion of additional WFE investment. This creates tremendous opportunity for Lam as greater etch and deposition intensity is needed to enable higher performance and more scalable device architectures.
To ensure we are best positioned to win long term, we have been making significant investments to broaden our product portfolio for processing at the atomic scale. Lam is the established leader in 3D NAND and we are well positioned to benefit from the move underway to 3D in other device segments. Gate all around, 3D DRAM and advanced packaging are important 3D inflections that are expected to drive strong SAM and share growth for Lam.
Advanced packaging is becoming vital to the performance, power and cost roadmaps of high-performance applications, including generative AI. Customers are increasingly adopting a wide variety of packaging schemes to enable logic and memory integration. Some of these schemes enable up to 50% improved memory density, 10 times improvement in bandwidth and 60% gain in power efficiency. Lam has a track record of excellence in the advanced packaging segment with a strong set of manufacturing proven products.
Overall, we have greater than 50% market share in deposition and edge solutions required for advanced 3D stacking of high bandwidth memory. More specifically, our SABRE 3D copper electroplating tool and Syndion etch system hold 100% market share across all leading memory customers for through-silicon via formation as a result of our long established expertise in etching and filling high aspect ratio geometry structures.
Overall, we expect our packaging SAM to double in the next five years. We also see our market position strengthening as we bring technology and productivity innovation to address emerging opportunities. For example, in the June quarter, we had a critical win for a new inter-die gapfill application at a key foundry-logic customer for their chiplet architecture. As this latest win ramps into production, we will have secured a leading share position for this application across the top three foundry-logic customers.
We won this application for a couple of reasons. First, we delivered higher productivity relative to the competition by leveraging our unique triple quad platform architecture and multi-station sequential deposition chamber design. This enables film deposition of greater than 20 microns in a single pass, allowing the customer to run more wafers between chamber clean steps. Second, we delivered superior on-wafer performance, including better film stress management, improved defectivity, and enhanced wafer to wafter uniformity versus the competition.
As we expand our position in newer high-growth markets like 3D packaging and specialty technologies, we have looked to leverage existing R&D and tools within the Lam portfolio to broaden our product offerings most efficiently. In the recent example, our long-established Kiyo etch platform has proven to be highly suited to applications which are critical to enabling new die-to-wafer hybrid bonding schemes.
By delivering better etch profile than the competition, we secured a key win in the June quarter at a leading foundry-logic customer. We are currently the tool of record for a suite of etches and resist strip steps at this customer and expect to start recognizing revenue for the new application in 2023. As this customer continues to shrink packaging dimensions with future hybrid bonding iterations, we have the opportunity to double our revenue with them over the next several years.
In another instance, Lam has been working with customers within the specialty segment to port key 300 millimeter etch solutions, including atomic layer etch, to 200 millimeter to overcome challenges in the manufacturing of gallium nitride devices. Adoption of GaN technology is accelerating across multiple applications ranging from high-efficiency charging devices for consumer electronics and automotive to 5G RF infrastructure. However, the fabrication of such devices is complex and requires ultra-low damage etch processes with atomic-scale precision.
With our suite of solutions, we can deliver a GaN etch process that improve surface roughness and other material properties that impact device performance. These solutions were developed as a result of Lam's deep understanding of the required technology, as well as key partnerships with customers and research institutions.
Longer term, we look to outperform in the trillion-dollar semiconductor industry forecasted for the end of the decade. At Lam, we are executing a strategy to create competitive differentiation by focusing on what we see as three key vectors of rising complexity. First is the technical complexity associated with enabling the transition to 3D device and packaging architectures across all market segments. Second is the support and workforce complexity arising from the expanding geographic footprint, as regionalization efforts build momentum. And third is the sustainability complexity as we responsibly manage the carbon impact of the semiconductor industry as the output grows to $1 trillion.
This quarter, we detailed one component of our strategy in a press release for our Semiverse Solutions portfolio of advanced simulation and modeling products. These products are designed to help engineers get to process solutions faster, develop new products at lower cost and collaborate across the global ecosystem with greater effectiveness. We believe that by combining the physical advantages of Lam's diverse global R&D and manufacturing footprint with the virtual development and digital twin capabilities of our Semiverse Solutions, we will be in an excellent position to innovate and outperform as our industry grows into the future.
With that, I will turn it over to Doug. Thank you.
Doug Bettinger
Excellent. Thank you, Tim. Good afternoon, everyone, and thank you for joining our call today during what I know is a very busy earnings season.
We executed well in the June 2023 quarter with revenue coming in above the midpoint of our guided range and our profitability metrics exceeding the high end of guidance. While we continue to work through a challenging WFE investment year, most notably with very low memory spending, our focus on improving our operational efficiencies is showing up favorably in our results. We're laser-focused on what's in our direct control and executing well.
Let me dig into revenue. June quarter revenue came in at $3.2 billion, a decrease from the prior quarter as we expected. Systems revenue declined quarter-over-quarter as we had reduced levels of deferred revenue with the improvement in supply chain constraints. We closed the June quarter with $1.8 million in deferred revenue on the balance sheet, which was a decrease of $165 million sequentially. We've got all the deferred revenue related to outstanding back order parts completely back to normal levels. As we discussed last quarter though, our deferred revenue balance is currently still at higher than historic levels.
The deferred revenue balance includes increased customer cash and advanced deposits tied to orders from newer customers. We expect a significant portion of these deposits to convert to revenue during the second half of calendar year 2023.
Let me now turn to the revenue segment details. Memory as a percentage of systems revenue in the June quarter was low at 27%, which was a decline from the prior quarter level of 32% and our lowest concentration percentage for this segment in the last decade. The DRAM segment within memory remained flat at 9% of systems revenue. And NAND came down to 18% of systems revenue, down from the March quarter level of 23%. We continue to expect NAND spending to remain at low levels for the remainder of the 2023 calendar year. And I just mentioned, NAND spending is at the lowest levels we've seen since the advent of the 3D NAND architecture.
Consistent with the prior quarter, we see strong concentration of systems revenue in the foundry segment with the June quarter revenue percentage at 47% versus 46% that we saw in the March quarter. Investments were biased towards leading-edge devices. There also continues to be robust spending to support more mature specialty nodes.
We had a record level of concentration in the logic and other segment with 26% of systems revenue in the June quarter, compared with 22% in the prior quarter. There was broad-based spending in areas such as advanced packaging, microprocessors, analog, image sensors and power applications. These are areas where we've demonstrated strong momentum with our technology solutions, and I'd just point out that they are broadly distributed geographically.
With respect to the regional composition of our total revenue, the China region was 26% of the total, which was up a little bit from the prior quarter, which was 22%. China domestic customers were the majority of the China regional revenue in the June quarter. The result was strong concentration of investments for our customers in the Korea and Taiwan region, which comprised 24% and 20% of our total revenues, respectively, in the June quarter. The US region declined from the record level that we saw last quarter, largely due to the timing of customer projects.
The Customer Support Business Group revenue in the June quarter totaled approximately $1.5 billion, which was a decrease of 7% from the prior quarter level and 8% lower than the June quarter in calendar 2022. CSBG represented 47% of our June quarter revenue. This is a historically high concentration percentage, driven largely by continued strength in the specialty node investments which are serviced by Reliant systems business. The Reliant and Spares businesses continue to represent the largest individual portions of CSBG revenues. Both our spares and service businesses had been negatively impacted by low fab utilization levels, particularly at our memory customers.
On the profitability side of things, our June quarter gross margin came in at 45.7%, above our guided range and up from 44% that we saw in the March quarter. The quarter-on-quarter increase was related to cost and efficiency improvements, as well as a favorable product mix. We continue to expect to make further progress on our operational execution as we go forward.
Operating expenses for June came in at $590 million, which was down from the $608 million in the March quarter, although it was a little bit higher than our original estimates coming into the quarter. Investing in research and development continues to be a top priority for Lam, with over two-thirds of our operating expense concentrated in R&D. We're focused on extending our product differentiation and ensuring that our competitiveness remains very strong.
The June quarter operating margin came in at 27.3%, which was over the guidance range largely because of our strong gross margin performance. The non-GAAP tax rate for the quarter came in low at 7.5% due to a more favorable jurisdictional mix of income and provision true-ups that occurred at the end of our fiscal year. We estimate the tax rate for the remainder of the 2023 calendar year will be in the low to mid-teens level. And as always, this rate will have some fluctuation quarter to quarter.
Other income and expense for the June quarter was approximately $7 million in expense, consistent with the prior quarter amount and primarily related to our strong cash balances, as well as higher interest rates. And I'd just point out, from a return on cash standpoint, we're now realizing higher interest rates on our cash balances than we're paying on the longer-term duration debt in our capital structure. And as we've discussed in the past, OI&E is subject to market-related fluctuations that will cause some level of quarterly volatility.
On our capital return activities, we allocated approximately $906 million to open market share buyback and we paid $232 million in dividends in the June quarter. We have $3.5 billion remaining on our Board authorized share buyback plan. For the June quarter, we returned 109% of free cash flow, and we're very much in line with our long-term capital return plans of returning 75% to 100% of free cash flow.
June quarter diluted earnings per share came in at $5.98, over the high end of our guidance range. This was enabled from stronger revenue, improved gross margin and that lower tax rate. Diluted share count was 134 million shares, on track with our expectations and down from the March quarter.
Let me now pivot to the balance sheet. Cash and short-term investments, including restricted cash, totaled $5.6 billion, which was flat with the March quarter. Days sales outstanding were 80 days in the June quarter. June quarter inventory turns came down from the prior quarter level of 1.9 times to 1.5 times. While we did bring inventory down slightly during the June quarter, the balance is elevated from historic levels.
As business volumes reduce, we need to cancel a significant volume of purchase orders with our suppliers. As we work through these cancellations, we're taking more inventory than we need in the near term. This ongoing process will keep inventory higher than we'd like this year, frankly. We do expect inventory to come down though, albeit more slowly than we may have historically been able to drive it to.
Non-cash expenses for the June quarter included approximately $68 million in equity compensation, $76 million in depreciation and $14 million for amortization. Capital expenditures were $79 million, which was down from the March quarter by approximately $41 million. Spending mainly centered on product development activities and our global lab infrastructure.
We ended the June quarter with approximately 17,400 regular full-time employees, which was a decrease of approximately 1,300 people from the prior quarter. Most of this decrease is related to the restructuring actions we took earlier in the calendar year, with the timing of the headcount reduction showing up in the June quarter. Overall, we're on track with our transformation initiatives to improve our operations. And as we stand at the end of the quarter, we incurred approximately $143 million year-to-date out of an anticipated $250 million estimated range for calendar year 2023.
Let me now turn to our non-GAAP guidance for the September ‘23 quarter. We're expecting revenue of $3.4 billion, plus or minus $300 million. Gross margin of 46.5%, plus or minus 1 percentage point. This improvement in gross margin is a function of our operational efforts as well as a beneficial mix. Operating margins of 28%, plus or minus 1 percentage point. We're purposely spending a little more in R&D in the second half of 2023. And finally, earnings per share of $6.05, plus or minus $0.75, based on a declining share count of approximately 133 million shares.
So, let me summarize. The results this quarter demonstrate, I think, that we're on a solid path to strengthen our operations and our profitability profile. We're improving and optimizing all elements of the company and solidifying Lam's business foundation to ensure we're well-positioned for outperformance when WFE growth resumes. And I'll just reiterate what Tim said. We continue to see WFE a little bit second half weighted this year.
Operator, with that, that concludes my prepared remarks. Tim and I would now like to open up the call for questions.
Question-and-Answer Session
Operator
Ladies and gentlemen, at this time we’ll begin the question-and-answer session. [Operator Instructions] Our first question today comes from Harlan Sur from JPMorgan. Please go ahead with your question.
Harlan Sur
Hi, good afternoon and congratulations on the solid execution. As we progress through this weaker period for the industry, right, the activity level can be quite noisy, but there will come a period of stabilization of fundamentals and business trends, right, fundamentals being the things that you can track, like your customers' utilization and business-wise, the level of pushouts, reschedules and cancellations. So, I guess, with that, has the team seen utilizations starting to stabilize, albeit at low levels? And has the rate magnitude of shipment pushouts, rescheduling also come down to more normalized levels?
Tim Archer
Yeah. Thanks, Harlan. Yeah. I think that we've -- it probably varies by market segment, but clearly we've heard some of our customers, even just recently talking about more utilization cuts within certain device segments. Clearly, I would say it's slowing down as the majority of the cuts that were made in the first half of the year are kind of taking effect. I do feel like, and I made a comment about, in some ways, normalizing our operations is allowing the business to be more predictable. I think the same thing is happening on the end markets and customer side as well, where people are getting a sense of how the cuts that have already been made and how the capex reductions that have already been made are starting to flow through.
So I do feel like we're in that position. And just as we, at this point, ticked a little bit up on WFE from our prior guidance of low to mid-70s to mid-70s, yeah, you're starting to see little strength in certain parts of the market, HBM being one of them. And as also we mentioned, some of the domestic China spending. So, I would feel like things are getting to the point where people are able to predict the business a little bit better than, say, six months ago.
Doug Bettinger
And Harlan, this is Doug. Maybe the only thing I would add to what Tim said is we haven't really seen utilizations getting any better, quite frankly. You'll remember on the call last quarter we talked about, you'll see that from us in our spares and service components of CSBG. And as we sit here today, you're not seeing that at this point. So it's at a fairly low level and we really don't see it getting any better in the near term.
Tim Archer
And the only other thing that we have said that when you think about especially utilization improvements in the memory space, we would anticipate Lam being the first to sort of benefit from utilization turning back on, just given our exposure in that space and the impact that it's had on our spares and service business through the first half of this year.
Harlan Sur
Appreciate that. And then, I also appreciate you guys highlighting the advanced packaging opportunity. I think it's quite timely right, because you guys called out a slightly better memory WFE trends due to strong HBM DRAM demand. So you guys benefit both in your core memory silicon tools and now your advanced packaging portfolio as you outlined, right?
But just given the overall sort of recent and strong demand pull for accelerated compute and AI capabilities, I mean this is driving advanced packaging demand across CPU, GPU, networking, memory, right? And so, your customers are capacity-constrained here. So do you guys anticipate your advanced packaging segment to grow this year? And can you just give us some sense on how big this segment is for Lam?
Tim Archer
Yeah. I don't think we've -- we haven't exactly sized this business externally, but we -- it is growing and, it’s in certain areas, especially as I highlighted things like etch for TSVs, copper plating for filling of those TSVs, pretty much anything related to the building of that advanced packaging structure through etch and deposition. So it is a strong area of our business right now, especially given the relative weaknesses that’s seen in the other parts of the market.
Harlan Sur
Thank you.
Tim Archer
Thanks, Harlan.
Operator
Our next question comes from CJ Muse from Evercore. Please go ahead with your question.
CJ Muse
Yeah. Good afternoon. Thanks for taking the question. I guess first question, Doug, last quarter you talked about expectations for 100 bps improvement at least exiting the year. You've clearly blown that away. So curious how much of that in the guide for September is due to mix versus kind of greater efficiencies that you've garnered? And what is kind of the new exit rate that we should be thinking about for calendar '23?
Doug Bettinger
Yeah, CJ. Actually, I think if you talk to the leadership team at Lam, we're all pretty pleased where we've been able to drive the operational efficiencies in the Company and its gone quicker, I think, than I expected certainly. We're obviously beyond that 100 basis point improvement. Having said that, though, I would point to -- we do have a component of the September numbers that are mix related. I'm not going to, like, quantify it precisely for you, but we're doing real well operationally. I'm not quite ready to guide December for you yet. Except, I'd say I think we're pretty pleased where we're at and we're well beyond that 100 basis point improvement that we talked about, I think on that January call, if I recall, the first time we sort of talked about it.
CJ Muse
Great, thank you. I guess the second question relates to CSBG. And how are you thinking about sustainability of Reliant? And then, as you kind of contemplate the timing of utilization picking up for memory, how do you see kind of a hand-off maybe from one part of the business to the other and how we should be modeling CSBG into September and beyond?
Tim Archer
Yeah. I think that as you pointed out, I mean, obviously there is a Reliant component that's right now is quite strong. I talked about sort of at specialty technologies, I mean we're seeing still quite strong demand for our products across all of those different segments. Utilization driving spares and service to much lower levels than we've historically seen. And so, I think as people talk and worry about roll-off in the specialty technologies, maybe that does come at the time that the utilization picks up in the leading-edge and the memory fabs.
But actually at this point, I would say we haven't yet seen demand for those trailing edge technologies to really be rolling over. I mean, demand is still quite strong for us, we're focused on new applications, new tool types and it's just not something that we're ready to forecast yet, although, again, as we've said nothing grows forever, but the broadening it feels pretty good to us right now.
CJ Muse
So we should be modeling CSBG up in September?
Doug Bettinger
CJ, I guided you to $3.4 billion in revenue, you can do whatever you'd like to do with the competition. You've got the total number.
CJ Muse
Thank you, Doug and Tim. Appreciate it.
Tim Archer
Yeah.
Doug Bettinger
Thanks, CJ.
Operator
Our next question comes from Krish Sankar from TD Cowen. Please go ahead with your question.
Krish Sankar
Yeah. Hi, thanks for taking my question. And I have two questions. First one, maybe for Tim or Doug. Kind of curious, at SEMICON West a few weeks ago, Tokyo Electron spoke about a cryo etch product and I remember you had spoken about it a couple of years ago. So I'm kind of curious how to think about your high aspect ratio etch market share. And my understanding was, lot of it is geared towards 3D NAND and obviously NAND spending is at very low level. So I'm kind of curious, number one, how to think about your high aspect ratio market share and the cryo etch product, et cetera in that route? And then I have a follow-up.
Tim Archer
Okay, great. Well, thanks, Krish, and I'll take it first. I think that relative to NAND, as you pointed out, I mean it's -- spending is at very low levels, down more than -- probably more than 75% year-on-year right now. But to this point of cryo etch, and I guess I would just say that at this point, Lam is the only company with cryo etch in high volume production. So it's an area where we do have strength, I mean hundreds of chambers of experience right now.
We're the leader in high aspect ratio etch. I know that at the recent conference, competition has been out talking about a DTOR position for one pass at a memory customer for future node. But I think as probably everybody on this call knows, customers regularly engage more than one vendor when you're in the R&D stage for all these critical applications. And in this case, our production decisions are still quite a ways off. I look at Lam and our competitiveness. I mean, we take all of that experience that we have from years of leading in this space. We've developed a robust technology roadmap. We're engaged with the customer on a variety of different hardware and process innovations.
But I think most importantly, when I think about our long-term competitiveness, we are still the only vendor that has this huge installed base from which every day we're getting learning about how to take processes from the R&D stage into high-volume production. And that's ultimately what really counts for affecting market share. And so, we've said in the past, we don't win every decision, certainly not always at the DTOR stage. But I think given where we are and where we've come from, we're very confident about maintaining leadership and market share in this space.
Krish Sankar
Got it. Got it. Very helpful, Tim. And then just a follow-up on the advanced packaging. I understand you did mention the packaging SAM could double. And you spoke about the Syndion for TSV etch and SABRE 3D for metal dep. And if I remember, when I went back, looked at my notes, in 2015, you said it will be like a $200 million to $300 million opportunity. And then I think last you folks updated this in 2020 where you had said Syndion and SABRE actually doubled since then. I'm kind of curious. I understand a lot of it goes to HBM and all those things. So from that old realm of advanced packaging, how to think about like the opportunity from here? And also for AI, my understanding is some of your ALD for high k metal gate where DRAM can also be used. So if I put it altogether, is there a way to quantify where your Syndion, SABRE revenue numbers are today and how to think about it? And also, how to think about ALD for high k metal gate? Thank you.
Tim Archer
Yeah. Well, you threw out a couple of numbers. We used to talk about a couple of hundred million dollar opportunity, doubling from there, puts it in the hundreds. I mean, it's growing from that point. I mean, it's becoming a sizable business for us. We just haven't quantified specifically for those two products. I think what you can take away is we talked about our market share position. We know -- I think everybody knows HBM and 3D chip stacking, those products also play into other elements of advanced packaging. It's a very fast-growing part of our business. And so, without giving you our forecast for those markets, I think you can apply what you think about how chip stacking and 3D packaging is building momentum. You know we have 100% market share. I think just -- we're very happy to see what's transpiring in that part of the market.
Doug Bettinger
Yeah. And Krish, I would just add. You're right about -- I don't know how long back, we quantified it as few hundred million dollars, not billions of dollars, just to kind of frame it for you. But it is a very fast-growing component of the business as Tim pointed out. So, maybe that's a good way to think about it. It's something when you listen to every talk and everybody is very excited about high-bandwidth memory and its enablement of AI, we're right in the middle of all of that stuff from a packaging standpoint.
Tim Archer
Krish, what I would say is in this environment, it's the -- those are the tool sets that people are actually pulling for in terms of accelerated deliveries. And so I think that gives you some sense of demand in that market.
Krish Sankar
Got it.
Krish Sankar
Got it. Thank you, Doug. Thanks, Tim.
Operator
And our next question comes from Tim Arcuri from UBS. Please go ahead with your question.
Tim Arcuri
Thanks a lot. Tim, there's been a lot of debate that I hear on NAND WFE. And obviously, you have the highest share of all the major vendors. And you used to give this model that suggested $14 billion to $15 billion per year to drive sort of mid-30% bit growth, and there was some sensitivity around it. But since then, a lot of the producers have been basically down ticking on long-term NAND bit growth, Micron is now saying low 20s.
So if you kind of use that old model, you get to something like $11 billion to $12 billion per year to drive low 20s bit growth. I know you've talked about more steps and so that number has gone up. But that's just optically, it's not a ton higher than the $9 billion or so that we're spending this year. And so I keep on getting a ton of debate in terms of just how much NAND WFE is going to come back. So I'm not asking you for a new model, but can you sort of like handicap what's changed vis-a-vis that old model in a market where bit growth seems to be lower going forward than what [indiscernible]?
Tim Archer
Well, I think there probably is equal debate about what long-term etch growth is in the NAND space with some of the new applications coming out. But I certainly am not going to get crossways with our customers in terms of their forecast versus ours. I mean again, it's a market where we have such a strong position that ultimately -- we look to satisfy the technology roadmaps and the volume requirements of our customers, and that's kind of where we are. So right now, we know this year, NAND is extremely low.
And without giving you our number, I mean it's -- I would say even if it gets back to something close to that model, it's a lot higher spending than where we are today. I think what's happened too is for Lam, specifically, is not only have we taken -- I mean, what might be different from that model of how much we capture of that SAM is the complexity of doing multi-tier stacking has increased our opportunity through a whole bunch of other steps that we've talked about in the past in terms of backside depth for stress management and gapfill -- ALD gapfills for plug fill, carbon sacrificial plug fills, lots of different types of applications.
And so I think even an environment where you didn't do a lot better than the model, Lam would do a lot better. and we do a whole lot better than where we are today in such a low NAND spending environment. So I think we will get Tim, a new model out once we see a better view of the end demand picture, but that's the best I can do for you right now.
Tim Arcuri
Thanks, Tim. And then, Doug, I guess just as a follow-up, how does your inventory situation play out? I imagine that it's still kind of a drag on gross margin, and you said it's going to come down, you said pretty slowly as you kind of get to the back half of the year. So can you quantify how much of a drag it is as you're -- as you still have pretty high costs, I would imagine, on these parts? And then when do you think it gets back to some sort of normalized level and it's not a headwind anymore to margin? Thanks.
Doug Bettinger
Yeah, Tim, I don't really think about it right now as a drag on gross margin, honestly. It will prevent improvement in gross margin as we have to consume some of the inventory sitting there, right, to the extent that we're negotiating costs of some of the stuff down -- the stuff that we have is already on the balance sheet, right? So you've got to consume that first. So that's the right way to think about it. I don't really think of it as a drag in the near term.
And Tim, it will come down, although what I tried to say in my script was it's going to come down a little bit more slowly than perhaps you've seen us be able to drive it to in the past because we canceled a lot of purchase orders, a lot of material in that as part of negotiating through that, you end up taking perhaps a little bit more than you need in the near term, we'll consume it over time. But it's just going to create a little bit of a delay in the reduction of inventory that you've seen us be able to do in the past. So it's going to be a little bit sticky, I think, Tim, through the year. And then I think it will -- it will come down through the year, but I think it's going to be a little bit sticky. And then you'll see it improve as we get into next year.
Tim Arcuri
Thanks, Doug.
Doug Bettinger
Yeah. Thanks, Tim.
Operator
And our next question comes from Brian Chin from Stifel. Please go ahead with your question.
Brian Chin
Hi, there. Good afternoon and thanks for letting us ask a few questions. Maybe just to backtrack on the increase to WFE this year of the, let's call it, $2 billion to $3 billion increase, is that mainly situated in second half? How should we think about, I guess, timing between calendar 3Q and 4Q? And I guess how much of the upside is HBM oriented?
Doug Bettinger
Yeah, Brian, I guess what I'd say is it is a little bit second half weighted, and the things we pointed to in our scripts were little bit of upside in China, China domestic and a little bit of high bandwidth memory. I don't think we're going to put a quantification between the third quarter and the fourth quarter of the calendar year, except to say it's a little bit second half weighted.
Brian Chin
Okay. That's fair enough. Maybe for my follow-up, clearly, there's a lot of focus on high bandwidth DRAM. But Tim, I was wondering, could you interject maybe your thoughts on the role of the role that SSD and flash storage might play in generative AI? And I'm sure that's factored into sort of that 1% penetration for AI servers in kind of $1 billion to $1.5 billion WFE investment?
Tim Archer
It is. And I think that was even in my answer to the last question was, I think there's still a lot of debate as to how much storage really gets driven in this, but we're a believer that there is some element of demand driven there. I think that when we think about DRAM, there's a couple of things at play in terms of what drives it. One is, of course, the HBM also increases die size, I think, as you know. And that leads perhaps even greater WFE increase associated both with the silicon side of that as well as the packaging side. I'm not sure that same dynamic exists on the NAND side, but in terms of consumption, it feels like there should be some. But again, I'm willing to sort of let this play out a little further to see what the new model should be. And we talk closely to our customers to see what they're seeing in the marketplace as well.
Doug Bettinger
Yeah, Brian, this is Doug. I don't know if this helps. But when I look at like the composition of some of these AI servers, nominally, there's roughly eight times the DRAM versus an enterprise class server and three times demand from a storage standpoint. I don't know if that will help you get your head wrapped around it. But DRAM is clearly a little bit more enabling on the compute side, but storage picks up also.
Brian Chin
Okay. Got it. And probably scale out of these language models probably also could increase that storage potential over time. But thanks for your inputs. Appreciate that.
Doug Bettinger
Thank you, Brian.
Operator
Our next question comes from Toshiya Hari from Goldman Sachs. Please go ahead with your question.
Toshiya Hari
Hi, guys. Thanks for taking the question. Tim or Doug. Just on China, I was hoping you guys could give a little more color in terms of what you're seeing on the ground. Doug, I think you said China was 26% of revenue and a big chunk of that was domestic. What’s sort of the makeup by application, foundry, memory and other? And I guess, importantly, how should we think about sustainability going forward? It's really hard to tell from our standpoint, just given the spend is so strategic, right? It's not necessarily tied directly to near-term economics. So it's hard for us. But in terms of what you're hearing from customers, what is the feedback? And how much visibility are they giving you going forward?
Tim Archer
Yeah. I think that it's split. I mean, obviously, on the domestic China side between some memory investments as well as a lot of specialty technologies, trailing edge focus on end markets, I think we're all familiar with automotive and industrial and those types of markets. As you mentioned, it's strategic. I mean, these are customers that are giving us long-term forecasts. Doug talked last quarter about deferred revenues are impacted by some of the down payments, at least within the visibility horizon that we need, we feel we have a pretty good view of a fair bit of investment that this continues in China, especially in those specialty technology markets where new fabs are emerging to try to capture some element of those trailing edge end markets.
Toshiya Hari
That's helpful. Thank you. And then as a follow-up, maybe one for Doug on gross margin. You talked about operational efforts and initiatives a couple of times. What exactly are you doing today? I know there's no ending to things like this, but which inning are you in, in terms of reaping some of the benefits? And you used to talk about Malaysia as a headwind a year ago, 1.5 years ago, where are you with the ramp there? And how should we think about the tailwind as volumes recover going forward? Thanks.
Doug Bettinger
Yeah. I think, Toshi, the way to think about it is business is down. Right now, we're not really growing anywhere. I think as we see growth come back, whenever that is, the pivot to that Malaysia factory, which has an opportunity to do more than it's doing today, will be largely what you see from us. Now maybe to help with the restructuring activities, we embarked on plans to kind of restructure the cost footprint of the company. And maybe the best way to think about it, I've given you a forecast. We think we're going to take charges of $250 million at this point. We've taken $143 million at the end of the last quarter.
So we still have some things we're working on. Although I would tell you the workforce portion of it is complete. We're done with that aspect of it. So that's behind us, we believe. So there's just some other things you're doing around looking at product, repositioning inventory, maybe a little bit of infrastructure stuff, we are embarked upon what will be a multiyear digital transformation initiative at the company, which will deliver efficiencies over the next several years. So there's a handful of things that are still coming in the future, I guess, is what I would describe.
Toshiya Hari
Appreciate the color. Thank you.
Tim Archer
Thanks, Toshiya.
Operator
Our next question comes from Atif Malik from Citi. Please go ahead with your question.
Atif Malik
Hi. Thank you for taking my questions. I have two questions on leading-edge investments. I thought you gave an interesting data point that 1% AI server adoption leads to $1 billion to $1.5 billion incremental WFE. When I listened to the major foundry in Taiwan, they talked about 50% AI semi growth rate in the next few years but we did not really raise any CapEx this year. And we believe AI server adoption is probably like high single digit, 8%, 9%. So what explains the discrepancy in terms of not seeing a lift in leading-edge investments from the higher server adoption? Maybe it's just underutilization in other end markets.
Doug Bettinger
I guess what I think about, Atif, is these are longer-term statements we're making. I think you're right about the composition of AI servers. In the short term, you're not really going to see a meaningful uptick in WFE, frankly. This is going to occur over the next several years. As more of -- and you're right, I think server volumes largely, I don't know, mid- to, I'd call it, mid-single digit percent of total servers. I think that grows over time. But in the short term, it doesn't really show up as quickly as you might think. It requires sort of capital investments to occur for future demand. That doesn't really happen in a meaningful way in the short term.
Atif Malik
Got it, Doug. And then as a follow-up, you guys are very well leveraged to get all around, particularly on the edge side and they have been waiting for this technology inflection for the last three years. Your peer in Europe is talking about pilot line orders in December quarter this year. And Tim spoke about having better visibility versus six months ago. So my question to you is, when do you see the gate all around opportunity kind of grow meaningful for you? Is it the first half of next year or second half of next year?
Tim Archer
I think it starts in line with the timeline that you just talked about and it's first -- through the first half and second half of next year. I mean these things tend to ramp and I think we are at the point where gate all around is going to become more meaningful. We'll see that in growth in some of the traditional products that are sort of lumped in with everything else. But then there are specialized products that we talked about, like selective etch those -- that suite of tools that's very specialized for applications like gate all around. We'll see faster growth in those segments, I think, as we move through '24 and on into '25. So yeah, maybe we're at that key inflection point. How that overcomes the rest of the market dynamics in '24? We're not ready to guide '24 yet, but some of those technology trends are starting to move in our favor.
Atif Malik
Thank you.
Tim Archer
Thanks, Atif.
Operator
Our next question comes from Joe Moore from Morgan Stanley. Please go ahead with your question.
Joe Moore
Great. Thank you. I wanted to follow-up on the China question. I guess, how are you seeing the impact of last year's export controls? Is that your China business has obviously done pretty well since then given the headwinds? Have there been -- have the restrictions been what you thought they would be? Is there anything that's different? And then there continue to be rumblings of additional controls. Do you have any visibility into what that could look like?
Tim Archer
Yeah, Joe, we had talked last -- at the end -- towards the end of last year, right around the October 7 announcement that the impact to our business this year would be $2 billion to $2.5 billion. We clarified earlier this year that because we were able to make some shipments for trailing etch memory that we had not -- we weren't sure whether we could, that impact was going to be closer to the $2 billion mark. That is still what we estimate the impact of last year's restrictions, the October 7 restrictions on our business, $2 billion of revenues.
Now in the same time, as China has shifted its focus towards the areas they can invest, these trailing etch, foundry logic and some of the trailing edge memory, we've seen that upticking a bit as we've moved through this year, and that's what we just spoke about. So I can't really speak to future regulations that may or may not be contemplated. We have a team in Washington DC that's regularly engaged with the government, make sure they understand our business and the details of the semiconductor industry, and that's about the best we can do right now.
Joe Moore
Okay. Thank you very much.
Tim Archer
Thanks, Joe.
Operator
Our next question comes from Vivek Arya from Bank of America. Please go ahead with your question.
Vivek Arya
Thanks for taking my question. Tim, I'm curious, what should be the signs we should look for, for when memory utilization starts to pick up? And when orders to you guys start to improve? So if you look at historical trends, should we be waiting for like one quarter of pricing improvement, two quarters of pricing improvement? Just what are kind of the external metrics we should be keeping an eye on to try and predict when memory WFE should start to pick up?
Tim Archer
Okay, a little bit difficult, I guess, just from the standpoint that, obviously, pricing improvement in memory, we tend to start to get customers interested in greater utilization of the fab. I think easier, I was going to answer relative to what you'd see in Lam. Obviously, once we start reporting an uptick in our CSBG spares and service businesses, obviously, it's an indicator that fabs are starting to be utilized more, and therefore, consuming spares and service. That's why I mentioned a little earlier that we are expecting to be really one of the first beneficiaries of a recovery in memory because so much of the -- there have been such deep utilization cuts in the memory fabs, that the spares and service side of our business has been impacted to a greater degree than we would have expected. When they turn those back on, that revenue should come back, and that will probably be your first indications from Lam's reported numbers of what's going on.
Vivek Arya
I guess that was really the origin of the question because from what I heard on the call, I think you were suggesting there is a scenario in which CSBG revenues do start to pick up sequentially in the next quarter. But then you mentioned that the visibility around memory utilization, right, is not there from an industry perspective. So is there still a delta between the two?
Doug Bettinger
Vivek, we did not suggest CSBG is picking up in the next quarter. That was not a statement anybody on the call made. So if we said something that led you did that conclusion, maybe we misspoke or perhaps you misinterpreted something else we said.
Vivek Arya
Okay. Sorry. Maybe I misheard. Then maybe for my second question, Doug, the gross margin upside in Q3, right, it was quite sizable. I don't remember the last time you had such a strong gross margin upside. I think you mentioned something along the lines of mix. Is that mix unsustainable? Like is that likely to reverse, right, at some point? What is the right way to think about the sustainability of gross margins at these levels. So let's say if you assume that your sales continue to grow sequentially, then can gross margins continue to be at plus minus these levels? Or is there something that would dramatically change in the mix to change the trajectory of gross margins?
Doug Bettinger
Vivek, I don't know if there's anything I'd point to that is dramatic. I suggested two things. One, our operational efforts in terms of just efficiencies and whatnot, it is maybe a little further ahead of where I thought it would have been. And I did point to favorable mix in the current quarter. You always see mix up and down, both products as well as customer mix. We're benefiting a little bit in the September quarter. I don't know if you'll see that same magnitude of mix benefit in December. The operational stuff, we will. I wouldn't run too far away from where we are right now into the near term anyway. But I would tell you, I'm very pleased with what we've been able to execute and the things that are in our own control, we've done extremely well.
Vivek Arya
Thank you.
Doug Bettinger
Thank you.
Operator
Our next question comes from Stacy Rasgon from Bernstein Research. Please go ahead with your question.
Stacy Rasgon
Hi guys. Thanks for taking my questions. For my first one, you talked about the bulk of the China $1.8 billion in deferred selling through in the second half. How much of that $1.8 billion actually is coming from China? And how much of that do you actually see yourself recognizing in the second half? And is it more in Q3, Q4? Do you have any idea on the timing of that?
Doug Bettinger
Yeah. I've never said how much of it is in China. I've described it as newer customers, right? Customers that are new that we're not exactly sure of the creditworthiness, we require cash in advance before we build the tools. I haven't put a number on how much of it is China. Although a decent amount of it is. And what I said on how much that turns to revenue, as I said, the majority of it turns to revenue in the second half of the year.
Stacy Rasgon
So, like, you have $1.8 billion of revenue that will be recognized in the second half from that deferred? Like, what would be the normal amount [indiscernible]?
Tim Archer
No, Stacy, there's a normalized level of deferred revenue that's always here, always has been here, and you should expect to always be here. It's probably $750 million to $1 billion on a regular basis, I think, is what you've seen from us in the past. I wouldn't expect it to be too different than that. We've just got a lot of this cash in advance payments right now. That's why it's elevated.
Stacy Rasgon
Got it. Thank you. For my second question, so you're talking memory WP down in the mid-40s. Can you give us a feeling for how you see DRAM versus NAND breaking out in that? And whether or not your views on both of those have changed over the last 90 days versus like where we were a quarter ago?
Tim Archer
Yeah, we've said that NAND is obviously by far worse. We said about -- about 75% year-on-year. So obviously, that was quite that. I mean DRAM then obviously less than down mid-40s. And I think that in terms of change, we've seen a little bit of improvement, as we've just talked about with HBM and some signs of, at least a little bit more optimism on our part that the DRAM would be the first to recover as a result of some of these trends. And again, just listening to our customers and talking to them, we've heard some customers talk about further cuts in NAND just in the last 24 hours. So...
Stacy Rasgon
Like, do you think utilization in DRAM is still falling?
Doug Bettinger
It's still at a pretty low level, Stacy.
Stacy Rasgon
Okay, got it. Thank you.
Doug Bettinger
You're welcome.
Tim Archer
Thanks.
Operator
Our next question comes from Blayne Curtis from Barclays. Please go ahead with your question.
Blayne Curtis
Hey, thanks for squeezing me in. Maybe just following back up on what Stacy just asked. I was just kind of confused, you raised the outlook for WFE by a couple billion dollars, and it seemed like it was memory, that was the inflection. But then when you talk about memory, it seems like NAND is kind of flattish. And I'm just kind of curious how much of the increase is there in DRAM? Or is really the growth you're seeing in September more kind of this China part that Stacy was just asking about?
Doug Bettinger
Yeah, Blayne, Tim pointed to two things. A little bit more in China and high-bandwidth memory as it's not a huge uptick in WFE, but those are the contributors.
Blayne Curtis
Got you. And then I guess, I just want to understand, there's two moving pieces in foundry logic, and you've seen some delays at the leading etch. Can you just some commentary between those moving pieces, kind of the non-China piece, what are you seeing for the back half of this year?
Doug Bettinger
We described in generally, when you look at WFE down, where it is memory is down mid-40s, and you probably know kind of where that is run rating suggested foundry logic, looks like it's down roughly 10%. And yeah, you're right, just been a little bit of a softening in leading-edge foundry logic. We talked a little bit about that last quarter. I'm not sure we're describing anything incremental. Perhaps we saw some things others didn't see sooner than they did. But that's how we see things setting up right now.
Blayne Curtis
Okay. Thanks, Doug.
Doug Bettinger
Thanks, Blayne.
Tina Correia
Operator, we have time for one more question please.
Operator
And ladies and gentlemen, our final question will come from Vijay Rakesh from Mizuho. Please go ahead with your question.
Vijay Rakesh
Yeah. Hi. Just a question on -- as you look at WFE for next year, any thoughts on how much some of these onshore greenfield fabs could contribute to CapEx for next year?
Tim Archer
Well, I think it's -- as we're just in the middle of this year, it's too early to talk about '24 from an absolute number. But clearly, as we move through the next couple of years, I talked about this regional -- these regionalization efforts building momentum. I do think that we'll see that becoming a meaningful contributor to spending, not only here but in other parts of the world.
Vijay Rakesh
Got it. And then on the AI side, obviously, you have mentioned packaging as a big roadmap there. When you look at CoWoS packaging, any thoughts on how much -- what your exposure is there and how you see that growing next year, I guess? Thanks.
Tim Archer
Yeah. I mean everything related to advanced packaging is growing right now. And I would say Lam's exposure kind of across the board in etching deposition is very strong. So I called out specifically HBM, but you're right. I mean when you look at the importance of things like CoWoS-2 and AI packaged system, our exposure there even broadens further to the types of tools that we sell in there.
And I mentioned this point, if people caught it, even in advanced packaging, we're having to invest in new technologies for atomic scale processing. And so you can see things like ALD being used in advanced packaging now. And so we're really in a world where ALD is being used for advanced packaging, ALE is being used for 200-millimeter GaN. I think that it is creating tremendous opportunities for Lam to leverage our really broad portfolio of technologies for all these new emerging growth opportunities. And it's pretty exciting.
Vijay Rakesh
Got it. Thank you.
Tim Archer
Yeah. Thank you.
Doug Bettinger
Thanks, Vijay.
Operator
And with that ladies and gentlemen, we'll be turning the floor back over to the management team for any closing remarks. And management, are you able to hear me? Ladies and gentlemen, we'll be closing today's conference call. We thank you for joining. You may now disconnect your lines.
