And welcome to the September 2019 Quarter Financial Call for Lam Research. At this time, I'd like to turn the conference over to Ms. Tina Correa, Corporate Vice President of Investor Relations. Please go ahead, ma'am.
Thank you, operator. 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 review our financial results for the September 2019 quarter and our outlook for the December 2019 quarter.
The press release detailing our financial results was distributed a little after 1 o'clock p. M. 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 and A includes forward looking statements that are subject to risks and uncertainties reflected in the risk factors disclosures of 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 today's earnings press release. This call is scheduled to last until 3 pm Pacific Time. A replay of this call will be available later this afternoon on our website.
With that, let me hand the call over to Tim.
Thanks, Tina, and welcome, everyone. In the September quarter, Lam delivered solid results. Our continued execution to commitments combined with our guidance for the December quarter increases our conviction that Lam is in a strong position to outperform as wafer fabrication equipment spending inflects higher. Doug will cover the financial results in more detail shortly, but I am especially pleased with the demonstrated earnings power of the company. At the midpoint of our December guide, calendar year 2019 diluted earnings per share will be the 2nd highest in our history despite the current industry cycle.
I would like to take this opportunity to thank our customers and partners for their continued support of Lam and our employees throughout the world for their contributions to these results. From an industry perspective, we have revised upward our view on 2019 WFE to the mid-forty $1,000,000,000 range versus our prior estimate of down mid to high teens percentage year on year, which implied a low $40,000,000,000 level of spending. We are beginning to see improvement in the memory market, led first by NAND. NAND demand dynamics are improving and oversupply conditions should continue to abate as we move through the December quarter. We expect to exit 2019 with a bit supply growth rate for NAND of approximately 30%, which is well below our view on long term demand.
And as a result, NAND inventories are expected to decline to normalized levels in the first half of calendar twenty twenty. While the timing of a memory equipment spending recovery is always hard to predict, we are encouraged that customers continue to manage supply growth even as we are starting to see favorable end market demand indicators. This is a sign of a healthy industry and a good setup for increased NAND spending in 2020. On the DRAM front, inventories have remained elevated and we do not expect them to reach normalized levels until the second half of twenty twenty. However, we see positive demand catalysts ahead in both the server and smartphone begin next year with increased adoption of new generation platforms from leading manufacturers.
For smartphones, major vendors are planning to launch additional 5 gs models, which is expected to drive content growth for the overall smartphone market in 2020. Turning to foundry and logic, spending in this segment has been strong throughout 2019 and based on recent customer commentary looks to remain so heading into next year. Diverse end market applications are driving higher levels of foundry and logic spending. Moreover, challenges in scaling functional blocks such as SRAM in logic devices are leading to increases in die sizes and these in turn are accelerating changes in device architectures and chip manufacturing technologies. Lam's growing position with key foundry and logic customers has positioned us to incrementally benefit from these secular trends.
Competitively, we are executing at a high level. Based on the midpoint of our December guidance, Lam's 2019 foundry and logic revenues are set to significantly outgrow announced customer CapEx plans. The share gains we are now seeing in the foundry and logic segment are the result of close customer collaboration and strong product execution over many years and multiple technology transitions. They are evidence of the benefit of sustained investment in R and D throughout industry cycles. Looking at the market as a whole, including memory, foundry and logic, we are on track in 2019 to deliver our best ever penetration and defense performance as measured by net forward looking 3 year revenue opportunity for application decisions made in this calendar year.
A key contributor to our strong penetration and defense performance has been continued focus on technologies that enable 3 d device architectures, which are becoming increasingly important to performance and cost scaling across all market segments. We invested early in 2 d to 3 d inflections. And as these transitions are occurring, we are seeing expansion in both our SAM and market share. Etch and deposition processes are critical enablers for 3 d scaling and we are investing aggressively to deliver the technology and productivity innovation required to satisfy customer roadmaps. As evidenced by our penetration and defense wins this year, we believe we are extending Lam's leadership in this space.
In 3 d NAND, we have successfully defended 100% of our memory hole dielectric etch positions and continue to be the supplier for this application at all 3 d NAND manufacturers. We are also winning 3 d NAND applications where productivity is the primary point of differentiation. Notably, Lam has been the 1st to deliver production proven edge yield solutions for etch. In this quarter, we used our Corvus tunable etch hardware on our Flex Dielectric Etch system to improve profile tilt uniformity and win an important productivity sensitive slit etch application. On the conductor etch front, we won a 3 d NAND application for a new vertical architecture that reduces die size and is a technical solution for lowering bit cost.
In deposition, we recorded an important 3 d NAND win for the Vector DT, which deposits backside films to control stress as layer counts increase. Another significant deposition win was for our Stryker ALD tool used to deposit high quality liners and gap fill as aspect ratios get higher. We also continued to extend our 3 d expertise and position outside of the 3 d NAND space, including in rapidly growing markets such as advanced packaging and heterogeneous integration. Over the last 3 years, the installed base for our Sabre 3 d electroplating system has grown by more than 70% and we are the leading electroplating supplier for TSV for DRAM, CMOS image sensor and logic devices. Our Sabre 3 d electroplating solutions embed best in class technology backed by years of high volume production experience.
With each successive win across our served markets, the installed base of Lam equipment continues to grow, resulting in an expanding long term revenue opportunity for our customer support business. To create value for customers over the entire lifecycle of tool ownership, we are actively developing upgrades and advanced services targeted at extending technical capability and increasing productivity from existing installed base assets. These offerings help our customers reduce their total costs and as a result, we continue to see growing demand. Revenues from our customer support business grew in the September quarter on a sequential basis and our Reliant business achieved record quarterly revenue for the Q3 in a row. We expect 2019 overall will be another growth year for our customer support business.
Looking at our year to date performance, we have made tremendous progress against our objectives of expanding our TAM, increasing our market share, and building our installed base business. Importantly, 2019 has been a year where Lam has strengthened its position in the Foundry and Logic segment. Also, with early indications of improving NAND demand and positive catalysts on the horizon for DRAM, we are increasingly optimistic that calendar year 2020 is setting up to be a year of outperformance for Lam as spending mix moves back in our favor. Thanks again for joining today. And now here's Doug.
Okay, great. Thank you, Tim, and good afternoon, everyone, and thank you for joining us today on what I know is a busy earnings season. We're pleased with Lam's performance in the September quarter. Our results once again exceeded the midpoint of guidance for all financial metrics. Operating income and diluted earnings per share came in at the high end of our guidance range as we remained prudent in managing our spending throughout the quarter.
Let me begin as I always do by talking about our revenue segmentation in the September quarter. The combined memory segment was flat with the June quarter at 64% of total systems revenue. We had a decrease in the September quarter in the non volatile memory segment moving from 46% to 38%, while DRAM increased to 26% from 18% of systems revenue. Spending in the NAND segment was focused on multiple nodes and we're beginning to see the 1st ramp of 128 layer structures. On the DRAM side, apart from the 1X and 1Y nodes, we are seeing initial investments at 1Z.
As I noted in our last quarter earnings call, foundry and logic spending was strong and we expect strength from this segment to continue through the remainder of the calendar year. The foundry segment represented 25 percent of our systems revenue in the quarter. Strength in this segment is related to spending on the 7 5 nanometer nodes. The Logic and Other segment was down slightly from the prior quarter level coming in at 11% of system revenue. We've continued to demonstrate strong progress in the foundry and logic segments, enabling us to maintain solid profitability levels during a period of depressed memory spending.
I believe foundry and logic spending will be even stronger in December. Revenues for the quarter came in at $2,166,000,000 which was again above the midpoint of guidance. Our revenue had a slightly broader geographic mix in the September quarter compared to the June quarter. Our top regions continue to be China, Korea and Taiwan. The China region quarter performance remain higher remains higher than our historic average concentration of revenue.
And similar to what I talked about last quarter, the majority of this came from indigenous Chinese customers across multiple segments. Gross margin came in at 45.4%, which was 40 basis points above the midpoint mainly due to customer mix. And as we've stated in prior quarters, you should expect gross margins to be a function of several factors such as business volumes, product mix and customer concentration and we expect to see variability quarter to quarter. We continue to manage our spending levels in the company as operating expenses in the September quarter declined to $431,000,000 which was down from $450,000,000 in the previous quarter. We remain laser focused on investing in research and development programs as we saw the percentage of spending in R and D increase quarter over quarter to 67% of operating expenses.
The December quarter, the guidance reflects total spending increasing back to the June level, primarily due to an increase in variable compensation expense. Our variable compensation fluctuates based on the level of quarterly profitability. I would also remind you that as we look ahead to the 2020 calendar year, you will see the normal seasonal spending increases related to the March quarter comes from things like payroll taxes. Operating income in the September quarter was $552,000,000 and operating margin was 25.5 percent at the top of our guidance range. Our September quarter non GAAP tax rate was approximately 11% which was slightly lower than our long term rate.
There will be fluctuations in the rate from quarter to quarter and we now expect our long term rate to be in the low teens level. Other income and expense was a total of approximately $11,000,000 in expense in the September quarter. The main components of other income and expense are interest income from the cash and investment balances we hold, offset by expense related to our outstanding debt. The total interest expense on all tranches of our debt is right now about $41,000,000 per quarter. You should expect that other income and expense will fluctuate quarter to quarter based on several market related items such as our deferred compensation assets, venture capital investments and foreign exchange.
We continued to execute on our capital return program during the September quarter. We allocated $234,000,000 to capital return in the quarter with $75,000,000 related to open market share repurchases and $159,000,000 in dividends. I would like to remind you that we continue to have an ongoing structured repurchase program that is expected to mature in the December quarter. This will continue to reduce our share count. We remain on track with our committed capital return.
We currently have approximately $3,000,000,000 remaining in our board authorized share repurchase program. Diluted earnings per share came in at 3 point $1.8 which was at the high end of the guidance range that we provided for September. We ended the September quarter with diluted shares for earnings per share at approximately 151,000,000 shares which is the 7th consecutive quarter where our diluted share count has declined. The share count includes a dilutive impact of approximately 5,000,000 shares from the 2,041 convertible notes. And I'll remind you the dilution schedules for the remaining 2,041 convertible notes is available on our Investor Relations website for your reference.
Let me now switch to the balance sheet. Our cash and short term investments including restricted cash increased slightly in the September quarter to $5,800,000,000 from $5,700,000,000 in the June quarter. Cash flows from operations were $464,000,000 which was offset by share repurchase and dividends. Year to date in calendar year 2019, we have had strong cash from operations performance and we're on track to end this year with 2nd highest level of free cash flows in the company's history. DSO increased to 69 days versus 56 in the prior quarter.
The DSO increase is related to the timing of customer payments occurring at the end of the September calendar month, which falls within our December fiscal quarter. Our inventory balance declined sequentially by $57,000,000 which is the 5th consecutive quarter where inventory balance declined. Inventory turns were 3.2 turns, which was just a little bit less than the 3.3 turns that we saw in the June quarter. Non cash expenses included approximately $43,000,000 for equity compensation, dollars 49,000,000 for depreciation and $16,000,000 for amortization. September quarter capital expenditures came in at $39,000,000 which was a decrease from $66,000,000 in the June quarter.
Our September quarter end headcount was flat with the prior quarter at approximately 10,700 regular full time employees. So now looking ahead, I'd like to provide our non GAAP guidance for the December 2019 quarter. We are expecting revenue of $2,500,000,000 plus or minus $150,000,000 Gross margin of 45 percent plus or minus 1 percentage point. Operating margins of 27% plus or minus 1 percentage point. And finally, earnings per share of $3.80 plus or minus $0.20 based on a share count of approximately 150,000,000 shares.
I'm pleased to share with you the results we've delivered throughout calendar year 2019 in what has been a challenging industry environment. As Tim noted, the supply demand environment is improving for the memory segments and we've made good progress in our foundry and logic positioning. Our installed base business continues to be on track to deliver a growth year in 2019. We're well positioned heading into 2020 and are optimistic about our future performance going forward. Finally, I'd like to announce our plans to host an Investor Day on March 3, 2020 in New York City.
Details on the venue and precise time will be forthcoming. Operator, that concludes my prepared remarks. Tim and I would now like to open up the call for questions.
And we'll take our first question from Atif Mehlich with Citi.
Hi, thanks for taking my question and good job on the execution. Question on the WFE. Tim, you mentioned WFE going from low $40,000,000,000 to mid $40,000,000,000 I want to understand if the majority of the increase has come from the foundrylogic segment in terms of the revised outlook?
Maybe a few modifications. I mean, clearly, as we've said, we've continued to see strengthening in foundry and logic. So that is a portion of it. Some very early indications of NAND spending increase we now guided for December quarter. And finally, continuing strengthening in the China WFE as we've kind of seen increased strength throughout the year.
So a combination of all those things led to our upward revision in WFE.
Okay. And just a follow-up. In your prepared remarks, you mentioned foundrylogic share gains above foundrylogic CapEx this year. How should we think about the outperformance of your foundrylogic business into next year and out given some of the elements like reuse and growing EUV steps?
Well, that's a great question. Basically, as we have said on a number of occasions, the intensity of etch and deposition we see continuing to increase at every technology node. Our SAM grows whether it be 10 to 7 to 5 to 3 on into the future, SAM increases even in the face of EUV increased usage. And there's a number of reasons for that. Patterning complexity continues to increase regardless of the introduction of EUV.
And so there are additional steps for deposition and etch. EUV itself introduces new requirements for hard mask, which gives land opportunity to participate from a deposition perspective. And EUV also we have talked about in the past introduces opportunities for new steps like and process like atomic layer etching that can be used to help increase quality and productivity of the EUB pattern itself. So, I think the way you should think about it is our SAM is expanding. Competitively, we're gaining share at some of these new layers and we feel we're well set up in both the foundry and logic space going forward as technology transitions.
Thanks.
Thanks, Arthur.
Thank you. We'll now take our next question from C. J. Muse with Evercore.
Yes, good afternoon. Thank you for taking the question. I guess first question, if we look back to 2018, it looks like your share of wallet for WFE with NAND is roughly 30%. So curious, as you think about moving to a rising layer count and your leadership in the highest spec ratio etch and the just announced deposition wins on the call tonight, How should we think about your share of wallet as we go to 128 layer and above into 2020?
Our SAM as a percentage of customer spend continues to increase with layer count. And that's for two reasons. 1, I mean, obviously, the simple fact of building and etching the higher stack, but also the fact that as I mentioned there are new steps and new opportunities that get created for dealing with issues such as in the vector DT case, the stress associated with those taller layer counts. So I think that our view is SAM grows as layer count increases.
Thanks, Tim. And I guess as a follow-up, Doug, I'm not sure if you spoke to the entirety of installed base revenues, but did they grow sequentially in September? And how should we think about the trend or what should we kind of model it to into 2020?
Yes, C. J, I think Tim actually mentioned in his prepared remarks that they did grow sequentially and it was the 3rd consecutive record quarter for the Reliant component of that business. I'm not going to quantify next year yet a little bit too soon. But what we said in the past and I continue to be very comfortable saying today is I have a hard time envisioning a year when the installed base business doesn't grow. It should grow every single year because Chamber Comp grows every year, it's growing this year even in a depressed memory spending environment.
And we continue to bring new advanced service offerings to market that we hope to enable us to achieve more and more of the customers' OpEx spend. So that's how you should be thinking about that, CJ.
Great. Thanks so much.
Thanks, C. J.
Thank you. We'll now take our next question from Harlan Sur with JPMorgan.
Good afternoon. Great job on the quarterly execution. This is the 2nd quarter in a row of the heavier China domestic mix. Do you guys anticipate this bias to continue into the December quarter and into next year? And similar to my question last time, is this, in your view, a focused effort by China to accelerate their semiconductor manufacturing capabilities given the trade tensions with the U.
S. Or more your China customers just having more confidence to move forward with their early memory and foundry programs?
Yes, Harlan, what I've tried to describe, in fact, I tried to foreshadow it last quarter in earnings that this is above normal run rate of local Chinese customer spend. But Tim did point out in his remarks that some of the strengthening in WFE came from local China. Probably now nominally somewhat above $6,000,000,000 in WFE. And as we look into next year, we absolutely think it will grow again next year, But it will ebb and flow. These big projects can be lumpy at times and it'll go up and it'll go down depending on when equipment ships into any one fab project.
Tim, you want to add anything?
Yes. I think that as I commented, we've seen strengthening in domestic China spend through the year, anticipate that continuing. Maybe the most important part of that story for Lam is that clearly a big portion of the new incremental spending in China is targeted towards the memory market. And obviously, our SAM and our share in the memory space is quite good. So from that perspective, we see China as an area of strength for us.
And Harlan, I forgot the one part of your question where you asked. Do we think anything pulled in? Actually, really don't. I don't think it would make sense that they would be pulling things in sooner because if you're concerned about our inability to ship being the reason they pulled in, if they can't get spares and service from us, they can't actually really utilize the tools very well. So, I don't believe that there was pull ins related to anything like that.
Great. Thank you for the insights, Darren. The team continues its strong design win momentum on non critical and or legacy technology nodes. Part of it, as you pointed out, is the Reliance Systems products continue to do well. We continue to hear from your customers that they're laser focused on productivity, throughput, uptime, footprint, all of the things that impact overall wafer costs.
Are you guys now in a position to at least give us some sense on how fast the non criticalrefurbished business systems business is growing and roughly the size of this business relative to your total revenue base?
Maybe you'll have to wait till Investor Day for that. No, we do plan we have promised for some time now to provide you more detail in those areas and we're just we will do that. I think maybe a couple of points. Your comment about customers being laser focused on productivity. I mean it's one of the reasons why we have been talking about it.
It takes some time for those products, for our efforts to really start to show up in new wins. But we're starting to see in this area that Doug talked about in advanced services where some of these intelligent database tools are really starting to help us reduce, for instance, troubleshooting time on the systems, reduce unscheduled downs on the systems. And those are things that the customers are pulling hard for because, again, it's productivity for tools that are already in place and relatively easy to implement. So, we're prioritizing productivity because it's in the best interest of the customer and the industry as a whole.
Yes. Thank you.
Thanks, Harlan.
Thank you. We'll hear now from John Pitzer with Credit Suisse.
Yes. Good afternoon, guys. Thanks for letting me ask the question. Congratulations on the strong results. Tim, I just want to go back to the market opportunity in logicfoundry for you.
If you kind of look at calendar year to date, that business combined for you is up somewhere between 30% to 40% year over year. I guess I'm just trying to get a sense of how we think about your market position, maybe in North America as we see 14 go to 10 go to 7 and maybe in Taiwan as we see 7 go to 5 go to 3. Can you talk a little bit about how much visibility you have on product tool of record on some of these critical etch and deposition steps and how we might think about your share opportunity as we migrate down these nodes?
Sure. I can do that without quite getting as specific as North America versus Taiwan on those nodes. So maybe I'll point you just to a couple of comments I made and then embellish those a little bit. As I said, within the logic and foundry space, it takes many years to establish yourself as the first the development tool of record and then ultimately see that roll out into volume buys as the production tool of record. And so when we talk about we gave a little bit of a new look into this penetration and defense forward looking 3 year revenue opportunity statement.
That is designed to give some indication as to how long it takes from the time we win one of those selection decisions. And you'll see that revenue opportunity take place over those following 3 years. That's definitely true in the foundry and logic space. And so I think that when we talk about improving opportunity at 10 and then 7 and then 5, you can kind of look out and say if like a 5 nanometer decision is revenue opportunities that now will roll in over the course of the next several years. So, that may not be quite the level of specificity you wanted, but it basically says things that we are winning now will actually be our revenue for the next 3 years plus in that space.
That's helpful. And then Doug, you mentioned in your prepared comments, you still have about $3,000,000,000 left on the buyback. And I apologize for the capital allocation call, but this most recent quarter, the buybacks dipped down a little bit, which I guess is understandable just given the level of uncertainty. Any color you can give us around the pace at which you might want to try to execute buybacks going forward and how we should think about kind of just remind us again the cash return policy you're trying to hit?
Yes, John. I mean one thing I would point out to you, if you look over the last 2 years, you will have seen several quarters where the cash we actually deployed moved down somewhat. That was because in the quarter before that we put one of these accelerated share repurchase programs in place. You have the same phenomenon going on this quarter. So even though the cash that we deployed in terms of open market repurchases wasn't all that significant this quarter, That ASR was still executing buying stock back and that ASR that's currently out there will complete in the December quarter.
So we'll be thinking about what we're going to do incrementally as we go forward. So far, let me just remind you what we talked about, it's been a while now, the last Analyst Day that we're committed to at least 50% of free cash flow return to shareholders. And obviously, if you look at our history, over the last 5, 6 years, we've done a whole lot more than that. And I've been silent on the timing to complete the current authorization and I'm going to kind of remain that way. I've said in the past, we'll be opportunistic and that's as much as I have for you right now.
Perfect. Thanks guys. Congratulations.
Thanks, John.
Thank you. We'll take our next question from Timothy Arcuri with UBS.
I guess my first question, Doug, is can you help us a little bit with the mix in December? I'm curious if the uptick in revenue is going to be more on the NAND side or more on the foundry side? Thanks.
Yes. What I said to him in my remarks was I expect December will continue to be a pretty strong foundry and logic quarter. So that's a key piece of what's going on.
Okay, awesome. And then
I just wanted
to drill down a little bit into the China piece. So China is about 25% of your revenues now on the trailing 12 month basis. So I guess I'm curious how much of this is going into indigenous projects versus the multinationals? Because I thought you said China was about $5,000,000,000 of 2018 WFE. And I think at that time you were talking about $3,000,000,000 of that $5,000,000,000 being domestic China.
And I think you just said that China is going to be over $6,000,000,000 this year. So can you sort of like level set us on how much of that $6,000,000 this year will be multi versus domestic and maybe how much domestic can even grow next year? Thanks.
Yes. Let me walk it back for you a little bit, Tim, because some things have moved around a little bit. When I look at what happened last year, local China ended up being, I don't know, dollars 4,500,000,000 of WFE roughly. So it ended up being maybe a little bit more than that 4 that we have been talking about. And the way we see it today, it's above 6, it's somewhat above 6.
Don't know for sure what next year is going to look like, but as we look at our analytics and the fab projects that are coming in and whatnot, I do pretty strongly believe it will grow next year. I'm not ready to tell you how much yet. We'll give you more clarity on our WFE view for next year on the next earnings call. But I think local China will continue to grow next year.
And Doug, just to clarify, those numbers are local China numbers, right?
That's right, Tim.
Yes. Okay, awesome. Thanks so much.
Yes, you're welcome.
Thank you. We'll take the next question from Toshiya Hari with Goldman Sachs.
Yes, thanks for taking the question and congrats on the strong results. Tim, you talked about early signs of a recovery in NAND in your prepared remarks. And Doug, you just mentioned that most of the growth or at least implicitly, I think you commented that most of the growth in December quarter should be coming from foundry and logic. So should we expect some of the NAND projects that we're all collectively hearing about should hit the March quarter from a rev rec perspective?
Yes. Maybe there's not such a big change here from what we've said. Ever since the beginning of the year, we said that memory spending would likely remain relatively weak throughout the year. And when I talk about early signs, it is the it's very early as in there. We're now starting to see some of those projects transpire as you just talked about, meaning a little bit of ordering occurring in the December quarter.
But we still maintain our view that we exit the year with this 30% supply growth rate and that customers are prudently managing the supply growth. But I do think those projects, it's kind of that early sign that's healthy industry and we're going to see growth in 2020. So exact timing, we really don't want to give you that right now, but it looks to be coming.
Got it. And then as a follow-up, I just wanted to hit the logic and foundry opportunity dynamic into 2020. Obviously, you guys have done a great job over the past couple of years in gaining share in both buckets. If we assume that spending in logic and foundry is largely flat in 2020 and if you assume EUV adoption continues to grow, based on what you guys know from a design win or application win perspective, can you still grow logic and foundry revenues in that sort of environment backdrop? Thanks.
Yes. That's a great question. And the answer is yes, we would expect to grow in that scenario based on a couple of things. 1, as you move forward, even on the same WFE, our SAM as a percent of that WFE should increase as transitions still continue to occur to more advanced nodes. Etch and deposition opportunity increases at each successive technology node even in the face of EUV.
So I think in the scenario described, we would expect to have a larger opportunity and we do believe that we're winning share as well and therefore we would grow.
Great. Thanks so much.
Thanks, Toshiya.
Thank you. We'll hear now from Krish Sankar with Cowen and Company.
Yes. Hi. Thanks for taking my question. I have 2 of them. First one for either Tim or Doug, more industry focused question.
You spoke about how NAND bit growth exiting this year is going to be 30%. Kind of curious like as your customers start ramping 128 layer NAND, if next year demand bit growth is below 30%, do you still expect the 128 layer spending to go through next year because it's more strategic low cost in nature or do you think it might get throttled back if demand will go to slower? Maybe I'll take
it and then Timnit you can add on. I mean, Chris, one, I don't think that's what's going to happen, but were that to be what happened, I think you would see a year that looked somewhat like this year in that most of the industry spending was allocated to node conversions as opposed to new capacity adds, because that lowers cost per bit, the economics are better and all of that. The other thing I would say and then I'll let Tim add on, when we look into early next year, I believe supply growth rate will continue to decline based on the investments that have occurred this year, right, because they have reduced this year. And so you're moving into a declining rate of growth as we look into the first half of next year. Tim, anything you
Yes. I think just to reiterate Doug's point, I mean, technology transitions, we think, occur every year simply because of the benefit to bit cost. And so, I guess I would say that in a year, which is not our view of declining supply growth, we would still see the majority of the spend in technology transitions. And I guess there's one other key point to think about and we've mentioned this a few times. In a technology transition from say 96 layers to 128 layers, the majority of that spend is for etch and deposition.
So, there's definitely an outperformance statement there if spend continues to be based just on technology transitions.
Got it. That's very helpful. And then as a follow-up, I think, Tim, in your prepared comments, you spoke about how you guys have successfully defended 100 percent of your dielectric etch market share. I'm kind of curious on the NAND side, is that a fair enough statement if you look past 128 layer higher than 128 layer DTOR tools or is it more up to 128 layers?
Well, I mean our statement is always as those decisions are made. So we always speak about accomplishments made. So I don't know of any decisions that have been made well beyond the 128 layer node. So, it's obviously something that we compete for at every node. But I guess what I'd point out is there's a very important piece of learning that has come from us running now millions of wafers through our dielectric etch tools.
And there was a question earlier about productivity, reliability, stability, defect performance. Those are all things that again we have millions and millions of wafers of experience. And so, we kind of come in with an edge every node that you have to compete for as the incumbent. So that, but specifically I think you could bucket most of those in that 128 nanometer and below node since their decision is already made.
Got it. Thanks, Tim. Thanks, Doug.
Thanks, Kish.
Thank you. We'll hear now from Patrick Ho with Stifel.
Thank you very much and also congrats on a really nice quarter. Maybe Tim first, a lot of the market questions have been answered on my end. But as you look at next generation memory technologies like MRAM, RRAM, can you just give maybe a qualitative view of how Lam is positioning itself to capitalize on those next generation opportunities?
Yes. I guess maybe the simplest statement that they make inside the company is we're not running from our leadership position in memory and that includes in these new emerging memory markets. And so, we're actively looking to develop new applications, new tools that meet the needs of those devices. We've spoken about one of them in particular so far this year, which is the INB Match tool that's used for MRAM. And we have established a very strong position for that particular application.
Phase change memory, reRAM, other devices, we're actively engaged in development with the leading companies. So it is a target market for us. In fact, one might argue that we feel we sort of we already sort of own that memory space and I think it's a focus.
Got it. Got it. Just to add on, if you really want more detail, you can go look at the transcript of the Flash Memory Summit where Tim and Rick Gautcho spent lots of time talking about how we view these emerging memory architectures.
Great. Thanks, Doug. And maybe as my follow-up question, you talked a little bit about your atomic layer deposition opportunity. I guess from a big picture perspective, how do you see the total available market for you in that? And do you see it expanding from where it looks like you have a strong position on the memory side of things?
Can this also expand on to the foundry logic side of things?
It already has expanded. I mean, ALD is not memory specific. So, we use these films also on the logic and foundry side. ALD, atomic layer etching, atomic layer deposition and atomic layer etching, these are technologies that will continue to be increasingly used as devices become smaller and smaller and structures become more complex. So, I think it's an expanding market opportunity.
And in fact, I think in coming years, you'll start to see inflections, for instance, from batch processing tools to single wafer ALD, just again for the type of wafer uniformity requirements of those processes going forward. So it's an area where we're growing and we feel confident about our development activities and the tools themselves.
Great. Thank you.
Thanks, Patrick.
Thank you. We'll hear now from Craig Ellis with B. Riley FBR.
Yes, thanks for taking the question. Tim, I wanted to follow-up on the comments that you made that calendar 2020 could be a year of outperformance for Lam. You just help us understand how much of that comes from moving deeper into the sweet spot of the foundry logic share gain that you've been talking about versus help that you would get from what appears to be a nice and very encouraging upturn that's starting to occur in NAND or maybe something else like SAM expansion?
Yes. I think for us to give you more detail on the breakout on that, you'll have to wait until we actually speak to our 2020 WFE outlook next quarter. But you kind of hit it. I mean, the way I think about outperformance in 2020, it's just what you said. 1st and foremost, everyone knows that we are highly leveraged to the memory market and we do believe that spending mix moves back in our favor next year.
And even in the face of stronger and continuing strength in logic and foundry, memory will improve to some extent. We've strengthened our logic and foundry SAM opportunity as well as share position and so that's a continued benefit as logic and foundry remain strong. And also as Doug mentioned, CSBG continues to grow as we expand our portfolio of advanced services in that space. And so, can't break it out for you how much each of those contributes, but each of them is an important part of the story in 2020.
Thanks, Tim. And then I'll do the follow-up to Doug. Doug, impressive trough to trough operating margin improvement of about 300 basis points. And within that, there's dramatic improvement with operating expense as a percentage of sales. And as you noted, there's been significant mix shift R and D.
The question is, how much further benefit is there in the model for increased operating leverage and how much room do you have to further drive expense towards R and D from SG and A?
It's something we've been working on for years. I mean, Tim was driving this years ago, honestly, and the rest of the company's rallied behind it. We're going to keep squeezing efficiency out of every SG and A dollar we spend. We want to be totally rigorous about that, so that we can allocate more of those dollars to R and D. I think that 67% last quarter might be an all time high for the company.
I haven't gone all the way back, but in my recent memory that was an all time high. We're going to keep at that. We're going to continue to try to get better. We do this everywhere. It's part of the culture of the company to be focused on continuous improvement.
I don't know how high we can get it, honestly, but we're going to keep working at it.
Thanks, guys.
Thanks, Craig.
Thank you. We'll hear now from Vivek Arya with Bank of America.
Thanks for taking my question. And I joined a little late, so I apologize if this was asked, but DRAM was quite strong in the quarter and I'm wondering what caused that and how sustainable is that strength?
When I look at it, it's primarily conversion spending. That's really what we've seen for the entire year is focus on node conversions. It will ebb and flow. The important thing when you think about DRAM is when we look at similar to what we were describing NAND exiting the year below where we believe demand growth to be, you got a similar story in DRAM. As we look at the investments that are occurring in DRAM this year, exiting the year supply growth is probably in the low teens.
And we believe long term demand growth is in high teens, maybe approaching 20%. There's a ways to go, Vivek, to continue to burn inventory out of the channel. I think that's going to take a little bit longer. But what I do know is at some point the inventory will be burned off and spending will grow more significantly. And quarter by quarter, it's just going to be dependent on what projects are underway at what customers.
Got it. Very helpful. And then as a follow-up, kind of looking at your cash flow generation, so fiscal 2019 from what I noticed, there was a lot of cash inflow from working capital if I'm looking at the model right. How should we think about the plus and minus from working capital this year? Thank
you. If you look at the first half of this year, calendar year, again, I never really think too much in terms of fiscal calendar year. I mean, cash was just super strong. I think we had 2 consecutive quarters around $900,000,000 in operational cash flow. And you're right, a lot of that came from working capital.
Now, what you normally see with the business when it's growing, it will consume working capital, meaning it'll take cash to build inventory and receivables and so forth. When it levels off or contracts, it will generate cash. And that's really what went on in the first half of the year. I mean, we manage cash quite proactively, but the practical reality of business ebb and flow is that's what happens. The right way to think about the free cash flow of the business is it's a couple points below where operational cash flow should be on a sustainable basis, but it will ebb and flow around wherever we're at in the business cycle.
Thank you.
Yes. You're welcome. Thanks for the question.
Thank you. We'll take our next question from Joe Quatrocki with Wells Fargo.
Yes. Thanks for taking the question. You talked about improvements in inventory that you've seen across NAND flash. I was wondering if you could talk about to the extent you can factor utilization rates in your installed base?
It's hard for us actually to put our finger on exactly what utilizations are. I mean we saw it pull back a little bit in both NAND and DRAM this year. And I think that was pretty well telegraphed from our customers. But better it's a better question to put to them. We don't always know exactly what their utilizations are going in.
Okay, fair enough.
And then
maybe a follow-up to that. You talked about exiting this year at 30% supply growth for NAND and maybe potentially going lower than that next year, the 1st part of the year. I guess the question is, how do we think about the level of capacity expansion needed to hit kind of the, I guess, expected demand growth for 2020?
Tough question to answer, Joe, to be honest. What's occurred this year in NAND largely has been conversions maybe with a couple exceptions. But if you look at NAND WFE through the year this year, that's what was going on. And you saw a declining rate of supply growth such that as we exit the year, obviously, we said 30% and I think long term NAND growth is in the high 30s, maybe 40%. To have 40% supply growth, you need to be adding a few wafers every single year.
And so it didn't happen quite so much this year and you've seen supply growth decline. I don't know if that helps you, but that's how I think about it.
Okay. Thanks. Thank you. We will take our next question from Mitch Steves with RBC Capital Markets.
Hey guys, thanks for taking my question. I really had 2, most of them have been answered. But from a high level, I mean, do you guys talk about kind of a better market for memory in 2020? So when do you think we'll see capacity upgrades? Or is this all going to be technology upgrades?
And then secondly, you guys gave a lot of color on NAND, but not so much in DRAM. So based on the commentary about server demand and kind of an increase there in Q1, does that mean you guys think that that's going to be kind of the bottom? Or do you think that exiting the year, we're going to start to see supply demand balance for DRAM?
Well, I think that what we said about DRAM maybe was that we think through the first half of next year, inventory remains elevated. And so it's really not I guess our view at this point is DRAM may be more of a second half story for next year. Obviously, we'll give far more color as we develop our full 2020 look. It's just that NAND is much more upon us right now simply because the issues are being worked through more quickly as a result of the supply constraints that have been in place this year and maybe also the demand catalysts for NAND that are occurring right now.
Got it. And is there capacity upgrades at what time in 2020? Is it like back half or do you think capacity upgrades begin earlier than that?
Yes, Mitch, too soon for us to talk in any detail about 2020. Hold off for 1 quarter, we'll give you a little better visibility on it. It's still a little bit too far away for us to get into detail.
Okay, got it. Thank you.
Thanks, Mitch.
Thank you. And that does conclude today's question and answer session. I'd like to turn the conference back over to Ms. Correa for any additional or closing remarks.
Just wanted to thank you everyone for joining us today. Appreciate it.
Thank you. And that does conclude today's conference. Thank you all for your participation. You may now disconnect.