FormFactor Inc (FORM) 2017 Q3 法說會逐字稿

Thank you, and welcome, everyone, to FormFactor's Third Quarter 2007 Earnings Conference Call. On today's call are Chief Executive Officer, Mike Slessor, and Chief Financial Officer, Mike Ludwig. Before we begin, Jason Cohen, the company's General Counsel, will remind you of some important information.

Thank you. Today the company will be discussing GAAP P&L results and some important non-GAAP results intended to supplement your understanding of the company's financials. Reconciliations of GAAP to non-GAAP measures and other financial information are available in the press release issued today by the company and on the Investor Relations section of our website.

Today's discussion contains forward-looking statements within the meaning of the federal securities laws. Examples of such forward-looking statements include those with respect to the anticipated effects and benefits of the completed merger between FormFactor and Cascade Microtech; projections of financial and business performance; future macroeconomic conditions; business momentum; business seasonality; the anticipated demand for products; our future ability to produce and sell products; the development of future products and technologies; and the assumptions upon which such statements are based. These statements are subject to known and unknown risks and uncertainties that could cause actual results to differ materially from those expressed during this call.

Information on risk factors and uncertainties is contained in our most recent filing on Form 10-K with the SEC for the fiscal year ended 2016 and our other SEC filings, which are available on the SEC's website at www.sec.gov and in our press release issued today. Forward-looking statements are made as of today, October 31, 2017, and we assume no obligation to update them.

With that, we will now turn the call over to FormFactor's CEO, Mike Slessor.

Thank you, Jason, and thank you, everyone, for joining us today. FormFactor delivered another strong result in the third quarter, with revenue and earnings at the high end of our guidance range. Our broad market positioning and diverse set of demand drivers continue to enable FormFactor to capitalize on industry momentum to deliver growth and strong profitability.

In the third quarter we achieved our highest DRAM probe card revenues since the second quarter of 2015; realized the highest third quarter sales in engineering systems history; and delivered the highest ever quarterly shipments to our largest customer, eclipsing the previous quarterly revenue high for that customer by 20%.

Consistent with comments on our last earnings call, we expect to finish 2017 by delivering revenues in the second half that approximate revenues we delivered in the first half. As we look forward, although our quantitative view of 2018 overall is still developing, we are seeing favorable market conditions in data center, mobile and automotive end markets, which we expect will drive meaningful growth in the first half of the year.

As we've discussed on previous calls, industry adoption of advanced packaging processes like integrated fan-out represent a significant market growth and differentiation opportunity for FormFactor's products, especially for our probe card products serving the mobile space.

To that end, I'm pleased to tell you that we recently received a significant customer order to support preparations for a major 7-nanometer mobile application processor project. We expect to build inventory and make initial shipments late in the current quarter, with volume shipments expected to begin in the first quarter of 2018.

This win represents an important follow-on to our initial 2017 10-nanometer penetration serving this key customer in integrated wafer level fan-out applications and demonstrates the differentiation of FormFactor's technology in advanced packaging applications. We are currently focused on planning and preparation to assure operational execution and product performance that will drive future wins and share growth with this strategically important customer.

Continued growth and innovation in the data center is also driving many of our businesses, with an example being the previously described all-time record shipments in the third quarter to our largest customer. This demand is primarily in support of the early part of this customer's 10-nanometer production ramp, although we are also helping enable 7-nanometer development and 5-nanometer pathfinding with our broad array of electrical test and measurement products.

As with many advanced node transitions, there is some short-term lumpiness in timing of customer demand, and in the fourth quarter we expect some digestion of the record third quarter shipments. However, as 10-nanometer volumes accelerate in the first quarter of 2018, we are also expecting and planning for a return to the record levels experienced in the third quarter as we support the release of leading-edge microprocessors that will power next-generation data centers.

Driven by a combination of data center, mobile and automotive content growth, demand for DRAM probe cards continues at strong levels. With each of the major DRAM manufacturers executing a combination of node transmissions and capacity additions, albeit with each on a slightly different cadence, the new design environment that drives probe card demand remained healthy and robust.

As you have heard recently from both our customers and other suppliers, overall DRAM bit supply growth appears to be roughly balanced with the content-driven demand growth, at least in the medium term, and we are optimistic this balance will continue through the first half of 2018.

Another trend contributing to our solid third quarter and extending into the fourth quarter is the strength of our engineering systems business. As a reminder, these systems enable customers to develop processes and improve yields for new semiconductor architectures, new chip designs and even brand-new technologies such as micro-LED displays and silicon photonics integrated circuits.

We are seeing particular strength in China, where we are supporting initial silicon debug and yield improvement initiatives in both memory as well as Foundry and Logic projects. These investments are supporting the initial phases of the aggressive growth and buildout plans for the emerging Chinese domestic semiconductor industry. This China-specific growth, along with the initial momentum in applications such as micro-LED that we have discussed previously, lead us to expect double-digit annual growth of our Systems segment results in 2017.

Another highlight for the quarter was the customer installation of our first product that combined FormFactor and Cascade legacy technologies. This new probe card architecture delivers heightened levels of productivity and performance to RF SoC wafer test. We've received positive customer feedback from the testing and evaluation of the initial units and we are now executing the initial production orders and investing in broader product market release activities.

Building on this initial success, our 2018 planning is increasingly focused on achieving longer-term revenue growth synergies from the next steps of product and technology integration while still retaining a sharp focus on cost and efficiency.

On the operational front, although the underlying third quarter in product and customer mix was substantially different from the unusually strong second quarter, we delivered gross and operating margins that are tracking towards our long-term financial model. As I mentioned earlier in the call, we are currently preparing our factory network to ensure we can efficiently execute on specific first quarter 2018 demand from our key customers. This demand continues to be resonant with the line of sight opportunities we have described in advanced packaging, mobile data and automotive ICs.

With the continued strength of the overall semiconductor industry, augmented by these incremental opportunities, we expect to deliver sequential revenue and profit growth in the first half of 2018 as we progress towards our target financial model that delivers $650 million of revenue and $1.50 of earnings per share.

I'll now turn the call over to our CFO, Mike Ludwig, for further details on our third quarter results and to provide insight into FormFactor's outlook.

Thank you, Mike, and good afternoon. As you saw from our press release and heard from Mike's comments, we had another strong quarter, delivering revenue and non-GAAP EPS at the high end of our guidance and delivering another quarter of positive GAAP net income.

FormFactor's revenues for the third quarter were $143.7 million, slightly down from the $144 million in the second quarter. Probe card segment revenues of $119.4 million decreased 2% compared to the second quarter, while system segment revenues of $24.3 million increased 9% or $1.9 million compared to Q2.

Within the probe card segment, Foundry and Logic revenues of $81.9 million decreased $6.8 million or 8% compared to our second quarter. The decline was attributable to a reduction in revenue from a customer who utilizes our probe cards and advanced packaging applications at leading edge nodes, confirming our previous comments that this business would be lumpy as we achieve long-term market share gains at this customer.

We saw continued strength and increased revenues from the data center vertical market in the third quarter. Foundry and Logic revenues comprised 58% of total company revenues in the third quarter, down from 62% in the second quarter.

DRAM revenues were $32.4 million in the third quarter, an increase of $0.9 million sequentially, as we continue to see a robust demand environment in the quarter. DRAM revenues comprised 23% of total company revenues in the quarter, a slight increase from the second quarter.

Technology node transitions, the strong data center demand environment and increased DRAM content per mobile handset continued to positively impact the market. Flash probe card revenues were $5.2 million in the quarter, an increase of $3.8 million over the second quarter. The majority of third quarter revenues were from NOR Flash applications. The $1.9 million and 9% in sequential increase in Systems segment revenues was driven by increased revenues from 200- and 300-millimeter platforms and thermal subsystems.

GAAP gross margin for the third quarter was $57.6 million or 40.1% of revenues compared to 42.9% of revenues for the second quarter. The third quarter includes $6.4 million of reconciling items, which you can find outlined in our GAAP to non-GAAP reconciliation table available on the Investor Relations section of our website.

On a non-GAAP basis, gross margin for the third quarter was $64 million or 44.5% of revenues, down from 47.6% in the second quarter. The expected decrease, as contemplated in our third quarter guidance, is due to lower margins in the probe card segment, which decreased to 43.1% of revenues compared to 46.9% in the second quarter. The segment had a less favorable product mix in the quarter as our Foundry and Logic revenues comprised a smaller percentage of overall revenues compared to the second quarter. Our factories continued to perform well at high production levels.

In the Systems segment the non-GAAP gross margin was $12.6 million or 51.7% of revenues in the quarter compared to 51.5% in Q2. The increase in gross margin resulted from a favorable product mix offsetting a stronger euro.

Our GAAP operating expenses were $43.4 million for the third quarter compared to $42.2 million for the second quarter. The third quarter includes $6.5 million of GAAP to non-GAAP reconciling items.

Non-GAAP operating expenses for the third quarter were $36.9 million or 25.7% of revenues, compared to $37.1 million or 25.8% of revenues in the second quarter.

As outlined in our financial model discussions in late June, we are achieving increased leverage on our SG&A infrastructure at current revenue levels, as our SG&A expenses were 13.2% of revenues in Q3 compared to 13.8% in the second quarter.

The non-GAAP effective tax rate for the third quarter was 4% compared to 3.8% for the second quarter, consistent with our expectations of a continued low effective tax rate while we utilize our previously generated U.S.-based net operating losses. We expect our non-GAAP effective tax rate for 2017 to be between 4% and 6%.

GAAP net income was $12.6 million or $0.17 per fully diluted share for the third quarter compared to net income of $0.24 per fully diluted share for the second quarter. We have been GAAP-profitable in each of the first 3 quarters of 2017 and expect our fourth quarter to be GAAP-profitable as well.

Third quarter non-GAAP net income was $25 million or $0.34 per fully diluted share compared to $29.2 million or $0.40 per fully diluted share for Q2. The third quarter results represent another important data point that indicates our financial performance is on a trajectory to achieve our communicated target model. Company non-cash expenses for the third quarter included $7.5 million for amortization of intangible assets, $4.6 million for stock-based compensation and depreciation of $3.5 million.

Moving on to the balance sheet, the company generated $12.4 million of free cash flow in the third quarter, bringing our cash flow for the first 9 months of the year to $49 million. In the quarter the company spent $9.8 million on principal and interest payments, including a $5-million accelerated principal payment and $6.2 million for capital expenditures.

Capital expenditures for the first 3 quarters of fiscal 2017 were $14 million, demonstrating a run rate well within our expected annual capital expenditures of $16 million to $20 million. In total, cash, comprised of cash, short-term investments and restricted cash, ended the third quarter at $135.7 million, $5 million higher than Q2. The balance of our cash, short-term investments and restricted cash exceeded the balance of our debt by $21.4 million at quarter end.

Turning to the fourth quarter non-GAAP guidance, we continue to experience broad-based demand across product lines and markets in both our probe card and Systems segments.

Foundry and Logic as well as DRAM probe card demand continued to show strength, as demonstrated by our record revenues from our largest customer and higher DRAM revenues in the third quarter, and a recent significant follow-on order from our advanced packaging Foundry and Logic customer. We do, however, expect to experience some seasonality in Q4. And as Mike mentioned, we expect a reduction in revenues from our largest customer in Q4 followed by renewed strength from this customer in the first quarter. Therefore, we expect revenues in Q4 to be in the range of $126 million to $134 million.

Consistent with the reduced revenues, fourth quarter volumes and factory utilization are expected to be lower than the third quarter; but with a product mix that will yield a standard margin slightly higher than the third quarter, we currently expect fourth quarter gross margins to be similar to those of the third quarter, in the range of 43% to 46%. We expect to realize fully diluted earnings in the range of $0.24 to $0.30 per share. Our Q4 non-GAAP guidance assumes consistent foreign currency rates. Achieving the midpoint of our fourth quarter non-GAAP guidance ranges results in fiscal 2017 revenues of $547 million, gross margins of 45% and fully diluted earnings per share of $1.24, demonstrating meaningful progress towards achievement of our long-term target model.

With that, let's open the call to questions. Operator?

(Operator Instructions) Our first question comes from the line of Craig Ellis with B. Riley.

To start off, just congratulations on the broad-based growth you're getting in the portfolio and the nice margin execution. The first clarification I had was just on the fourth quarter guidance. With respect to segment activity and looking at some of the gives and takes, is -- are we to infer that beyond the Foundry and Logic business that their segments would be flat or up, or is Foundry and Logic just called out because of its size?

It's Mike Slessor. I think, looking at it first order, obviously with a business as diverse and having as many different demand drivers as we currently now have in the FormFactor portfolio, there's quite a few puts and takes. But if I were to try and take the highest-level view of it, it really is a quarter-on-quarter reduction in Foundry and Logic probe card demand associated with digestion of record shipments to our largest customer in the third quarter. As we said, we're holding capacity and have forecasts and indications that we would expect to return to nominally the third quarter levels in the first part of 2018. So we're pretty confident as that customer ramps 10 nanometer that we're going to see some strong demand and growth there. I guess the other point I'd make about the fourth quarter and our largest customer is, we still expect to operate essentially at the double demand level, doubled over 2015, that we've raised our baseline capacity to. Again, as they ramp 10 nanometer, release new designs, we're expecting to see some growth on top of that. On DRAM, there's probably a little bit of calendar seasonality that's one of the takes, compared to a very strong second and then third quarter. But as we mentioned in the prepared remarks and you've heard from other people, the overall DRAM environment from a structural standpoint still feels very strong. So a little bit of a seasonal downtick in the December quarter for DRAM, but nothing significant, and we expect to see it continue over the longer term to be relatively strong from an overall environment perspective.

That's helpful. Then moving on to some of the commentary around specific segments, the company's done quite well with the 10-nanometer program this year and you've got follow-on activity at 7 nanometer, it sounds like with the same customer. Can you talk about the potential to diversify that business over time, and as that happens what does it mean for growth in the business?

Yes. Mike Slessor again, Craig. So as we said on the call, we have received a significant order to support 7-nanometer production of a mobile application processor project using integrated fan-out technology. We view this as a very important follow-on and validation to our initial 10-nanometer win at this customer. But I think, consistent with a lot of the other suppliers, we're beginning to see the very initial stages of diversification outside both one foundry and to multiple fabless customers in adopting the integrated fan-out technology. For sure, it's still early innings in this, and the majority of the wafer volume and our probe card volume associated with supporting integrated fan-out packaging in particular is pretty concentrated with one foundry and one specific application processor design, but we're now beginning to both have discussions and work on chip designs for both other foundries and other fabless customers that utilize advanced packaging technology.

That's helpful. And then the last one for me, and it'll probably stay in your court as well. Helpful to get the look-ahead and the comments on meaningful first half growth. Just clarifying that that was a comment that was intended to be half-on-half, or is that year-on-year? And related to that question, some larger front-end companies have talked about mid to high single digit growth next year. If we had revenue or if we had industry spending growth that was that healthy, against that backdrop, can you talk about the opportunity you would have against the 3 line of sight opportunities you outlined at the strategic update, and what could that mean for top line growth in 2018 for Form?

Sure. I think, for our business, we've tried to give some view on certainly the first quarter and into the broader first half of 2018. As I said in the prepared remarks, we're still developing our quantitative view of 2018 overall. But with the position we're in as the market share leader in our operating segment, for sure the industry growth and overall customer spending is going to be one of the key variables that drives our growth. For now, we're pretty optimistic about for sure the first quarter and into the rest of the first half. We suffer from a bit of lack of visibility into the second half. But given our market footprint, our relationships with key customers and our penetration in supporting some of the key growth initiatives in the industry, whether they be advanced packaging or automotive ICs, as long as the industry stays healthy we feel pretty comfortable about growing in 2018.

And our next question comes from the line of Patrick Ho with Stifel, Nicolaus.

Congrats on a nice quarter as well. Maybe Mike, first off -- in terms of your capacity and your ability to cater to your customers, a couple of years ago you did have to ramp up capacity for your largest customers. Given the current demand strength across both Logic and DRAM, how do you feel about the company's total capacity and the ability to keep up with this elevated demand levels that you're seeing across the board today?

Yes. And so Patrick, a good question, because I think across the industry most of us are continuing to incrementally add capacity, yet still operate with the degree of flexibility that we can modulate the capacity up and down as we need to. The comment I'll make -- one of the lessons we learned back in the late part of 2015 and early part of 2016 was that we really had to create more fungible capacity, and the ability to use our factory network to serve these different elements of industry demand has given us a little more flexibility and a little more leverage in being able to scale up and down. If you look at the CapEx guidance we've provided for the year, we do believe we're going to be within our $16 million to $20 million CapEx guidance, so we are spending money on increasing capacity and we are adding direct labor incrementally around some different opportunities, but the fundamental capacity demand balance in our business still feels pretty healthy. We're trying to be as responsive as we can in moving things around and holding world-class lead times to make sure that we can continue to lead and compete strongly for business in the industry.

And I think you can see that what we've said in the last 2 quarters at, we'll call it, elevated revenue levels, that our factories have really performed well with respect to -- we'll call them operation metrics, whether that be scrap, on-time delivery, quality. So I think we feel like we've got, as Mike said, a factory network in place that can really responded well to elevated revenue levels.

Great. That's helpful. And maybe as my follow-up question, in terms of the DRAM probe card market, given your broad-based customer base and the different ones that you serve, have you seen any transition over the last few quarters to, say, below 20 nanometers, or is still 20 nanometers kind of the sweet spot that you're seeing for your products today?

Patrick, Mike Slessor again. I think it's different, and slightly different for each of the different customers. Obviously one of the larger DRAM manufacturers has made a pretty bold move to move to the 1X nanometer node and drive significant wafer volume there. One of the other manufacturers, getting there, and really investing in both the innovation and capacity, with the third major DRAM manufacturer maybe running a slightly different playbook. So I think we've seen at each customer a different mix of tactics and timing as they continue to move forward with those node transitions as well as overall bit and wafer growth. So I think in 1 of our leading customers, for sure, a significant move to the 1X nanometer node and driving volume for them and volume for us in terms of probe cards onto that node; maybe less so at the other 2, but certainly something to look forward to as we move into 2018.

That's great. And maybe a final question for Mike Ludwig. In terms of the cash flow generation, you guys have done a really, really good job since the Cascade deal closed. Is the primary focus still on these accelerated debt repayments? Is that going to be the primary focus as we look at 2018 as a whole?

Yes, Patrick, it will. So we are again very focused on delevering the company and accelerating our debt paydown to the extent that we have a portion of our loan that is unhedged. So we definitely want to get that portion paid down, and that will take us through -- or that will take us into the second half of 2018. So I would suggest that, through the first half and into the second half, that will continue to be a focus of our cash generation and use of some of that free cash flow, in addition to again bolstering our balance sheet in anticipation or looking for M&A opportunities.

Our next question of Tom Diffely with D.A. Davidson.

So I guess first quarter is on the significant customer order for the 7-nanometer project. So I would have assumed if you use the term significant it would have been a high-volume product that was going to the market; but it's my understanding that that 7 nanometer's still kind of in the pilot line stage, so I was wondering if you could reconcile those two.

Sure. Well, it goes back a little bit, Tom, to maybe Patrick's comments about having capacity in place and making sure that, from our customer perspective, that they have the certainty of capacity and delivery in place. I think one of the things we see happening, especially with the more strategic relationships with key customers, is them making an attempt to secure capacity in place and make sure that we have the commitments -- FormFactor has the commitments from them -- to place that capacity in place, to have it reserved and to be executing on the initial ramps of these things. So if things were going on a PO-by-PO, unit-by-unit basis, I'd agree with you. We'd be looking at somewhere like probably mid-2018 for the orders to come in. I think we're operating in a bit of a different mode, with everybody trying to ramp significant capacity at leading-edge nodes, where technically things are certainly difficult, and our customers are doing whatever they can to reduce risk.

Okay. So you see this capacity fungible for both the 10 and 7 nanometers, and whatever projects come up, you'll be ready for them.?

We do.

Okay. Great. And then I noticed the nice little step-up in the flash business on the NOR. Anything in particular going on in the NOR market that we would expect to drive growth going forward, or is this kind of a one-off?

Well, I think, as we've described before, our overall flash revenues, whether they be NANDs, planar NAND, 3D NAND or NOR, are pretty lumpy because we have a relatively small market share position given our choices to prioritize against some of our more profitable and higher-return opportunities, at least in the short term. A lot of the NOR growth in the quarter or step-up in the quarter has been associated with some of the China activity I talked about. We've had a significant amount of NOR activity in the emerging China memory ecosystem. And I think overall some of our more classic NOR customers have released some new designs and continue to innovate and try and drive some progress on the NOR Flash side as well. More specialty applications than mainstream storage; but as you noted, a nice bump-up in the third quarter.

Yes. No, there seems to be strength coming from all sorts of different areas right now. So if I look at or listen to some of the comments you made about the record levels of DRAM -- couple-year record here, and the fact that only one of your customers was really going all out on the design activity, it sounds like there's still potential for upside in the coming quarters if some of these other customers get more aggressive on their 1X ramps.

Well, yes. I guess I want to make sure I'm clear on the distinction. We still have all 3 major DRAM manufacturers releasing new designs on a variety of nodes, whether they be 20 nanometer, 1X nanometer or even now the leading-edge 1Y nanometer, very early production. So customers, as I said in the prepared remarks, kind of operating on slightly decoupled cadences for their node releases and design releases, but we see pretty broad-based design activity across all of the major DRAM manufacturers here. With one of them for sure and probably arguably a couple of them, a lot of the wafer start volume is still ahead of us on the 1X nanometer node.

Okay. Great. And then a question for Mike Ludwig. If you look at the tax, that 46% -- is that a good number to use going forward as well?

I think 4% to 6% is a good effective tax rate to use from a cash basis, at least until we utilize all of our U.S.-based NOLs, which I think will be into the early 2020s.

Okay. What is the current level for NOLs right now?

We expect to end the year at somewhere around $250 million of NOLs.

Our next question comes from the line of Edwin Mok with Needham and Company.

So first question -- I want to go back to that 7-nanometer orders that you talked about. Is there a way you can think about -- does it mean that -- is this order a -- meaning that you're gaining share at this particular opportunity, or -- and relative to what you have seen on the 10 nanometer this year? Would it be a higher volume from this customer or both, or is there any way you can kind of talk more -- in a little more detail about this particular order?

Sure. Edwin, Mike Slessor again. Look, as Tom noted, it's pretty early in the 7-nanometer production ramp. But I think the fact that we're participating at this early stage in the 7-nanometer node with this customer is an indication that we're building that relationship and building share after that first 10-nanometer win last year. Obviously we were a little late to the 10-nanometer party relative to -- at that customer, relative to some of the other suppliers in the overall test and packaging ecosystem. But that represented us getting in the door, and I think this now maybe is indicative of a more usual cadence on being engaged early in the development, early in the pilot production ramp, and then having the opportunity to go compete for, win and earn higher business when it gets into full production, probably more in mid-year 2018.

Okay. Great. That's extremely helpful. And then on -- for your largest customer you mentioned you doubled your capacity and they basically ramp up to that level, and I remember earlier in the year talk about how they are (inaudible) producing in both 14 and 10 nanometer. If we look forward, they are more likely to transition more to 10 nanometer. Is there any risk that that demand might come down from that double level or do you think that's kind of a sustainable level going forward?

Well, we've said several times that we think that doubled level -- and doubled with respect to 2015 -- is a sustainable level. And even with the step-down I described in the prepared remarks and with Craig's question, we still expect to be operating in the fourth quarter at that doubled level. So we feel like, even though we're going through this initial stage of capacity digestion for the 10-nanometer ramp, we're still going to be operating at the doubled level here in the fourth quarter and we expect to come back up to levels that approximate what we delivered in the third quarter as this customer really starts to ramp 10 nanometer in the first part of 2018.

All right. Great. That's helpful. Last question I have was on the margin side. I guess, 2-part question. First is, with Foundry mix coming down this quarter and lower utilization, why do you see that -- why do you think that your gross margin can actually go up, or I think you said flat to up this quarter. (inaudible) through the midpoint, so flat. And then also on the system gross margin, I remember a year ago, that business has -- kind of has been trending to the high, or let's just say the mid-50s in terms of gross margin. The last 2 quarters is kind of trending at below 52%. With that business growing, [should it] (inaudible) that you can improve the gross margin of that business?

Yes. So the first question, Edwin, in terms of the -- why do we think we can hold margins flat in the fourth quarter with a particular Logic customer -- Foundry and Logic customer, really kind of going down, and then what does that mean for the overall business? Again, it really becomes very much a product mix and customer mix discussion. And again, based on what we see right now as product and customer mix in Q4, even with the one change that Mike had talked about, we still see that as a pretty strong product/customer mix in the fourth quarter. Therefore, we believe our standard margins at this point in time will be better than what we had with our third quarter mix. So that's the reason why we still believe our gross margins in the fourth quarter should be consistent with those in the third quarter. For your second question with regard to the Systems gross margin -- so earlier this year we actually had -- I think in Q1 we had a margin of about 54% in that business and since that time anywhere from 51% to 52%. And so one thing that has definitely impacted that is the strength of the euro. So most of those revenues are denominated in US dollars with a lot of cost denominated in the euro, so that definitely has had an impact of a couple percentage points in each of the last couple quarters, so I'd say that is primarily what we are seeing in terms of impacting that particular margin. We certainly are not seeing it from the standpoint of reduced ASPs per se.

Our next question comes from the line of David Duley with Steelhead.

As far as fan-out goes, you mentioned earlier you have basically one large customer or one large foundry and one large customer that have ramped up on fan-out. Could you give us an idea how many more you would expect to join the party in 2018? We've heard from other companies that there are certainly several customers who are ramping up fan-out production. I was just wondering what your perspective was.

Yes. So a great question, David, and one that we're obviously excited about the answer to. And I think you have to look at it in 2 dimensions. Since this is primarily -- since fan-out is primarily a technology that's directed towards the mobile space and the majority of mobile chip designs are produced in the fabless foundry ecosystem, I think you first have to look at the foundries or OSATs that are investing and capable of producing integrated fan-out from a factory or foundry perspective. And then you have to look at the utilizers of that technology -- the fabless companies, or the mobile chip designers. I think we're seeing increases on -- in both of those dimensions. For sure, everybody who is trying to play at the leading edge of the foundry space is investing and developing and trying to qualify wafer level or panel level fan-out technologies depending on the different flavors. And so we're seeing -- there's not that many foundries, but we're seeing everyone who's trying to lead at 10 nanometer and below from a foundry production standpoint start to work through the initial characterization, if they're not already there, of their fan-out processes. On the fabless side, we see multiple customers now releasing designs that we're beginning to design and in some cases build probe cards for, that utilize the fan-out technology. So I think we're seeing growth beyond the single foundry and the single customer; seeing growth in both of those dimensions. Hard to say whether it all results in significant volume in 2018. But it's definitely moving in the right direction. And I think the compelling customer value statement and performance associated with integrated fan-out would lead you to believe that this is really only going to move in one direction. Now it's not going to absorb all the wafer starts in the world. It's probably pretty restricted to mobile in the short term. But it's got a long way to go before it saturates that.

So I guess to summarize, you might see a couple of foundries investing -- a couple of incremental foundries investing several test and assembly houses and perhaps some big IDMs.

Yes. I'm not sure about the big IDM piece. I'd say -- what I'd say is, the fabless customers, you're going to see several big fabless customers incrementally adopt this technology for their products, utilizing the foundries and OSATs that you denoted earlier.

Okay. And just remind me, how much more probe card-intensive is fan-out versus the -- a standard part that's not built on that process?

Well, it's -- there's certainly increased test intensity associated with not just the fan-out process -- and as we talked about in our June analyst call, one of the things that drives increased test intensity is because you're driving the packaging process now to the wafer level, you want to have pretty high confidence that the die you're packaging together are good. And so you're doing a more stringent job of testing. That drives probe card volumes up, because test times are longer. I think the other piece that's superimposed on that is wafer level fan-out packaging is really only happening at leading-edge nodes, and we're early enough on leading-edge nodes that the inherent yields aren't great. Any time the inherent yields aren't great, you're going to do an awful lot of tests to screen out bad die to make sure you're not packaging bad die. And so at least at this stage of the game, you've got this kind of constructive behavior or additive behavior where you're a leading-edge node that requires a lot of wafer test and you're packaging on a technology that demands a lot of wafer test for essentially known good die.

Okay. Final question from me is do you expect any new DRAM fabs to be built in 2018 outside of China?

Well -- so there are definitely DRAM facilities and DRAM shells that continue to be upgraded and produce more bits, whether that's through shrinks, whether that's through incremental, although things are getting pretty tight from a wafer start perspective. I don't think there's going to be any major new shells that come online. But the existing investments that our customers already have outside of China would seem to be continuing to upgrade and produce more bits. So to drive DRAM supply, at least at the current -- people have been talking about high-single-digit quarterly bit growth rates. I think our customers are continuing to squeeze out capacity from their existing fabs or facilities. In 2018 I don't think you're going to see any brand-new step function facilities come online.

I said that was my final question. I have one more. I'm sorry. You -- just a clarification. You talked about how you have to have known good die to do this fan-out process, which means that you're -- it's more test-intensive and probe-card-intensive, to make sure you have known good die. Do you also -- once you attach these known good die to a new wafer and start to do the fan-out process, do you do another probe test there before they're cut up?

So that's one that really depends on the customer flow. For things like high bandwidth memory, again generalizing advanced packaging a little bit more beyond just wafer level fan-out, there are multiple probe steps as you stack this chip up. Most of the wafer level fan-out applications that are being contemplated, and for sure the one in high-volume production, is a combination of 2 chips. So you really only have to do the 2 individual probe steps. As these integration schemes get more complex and integrate more die, we would expect to see, just like we do in HBM, the stacked DRAM part, multiple probe insertions for one finished chip.

Okay. So the intensity levels will -- should continue to go up as this stuff is adopted.

We believe so, yes.

(Operator Instructions) Our next question comes from the line of Jagadish Iyer with Summit Redstone.

Two questions, Mike. First, just on the significant order that you talked about -- so how should we be thinking about, if we have to quantify it, of where you were at 10 nanometer with this customer for advanced packaging vis-a-vis the 7 nanometer? And assuming that the wafer starts out kind of flattish, how would you kind of quantify the opportunity? Is it probably 10% more than what you had at 10 nanometer, or is there some more color that you can give? And then I have a follow-up, please.

Yes. Well, as I said, Jagadish, in response to Edwin's question, we certainly have the opportunity to make it bigger. We are engaged much earlier in the node transition cycle at 7 nanometer than we were at 10 nanometer. And I think so long as we execute and deliver quality product that continues to perform on the test floor, there I think is a fair chance that we can make this opportunity larger than our opportunity -- than what we realized at 10 nanometer. There's still a lot of hard work in front of us, and as I said in the prepared remarks our team is very, very focused on making sure we lay the groundwork to get these right, as we do, and I'm confident that we can. We're going to be in a position to make this business -- I think 10%'s a reasonable estimate for where we are right now, but materially larger than it was in 2017 for us.

Okay. That's fair. That's excellent. Then I have a follow-up. A lot of the front-end companies have suddenly disclosed how they've been receiving a lot of Chinese orders. So I know you guys break out the Asia Pacific part of it, but I was wondering, how much do you think is the contribution from Chinese memory customers in 2017, and how do you see that going into 2018?

Yes. I think the important thing to remember about our business is, because we're probe cards testing production wafers, we're always going to lag not just the orders the installation and qualification of equipment by somewhere between 6 to 9 months typically. And so you can think about the equipment guys and them first receiving the orders and then delivering the equipment, qualifying the equipment and ramping the equipment as a bit of a leading indicator for where our business is going to be. Obviously we remain very engaged in the local China, not just memory, but overall semiconductor ecosystem. I think the commentary we provided on engineering systems growth in helping these customers debug these initial memory chips is also a good leading indicator that there's some real activity going on there. But from a production probe card volume, we right now see it as probably a late 2018 event for us, once all this equipment gets, again, not just ordered but installed, qualified and running production wafers that then need to be probed.

And our next question of Ed Roesch with Sidoti & Company.

I think I have just 2 questions remaining, but the first one is with your probe card business set to grow, it looks like low teens this year, I want to check in with you, Mike, on the market itself which had been forecast to grow at a 7% CAGR. And would you say that your outperformance is primarily on the side of market share gains, or do you think that the market itself is well outperforming that 7% rate, and what would be some of the drivers there, if it's AS ASP's more, or volumes? If you could comment a little bit on what you sense from that.

Sure. We operate in kind of a niche market. Probe cards is covered by a handful of analysts, and not really in the real-time way that, for example, front end wafer fab equipment are. So I'm operating with a bit of a visibility problem on the overall market. Having said that, given that we lead in overall market share, our view is that the market has been stronger than it was going to be assumed to be at the start of the year when the 7% assumption came out. So I think it's fair to say that the market has grown faster or more significantly in 2017 than we initially expected it to. Having said that, though, I also believe that we've gained share on top of that expansion in served market that we've seen. We're going to have to wait until all of the surveys get rolled up and we get our results in early April 2018, because it's still -- outside the top 3 suppliers, there's a pretty long tail of suppliers. And so there's a lot of work that needs to be done to really understand what the market looked like and what different share positions there were. Given how we've executed on some of the line of sight opportunities, given how our large customers have grown, we do expect that we've gained share. I think the other data point outside of the probe card business that we've seen is the -- we're expecting double-digit growth in our engineering systems business. And that's a business that's historically grown at sort of a few percent. We've talked about the dynamics around China, around micro-LED, around silicon photonics. Those are still applications that we believe are in front of us, but we're pleased to have delivered some significant growth from that business as well this year.

Okay. Thanks. And then on the DRAM side on a trailing four quarter-type basis, you've an annualized run rate of about $130 million at the peak back in 2015, and it looks like you could approach $125 million with a pretty strong fourth quarter here in 2017. So I guess setting aside some of the minute differences, but would you expect the peak in this cycle could touch that $130 million that it did back in 2015 or do you think it could exceed that? What are sort of your general thoughts there, please?

So general thoughts on DRAM -- I do think the overall volume, given the bit growth, the wafer growth and some of the advanced packaging trends like high-bandwidth memory that drive more test content, we could certainly approach, and maybe even if there's a little more investment on some leading-edge nodes, exceed that peak of 2015. Again, most of that revolves around the supply/demand balance and the continued investment. I don't think it's any surprise, given the commentary from either our customers or some of the big suppliers, that DRAM's in a bit of a sweet spot right now, where people are growing bits but the demand is continuing to consume that bit growth. Now as long as we stay in that situation, I do think we can approach or maybe even sneak over those 2015 run rate levels.

And our next question comes from the line of Atif Malik with Citigroup.

I just have one question and a clarification. Mike, the capacity digestion comments are only with respect to one customer, correct?

That's correct. We delivered 20% higher than our previous record revenue to that customer in the third quarter. As they work through and digest some of that capacity in the fourth quarter, we expect that to be, if you like, isolated or concentrated behavior with them.

Okay. And then I think that particular customer trimmed its CapEx for this year and is showing this different financial discipline than it has in the past, and did not comment on next year's CapEx. So what is the confidence that's driving the outlook that Q1 could kind of return to higher levels for that particular customer?

Yes. Well, with all of our key customers we work pretty closely on capacity, capacity planning in their demand forecasts, and obviously our discussion about the 7-nanometer foundry customer were in many cases receiving orders for Q1 delivery already, as our customers try and risk-reduce. So I think the combination of the relationships we have, the dependence some of our leading customers have on us to meet their ramp plans and therefore the forecast and in some cases PO commitments we get from them to make sure that capacity's in place is part of what's driving, if you like, the confidence, or our behavior to continue to prepare for increased levels back in Q1 and potentially through the first half.

And that concludes our question-and-answer session for today. I'd like to turn the conference back over to Mike Slessor for any concluding remarks.

Thanks for joining us today, and thank you very much to the worldwide FormFactor team for delivering solid results again in the third quarter. Our positive momentum continues and we remain focused on driving a strong finish for 2017 and accelerating into 2018. Thanks a lot.

Ladies and gentlemen, thank you for your participation in today's conference. This does conclude the program and you may now disconnect. Everyone, have a great day.