PDF Solutions Inc (PDFS) 2016 Q4 法說會逐字稿

Good day ladies and gentlemen, and welcome to the PDF Solutions, Inc conference call to discuss its financial results for the fourth quarter and full year, ending Saturday, December 31, 2016.

(Operator Instructions)

As a reminder, this conference is being recorded.

If you have not yet received a copy of earnings press release and the Company's fourth-quarter and annual commentary, it has been posted to PDF's website at, www.pdf.com. Some of the statements that will be made in the course of this conference are forward-looking, including statements regarding PDF's future financial results and performance, growth rate, and demand for its solutions. PDF's actual results could differ materially.

You should refer to the section entitled Risk Factors on pages 12 through 9, of PDF's annual report on form 10-K for the fiscal year ended December 31, 2016, and similar disclosures and subsequent SEC filings. The forward-looking statements and risks stated in this conference call are based on information available to PDF today. PDF assumes the obligation to up date them.

Now would like to introduce John Kibarian, PDF's President and Chief Executive Officer; and Greg Walker, PDF's Chief Financial Officer. Mr. Kibarian, please go ahead.

Thank you, and welcome everyone. As you may have noticed, besides providing a press release for this quarter's financial results, we have included on our website a presentation which summarizes the financial information for the quarter and year. We hope this'll make it easier for our shareholders and analyst to see the results instead of having to write down the numbers as Greg reads them.

Today, I will start the discussion with brief highlights of the fourth-quarter and full-year 2016. I will also provide a summary of today's performance against our 2016 goals. I will then talk about our plans for 2017, and review our perspective of the general semiconductor market.

Overall, 2016 was the best bookings year in the history of the Company, and by a significant margin. The booking strength was reflected across our entire solutions and product offering, including IYR, Exensio, and our newest solution Design-for-Inspection.

During the fourth quarter, we added many new deals such as the following. A 14-nanometer IYR engagement for early technology development, that also included DFM templates in DFI with a foundry customer in Asia. An amendment to an existing 28-nanometer IYR engagement with an Asian foundry. A 40-nanometer IYR engagement with an Asian semiconductor company. A 10-nanometer DFM engagement with a US fabless company, and multiple Exensio contracts with several foundries and fabless companies.

The 14-nanometer contract signed this quarter, like the 7-nanometer contract signed earlier in the year, with a different customer, starts from the early development of the manufacturing node. It includes DFM templates and template ties to software, which were bought and being used by the foundry's internal IP provider, and third-party IP providers, to create foundation logic cells.

Customers are recognizing that IP built from templates demonstrate manufacturing benefits. In this 14-nanometer program, DFI is used to accelerate the early technology development. As we moved through 2016, we started to recognize usage for DFI with the eProbe-150, that we had not foreseen when we began the program. This has accelerated the program's early commercial adoption.

By including DFI in our customer's early R&D programs, we believe we can accelerate their process development, and also validate and demonstrate the on-chip instruments, which will be important for mass production control. This is the third commercial customer of our DFI solution, and we continue to discuss the adoption of our DFI solution with several other foundries.

Our Exensio Big Data analytics solution, in addition to hitting record bookings in the quarter and the year, greatly expanded its customer base, horizontally, deep into existing customers and new fabless and device manufacturers, as well as vertically into OSATs and system houses.

Our development of DFI and Exensio have allowed us to create new offerings for our customers that grow our total available market, add more IP hardware and software to our solutions. As Greg will discuss, this has helped us grow our solutions revenue at a rate faster than the past, as well as improve the gross margin of the solutions revenue, as customers recognize the value of the [high that margin's] software, IP, and hardware components of the solutions we are delivering.

In total 2016 revenue, at $107.5 million, was also a company record and represented a 9% increase year-over-year. This increase in revenue was driven by the strength of our solutions revenue, which grew 21% year-over-year to $77.2 million. Offsetting this gain was 11% decline in our Gainshare revenue, year-over-year to $30.3 million, due to decline in some customer's utilization rates.

We started 2016 with four key strategic objectives. One, deploy our Yield Ramp Solution for leading-edge fabless and foundries, which at the time we anticipated would expand during 2016 into 10-nanometer and 7-nanometer. Two, drive adoption of DFI, by increasing the number of chips with on-chip instrument measurement structures, installing early tools at multiple fabs, and demonstrating benefits, all while continuing development our systems.

Three, expand our overall business activity in China by supporting the entire semiconductor ecosystem there. And four, assist our key foundry and fabless customers by in applying Exensio-based solutions to improve their operational efficiencies, both on new products as well as mature products.

We have successfully achieved these objectives. We are engaged with multiple customers at 10-nanometer and 7-nanometer. We began commercialization of DFI, installing two systems in 2016, and closing a contract for a third. There are now over 50 chips that have DFI instruments at multiple foundries. We continue to see great reception at our fabless customers for DFI.

We signed multiple engagements with several key Chinese foundries and fabless companies. Collectively, these agreements include all our components of our solution, from DFM Templates to IYR, to Exensio and DFI.

And, for Exensio, our business activity was strong. We now have over 90 customers, moreover, at the press release -- earlier in the year, with Fairchild indicated, customers are enjoying the benefits of operational efficiencies on mature products. Our User Conference was well attended, and the keynote from SPIL, demonstrated the benefits of Exensio are being enjoyed across the supply chain.

Finally, for our more-than-more foundry customers, they have benefited from solutions that include CVs, and Exensio to improve their manufacturing effectiveness. All of this activity has served to greatly increase the number of customers utilizing our products and solutions, expand our relevance that the entire lifecycle of nodes, and across more of the semiconductor supply chain. We believe this is greatly increasing our addressable market.

Turning to 2017, from an industry perspective, we expect to see continued growth of volumes at 14-nanometer. Also, we are lucky to see the start-up of 10-nanometer production during the year. The 28-nanometer node will trend up over the long term, but be lumpy during 2017, as some customers are transitioning capacity to newer technology nodes, and some Gainshare periods end while new ones start up.

Our 2017 objectives are a continuation of our 2016 objective, as we believe our new products and market expansion represents a multi-year growth opportunity for us. Overall, in 2017, we intend to further broaden our IYR customer base, and exploit the new Chinese opportunities in logic as well as other process technologies.

Additionally, we intend to expand our Exensio customer base, both vertically and horizontally, as the electronics industry moves to manufacturing 4.0, by leveraging our Big Data analytics software capability. We intend to build on the early traction of our 150 series systems for DFI, with further commercializations and to complete the development of our 250 systems.

I look forward to speaking with you during the year, and letting you know how we are doing against these goals. I will now turn the call over to Greg. Greg?

Thank you John. For everyone, as Jon pointed out, we have posted in the Investor Relation's section of our website, a management report for the quarter and the year which gives detailed comments regarding the financial results of PDF. Given that, I'm going to focus my verbal comments today on the non-GAAP financial results for the year, and some key messages reflected in those results.

First, as John stated we had record bookings for both our IYR and Exensio offerings, which drove our solutions revenue growth to just under 20% for the year. Additionally, we ended 2017 with the largest solutions backlog in the history of the company. This record performance allowed us to grow our total revenue year-over-year, despite an 11% year-over-year drop in Gainshare revenues, as John pointed out.

Second, beginning in 2013 and 2014 we set out to substantially increase our position within the semiconductor industry in greater China, to take advantage of the growing investments within that geography. Since that time, Asia-Pacific has increased from 15% of our total revenue, to 42% of our total revenue in 2017. That is a significant change in the revenue structure of the company. We have seen, however, on the downside of that, an increase in our Accounts Receivable as the payment periods and collection issues out of China are a little more difficult than in the US or Europe.

Looking at 2014, we began investing in a new technology which we called Design-for-Inspection, with the goal of significantly increasing our addressable market and extending our relevance throughout the entire node life for any process technology. Our original plan was to introduce that were first-generation development platform by 2016, and subsequently introduce our second-generation in-line solution with first commercial shipments in early 2018.

Driven by customer demand, however, we were able to commercialize the first-generation platform and in fact generated some material revenues in 2016. During 2016, we were also able to significantly increase the capability of our Exensio solution, and expand its addressable market beyond foundries and fabless customers to test in assembly.

In order to take advantage of this expanded market, we increased our investment in our sales and support teams in the field during the year. As a result, we added many more new customers during the year, and expanded the overall marketplace, and also expanded our position with existing customers.

Looking at revenue, even though the mix between Solutions revenue and Gainshare revenue shifted away from Gainshare by more than 11% since 2013, we've been able to effectively sustain our gross margin levels. We have been able to accomplish this by improving our Solutions margins over that period, driven primarily by higher software content in many of our deals which have higher margins.

While our large R&D investment in DFI has limited our profitability over the past two years, we have been actively managing our other expenses to offset this impact as much as possible. As a result, we have remained profitable during this entire investment period, and have not had to utilize a significant portion of our balance sheet cash.

In summary, as John has stated, during 2016 we were able to successfully meet the core strategic objectives that we set out to accomplish. In executing against these objectives, we added many new customers through the expansion of our Exensio business, significantly grew our position in China, opened up a new market with our DFI solution, and developed many solutions for the more-than-more, or older (inaudible) process nodes.

We feel that we are well on our way towards achieving our goals of increasing the addressable market for PDF, and significantly reducing our overall business risk, particularly as far as customer concentration. In the management report that I referred to earlier, there is a discussion of our 2017 outlook for both revenue and spending. That is available on the website.

With that, I will turn the call over to the operator for Q&A. Operator?

(Operator Instructions)

Jon Tanwanteng, CJS Securities.

Hi John and Greg. Good afternoon, and thanks for taking my questions.

Sure.

Hi, Jon.

Can I ask, what you have built into your 2017 outlook in terms of DFI machines or revenue, and if that includes any shipments of the gen-2 machine, or not?

This is John. We do anticipate completing the development of the generation 2 machine, and being in a position to ship it to a customer in the second part of this year. But, we have not forecasted any revenue associated with the gen-2 machine at this point. All the revenue that we forecast in 2017 is based on the generation 1 machine.

Got it. And the, if I can go out a bit further, do you envision growth accelerating meaningfully into 2018, if you are successful with the 250 series machine? And how should we think about that from a quarterly run rate, in terms of number of machines sold? Or any kind of metric that?

Yes. So, I think it is a bit early to specify how many machines we expect to sell, or install, a quarterly basis. But obviously, this has been a big investment for us. We do anticipate, if we are successful, meaningfully accelerating the rate at which we grow the business, as you get out in 2018 and beyond. Greatly, in part, because of DFI and the solutions that the 250 system will enable.

What has been a positive for us, is that actually the 150 system, has enabled a number of very powerful applications for our customers, which is why we're generating business and traction from it as well. And that could also continue into 2018. There are some applications towards the150 we believe is particularly well-suited for, and its costs from our perspective are more modest than the 250. So we would be inclined to continued use the 150 as we go out into 2018 as well.

Got it. That's helpful. And then, just switching over to the Gainshare side. Your largest customer just announced a 20% increase in their production capacity. How is that factored in to your outlook at all, in 2017? And does that impact the Gainshare?

Customers announce capacity expansions frequently. Those usually take two years from when they announce them to when you see first meaningful -- about a year to build the building, some number of months after that to install, and then your getting to about usually about 3,000 to 5,000 [micro-starts] after that.

We always look at those things with a very careful eye. They quite a while for them to materialize, and in the industry we're in, it's quite cyclic, so sometimes before they are able to get those things up and going something can go awry. If you look at that -- unrelated to us, but if you look at the announcement about the plant in Arizona for Intel, that was announced, I think, multiple times over these last five years. We are excited about that, but we factor it pretty cautiously.

Okay, great. And then, Greg, just on the SG&A in the quarter, there was a pretty big sequential jump. Was that all the additional sales you're talking about? Or is there something else in there that we should be thinking about? And is that the run rate we should build from going forward?

Yes. The majority of it was related to accelerators been reached by sales guys on their bookings. It is a one-time jump, as far as the spike, and then we will be in more of a revenue-stable, revenue-driven growth after that.

Okay, great. Thank you, very much.

Tom Sepenzis, Northland.

Hey. Thank you for taking my question. I'm just curious, with 90 Exensio customers, can you give us some sensitivity, or in terms to the rate contribution in 2016 and 2017 from Exensio? Or is that really just being used to bring customers into your other products?

John, do you want to talk to that?

Sure. Tom, there's a long tail on the Exensio customer. So, we have customers where their contracts represent millions of dollars per year in revenue for us, so the bookings may be anywhere from let's say $3 million to $5 million on a multi-year time-based license, to customers that are using just some small fraction of the system, and their annual spend with us is $20,000 or $30,000 a year. It really is quite a wide spectrum.

And then, there are a number of customers, as I alluded to in my prepared remarks, foundries who are using a combination of CVs and Exensio to drive manufacturing operations improvements on those 55-nanometer and 90-nanometer, and those are multi-million dollar per year contracts, larger than the (technical difficulties) customers. They constitute multiple elements.

Collectively, it represents a meaningful part of our Solutions revenue, and that is why the gross margins are going up. Even in the case of DFI, DFI ships with an Exensio big Data module for doing all of the analytics. So, it is hard to split it out per se, but it is growing as a factor of our Solutions Revenue, and hence the gross margins are going up.

And, you could ask, well, why do you keep around the smaller customers? Well, we find the smaller customers often get acquired by larger companies. And when customers, through this consolidation, a customer(technical difficulties) we have a much more complicated supply chain, particularly our fabless customers and our analog asset-light manufacturing customers.

They find that, many parts of the organization are already familiar with Exensio, and so we have been able to get customers to standardize on our platform, in part because many of their acquired elements are already familiar, either with Exensio-Yields or Exensio-Tests or Exensio-Characterization, or one of the modules. And we build from there. We have had, during 2016, a number of the bookings were from customers that were merging, standardizing on Exensio and increasing the run rate as they deployed it across. They're now a larger enterprise.

Great. Thank you for the color there. And then, just given the annual guidance that you had on the website, it seems that the return on the investment from DFI is really going to be more of a 2018 event rather than a 2017 event. Am I reading that right? Or is it something that starts to -- does revenue start to grow a little bit quicker than the operating expenses in the second half of this year?

Yes, this is Greg. Go-ahead John.

Tom, we anticipate DFI to grow throughout the year in terms of its revenue contribution, and be more significant in 2017 than it was in 2016.

Great. Thank you, very much.

Christian Schwab, Craig-Hallum.

Good afternoon, guys. So, as we look at the operating expenses, we did $43 million in non-GAAP operating expenses in 2016 and we talked about low- to mid-teens revenue growth in 2017, but we said we would be spending more money. Is that -- should we think about the growth rate of OpEx to be about half of revenue?

Yes. This is Greg. Without specifically commenting on how much, I think you are right. There is some leverage in the model. What we said was, with the exception of the spending on DFI R&D development, because we're still developing the 250 series platform, the rest of the Company's revenue -- expending growth, should be commensurate with revenue growth.

What that means, is when you look at the field organizations, and the cost of sales, and the sales expense, those will grow pretty much close to the revenue growth rate. But some of the overhead, particularly in the G&A organization and things like that, won't grow at that rate at all. What that means, overall, is that spending will grow. We hope to get a few points of margin leverage at the end of the day, though.

Just to make sure I am thinking about this correctly. If we, -- just absolute numbers. If we have 14% revenue growth, we should have 7% OpEx growth, and we may get a point or two of gross margin improvement?

Yes. Okay. One to two points of gross margin. Perfect. And then, as we look to 2018, and thank you guys for discussing that. What should we be monitoring, or listening for, looking at, throughout the course of 2017 for all of us to get increased conviction that the DFI-2 product will create meaningful or measurable revenue in the 2018 and beyond timeframe?

Let me take that one Christian. I think a couple of things. First of all, we become increasingly more confident that we can drive revenue even off the eProbe-150. So, we expect in 2017 for growing revenue contribution from the 150, we have started setting goals to have those systems driving revenue for this year, and being a material contributor.

Because of the ratable nature of the DFI business, those revenues, for many cases the customers continue out through 2018 as well. So, we expected that to be a significant fraction of the business. You should be listening for us talking about additional customers deploying DFI on the 150, and expanding the number of 150 systems they have in their facilities. That would be a good indicator to 2017.

Number two, you should be listening for us talking about the number of fabless customers putting content inside their designs, because really that is the way that we see the market. That should propel throughout 2017, and we anticipate expanding the number of fabless customers and IP providers that are using DFI in their design elements. You should be listening for that.

Number three, you should be listening to us talk about the progress we're making on the 250 in terms of being able to complete the development, and then start shipping at the end of the year to customers that want 250 system for applications that the 250 are particularly suited for.

That will be a lagging indicator. You want to see the other two indicators before you see that one. You should see all three.

Perfect. Was there a fourth, or was there just the three?

No, that is it.

Okay, perfect. Just, because there were so many comments said earlier, and I have notes all over. So, remind me, right now, how many DFI 150 customers do we have as of today?

Today, there are three. We signed in the fourth quarter, for the third customer on DFI.

So, we have three DFI customers. And then remind me, of how many fabless content design wins do we have right now, today?

As I said there are over 50 chips, combination of test chips and full-product chips, that have DFI content in them.

Perfect. Thank you for that. As we look to, 2018 -- one other question -- if we didn't have DFI initiative and success in the eProbe 150 et cetera, what with the revenue growth for 2017 be?

It is very hard to, one, estimate that, because we have be coming off a little bit lower base. But also, when you look at these DFI deals, in many, many cases, they are combined deals with IYR, and even Exensio in most cases.

When you look at that, depending on the nature of that deal, a lot of it is reported in the Solutions revenue space as opposed to any standalone DFI -- I mean, in the IYR revenue space. So it would be really hard to predict that, but clearly without DFI, our growth would probably be a little bit slower, but you cannot really identify that.

Okay, and then, just lastly, if, given the initial success of the 150 and then the DF-2 coming, can you quantify what you think the revenue opportunity could be in a broad range in 2018, 2019 or 2020, should we be modestly successful, or not very successful to massively successful?

That's a great question, Christian. We talked about that at our Analyst Day back in November. As we -- I think we talked about it, if we're even able to, let's say, seed the world with on the order of 10 systems, we anticipate that would drive $10s millions of revenue per year on a ratable basis. That would greatly improve the gross margins and the profitability of the Solutions revenue.

Because, as we said, in Analyst Day, our DFI 150-based Solutions systems generate between 70% and 80% gross margin points today and we expect that to grow over time. That would be, DFI is a very useful tool for customers but it is not a home run, so that is the low-end case. If you then go back and say, DFI is able to be a full in-line production control application for customers, and valuable over some number of layers to the manufacturing process, then as we discussed in the Analyst Day, the market opportunity for DFI is many times larger than the business is today.

At what rate to grow into that, over let's say a three- to five-year period, depends greatly on what the invest levels in the nodes, your ability to penetrate multiple layers, but then that would suggest that DFI would contribute more revenue than the Company is today. [You can have a sip]

Okay. So, if one wanted to take a multi-year outlook on the Company, not just looking next quarter, or next year, we would assume that with the information we have in hand, that we should at least have modest success with the DFI and DFI-2, which should allow us to continue to drive low- to mid-teens revenue growth.

I don't want to put words in your mouth, but let's just say for, at least the next few years, with the opportunity for gross margins to expand at the same time. And should it become a home run, then it is a home run. Is that fair?

Yes. That is a great way to look at it. The way I think about it, right? Because if you remember, if you go way back, with [Kevin and Tommy, bootstrap] PDF. So I have always had this conservative way of looking at things. The excitement that I have about the success in 2016 is, I feel pretty good that we're going to do better than the bottom case on this, because we see applications, where even the 150 is really valuable for customers. We think it is quite a useful thing, and we can get to that case where it is going to allow us to grow our Solutions revenue, as you said, in the mid-teens, improve our gross margins. And as we proof out the applications that drive even more usage, some of which that are enabled on the 250, some of which we think are even enabled on the 150, we believe we have a chance at growing the business even greater than that.

For years, if you remember, we've told the investors Solutions is a single-digit growth business, and Gainshare is a double-digit growth business. Now what we're really tell you all, is we do see really solid growth in the Solutions business, as well. Right? We into 2016 saying it was a mid-teens growth year, and it exceeded that if you look at the way 2016 ended, from a Solutions growth standpoint.

My last question, if I may, on Gainshare, you guys talked about revenue increasing greater than 10% in 2017, driven by the continued ramp of 14-nanometer volumes. Does that have to do is an expansion of customer base, or does it have to do with more of the well-telegraphed production capacity expansion of your largest customer?

When we factor in our analysis, we always -- and it is part of the reason why we gave the caution on the 28th. We always factor the stuff that is volume expansion of a known facility with one customer -- customers that have already design once, that have been in that factory, with qualified deals.

We always build that in, in our modeling. And the things which involve startup of new nodes, startup of new products, startup of new facilities we always discount that, greatly. So, our confidence in the year, that we would communicate, would be based on facilities where we felt the volume was going to grow because they have done their homework in 2016 and before.

Alright. Great. No other questions. Thank you.

Tom Diffely, DA Davidson.

Good afternoon. So, maybe following up on the last few questions on DFI. John, I think you mentioned you that the 150 had a pretty long life ahead of it, even if the 250 is out in the marketplace. What is the difference in the type of application you would expect the 150 to be doing instead of the 250? It seems like the 250 is so much faster, that it would dwarf the capabilities of the 150.

It does. What we started seeing customers doing in 2016, was -- we had anticipated the 150 really only being useful for test chips. So we said, you're going to use it for early R&D. That is why we coupled it with the IYR for customers, because it gives them a shorter, faster learning loop on some of the layers, that if you want to test electrical you need to process many layers, from then on, for the middle of line, you want to learn quicker.

What happened was, customers liked some of the structures. We started seeing ways that we could put them inside the scribe line on their early customer [tape-offs]. The customers at 14-nanometers, 10-nanometers, and 7-nanometers, the alignment of the layers becomes a really big challenge. And the process window -- in other words, the tolerance to that alignment is very different depending on the combination of layers.

With a DFI, you have a way of electrically measuring the amount of misalignment you have got, as well as the width of the process window. We were able to show that it correlates with product yields to customers.

And so, now you can take this application and use that data to improve your lithography control loop, as well as your upstream etches and depositions, in terms of widening that process window. We start to see production control applications, for being able to measure a modest number of test structures, in the millions instead of billions, that is actually value for customers.

Now we've got customers that are using that data and turning that into their factory control databases, to improve their manufacturing control on their leading-edge nodes. That is what has gotten us to say, wow, there's an application we didn't anticipate. We didn't think people would be using the 150 on volume production chips. We thought they would only be using it on test chips. It turns out they can use it in volume chips.

Is that data for valuable for the foundries, for the process, or for the fabless guys, trying to get the feedback loop for the next design?

The test chip work we thought would be useful for the fabless guys, and it has been because they learn about specific structures that may or may not be easy to manufacture. The scribe-line application that I described, is really for the foundry to improve their control.

Yes. Okay. So, based on that, and that success early on here. Would you expect to ramp up shipments of the 150, then, over the next year? And if so, are you capacity constrained at all? Or is it just a matter of educating the customer of what you can do?

We believe we can keep up with the demand, so far, on being able to build out the systems. Our real bottleneck right now, Tom, is we have got to design those structures, put them on a customer's chip, collect the data, show them how they can be used for production control, and that is, it is a very powerful way of getting a feedback loop.

I always joke with customers, that, it seems like we go around the world proving that the physics is the same in every geography. But as much as I crack that joke, the customers still expect, let's do a pilot, let's tape out a DFI scribe line, send it away for the PDF, PDF will test it in our clean room, show them what they can do, and then that begins the dialogue on how they could then deploy it inside their facilities.

And we have a number of those ongoing right now with customers. That is really the long-lead call, on how fast we could move in 2017.

Okay, and is most of that work done at the 10-nanometer node right now, or is 14-nanometer? Or how does that change when you go from 14-nanometer, 10-nanometer, to 7-nanometer, as far as the demand for your tools, you think?

We have activity going on, on all of those nodes. We also have some activity with customers on 28-nanometer, and we have some activity on advanced memory technologies, with DFI, in discussions with customers and pilots.

We think the application space is broader than just leading-edge, but we are testing that. I don't think we have the confidence to say that it is, right now, Tom. We will be very excited if we could prove that it is valuable for a lot of those other non-leading-edge nodes.

Okay, and then last question on that. Is one tool per fab enough for this type of measurement?

We don't believe it would be. But, we're still early in proving that out. That is another big -- we set a goal for ourselves, internally, to prove this year that that's not true, that you need multiple to be able to do this work, but we haven't proven that yet.

Moving over to China, obviously, a big potential market. How big is that today for you on the Solutions side? And based on all of the incoming inquiries, how big do you think that can get over the next couple of years as a percentage of your business?

As it Greg pointed out, in his prepared remarks, Asia, overall, has gone from about 15% of revenue in 2013, to, I think, 40%-something in 2016.

Yes, 42%.

A lot of that growth is China, Tom. China, and to a lesser extent Taiwan. How big can it be?

I think I said in our Analyst Day, if you just look at the number of opportunities in China, and compare it to the yield ramp business we have outside of China, if they actually build out what they say they are going to build out, and I know there's a lot of skepticism in the semiconductor market about whether all of these plans are really going to materialize. That opportunity is as big as PDF's.

There are as many fabs as we're collecting Gainshare from to today, that are being planned. They are, greatly at 28-nanometer and below, where we think we have a lot of value to offer them, and electrical characterization, we think is particularly important.

They, with their fabless and system companies, we have had a lot of activity with Exensio to control their operations. So we see, really as I said in my prepared remarks, we see the entire Chinese semiconductor ecosystem, because it is being built new, as -- PDF existed before it greatly existed, they can really leverage PDF in ways that it has always been hard for us to slowly convince US fabless customers how to use PDF more broadly.

We are very excited about that opportunity. We think it could be as big as PDF is. And the growth rates would suggest that it is on track to do that.

I guess if you look at your guidance for Solutions for the year, in only high-single digits, when both DFI is ramping nicely, and China is ramping nicely, is there a segment in there that is not doing well that is eating away at some of the growth rate?

If you go back to 2016, Tom, this time last year we said we thought we would do low- to middle-teens, if I remember correctly, for 2016. And as we went through the year we became more confident with the bookings, and we slowly increased the actual performance, and so, the Solutions revenue growth rate, I think, was closer to 20%.

As we look out this year, if you talk to all of our team, we see the bookings for this year looking to be very strong and potentially another great year, like 2016 was. But, there is always -- our contracts always take a while to negotiate and sign et cetera, and it slips one or two quarters, and then that greatly changes your horizon.

So, at this point of the year, it is prudent to forecast it to be not as good as growth as 2016, even though, as you point out, we have a lot of irons in the fire that we believe could make it a better year than 2016. But it is early to say -- to go and put a stake in the ground and say it will definitely happen.

As for his Gainshare goes from China, that will lag a little bit, here? Maybe a few years out?

No. We anticipate seeing Gainshare from China starting in 2017. But, again, because it is new, we tend to forecast that to be a little bit behind what the customers' own models would suggest.

So, we anticipate by the end of 2017 to be collecting Gainshare from our customers in China -- or at least starting to.

Okay. And then, just to go back to your big-picture, long-term view, you said earlier that DFI could double the revenue at higher margins than you currently have. It sounds like China could double revenue, at maybe slightly lower margins than you currently have. But net/net, the combination of those two is pretty explosive on a multi-year basis

Yes. We're very excited about the investments that we have made. I think if you look back at the materials we prepared for the Analyst Day, we chose to do the Analyst Day then because we felt that there was a lot of exciting things going on at PDF, and we have made -- we've had a lot of very patient investors, that have been with us as we've a lot of these investments over multiple years.

As we are saying today, we are not done with the investments, but we're starting to see a lot of the benefits. You can tell in the 2016 numbers, and see that more in the 2017 and 2018 numbers.

And then, just two for Greg. On the downside, you said collections in China are a little bit more difficult. A little more color on that would be helpful, just to know what kind of risk you might have.

Yes. When you look at it, number one, just on regular normal business transactions, payment terms in China tend to be longer than you would see, normally, in the US or Europe where you are looking generally at 30 to 45 days. On top of that, to date, our business in China transacts on a dollar basis, so all of the cash related to revenue comes back to the US, which means you have to go to the currency process in China to get hard dollars out, which takes some time.

Now, we're looking at, as part of our overall business strategy, should we look at changing some of that for part of the revenue streams, which would make it easier and mitigate some of that difficulty in getting collections.

We haven't seen this impact our actual ability to collect. It just strings out the time. We factor that in when we look at cash flows, and things like that. It is just a fact, if you are going to transact dollars out of China, it is going to have some sort of extended collection period on it.

The last one is just on the SG&A line, that you said took a little step function of in December because of payments in sales. Is that something -- it is unclear does that come back down and then ramp with sales, or percentage of sales growth? Or does it stay at this level and ramp from there?

It should come back down a little bit as we get into Q1, and then ramp from there with sales revenue.

Okay, and then I guess, one more. On the research and development side, is that mainly people cost, or are there tools in there that you'll peak at some point later this year after the 250 is out?

It's actually starting to move from 2015 and 2016, where it was a combination but heavily weighted with people cost, to, as we go through 2016, toward 2017, it is going to be more on the third-party vendor cost, as we're moving from design and engineering to actual prototyping.

So, just NRE, then?

Yes, a lot of NRE, some capitalized software development, and things like that. We have been saying we expected R&D to increase by $300,000 to $500,000, each quarter over the last six quarters or so. I think that by the time we get to the second half of the year, that will probably slow a little bit.

Okay. Well, thanks for all the detail today.

Sure.

Mike Cotongo, Cardinal Capital.

Hi. This is Gene Fox and Mike Cotongo. Just a couple of questions.

When it comes to your Solutions guidance, do you -- it sounds, John, like you probably do a very detailed one and then, I would say, adjust it based on contracts. But, a couple of questions. Are you assuming more than the three customers for DFI-1 in the guidance? And as it relates to that, when do you get visibility or the year? Is that something by the middle of the year, do you have enough contracts so you should have a pretty good idea where you are going to end up?

Yes, that is a great question, Gene. So, when we built our model and guidance for this year, we assumed no more revenues other than these customers -- many renewals that these customers would need to do. So, the first three systems surviving in the world, and that is about it.

That of course is not -- we're spending to do more than that, but as you said, when you're in the middle of a pilot in the beginning of the year, you do not want to go and forecast that you are going to be successful, and then assume success. You like to be little bit more conservative.

So as you get to the first half of the year, you have better visibility about what you already booked, where you are in negotiations with customers, et cetera. If you go back and look at the transcripts in 2016, you will see the confidence that we were going to do something with DFI in 2016, increased as we started -- we were in effect negotiating those contracts -- the second contract in the first quarter, although it didn't sign until the second quarter.

So, our confidence level went up as we went through 2016. And we anticipate the same clarity will come in 2017 while we get through the first half of the year. And remember, what we are talking about here is bookings, not so much revenue.

So revenue -- when you make a big booking in PDF, it has a very de minimis impact on that quarter's revenue, typically. That tends to build the base for future more than it changes that quarter's numbers.

So, your point, John, is, even if, as you sign up more customers for DFI-1, there is just a lag between the signing and when you see material economics.

Yes. You start seeing -- if you look at when we signed that contract in Q2, it was really Q3 before we started seeing some material economics. It did not start getting a steady-state, I think, until Q4.

It takes a while. It is a very ratable business model. The good news is of course, it then builds.

Understood. John, can you give us a little more granularity into where we stand on the DFI 250 in terms of -- update us on what was accomplished in the quarter, and maybe a little more granularity on what you need to accomplish over, at least the first quarter, and then anything else you are willing to share for the year.

We anticipate in the first half of the year, being able to run -- have a system fully come together and be able to run customers' waivers. And then, do that work with customers in the second half of year. So, the first half of year is where they mostly lift up, by the second half of the year we start engaging with customers on the 250.

By and large, all of the components are built, and are being tested in some combinations of them being put together. And of course the software layer which controls all of that, which greatly benefits from our investments that we've made in the past on the 150, because we used a similar software approach.

So, 2016, for those people that came out to our clean room, you would have seen some components. If you were to come into over clean room today, you would see what looks like a machine with the skins taken off. Because there is no point putting skins on it, because all the components, with the exception of, I think, one have been put on the machine and are being integrated and tested within each other.

John does that mean you've basically chosen all of the suppliers, and are we locked in on that, or we still have more work to do?

We basically have been working with all of the suppliers. I do not believe there's any real changes going on. Nothing substantial, maybe we're changing a different wire supplier, or screw supplier, that I do not know about, but for the major components, those have all been locked in. And their systems have been shipped to us, and we're working with them at their facility.

Last question, from a technology standpoint, John, with the big obstacles on the 250 that we need to solve? Or is it literally just putting it together making sure it works? It's more of a get it, put together, make sure it does what we say it is going to do, as opposed to having to solve problems?

Gene, from my 10,000 foot level, there is always system integration challenges. I always joke that engineers' idea of the perfect 30-day teamwork project, is on day one, split the project up into parts, come back on day 29 and it all comes together. There's always integration challenges.

I'm always nervous about engineers working on big systems to make sure they have thought to all of the system choices. And those are always your last to resign risks. The system itself, without getting into the details that would really tell people how we're actually getting to a billion instrument measurements an hour, the system itself has a number of components which are very different. We have thought of it as, it's like an e-beam writer, but with a column that is more like a voltage contrast machine.

We have been retiring risks on our ability to achieve the alignment tolerances that we need, and we did a lot of things to make that happen. And we were retiring risks on being able to see the field of view that we need to see to be able to make that, seat in the system and get the throughput.

Those are probably the two biggest areas, retiring risk on alignment and retiring risk of the ability to see a large field of view. I can't really tell you exactly where we are with those things. We are pretty far along on them, though.

John, congratulations and great job. We, obviously, are very excited.

Thank you. Thank you for your support, Gene.

Brian Freckmann, LS Capital.

Hey guys. How are you? Almost every one of my questions has been answered, so this is just kind of a -- on DFI -- I know this is more speculation than anything else, but in your estimation, you have three wins with customers right now. In 2017, does DFI come from land and expand, you get multiple orders from certain customers, your best guess? Or do you think you grow DFI by getting all new customers?

Yes, you're right. It's a big speculation. We're trying to do both of those things, Brian. I would like us to be successful at both.

In other words, expand the number of applications, the number of systems customers need at our existing -- the early wins that we've had, because we're still early in those customers, and we certainly don't feel like we're anywhere near penetrated. And, as well as, achieve new customers to start deploying their first systems.

We would like to do both of those things throughout the year. Which one, -- you can't be capping which one we have a chance of doing. I do not know that. My guess would be as good as yours.

I guess on that note, maybe better understanding the sales cycle, at some point you're midway through the first quarter, you have some estimations of wins and sales cycles. I guess I was thinking, from that, if you looked at the current sales cycle for the 150, is it new, or is it current customers?

Part of the reason why we talk about the number of fabless customers, or the number designs that have our contact -- our biggest channel is the fabless companies. The fabless companies really believe that their designs can make inspection harder or easier.

So, they are very willing to put content on their -- technically, their test chips at first, to make the inspection problem easier. And they do that with us. Sometimes, before they even tell their foundry. And then after that test chip is taped out, they will go back to the foundry and say, hey, we put this new design for inspection content on our chip. We would like you to send away for the PDFs of PDF can show you what it can do.

If you go back and look at our second customer, really that is how we got that customer in 2016. It was a fabless customer that really was the pioneer, said this is a great thing, we should be doing this, and included the content on their test chip with that [marge] foundry, and then the foundry sent us a way so we measured it, and we showed the results, and then we negotiated a contract. But from that first tape out, to when we got the results, until we signed that contract, it was about a year.

We seed the ground by working with these fabless entities, and then we go back and harvest when the chips -- when they direct a chip from their foundry to us.

Okay. That's a good update. Alright, thanks.

Andrew Weiner, Samjo Capital.

Hi. Good afternoon, guys. So, following up on that question, I guess, the way I would ask it is, you have two150s as customers, you've got an order for a third -- what is the current build plan for 2017? How many do you expect to be building or have in inventory? To give us a gauge of what you think the opportunity set is over the next, call it 12 to 18 months.

We have another six or so, that we have built or are partially built at this point. We do anticipate at least a couple of them staying in our facility for the purposes of doing further engineering, R&D, and pilots et cetera, to do some of the work that I described in the last question with Brian. We know we have the ability to build more capacitors here. (technical difficulties) That would give you a baseline, Andrew.

And then second, with respect to the fabless, we have had conversations where the six or seven biggest fabless drive a disproportionate amount of the advance-node volume our way for purchases, and obviously they will have a significantly bigger influence on what foundries will or won't do. To what extent have we been successful in getting on-chip instruments in those top six or seven fabless?

I believe right now, four of the top six, either on test chips with full-product wafers, have DFI instruments.

And then, just following up on an earlier thought process with respect to the spending this year versus the leverage. Greg, I think you had described at the Analyst Day getting EBITDA margins back to 40%plus. I think the 150 million or so run rate. Is the way to think about that in the context of 2017, 2018 is that you are ramping the solutions business in China very quickly, but you will see only a small partial-year impact of the Gainshare component? But that 2018 Gainshare in China should grow very rapidly, and become a meaningful piece of the overall gain share?

Yes. I think that is true, and you combine that with some level of DFI growth as we go into 2018, from both the 150 systems, which you start to get full year's worth of revenue off of, plus any new revenue coming in from the 250 series. That basically, should push you back toward that level we were talking about approaching a 40% number.

Along those lines, I will make a final comment on 2016. From the analyst day, basically you could see really starting in 2014, going through 2018 to get back to the original model, we were going through about a four-year investment cycle. At the end of 2016, we are halfway through that.

I think one of the good indicators of how effective some of the things that we're doing are, is to look at one of the things that we often talked about before, which was our customer concentration and if you look at that -- and we look at it as a number of customers that are 10% customers, and how much of our total revenue are they generating. If you look at that, over time, that really peaked in 2014 at nearly 80% of our total revenue was coming from our few 10% customers.

If you look at that, now as we exit 2016, that number is down to almost 50%. That is a great widening of the customer base, where we are having significant revenues coming out of multiple geographies, multiple customers, and multiple technologies, now.

So, halfway through our investment cycle, we're seeing some very positive results and we're seeing some real impact on the risk position of the company. We have got another two years to get back to where we want to be, but we are on that path.

And then, just with respect to Exensio, is there a, some sort of quantitative measure, whether it is growth in bookings or something that would provide a little bit better color, as to the rate of growth or the scale of that business?

Yes. If you'll recall, we used to, we talked at the end of last year and early this year, about over two years, being able to double the scale of Exensio, from where was back in 2014, I believe in into 2015. And, we are definitely on track for that.

Without -- we don't talk about the individual revenue levels because so much of the business is mixed with IYR and DFI, but when you look at the bookings levels, and how that plays out into revenue streams, we're definitely on track for that kind of growth rate.

Okay, and when you say on track, so that means, as you look forward, when you think about the opportunity, do you feel those growth rates are sustainable for the next couple of years, despite the larger scale?

I feel that, we're talking about a two-year doubling -- doubling the scale of the business over two years. We are halfway through that, and we are on track to make that number.

Great. Thank you.

At this time there are no more questions. Ladies and gentlemen this concludes the program. Thank you for joining us today.

Thank you everyone.