科磊 (KLAC) 2017 Q4 法說會逐字稿

(technical difficulty)

operator today.

At this time, I would like to welcome everyone to the fourth quarter fiscal year 2017 earnings conference call.

(Operator Instructions) Thank you.

Mr. Ed Lockwood with KLA-Tencor Investor Relations, you may begin your conference.

Thank you, Christine.

Good afternoon, everyone, and welcome to our conference call.

Joining me on our call today are Rick Wallace, our President and Chief Executive Officer; and Bren Higgins, our Chief Financial Officer.

We're here today to discuss quarterly results for the period ended June 30, 2017.

We released these results this afternoon at 1:15 Pacific Time.

If you haven't seen the release, you can find it on our website at www.kla-tencor.com.

A simulcast of this call will be accessible on demand following its completion on the Investor Relations section of our website.

Today's discussion of our financial results will be presented on a non-GAAP financial basis, unless otherwise specified.

A detailed reconciliation of GAAP to non-GAAP results can be found in today's earnings press release and in the investor presentation on KLA-Tencor's investor relations website.

There, you'll also find a calendar of future investor events, presentations and conferences as well as links to KLA-Tencor's SEC filings, including our annual report on Form 10-K for the year ended June 30, 2016.

In those filings, you'll find descriptions of risk factors that could impact our future results.

As you know, our future results are subject to risks.

Any forward-looking statements, including those we make on the call today, are subject to those risks, and KLA-Tencor cannot guarantee those forward-looking statements will come true.

Our actual results may differ significantly from those projected in our forward-looking statements.

With that, I'll turn the call over to Rick.

Thanks, Ed.

We're very pleased to report that June marked yet another exceptional quarter for KLA-Tencor, with shipments, revenue and earnings per share each finishing at or above the range of guidance.

New orders exceeded the $1 billion mark for the second time in the past 3 quarters.

Shipments were a record $971 million, and we ended June with backlog of $1.8 billion, which was also an all-time high level for the company.

These numbers were once again driven by KLA-Tencor's ongoing market leadership and strong execution of our strategic objectives and are also reflective of the broad strength of today's overall WFE industry environment.

One factor that may not be evident in the recorded numbers, but which is key to our positive outlook for continued strength in the overall demand environment, is the growing momentum in orders we've experienced in the first half of calendar 2017.

In fact, since January, we have adjusted K-T's internal order forecast for CY '17 to the upside by almost 20%, reflecting strengthening customer demand driven by new products across each of our end markets and leading to another excellent year of growth for the company in 2017, while setting the stage for continued momentum into 2018.

Judging by the positive announcements recently of overall industry growth projections provided by our peers and other market participants, the strength KLA has experienced today is also broad based, which bodes well for the industry as we move ahead in the year.

Turning now to a discussion of some of the key areas of near-term focus for the company as we execute our growth strategies.

As a market leader in process control, K-T plays an enabling role in driving our customers' growth strategies and long-term success.

We do this by investing at a high level to drive innovation in process control and supply our customers with the advanced inspection metrology products and services that they require to advance their device technology road maps and also resulting in expansion of the serve market opportunity for KLA-Tencor in the most critical applications in process control.

Along with the ongoing focus on investment and innovation, we're also continuously executing our strategies for operational excellence and converting our growth and market leadership into industry-leading margins and cash flow, providing the fuel for future investment and resources for the ongoing strong cash returns to stockholders.

Drilling down a bit on the factors that are driving KLA-Tencor's performance, I'll focus on 3 key drivers of our growth.

First, expansion of the serve market opportunity in memory, driven by new products.

Today, we are developing 5 new technology platforms, 2 of which are targeted specifically at memory.

And we're expecting higher process control adoption and record revenue from memory customers in CY '17.

Growth from memory customers is already contributing to the record business levels we're experiencing in our bare wafer inspection and metrology markets, and we expect revenue from memory customers to continue to grow with the market.

Second, growth in China, which is seen by many as the next major new market inflection for WFE.

We believe China could represent a generational opportunity in terms of investment and growth for our industry.

Business levels in China for KLA-Tencor are expected to more than double in CY '17.

China represents a tremendous growth opportunity for KLA-Tencor and for process control, as our customers in this region are adopting process control at a high level to speed yield learning in their development efforts and accelerate market entry.

And finally, services.

Growth in the service business is historically driven by the expansion of our installed base of systems in the field.

And this represents an annuity growth model to complement our systems business.

Services growth is also driven by expanding investment in the trailing edge.

This is resulting in legacy node fabs running at full utilization and putting pressure on the equipment base to service, refurbish, remanufacture and upgrade process control systems as our customers continue to derive value from these assets.

Quarterly service revenue topped $200 million for the first time in the June quarter, and we're modeling an annual revenue growth rate of 7% to 9% in services for the foreseeable future.

To sum up today's results before I turn the call over to Bren, June was another record-setting quarter for KLA-Tencor, demonstrating the company is executing well and the strong industry environment.

Given our backlog and with the momentum we are experiencing in the marketplace, 2017 is shaping up to be an outstanding year for us.

We're delivering a steady cadence of leading products, expanding our serve market and delivering strong profitability and cash flows.

We're growing our revenue with memory customers, and we expect to see very strong growth in China, as customers view process control as a competitive advantage in their market growth strategy, and they look to KLA-Tencor as the market leader and partner.

I'll now turn the call over to Bren for his comments.

Bren?

Thanks, Rick.

And good afternoon, everyone.

As Rick highlighted in his opening remarks, the June quarter represented another outstanding period of financial performance and operational execution for KLA-Tencor.

Shipments were a record $971 million, finishing above the range of guidance.

And revenue and GAAP and non-GAAP diluted earnings per share each finished above the midpoint of the range of guidance in the quarter.

This result was driven by strong demand across our product portfolio as well as solid execution in cost management in our engineering, manufacturing and service operations.

Revenue was $939 million in the June quarter.

GAAP diluted earnings per share was $1.62 in the quarter, and non-GAAP diluted earnings per share was $1.64.

In our press release, you will find a reconciliation of GAAP to non-GAAP diluted earnings per share.

With the exception of when I explicitly refer to GAAP results, my commentary will be focused on the non-GAAP results, which exclude the adjustments covered in the press release.

Now turning to highlights of the June quarter demand environment.

Although we have discontinued guiding quarterly orders, beginning this quarter we will begin to offer greater detail on our shipment results and guidance to provide more information and color on the current business environment to give investors insight into industry trends and KLA-Tencor's performance.

For historical shipments mix data, please refer to the supplemental information posted with today's press release on our website.

So as a reminder, the following details are related to shipments.

Total shipments in the June quarter were a record $971 million, up 7% on a sequential basis and finishing above the guidance range.

Foundry was 64% of shipments in June, driven by an anticipated broadening of the customer base for investment in 10-nanometer production and 7-nanometer development and by continued investment in legacy technology nodes.

We are currently modeling foundry shipments to be approximately 42% of the total in the September quarter.

Memory was 32% of shipments, with delivery evenly split between DRAM and NAND.

We are currently modeling shipments to memory customers to be about 43% of the total in the September quarter, with NAND representing about half of the memory mix.

Logic was 4% of shipments in June and is currently forecasted to be approximately 15% of the September quarter total.

In terms of the approximate distribution of shipments by product group.

Wafer inspection was 50% of shipments.

Patterning was 25%.

Patterning includes orders from our reticle inspection business.

Non-semi, which includes our back inspection business, was approximately 4%, and service was 21% of shipments.

Looking forward, we are modeling September quarter shipments to be in the range of $945 million to $1.025 billion.

Current build plans are supporting quarterly shipment levels of $900 million to $1 billion, and we expect this trend to continue into calendar year 2018.

This outlook has strengthened since the earnings call back in January, consistent with the upside in orders we've experienced year-to-date.

In fact, compared to the outlook we provided back at our earnings call in April, the CY '17 order forecast is over $225 million higher than 3 months ago.

And shipments for the second half of calendar year 2017 are expected to be up mid-single digits compared with the first half, inclusive of the stronger-than-guided performance in the June quarter.

Turning now to the income statement.

Revenue was $939 million in June, finishing at the upper end of the range of guidance.

We expect revenue to be in the range of $910 million to $970 million in the September quarter.

Non-GAAP gross margin was 63%, in line with expectations for the quarter.

The strong gross margin performance in June is consistent with recent margin trends in terms of mix of product business and operating leverage in our manufacturing and service operations.

Looking forward to the September quarter, we expect gross margin to be in the range of 62.5% to 63.5%, flat at the midpoint compared with June, as we expect a similar mix of product revenue, with slightly higher manufacturing costs, offset by an improved service mix quarter-to-quarter.

Total non-GAAP operating expenses were $238 million in June, and non-GAAP operating margin was 37.6%.

This operating margin result is in line with the updated target model for annual revenue levels in the $3.6 billion to $3.9 billion range, as we outlined earlier in this quarter.

We are modeling operating expense levels to be approximately $250 million in the September quarter due to increasing R&D headcount and prototype material expense for current programs as well as approximately $2 million to $3 million of incremental operating expense associated with the acquisition of Zeta Instruments, a privately held company that designs and manufactures optical profilers and defect inspection systems for the advanced wafer packaging, LED and data storage industries.

We completed this small transaction in early June.

Our non-GAAP effective tax rate was 21% in the quarter, just below our long-term planning rate of 22%, reflecting the higher mix of revenue from products developed or manufactured offshore and other discrete items impacting the tax rate.

We are modeling the September quarterly tax rate at 20% due to a nonrecurring tax benefit expected in the quarter.

However, you should continue to model a long-term tax rate of 22% going forward.

Finally, non-GAAP net income for the June quarter was $259 million, and we ended the quarter with 158 million diluted shares outstanding.

I'll turn now to the highlights on the balance sheet and our cash flow statement.

Cash and investments ended the quarter at $3 billion, an increase of $313 million compared with the March quarter.

Cash from operations was a record $463 million in June, and free cash flow was $452 million.

As mentioned earlier, we completed the acquisition of Zeta Instruments in the June quarter.

Total purchase consideration was approximately $37 million, including cash paid of almost $32 million at closing.

And finally, during the quarter, we paid an aggregate of $85 million in regular quarterly dividends and dividend equivalents for fully vested restricted stock units.

In conclusion, KLA-Tencor's results in June reflect our market leadership, the critical nature of process control in our customers' growth strategies at the leading edge and in legacy design rules and our industry-leading business model.

This, fueled by a record total backlog of $1.8 billion as of the end of the June quarter, position the company for another year of growth in a strong and stable overall WFE industry environment.

This performance demonstrates the company's market leadership, the strong customer acceptance of our portfolio of solutions addressing the most critical yield requirements of leading edge and our focus on operational execution across our business.

With that, our guidance for the September quarter is shipments in the range of $945 million to $1.025 billion, revenue between $910 million and $970 million and GAAP diluted EPS of $1.48 to $1.72 per share as well as non-GAAP diluted EPS of $1.50 to $1.74 per share.

This September quarter EPS guidance range assumes a 20% tax rate in the quarter.

This concludes my remarks.

I will now turn the call back over to Ed to begin the Q&A.

Okay.

Thank you, Bren.

At this point, we'd like to open up the call to the questions.

(Operator Instructions)

All right.

Christine, we are ready for the first question.

(Operator Instructions) Your first question comes from the line of Farhan Ahmad from Credit Suisse.

My first question is regarding EUV and what trends you are seeing in the business there.

ASML recently talked about potentially using single lithography edge, which I would assume would require a lot of inspection as well because there is no pattern inspection.

So maybe the need for the wafer inspection is at a higher level.

So maybe can you just talk about EUV trends that you are seeing in the business.

Yes.

Farhan, it's Rick.

Thanks for the question.

We do see a lot of work around EUV still from our customers working in the pilot phase and the technology development phase but a lot of work whether it's in reticle qual, which is still something that there's a lot of challenges getting defect-free reticles, and then in terms of lithography qualifications when they're testing out the litho cell.

But as you know, we're ways off from production, so it's pretty hard to anticipate exactly what the production use cases are going to be.

But we've certainly seen an increase amount of interest in our customers and kind of continuing our belief that we'll see HBM by 2019 and 2020.

Got it.

Rick, you mentioned that you guys are working on 5 new platforms and 2 of them specific to memory.

How should we be thinking about the ramp of these products?

And are any of these products something that will affect the revenues in 2018?

2018, probably by the mid, later part of calendar year, we'll see some impact.

I think -- we've talked before about the overall increase in memory intensity.

And I think if we look out, we're modeling about a 200 basis point increase in the intensity as we ramp these new products over the next couple of years.

We'll already have a record for memory in calendar '17, and we continue to see strength as we go forward.

Farhan, it's Bren.

I think the only thing I would add to that is we're still waiting for these products and that contribution.

But we have from a 3D NAND perspective, from planar NAND, we have seen intensity improve.

If you think about the peripheral products around bare wafer inspection and metrology, where wafer flatness is really, really critical to advance layer development in 3D NAND, film measurement, CD measurement, laser scanning, opportunities around defect inspection, some of the bigger challenges are still out there that these platforms hopefully will address.

But we're encouraged by what we're seeing so far.

And as Rick said, it's a big part of our calendar '17 view.

Your next question comes from the line of Harlan Sur from JPMorgan.

As it relates to the financial model, if I look at your free cash flow generation for the quarter, it was -- was it 400 -- what was -- was it $452 million in free cash flow for the quarter?

Sorry, if I missed that number.

Yes, that's correct.

It was pretty high.

I mean, as you know, free cash flow can be pretty lumpy.

But the linearity of shipments enabled a record collection quarter, so that was a big part of it this quarter.

So we're real pleased with it and...

That's a 48% free cash flow margin.

So maybe you can help us level set, Bren.

So in an environment where you're sort of operating sort of 36% to 38% operating margins, how do we think about the free cash flow margins underlying that?

Well, it's a great question.

And I think as our business has been so stable at these levels, we're not having to make significant investments in working capital.

Obviously, we invested a lot in ramping inventory to prepare to be able to ship it, $900 million to $1 billion, where we are today.

And so now you're seeing modest increases in inventory to support that activity.

So then as a result of that, given our capital position and so on, we see a fair amount of the operating margin dropping through.

So the way I think about it, I mean, when I look at calendar '17, I see free cash flow probably in excess of $1.1 billion, and we're talking about revenue levels of $3.6 billion to $3.7 billion if you take our guidance and expectations around the December quarter.

So you're in the 30th percentile range is -- I think if I do the math quickly.

So I think that's probably how we ought to think about it.

I think if you see an inflection, obviously, we'll have to invest into that.

But the resiliency of the model is pretty strong.

The margin profile of the business is good.

And I think that a lot of the trends that I've outlined around gross margin is, I think, are -- is fairly sustainable going forward.

So we feel pretty good about the model moving forward and the cash flow generation that comes from it.

Yes.

And then on China domestic memory.

There are a number of programs that are going to start to put up either development lines or reproduction capabilities in 2018.

You guys typically have a bit longer lead times.

So I'm wondering if you're starting to see orders or at least tools showing up in your second half pipeline to support some of these early China-based memory programs.

And as you've highlighted previously, China's spending intensity for process control in general tends to be a bit higher.

And I'm just wondering, on these new memory programs are you seeing that sort of higher spending intensity level?

Yes.

Harlan, as we mentioned in the call, the orders, the new bookings for the quarter were very strong, driven in part by activity in China, as you say, in anticipation of having deliveries late calendar or maybe early in '18.

So yes, we are seeing that and agree with your suggestion.

Part of what we see overall in the strength of China going forward is broad participation across the board.

And these projects are absolutely real, and they're backed with funding, and we're pretty excited about the opportunity there.

Yes.

Harlan, I think the only thing I'll add to that is, is that the first part of -- well, all of '16 and part of '17 was very foundry-centric.

We thought most of the memory bookings we would start to see in calendar '18.

One of the things that we've been encouraged by is the strength of what we've seen from an order perspective both in the June quarter but also what's in the funnel over the next couple of quarters.

So those are -- there are tools that are slotted to ship early in 2018.

So from a shipment and revenue perspective, they're into the next year.

But as Rick said, those projects are clearly real and are moving quickly, and we're preparing to ship into that in the first part of next year.

Your next question comes from the line of C.J. Muse from Evercore.

I guess, first question.

As you think about China and their emergence as a key spender here on equipment, historically, you guys typically see roughly half your revenues early on around R&D and then half as you start to ramp.

Curious if that's how you see spending patterns there or might it be different?

Well, I think that there's not as much R&D, first start with that.

A lot of these are established processes.

So I think we're seeing it maybe more pilot and then when you're talking about the ramp-up.

So not quite the same profile as you would have in maybe a more established company.

Most of these are ramping trailing edge or not leading technologies.

So a little bit smoother, I think, from that regard but definitely front-end loading the metrology capability just to be able to ramp the equipment.

So for sure, we're going to be one of the earlier ones to see it.

And then, I guess, as they ramp, would you expect that ramp to be similar to what you see with nondomestic Chinese customers?

Probably a little stronger, simply because the scale of these fabs tends to be smaller.

And I think from the standpoint of -- on a relative basis, I think that you would see the concentration being higher.

I mean, when our participation really slows is when you're talking some mega fabs that have gotten many, many years of experience.

These fabs are newer, and they're greenfield, so -- the ones we're dealing with right now.

So I wouldn't expect it to fall off the way it might have in a more traditional [mega fab] -- giga fab.

That's great.

That's helpful.

As a follow-up, as you think about calendar '18, it looks like you're working from a backlog of 7 months today.

And other front-end guys, I presume, including yourself, have fairly good visibility into Q1.

How are you thinking about the trajectory for revenues into 2018?

Yes.

C.J., I mean, it's a little early.

But you're right.

I mean, given the backlog position we have and what we expect in terms of the order outlook over the next 6 months, we'll see where WFE is.

I think we've got a view that a lot of the dynamics that are driving the industry today continue into next year.

Obviously, NAND flash will probably be a higher level of investment next year.

China's probably bigger.

DRAM is probably flat to a little bit lower.

Foundry is probably on the margin a little weaker.

So I think as we look at all that, we see sort of this continuation of these trends.

And so we're sizing the company and modeling similar levels of output.

So I don't think it's -- right now, I'm not seeing anything that leads me to believe that it's not flattish or a little bit better than that.

Your next question comes from the line of Romit Shah from Nomura Instinet.

I also wanted to ask about WFE spending.

Last fall, it seemed like the industry wasn't contemplating anything more than $35 billion WFE through -- really through next year.

And today, we're on track to see more than $40 billion in spending in 2017.

So basically you've seen, over the course of 9 months, an incremental $5 billion plus in WFE relative to what people were thinking back then.

And I know that you along with your peers are positively biased for next year.

But I guess my question is how do you get comfortable that the increase that we've seen so far over the last few quarters hasn't been borrowed from the future?

Yes, that's a great question.

I think a good way to think about it is where it's coming from.

One of the biggest changes in the profile is relative to what's happening in China overall.

And that investment is not really against leading edge.

And so that's less about coming from the future and more about -- it may affect some other people's trailing edge once it comes up and be a revenue battle for some of our customers.

But it's kind of new capital.

And we'd all heard about China's spending for many years.

It just hadn't materialized.

And now we're seeing a lot of those projects.

So that's part one.

The other part for us is the broad participation across the 7- and 10-nanometer logic in the foundry space.

And so that's a little bit different too and just a bigger commitment to that node than we saw in the last cycle.

But of course, there's a cyclicality associated with these businesses.

So I think there'll be ebbs and flows.

For example, NAND probably softens in 2018 because of all the capacity that's coming on or at least stays flat.

And your backlog, I thought C.J. mentioned it was like 7 months.

Is that -- I mean, is that pretty locked in?

What portion of that is actually cancelable?

Well, I mean, normally our business runs -- it's a little bit hotter now because the order strength over the last couple of quarters has been pretty significant.

I mean, we've had 2 out of 3 quarters over $1 billion.

So we'll start to ship down on that.

I mean, we generally run between 5 and 6 months.

I guess, any of the backlog -- I mean, what customers do is they don't usually cancel.

What they do is they push, right?

That's usually the indication.

You'll see them start to push orders out.

And then at some point, then they'll start to convert those to newer versions of the products they had as they get ready to receive those tools.

But if you look at what's loaded in the factory, I would characterize the environment as customers generally pulling to get the tools quickly.

And so far, we feel pretty comfortable with what we're seeing.

We spend a lot of time looking at indicators and all the different things that are out there.

And we feel pretty good about the underlying fundamentals of what we're seeing today.

And a lot of what's going on in China is -- I'm not sure is really contributing to the global sort of supply-demand dynamics, at least at this time.

And so that's future investment for the long run.

And so I don't think that, that necessarily worries us from an overheating perspective at least early on.

Okay.

And if I could just sneak in one more.

You talked -- Rick, talked about NAND and the foundry logic.

How does DRAM fit into the 2018 equation?

What we were saying -- I mean, DRAM has bounced back this year from '16, 1. The leader didn't invest a lot in '16, so we've seen some investment.

I think -- look, it looks like it's an undersupply.

So I don't think DRAM is going to be up as we move into next year.

Maybe it's flat, maybe it's down a little bit.

I don't see it changing a lot.

But hard to see how it's up a lot from where it's at right now, at least just generally how we see it.

Your next question comes from the line of Toshiya Hari from Goldman Sachs.

I had a follow-up on China as well.

Unless I heard you incorrectly, Rick, I think you talked about your Chinese business in calendar '17 doubling year-over-year.

I was wondering if that included some of the multinational spending or was that a pure local number.

And if you can talk a little bit about the preliminary outlook into 2018 for your China business, that would be helpful.

Yes.

The doubling we're referring to is native Chinese investment.

We also have the multinational, but we're talking about what's going on there.

And in terms of 2018, if you think about some of the ordering we're seeing is for the first phase, what these customers will tell you is there's a subsequent phase planned.

So we think that China continues to be strong as we go forward into 2018.

But whether or not -- internally, we discount the numbers that we've heard in CapEx from China over the next 5 years, and it's still pretty strong if you even discount them in half based on what some of these players are saying they're going to do.

Great.

And then as a follow-up, I had a question on your bare wafer inspection business.

I appreciate it's not necessarily a big part of your business today.

But if and when some of your customers decide to expand capacity, I think it could be a pretty nice business for you guys given the gross margin profile.

Can you talk a little bit about what you're seeing today from those customers and what the outlook is into the second half and potentially to 2018?

Sure.

Yes.

No, it's actually -- it is an important business for us.

It's a strong business.

And not only that, what we're seeing is demand both from wafer manufacturers but also from customers.

I see customers who are qualifying tools for ramps.

It is an important part of the business.

It is part of the uptick, and we're seeing in orders, and it will come in revenue as we go through.

So it is -- it has been strong, and that's both for defect and also for flatness.

So it's an important part of our story going forward.

Your next question comes from the line of Patrick Ho from Stifel, Nicolaus.

Rick, the last time you saw this kind of really surge in sustained business was when the 28-nanometer node rolled out.

Maybe a 2-part question there.

Do you believe based on your customer conversations that the 10- and 7-nanometer node will be as large as 28?

And secondly, can you -- given the strength that you've seen in the last few quarters, what is the capital intensity change or increase from 28 to the 10, 7?

Yes.

Great question.

I think that the first part of that answer is there certainly is a lot of broad investment for the 10 and 7. What's interesting about it though if you unpack a lot of our orders, the strength in our orders is really not as much from 10 and 7 as it is what we're seeing in memory and some of the other segments that we're in.

So it is strong, but we haven't actually seen the same kind of surge we saw in 28 as we're seeing in 10 and 7. That said, it's pretty broad right now.

And we think there are at least 3 maybe 4 players that are going to be involved in that, that we'll see going forward.

We are hearing from customers they think the size is going to be similar.

But it's still early for 7. And right now what we're seeing is pretty intense investment around the 10 node.

Great.

That's helpful.

And maybe as my second question, in terms of the services business, where you've also seen the revenues ramp over the last few quarters, given the process control isn't like a process tool like edge that basically eats itself up, what type of offerings or what type of solutions are you offering customers on the services end that will help grow that business to that 7% to 9% growth rate you talk about?

Yes.

Patrick, it's Bren.

It is a different business than our process tool peers.

And so what you typically have is -- so 75% of that revenue stream is contract, right?

So we have customers that buy service contracts that -- with different levels of coverage across either certain tools or broadly across a fab.

And so most of that revenue is repeatable, and it allows us to staff and rightsize those fabs to maintain good utilization but also to get pretty good predictably about part failure and so on.

So -- but what we end up selling is -- we replace parts, right?

And so then there's the parts.

And as those parts fail over time, and you have lasers that have lights -- use of lights and so on.

So that's really the biggest part of the business.

But it isn't traditional break and fix and that -- it's billable.

It's really contract.

So it really works for us and works for customers because we can keep the tools up and keep them optimized and run preventative maintenances and those kinds of checks on them over time to keep them running, and it works out on both ends.

One of the other things is the decay rate.

You think about service contracts on tools and how long you service them, part of what's happened is people are keeping them in production longer.

So if you think about fully utilized, older-generation fabs that for many of our customers are just printing money because they're fully depreciated, the tool life is longer than we expected, which means the decay rate on the back end is slower, which is part of the contributor to the growth because you have stuff coming in and you don't have as much stuff coming out.

And that's really the benefit we're seeing from IoT.

Your next question comes from the line of Edwin Mok from Needham & Company.

So first question, I want to try and get an update in terms of your view about Gen 5. We talked about it before, and I think right now you talk about mostly still being used on rebuild.

How do you kind of think about that tool onto production?

Do we still have to wait until 5 nanometer?

Or is it you'd be the big kicker for that?

If you can give us some color on that.

Well, the good news is we're kind of doing what we wanted to do with Gen 5 in terms of we're having broad shipments across a number of customers.

And we're basically on a 1 a month shipment rate, and so we continue to do that.

I think the time when we would ramp that up is when we'd see it going into larger scale production.

So there are some customers that are talking about production for it, but it's a limited number of layers.

And when that happens, then we'll see multiple tools.

But right now, we're seeding the market.

We're penetrating the market.

We -- it's a tricky supply chain situation, so we're keeping it at 1 a month right now as we go forward, and we'll ramp up.

It's hard to say.

Preliminary results are good.

There's a lot of use cases for customers.

But we stick to what we said before, where we think it really is a tool that ramps into production out a couple of years, high volume really when EUV is kicking in, so you'd say 2019, 2020.

Okay.

Actually, that's helpful color.

And then on the server side, I have a question actually.

So one of your peers or some of the company in the supply chain, right, they would charge customer or they would write a contract with customer where they do almost a yield-based or performance-based contract with customer.

For example, I think [Fima] used to do that.

So I'm just curious, is that something that you guys entertain or is -- was in the 75% of your services contract?

Are you guys trying to look at doing that in your service contract?

Or is it still mostly fixed type of contract?

So we have a lot of key metrics that were -- that we agree to with customers, largely around availability of the tools and response time in service.

And so we've historically done that.

And of course, as these tools become more mission critical for our customers, that's part of the arrangement that we have.

So when it comes to extending the life and working very hard to make sure we don't disrupt production, we already do that.

That is part of the benefit that customers have for going contract.

There are things we're doing to enhance the business we talked about in the last question.

And some of them have to do with older fabs, where we continue to offer enhancements to the tools.

A lot of tools that have been in the market for a long time have upgrades available for customers, and we'll continue to offer those as we go forward.

So I'd say we already have that as part of the offering.

The next question comes from the line of Jagadish Iyer from Summit Redstone.

Two questions.

Rick, first, historically, September quarter has been a big one for you guys, but you have a very guidance.

So the question is, are there any greenfields that are driving that?

And as a part of that question is that can you provide some color on how many greenfields you have for this year?

And are there any thoughts on for next year?

And then I have a follow-up.

Yes.

Jagadish, it is a little atypical to see the strength in September compared to the last few years.

So we're encouraged by that.

I don't have the actual number in front of me.

I think we'll be shipping a fair amount.

I think from a -- the second half has a fair amount of memory in the mix of shipments.

There's obviously some additional 10-nanometer investment on the foundry that's shipping as well.

So most of the China stuff which is greenfield is not shipping for a little while.

So I guess, that's how I'd characterize the shipment profile next quarter.

But as I said in the prepared remarks, strong and stable, and we see a second half shipment profile that's up a little bit versus the first half so -- evenly weighted across the year and probably more memory-centric in the second half of the year versus what we saw in the first half.

Yes.

One other interesting attribute of our business environment, we looked at this recently, and the concentration from the top customers is actually down.

We actually have broader customers.

And when we look at the top 10, the percent they make up of the business has decreased in this calendar year.

It was slightly down last year, and we anticipate that broadening.

So the good news is I think we have more customers and more locations and more offerings for them.

Okay, excellent.

And just as a follow-up.

You talked about, Rick, in your prepared remarks about new products and particularly for the memory segment.

If -- hypothetically, if you assume that WFE is up, say, 5% for next year, would it be fair to say that you could outperform WFE as these new products start to bear fruit?

Actually, I'll let Bren take that one.

Yes.

I mean, one thing about next year is it looks like next year's WFE mix is a little more memory-centric.

And as you know, the process control intensity in memory while it is getting better isn't near foundry.

So $1 billion of WFE depending on the segment is not created equal for us.

So -- but we've seen some improvement on the memory side, which we're encouraged by.

I think these new products given the timing of when they'll ship and go to market and we'll actually start to see revenue given the valuation process and so on, I think in the second half of the year we might see some revenue.

But I think it'll be a pretty small amount.

But as I look at our plans for next year, I look at the funnel, I don't see any reason why KLA shouldn't grow in line with -- at least in line with the market as we move into '18.

(Operator Instructions) Your next question comes from the line of Atif Malik from Citi.

Rick, good job in kind of refocusing KLA on memory in China after the merger breakup.

My question is on 3D NAND yield.

Can you help us understand where the 3D NAND yields are on Gen 1 NAND at 48 layer, just kind of blended across all (inaudible) and where the yields are on second generation in 64-layer NAND?

And is there a reason to believe that there's something structural about migration from 48 to 64 to 96 which will keep the yield kind of low for a sustainable period of time?

Well, as you know, one of the biggest challenges is one of the ones we're trying to address, which is the ability to catch defects in the process.

So as a result, when customers have process problems, they have to go through some pretty rigorous engineering analysis to unpack those problems.

So you see 2 things.

One, you see actual wafer yield as sometimes -- probably, it is averaging lower than what you would see on 2D.

And it's probably -- depends on customers somewhat how that is, but it is lower.

But it's still pretty good.

And you see that because they're having success selling them commercially.

But the other thing is you see line yield bust, which is a little different, which is when you have a number of wafers that are actually bad and have to be scrapped.

And for many customers, that's very disruptive to their business.

So I'd say that, yes, structurally it's very hard to build these.

Especially as you're going up in layers, the devices -- for 2 reasons.

One, the process technology is very tricky.

And the second one, you don't have a lot of visibility into it.

And it's kind of like the old days of semiconductor manufacturing when we didn't have as much in line and people had to use destructive means or short loops to try to figure out what was wrong with their process.

So there is opportunity.

I think we've repeatedly said that if we could help with solutions that help customers gain a few points of yield, that's easily justifiable on their end.

But I would say yields are in the range of probably -- if you'd see them over 90% in 2D, you'd see similar yields at, say, 10 points lower, say, 80% or so for -- and maybe they're even higher than 90% in 2D.

So you'd see a degradation as you go to the 3D.

Very helpful.

As a follow-up, the bifurcation in leading edge foundry versus lagging edge is expanding.

UMC [declared CapEx] and have decided to actually focus on the lagging edge.

So when you're talking to the Chinese domestic foundry customers, are they asking for 28-nanometer equipment?

Are they asking for the latest 10-, 40-nanometer equipment?

It varies.

I mean, they definitely have ambitions to advance nodes down the road.

So you do see some interest in the advanced, and you'll see customers doing, in that case, in the foundry some R&D or at least some of them doing some.

But by and large, it's the older tool sets because they do have capital costs to manage, and they've got to try to hit the cost per wafer target that everybody else has.

But you do see some interest in the advanced.

And they also want to have a path and a road map.

The good news relative to our equipment is we have upgrades available on things like our GEN 4 wafer inspection, as an example, where we can continue to add value to a tool set that they buy.

There are no further questions at this time.

Mr. Ed Lockwood, I turn the call back over to you.

Okay.

Thank you, Christine.

Thank you all for joining us here today.

Just a reminder, there's an audio replay of the call, and it will be available on our website later on this afternoon.

Operator, this concludes our call.

Thank you.

This concludes today's conference call.

You may now disconnect.

Thank you.