Good day, welcome to the Lam Research Corporation December 2017 conference call. At this time, I'd like to turn the conference over to Satya Kumar, Vice President of Investor Relations. Please go ahead.
Yes, thank you, good afternoon, everyone. Welcome to the Lam Research quarterly earnings conference call. With me today are Martin Anstice, Chief Executive Officer, and Doug Bettinger, Executive Vice President and Chief Financial Officer. During today's call, we'll share our outlook on the business environment, review our financial results for the December 2017 quarter, and our outlook for the March 2018 quarter. The press release detailing our financial results was distributed a little after 1:00 P.M. Pacific Time this afternoon. It 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 includes forward-looking statements that are subject to risks and uncertainties reflected in the risk factor disclosures of our SEC public filings. Please see accompanying slide 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 the GAAP and non-GAAP results can be found in today's earnings press release. This call is scheduled to last until 3:00 P.M. Pacific Time, as always, we ask that you limit your questions to one per firm with a brief follow-up so that we can accommodate as many questions as possible. As a reminder, the replay of this call will be available later this afternoon on our website. With that, let me hand the call over to Martin.
Thank you, Satya, thank you, everyone, for joining us today. I will now provide an overview of December quarter results, summarize calendar 2017 relative to the multi-year outperformance opportunity for Lam, provide a first outlook for 2018. As previously noted, we plan a comprehensive investor analyst event in a few weeks, so our prepared comments today are crafted in that context. The December quarter was a strong conclusion to an extraordinary year with record performance for Lam reported in shipments, revenues, and non-GAAP EPS. As noted in our earnings release, that positive momentum is expected to continue into calendar 2018, with our March quarter guidance indicating midpoint shipments at $3.175 billion and revenues at $2.85 billion, both levels of strategic relevance and levels of business not previously seen in our history.
While scaling the company at essentially a 50% pace in calendar 2017 was not our expectation one year ago, competitive business model flexibility and a commitment from our employees to the success of our customers was everything about how we responded to the incremental opportunities presented. Without the tremendous effort of the team and the support of our customers and partners globally, these accomplishments would not be possible. It is my privilege to acknowledge this great work and express my personal gratitude. At the same time, it's important to reinforce the prominence of our culture and values, and specifically our focus on increasing customer trust and codependency through the delivery of enabling technology, productivity, and speed now and in the years ahead.
As an established multi-product and services company delivering results on wafer to the semiconductor industry through complex systems engineering innovation, our primary objective is to promote our customers' success through an increased proportion of their spending that we actively compete for and through differentiation to win and sustain more of that business over time. In 2017, our reported shipments were approximately 21% of WFE versus 18.5% in the prior year, and that compares with 14.2% in the first full year after the Lam-Novellus merger in 2013. Reflecting on 2017, Lam delivered a remarkable sixth consecutive calendar year of growth and industry outperformance, a byproduct of high-quality strategies and execution jointly with our customers. Over that period, Lam has grown shipments by a cumulative annual growth rate of 24%, approximately three times the rate of WFE growth.
In 2017, through a commitment to invest more in our future and to sustain the economic success for our customers and suppliers in partnership with Lam, we grew our shipments by approximately 50%, well above WFE growth of 30%, and reported record non-GAAP earnings per share of over $13, almost double the prior year. With fundamental industry trends that we believe provide the platform for sustainable semiconductor ecosystem investments and growth, we remain focused on making the necessary investments in our company to increase the probability of achieving our long-term vision, we plan for another expansion in the proportion of WFE that we actively compete for and win in 2018. Said differently, we plan for continued outperformance. 2017 recorded more than two percentage points of shipments-based share gains across the equipment segments that we serve.
Our market share performance was broad-based, with strengths coming from areas that are most critical to the success of our customers, with a focus on technology enablement across the full suite of industry inflections anticipated as relevant for our customers for the next five to 10 years. We had very strong momentum in atomic layer deposition products, which have been a focus area for new product introductions for several years. We more than tripled our shipments of ALD products in 2017. We had record shipments in dielectric etch, with more than a tripling of the installed base to over 1,000 process modules for our Flex G series products, which address the most challenging aspects of 3D NAND scaling. We had a record year for conductor etch with application wins across foundry and NAND markets. We strengthened our position in tungsten CVD and ALD, as well as Electrofill deposition.
We had a strong year for clean products with a 50% growth in our EOS clean product installed base. As you know, broadening our customer base has been a key part of our strategy in the last decade, and we are pleased to see it deliver results. With record shipments in logic for the year, strength here was driven both by ramp of etch application wins at leading-edge logic, as well as trailing edge logic shipments to a broad set of customers for IoT applications. There should be no question about the fundamental relevance of the Lam product portfolio in etch and deposition to the device architecture and manufacturing processes planned by our customers broadly now and in the long term.
We have much more comprehensive engagements across all device segments than ever before, creating a platform of learning and optionality that enhances our competitive differentiation and the quality of our earnings. Increased strategic relevance to our customers' success creates expectations and also significant opportunity for the company, not least relative to the performance of more than 50,000 Lam process chambers in the global semiconductor installed base today. Our customer service business group grew profitable revenues in 2017 faster than the pace of our installed base growth. More importantly, they took custody over the delivery of a portfolio of productivity solutions that we believe created more value for the customer and enhanced competitiveness of Lam than in any prior year. Over the last decade, Lam has transitioned from an etch company highly levered to memory customers, to a multi-product and services company with broader customer exposure across all device segments.
That transition speaks to the quality of technology leadership. It speaks to cycles of learning and also business momentum. As device performance is increasingly driven by adoption of new structures and new materials, as the traditional scaling solutions are limited by physics or economics in high-volume manufacturing, enhancing our etch clean and deposition portfolio with disruptive and value-creating technology is a strong focus. Illustratively, shipments for new memory solutions, which have extremely demanding requirements, more than tripled for Lam in 2017. Our strategy for differentiation extends well beyond technology enablement for unit process and co-optimized or integrated products. Our fundamentally collaborative culture and values create an unmatched potential as a participant in the semiconductor ecosystem. We are committed to harness the holistic value of Lam for the benefit of all stakeholders. Now turning to the overall equipment spending trends of 2017 and our expectations for 2018.
We believe that WFE investments in 2017 ended at a little over $47 billion, which was up approximately 30% from the prior year. Slightly more than half of the investments made in 2017 were for memory and were weighted primarily towards NAND flash. Semiconductors are fundamental to enabling the generation, transmission, storage, and analysis of increasing quantities of data, which has the potential to transform every segment of the global economy. We remain optimistic about the potential for continued global economic stability and growth. We anticipate healthy demand for semiconductors across the full spectrum of consumer and enterprise markets. We expect year-over-year growth in overall WFE in 2018 in the low double-digit percentages, led by DRAM and logic investments, and supported by the fundamental demand drivers for silicon and the confidence we have in the continued motivation for investment discipline by all key market participants.
We still anticipate that approximately 85% of the year-over-year spending growth is memory-based. Preliminarily, we plan for double-digit growth in overall memory WFE in 2018 with a greater balance between DRAM and NAND spending than was true in calendar 2017. Sustained to slightly increased investments in NAND overall is the year-over-year headline for 2018. More of the incremental growth is from DRAM, where supply and demand balance tightened last year. Importantly, in DRAM, technology trends are resulting in meaningfully higher levels of investments required to drive similar levels of bit supply growth compared to historical levels. We do expect a modest market demand-led expansion of the wafer starts per month installed base in DRAM, particularly supporting data center content expansion. We align to the consensus view on bit growth in both markets.
On the demand side, NAND continues to be increasingly qualified into the data center, and density, not units increases, continue to be the important driver for clients mobile and PC devices. We continue to see Lam's addressable markets, net of customer yield and productivity in NAND growing faster than NAND's WFE as customers transition from 64 layers to 96-layer investments starting late in 2018 and into 2019. We expect growth in non-memory CapEx led primarily by logic investments in both leading edge 10 nanometer as well as trailing edge 28 nanometer and above investments, offset by flats to slightly lower investments in leading-edge foundry from reuse, among other strategies. As a headline, 2018 looks to be a very solid year of demand-led equipment investments by our customers within the exciting context of increased prominence of silicon as a foundation of global innovation and presents enhanced opportunity for Lam.
I would like to touch briefly on the topic of capital allocation in light of the passage of recent U.S. tax reform legislation. Doug will address more details later, but at a high level, the guiding principles for capital allocation at Lam remain unchanged. First and foremost, invest in the long-term profitable growth of Lam. Deliver increased optionality to create value from the strengths of Lam with accretive investments in our core and non-core markets. Return excess cash to our shareholders through strategies that include sustaining and growing dividends over time and excess capital redistribution to shareholders as appropriate. We are obviously inspired by our opportunity and proud of our operational execution to date. On this last subject of tax, it is worth emphasizing that we implemented a comprehensive global strategy in 2003. We have compelling value created from this recent legislative action.
Illustratively, strategic optionality is accelerated with cash available to U.S. operations increasing from $1 billion to $5 billion. In the hypothetical event of 100% U.S. repatriation, we save $1 billion of cash from the tax rates previously applicable. 85% of our gross cash balances are now available for repatriation, and that represents $25 per share or 12% of the Lam share price currently. How much and when we repatriate and plans for capital allocation are framed by a broad set of strategic choices, which we will discuss in our upcoming investor and analyst meeting. Unrelated, you will have seen another announcement today about Lam leadership, specifically the promotion of Tim Archer, our Chief Operating Officer, to the position of President and Chief Operating Officer.
This promotion is recognition of Tim's contributions so far on the journey of Lam, it is a commentary on the partnership that we continue to share. Except for some tweaks in emphasis, our fundamental responsibilities will remain unchanged. Our commitment and passion for seeing Lam realize its full potential only strengthens. The full complement of global leaders at Lam have made possible our vision. Tim and I both look forward to the challenges and opportunities in the years ahead. In conclusion, we're optimistic about our opportunity, our strategies, and the potential for continued outperformance for Lam. We look forward to sharing more aspects of our vision with you in March. With that, let me turn the call over to Doug.
Okay, great. Thank you, Martin. Good afternoon, everyone, and thank you for joining us today. We ended calendar year 2017 with strong performance, exceeding the midpoint of our guidance for the December quarter on all of our financial metrics. In addition to the milestones that Martin mentioned during his scripted remarks, during calendar year 2017, we generated over $2 billion in cash from operations. That was an increase of 37% compared to calendar year 2016. We returned $2.193 billion to our shareholders through share repurchases as well as dividends. We're very pleased with what we've achieved this quarter as well as this calendar year. Shipments continued at a healthy level in the December quarter, totaling $2.632 billion, which was a little above the midpoint of our guidance.
Memory shipments increased in the quarter with the combined memory segment representing 77% of total system shipments, that compares with 66% in the previous quarter. Customers continued to ramp NAND and other nonvolatile memory technologies, which made up 53% of the system shipments. This was up from the 49% level that we saw in the September quarter. The NAND investments were directed towards meeting increased bid demand through a combination of capacity additions, 2D to 3D conversions, as well as 3D installed base layer count growth. DRAM shipments made up 24% of the system shipments, this was up from 17% in the prior quarter. This was the highest dollar value of DRAM system shipments since the March 2015 quarter. Strong DRAM pricing supports these investments, which continue to be largely directed towards 1X nanometer conversions.
December quarter foundry shipments were 15% of system shipments, down from 21% in the prior quarter. Foundry spending was primarily directed towards 7 nanometer pilot capability, as well as 10 nanometer volume production projects. Finally, the logic and other segment accounted for 8% of system shipments. That was down compared with 13% in the last quarter. Revenues for the quarter came in at $2.581 billion, above the midpoint of guidance and up 4% from the previous record high level that we saw in the September quarter. Gross margin came in at 47.6%. As I always mention, you should expect to see some quarter-to-quarter variability in gross margins due to multiple factors such as product mix, customer concentration, and overall business volumes. Operating expenses in the December quarter increased to $449 million, coming in at 17.4% of revenue. That compares with 17.7% in September.
Approximately 63% of the OpEx spend in the quarter was allocated to R&D, which was around the same ratio as we've held for the calendar year. While we continue to see opportunities to further leverage OpEx, we will continue to increase our funding of strategic R&D programs to maintain our technology as well as productivity leadership. These investments are critical to meeting our commitments to customers and to our objective of growing the company at a faster pace than the industry. The growth in the size of our served available markets and the market share gains that Martin mentioned are the result of investments we've made and work that we've done over the course of the last several years. Operating income in the December quarter was $779 million, which was up from $733 million in the prior quarter.
Operating margin came in at 30.2%, up compared to 29.6% in the September quarter, again, a little above the midpoint of the guided range. Our tax rate for the quarter was approximately -1%, down compared to 14% last quarter. The primary reason for the reduction in the non-GAAP tax rate is a reduced statutory U.S. rate combined with the delayed implementation of U.S. tax on foreign earnings. This is attributed to the application of the recently enacted tax changes for a non-calendar year filer. The new tax law applies to tax years beginning after December 31, 2017, which for Lam, will be in the second half of 2018 onwards. Excluded from the non-GAAP tax rate is a one-time charge of $757 million. This is primarily made up of tax on accumulated foreign earnings to be repatriated. The cash tax payments will be made over the next eight years.
A tax rate in the mid-single digits for the first half of 2018 and mid-teens in the second half of 2018 would be reasonable for you to include in your models. Based on a share count of about 182 million shares, earnings per share for the December quarter totaled $4.34, above the high end of our guided range. The biggest contributor to the earnings per share being above the range was the favorable impact on the tax rate. Without the impact of the tax rate change, we would've been approximately $0.10 above the midpoint of guidance. The share count includes approximately 18 million shares of dilutive impact on a non-GAAP basis from the 2018 and 2041 convertible notes. I just mentioned, with the increase in our stock price, we've continued to receive requests for early conversions of both the 2018 and 2041 convertible notes.
Conversions that settled in the December quarter totaled $44 million, of which $32 million are related to the 2018 convert and $12 million to the 2041 bond. To date, we've also received requests for conversions totaling $175 million that will settle in the March quarter. I'll remind you, dilution schedules for the updated notional amounts in the 2018 and 2041 convertible notes are available on our investor relations website to help you with your modeling. We continue to execute on our capital return program. As you may recall, we announced in November a new $2 billion buyback authorization. In the December quarter, we spent approximately $1.1 billion on share repurchases, largely through an accelerated share repurchase program, and we paid out $73 million in dividends to our shareholders.
Consistent with our prior comments, we're committed to growing the dividend over time, and in November, we also announced an 11% increase in our quarterly dividend to $0.50 per share, which was payable earlier this month. Let me move to the balance sheet. During the December quarter, we initiated a $1.25 billion commercial paper program, and as of the end of the December quarter, we had $800 million outstanding. We ended the quarter with cash and short-term investments, including restricted cash, of about $6 billion. Approximately $5 billion of the 6 was domiciled outside of the U.S. This is the amount that becomes more flexibly available with tax reform. Cash from operations was $29 million, down from $858 million in the September quarter.
The low cash generation during the quarter was the result of the timing of certain receivables, as well as the growth in inventory to support the strong shipments we see in the March quarter. I would just mention, we collected approximately $700 million at the end of December, which was after our fiscal quarter had closed. Day sales outstanding increased to 80 days versus 56 last quarter. Inventory turns came in at 3.6, compared to four in the prior quarter. I just mentioned, I expect the March quarter will be a very strong cash-generating quarter. Deferred revenue at the end of the quarter were $1,114 million, which was up from $938 million last quarter. This number excludes $289 million in shipments to customers in Japan, which will revenue in future quarters.
Can I just remind you that these Japan shipments remain as inventory carried at cost on the balance sheet? I should mention that we will be adopting the new revenue recognition standard, ASC 606, starting in the second half of calendar year 2018 to coincide with the start of our fiscal year 2019. I'll have more to tell you about this as we get closer to the second half of the year. Company non-cash expenses during the quarter include the following: $42 million for equity compensation, $41 million for amortization, and $39 million for depreciation. Capital expenditures were $85 million, which is up from $60 million in the September quarter. This growth was driven by factory and lab expansions as well as lab tool investments. CapEx for the year came in at $224 million.
We do expect to see a further increase in CapEx in calendar year 2018 as we continue to expand our manufacturing network and support growth in strategic R&D programs. We ended the quarter with approximately 10,200 regular full-time employees. Majority of the headcount was added in the factory and field to support the growth of our business. As Martin mentioned in his remarks, company is enjoying significant business momentum heading into the March quarter. Our non-GAAP guidance for the quarter is as follows. We're expecting shipments of $3,175 million, plus or minus $125 million. We're forecasting revenue of $2,850 million, again, plus or minus $125 million. We're forecasting gross margin of 46%, plus or minus one percentage point. I do expect to see a stronger gross margin percentage in the June quarter as compared to March. We're forecasting operating margins of 29%, plus or minus one percentage point.
Finally, we forecast earnings per share of $4.35 ±$0.15, based on a share count of approximately 181 million shares. Before we move into Q&A, I would like to provide an update on our plans for our Analyst Day. We will be hosting this event on Tuesday, March 6th in New York City. We will be providing additional details regarding the event logistics in the next week or so. Entering 2018, we continue to be pleased with the positive momentum in our business and the progress we are making towards our ongoing business objectives. We remain optimistic in the strong demand for our products and services, and we look forward to sharing a comprehensive update on the company's plans at our Analyst Day in March. Operator, that concludes my prepared remarks. Please open up the call for questions.
Thank you. If you would like to ask a question, please signal by pressing star one on your telephone keypad. If you are using a speakerphone, please make sure your mute function is turned off to allow your signal to reach our equipment. Again, it is star one. We will take our first question from C.J. Muse with Evercore. Please go ahead.
Good afternoon. Thank you for taking my question. First question, was hoping you could provide some help around the cost-down efforts in the NAND market. As you think about moving from single-stack to multi-stack QLC, perhaps some other techniques that you are helping your customers with, would love to hear how you think about the bit guys and their ability to drive down their costs such that the hard disk drive market becomes achievable to penetrate. That would be very helpful.
That is a comprehensive question, C.J. Maybe I should start with our assumption that we communicated in the analyst meeting a year ago. We actually, when we articulated our perspective on the opportunity for Lam, and we gave this reference of $70 billion of investments, our assumption was actually that SSD would not be disruptive to hard disk drives. Our assumption was that the demand for storage and memory in the world would essentially be serviced by this new mode of non-volatile memory, or this new product of non-volatile memory, and that hard disk drives would just kind of be capped out. That was the modeling assumption that we presented to you. To the extent the scenario plays out in the way you articulated, you should presume that is kind of upside to the commentary of the company because of what I just said.
NAND is probably the poster child for a market with elasticity of demand, the industry has universally adopted a new technology. It is scaling that new technology in terms of planar to 3D transitions and layer counts. When our customers make investments, they make it because they have performance enhancements on device, and they have cost roadmaps that create opportunities for incremental demand to get kind of created. There's an awful lot I could say about the subject. The headline that I hope everybody retains here is we think the world of silicon, and we think the world of non-volatile memory specifically has tremendous opportunity in the years ahead as a byproduct of this innovation that's emerged around the value of data in the world.
The good news for Lam, from a lot of hard work and a lot of partnerships with our customers, we positioned a portfolio of products and a product pipeline that's focused on enabling that roadmap.
That's helpful. Appreciate it. I guess as a follow-up, as you think about your 1.7x outperformance in 2017, it looks like mix is going to be there for you again in 2018, would love to hear your thoughts in terms of how, whether you could rank order or not, how you would think about perhaps gains in share and logic, service growth, and other areas where you're seeing increased patterning, that will drive faster than expected industry growth for you guys. Would love to hear your thoughts on that. Thank you.
Thank you. Appreciate the question, C.J., maybe I can just restate something I said in my prepared remarks, because it's a very important disclosure from the company relative to our focus, our definition of success and our priority. I share that because I have some anxiety that the investment community gets trapped in fairly incomplete disclosure from the totality of the industry on market share, hence our focus. Our focus is the following. Our primary objective is to promote our customers' success through an increased proportion of their spending that we actively compete for, and through differentiation to win and sustain more of that business over time. That means that our focus at a product line level and a business unit level and a segment level has to be on the share that we have.
What really counts is where that share is. Our focus is on the most critical applications, the most valued applications for our customers, and the sustainability of that share through the economics that are associated with this. The most basic measure of performance for any company in any market is the one that I characterized, and we will continue to answer these questions on market share by a focus around that kind of basis. Our headline for last year was 21% of WFE against 18.5 in prior year, against 14.2 in the first year after bringing Lam and Novellus together. Momentum is really good. The technology inflections still have tremendous opportunity for us.
More or less, the headlines that we communicated in NAND Flash at the analyst meeting are only, I would say, a little better one year on. We'll spend more time talking about that in a few weeks. In the world of patterning, I think the headlines that we communicated in our analyst meeting a year ago are as valid today as they were then. That includes the various scenarios of adoption of EUV. That includes the various technology advancements that have played out since that time in multi-patterning process flow.
Thanks, C.J.
We'll take our next question from Romit Shah with Nomura. Please go ahead.
Yes, thank you, and congratulations. The March shipment number is obviously a very big number, and I feel like there could be some skepticism around that shipment number and how strong it is and whether or not that's really sustainable. So I was hoping you could comment on just if we look throughout the year and just triangulate between the outperformance you delivered in 2017 and your outlook for low double-digit WFE growth this year, it would seem like the March shipment guidance is a number that you could sustain throughout the year. I wanted to get your feedback on that.
Yeah. We're not in the game of guiding every quarter of the year or even the year with any specificity. To the spirit of your question, I hope there isn't skepticism around the quarter itself, because I hope at this point we've got a track record of actually delivering on the kind of quarters commitments relative to sustainability. Obviously, this is all about fundamental demand, and if you sign up for the conventions that I described in my prepared comments, the convention that silicon has established itself quite differently in terms of potential and a world of data, than you would sign up for sustainability of investment. Now, you might get ebbs and flows one quarter after another or one month, but the fundamental headline that we believe in and we're investing to create value in, is a scenario of sustained and disciplined investment by our customers.
At a more discrete level, just to put some more substance on it for you, our expectation is that, relative to calendar 2018, we see it as not perfectly, but a reasonably balanced year for the industry and for the company, both. I would say that we see memory could be slightly first-half biased, and foundry could be slightly second-half biased, with logic as fairly neutral in the year. As was proven last year, everything I just said could be entirely wrong, and probably is precisely wrong, but it's the best we have for you today in terms of trying to give you some color around quarter to quarter or half to half.
I appreciate it. I'm curious, just as my follow-up, how China plays into your forecast for this year, because there's obviously a lot going on between domestic and global and memory and foundry. How does China play into your outlook?
China, obviously for the domestic community and the global players that are investing there is a big deal, and the headlines in terms of new fab projects are exactly the same today as they were three months ago when I detailed them. The only incremental disclosure today, obviously, is there are more projects that we're tracking. When we look at WFE investments in calendar 2018, we think that the domestic community combined all segments. The domestic community in China, DRAM, flash, and logic foundry probably invest slightly less than $5 billion, which is a little less than 10% of global WFE in the way we've described it to you. That's up between $1 billion and $2 billion, maybe $1.5 billion compared to calendar 2017. It's an increasingly important component of spending.
It's a big focus for us in terms of building infrastructure, supporting our customers' success and their visions, and putting in place qualified people to support them. One last piece of data relative to our shipments. When we look at the Lam shipments in China, two-thirds of the shipment dollars will be probably for the global players investing there, one-third for the domestic community.
Helpful. Thank you.
Thanks, Romit.
We'll take our next question from Mehdi Hosseini with SIG. Please go ahead.
Well, thanks for taking my question. I want to go back to some of the structural changes, especially with 3D NAND. I want to get your view how we should think about a migration from 64 to 96. In that context, would a new fab and a change in the layout have any impact in architecture, especially as 3D NAND could lead to higher manufacturing cycle time, so perhaps a new clean room could help with that, versus converting existing planar NAND? Is there anything or any changes between these two set of variables that we should add to our thesis?
Yeah. I don't know I would characterize the kind of line layout complexities from 64 to 96 much differently than I would from the other versions that have preceded this. I don't think I would characterize it as more challenging or complex than a planar as a 3D conversion. I think your question is a very important one and there are, I would say, fairly dynamic customer plans around how best they service the opportunities in the marketplace that they see. We would expect to continue to see this portfolio of investments from greenfields to the planar to 3D scaling and 3D scaling vertically itself. We are, as we've said a number of times, we're agnostic actually in terms of the opportunity between the planar 3D conversion and the new fab addition from a business perspective.
I would expect, based on what I know today, that somewhere in the range of two-thirds of the investment by the industry is conversion related and maybe a third is additions. That's a proxy that will change for sure as discrete opportunities show up for customers and as they optimize their plans.
Sure. Just quickly as a follow-up, as we think about the longer term, how should we think about the service component, especially with some of the new areas like ALD? Would that carry different service attach rate versus the rest? I know Analyst Day is coming up, but is there anything you can share with us?
Yeah. Probably it's a wait to the Analyst Meeting. There are some very important headlines, I think, for the company around the pace of growth in excess of the installed base and the product portfolio and the segments that we're growing are part of that story. The emphasis that we're investing in around advanced services is part of that story. I think there's something quite unique about where we are in the ecosystem that presents opportunity as well. It's definitely an increasingly valuable annuity to the company, and it's an increasingly important part of creating and delivering value to our customers.
Got it. Thank you.
Thanks, Mehdi.
We'll take our next question from Joseph Moore with Morgan Stanley.
Great. Thank you. I wonder if you could characterize the service and supplies business, and it seems like that's been consistently holding the same proportion of your overall revenue. How is that business kind of growing as fast as the systems business, and how sustainable do you think that growth is?
It's actually been growing faster than the installed base, which is a kind of statement of the breadth of the product and services portfolio that we're making available to our customers and we're investing in creating. Dialogues around artificial intelligence and machine learning are as relevant inside of our company and as relevant in terms of developing products and services for our customers as it is for our customers, and in turn, their customers and the consumers. Don't disconnect the AI and the big data analytics headlines from the realities of our company as well. It's a broadening portfolio. It's growing faster than installed base.
In a period where we've had extraordinary growth in our systems business, it's a testament to the strength of our customer service business group that we're able to say what we're saying and preserve, in essence, the same percentage of revenues coming from that business. What that means in terms of modeling purposes, if you happen to choose that you're going to model less growth in calendar 2018 than was true in 2017, which is essentially what we've guided, then the value of that installed base business kind of goes up over time. I think it's a great commentary on value, sustainability, and quality of earnings.
Joe, I always tell you, it's my favorite part about the company's business model. It's an annuity that just keeps going, keeps growing, keeps generating enormous amounts of cash. It's a great part of what we do.
Great. Thank you very much.
Thanks, Joe.
We'll take our next question from Toshiya Hari with Goldman Sachs. Please go ahead.
Great. Thanks for taking the question, and congrats on the results. Martin, I had a question on EUV. I guess as we approach the timing of insertion, particularly in the foundry space, and you learn more about the recipe that your customer might follow, has your view on your opportunity set as it relates to patterning changed over the past 6-12 months, or has it remained kind of the same?
It's remained the same.
Okay. Fair enough. The second one, on the tax rate. Doug, I think you guided fiscal year second half, the tax rate to the mid-teens. Should we consider that rate as the new normal for the company, or should we expect some variability going forward post tax planning or whatever it may be?
Yeah, Toshi, if I was modeling out beyond the second half of 2018, I'd probably keep it pretty steady with what I gave you, which is mid-teens for the second half. It always changes a little bit depending on where the geographic distribution of our business ebbs and flows, so it's not going to be precisely that. That would be the best planning horizon or the planning number that I'd give you. Just to remind, make sure everybody heard, first half of calendar 2018, mid-single digits. Second half, mid-teens.
Thank you.
Yep. Thank you.
We'll take our next question from Farhan Ahmad with Credit Suisse.
Farhan, I think you're on mute.
Hello, can you hear me?
Yep, we can hear you now. Go ahead.
My first question is on memory in NAND DRAM. What level of bit growth do you expect from the CapEx outlook that you have for the year?
I think we kind of sign up for the consensus on bit growth. The low 20s for DRAM and mid to high 40s for NAND for calendar 2018. Obviously, the server phone content in DRAM is a decent kind of content headline. The same is true for the smartphone bit growth for NAND. I think one of the illustrative headlines that we should all be conscious of is the average smartphone today has got maybe 4 GB of NAND content and the best phones have 256. 40, sorry, I should have said 40 GB of content and the best phone's 256. When you think about applications innovation, as you think about the content headlines, not the unit headlines, that's one of the most fundamental drivers for the business prospectively.
Got it. Then, one question on the services. You had very strong growth in your systems business this year. Does that mean that we should see much stronger growth in your services business over next one year as these systems come off warranty?
Well, I'll say the objective of the company, as we've said many times, is to grow the service business with spares and service and upgrades business faster than we do the installed base. You're right to say there are warranties. When warranty periods end, you roll into that space, and we try to position for very strong market share and sustainability. I would tend to lean towards a very positive answer to your question.
Thank you. That's all I have.
Thanks, Farhan.
We'll take our next question from Harlan Sur with J.P. Morgan.
Good afternoon, congratulations on the well-executed quarter and the outstanding performance in 2017. You guys have driven great leverage on the OpEx with the scale and the revenue growth. If we assume that WFE is sustainable above $40 billion-$45 billion-plus going forward, seems like 28%-30% Op margins, plus or minus, being sustainable. More and more investors are focused on free cash flow and free cash flow margins. How should we think about the sustainable free cash flow margins at, say, the 28%-30% operating margin level?
Yeah, Harlan, we're front-running an update to the model, which we'll give you in early March. Generally speaking, we would expect operating cash flow to be close to operating income. Now, if you look at what happened in the last year, the business grew so rapidly that we actually did less well than that because the consumption of working capital as the business scaled up. We needed to build more inventory. Receivables grew along with the level of revenue, so we didn't do quite as well as that. On an ongoing basis, assuming kind of a steady state of business shipments and revenue, that's about where we should be. Our CapEx is nominally 3%-4% of revenue, so that's your free cash flow answer there.
Got it. Thanks for the insights there. Shipments are up 40% in December versus the year-ago quarter, versus two years ago. They'll be up strongly again in March. Obviously, my congrats to the operations, logistics, and manufacturing teams. Just given the significant strength, are customer lead times increasing, and any way to quantify the increase?
Yeah. There's a way to quantify everything. We don't disclose lead times for obvious competitive reasons and respect for the roadmaps of our customers. What I will tell you is, obviously, it takes a lot of work to do what we're doing. It's much more difficult. I think for those of us that have participated in the scale of growth that we've had, we've all been very challenged
By that scale of growth. I would say that from a benchmarking perspective, as best I can tell, which is why I spoke to flexible business model and commitment from employees in my prepared comments, we've executed as well, if not better than most or maybe all. It's a big focus. It gets more difficult and certainly 50% growth here is quite extraordinary challenges for operations and also field resources.
Yeah, I'm always amazed, Harlan, when I sit through the quarterly business reviews with the supply chain, the global operations team, as well as the field, just how well they're executing. It's been actually quite amazing when you see how it's all happened.
Yep.
Thank you.
Thanks, Harlan.
We'll take our next question from Sidney Ho with Deutsche Bank.
Thanks for taking my question. My first question is on DRAM. There has been a lot of chatters that the strong DRAM that the market expects and what you alluded to for next year could be because of capital intensity of DRAM going up, but it could also be yields are suboptimal. What is your view on that?
Well, my basic view is, I don't think customers spend money if a yield is suboptimal. I think that's a very important kind of test for them, making their investments and I guess that's all relative. We have characterized, obviously today, that the relative investments the industry has to make are a little higher for a bit out in the context of the technology conversions that are playing out today and this year compared to last year and the year before. I would also make the point that it's still an investment that is dominated by conversion. It is still an investment that is made with extraordinary discipline by our customers managing supply and demand balance. I would also say that the capital intensity data point is actually a lot less than the points in history that have triggered kind of corrections.
I think if you look at memory today, the capital intensity for the last four years, including the outlook for calendar 2018, is not even two-thirds of the level seen in 2007. It's a dramatically different profile of risk.
Okay, that's helpful. My follow-up question is, if you look at your forecast for WFE, what are some of the biggest swing factors in your forecast that can move above or below your forecast of low teens? Just wanted to see if I can get something out of you to be more precisely wrong.
Nice, Sidney. I appreciate that.
Good luck with that. Well, macro positive, negative is always there and always will be. This is a pretty extraordinary time for the industry, I think holistically, because there's an irony here of consolidated participants, and a diversified kind of statement of demand. It gets maybe more difficult, not more easy to forecast the future at a device demand level. You have an industry that is, I think, singularly invested in trying to take advantage of this opportunity and contribute to broader tech economy and broader innovation. It's all about creating incremental demand and the memory storage, data center, AI, big data, analytics, advanced process. It's all relevance, right? We live in a world where devices and silicon that support connectivity are critical to the roadmap.
Devices that create the capacity to store efficiently and cost effectively all that collected data are critical. The high-performance computing that is relevant to take all of that data and do something with it that creates value either in an enterprise or in a consumer context has to kind of show up. I think it's a very integrated system, which is perhaps why I answer the question maybe less precisely than you would like me to, because I think we live in a world where this is much less about discrete segments of a semiconductor industry as we traditionally described it, and much more about the integration of technologies from a chip integration point of view, from an embedded memory and logic solutions perspective.
We live in the world of systems design, systems architecture, it all has to be there and all is part of answering this question around incremental opportunity or incremental risk.
Okay, great. Thank you very much. Congrats.
Thanks, Sidney.
We'll take our next question from Craig Ellis with B. Riley. Please go ahead.
Yeah, thanks for taking the question. Martin, I appreciate the comments on the positioning for 2018's memory spending and DRAM starting to catch up and balance with NAND. Could you extend your view a little bit further and comment on maybe the multi-year trend for DRAM? I know the company's typically been comfortable talking about longer term, such as your five-year NAND spending view. Could you give us a longer-term look with what you see at DRAM?
I thought you started with NAND and you ended with DRAM.
You kind of moved around
Which one did you want to talk about?
DRAM. NAND was simply used as an example.
Okay
something you've done in the past for the longer-term view.
I see. The first thing I'd say, it's a content story, not a unit story. I don't think we're going to stand up and say, "Hey, we got this great view that all of a sudden PC units and cell phone units take off again," because I don't think that's the headline. It's all about content. There are two drivers of bit growth in DRAM. One of them is the server, which is the most comprehensive demand driver, and the second is expansion of phone content. I would say, I'd probably be stronger than conservatively optimistic. I'd say we are optimistic about those long-term trends, and we think they're fundamental to answering the questions about investment in silicon capacity and the investment that we make in our business.
That's helpful. The follow-up, I'll shoot it over to Doug. Doug, as you look at the company and the flexibility that you have now with cash, can you talk about what that means for debt levels and capital structure, and if we should expect that that would stay at current levels or if there's potential for whatever reason for debt to rise from here as a percent of your capital structure? Thank you.
Yeah, Craig, I'm going to decline to answer that one, but I will promise you that when we get to the March analyst day, we'll have lots more to tell you. As you might appreciate, all the tax changes are kind of late-breaking. We're talking internally about it, debating some things. We'll have more to tell you in early March, so I'm going to punt on this one for now.
Can I take a swing at another one then?
Yeah, go ahead.
Okay, I'll pop it back to Martin. Martin, this is more of a maybe I'll have better success. More of a qualitative question. I'm interested in knowing the degree to which your discussions with customers have changed over the last few years as capital intensity has stepped up dramatically and as there's more signs that industry is on a steadier growth profile rather than boom bust. What's changed with the duration of the discussion, the way that you may be collaborating across different types of technologies, et cetera?
Yeah. Actually, maybe that last kind of sub-question is actually a great response. The scale and scope of technical challenges and business challenges for everybody in this ecosystem is much more difficult today than was true five years ago. The companies that I think excel and sustain in the long term have attributes of partnership and collaboration and codependency that are quite fundamental. That's why almost every single quarter I say something about culture and values because that's a fundamental element of competitive differentiation. If you are not trusted, if you are not legitimately collaborative, if you're not sincerely invested in the success of your customers, then your opportunity to sustain your business is limited.
Thanks, guys.
Yeah, thanks.
That does conclude our question and answer session for today. I would like to turn the conference back over to Satya for any additional or closing remarks.
Yes. Thank you once again for joining us. We look forward to seeing you at our March Analyst Day in New York. We'll be sending you more details shortly on that. Thank you.
Once again, that concludes today's presentation. We thank you all for your participation, and you may now disconnect.