U.S. Markets closed

Edited Transcript of PDFS earnings conference call or presentation 27-Jul-17 9:00pm GMT

Thomson Reuters StreetEvents

Q2 2017 PDF Solutions Inc Earnings Call

SAN JOSE Aug 21, 2017 (Thomson StreetEvents) -- Edited Transcript of PDF Solutions Inc earnings conference call or presentation Thursday, July 27, 2017 at 9:00:00pm GMT

TEXT version of Transcript

================================================================================

Corporate Participants

================================================================================

* Gregory C. Walker

PDF Solutions, Inc. - CFO and VP of Finance

* John K. Kibarian

PDF Solutions, Inc. - Co-Founder, CEO, President and Director

================================================================================

Conference Call Participants

================================================================================

* Christian David Schwab

Craig-Hallum Capital Group LLC, Research Division - Senior Research Analyst

* Jonathan E. Tanwanteng

CJS Securities, Inc. - Research Analyst

* Thomas Robert Diffely

D.A. Davidson & Co., Research Division - MD & Senior Research Analyst

================================================================================

Presentation

--------------------------------------------------------------------------------

Operator [1]

--------------------------------------------------------------------------------

Good day, ladies and gentlemen, and welcome to the PDF Solutions Inc. conference call to discuss its financial results for the second quarter ending Friday, June 30, 2017. (Operator Instructions) As a reminder, this conference is being recorded.

If you have not received a copy of the corresponding press release, it has been posted to the 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 rates and demand for its solutions. PDF's actual results could differ materially. You should refer to the section entitled Risk Factors on Pages 10 through 17 of PDF's annual report on Form 10-K for the fiscal year ending December 31, 2016, and similar disclosures in subsequent SEC filings. The forward-looking statements and risks stated in this conference call are based on information available to PDF today. PDF assumes no obligation to update them.

Now I'd 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.

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [2]

--------------------------------------------------------------------------------

Thank you, and welcome, everyone. If you've not already seen our earnings press release and financial results presentation for the quarter, please go to the Investors section of our website, where they are posted. Our discussion today will start with a summary of the business for -- in the quarter, followed by an overview of the general environment. We'll then summarize some developments in our Exensio Big Data analysis platform and progress we are making with Design for Inspection for the quarter. I will then summarize my remarks and turn the call over to Greg.

Revenue in Q2 was flat compared with Q1. Solutions revenue was down quarter-over-quarter due to delay in expected contract closures. We believe this delay was a result, in part, of the recent softness at some of our foundry customers that we'd discussed in the Q1 call. To be conservative, we anticipate most of these contracts closing over the next 2 quarters. Gainshare, however, recovered in Q2. Although existing 28-nanometer, of all the nodes, gainshare remained soft, initial contributions of 28-nanometer gainshare from China and strong 14-nanometer volumes as well as our achieving R&D milestones resulted in significantly higher gainshare quarter-over-quarter.

In Q2, we closed the following significant contracts: an Integrated Yield Ramp contract for a new 3D NAND customer in Asia, an Exensio renewal for an existing foundry customer in Asia and an Exensio renewal for an existing analog components customer. The contract we signed for 3D memory was the result of a successful pilot that was completed in Q1. This contract includes CVs and Exensio for use in development of a 3D NAND product and process. Like our foundry contracts, it has gained share based on achieving technical milestones. We are very excited to begin deploying Integrated Yield Ramps for 3D memories as we see both standalone NAND and embedded nonvolatile memories as good applications for PDF's technology. For these technologies, a logical characterization can be more important than it has been for conventional planar technologies.

Turning to the general environment. In our core market, logic foundries, we saw a mix of customer confidence, with some relatively bullish on their business and some less confident. 28-nanometer capacity at nonleading foundries continue to be weak. We saw some leadership changes at some of those foundries in the last quarter. Overall, we continue to believe that our business with our customer should strengthen. We have seen significant presales activity in Integrated Yield Ramps for logic and nonlogic applications, Exensio pilots for fabless and fabs as well as DFI deployments.

Now moving to Exensio, our Big Data analytics platform for manufacturing. We had 2 significant events in July after Q2 close. First, we had an Exensio users conference with participation from 34 companies, including foundries, fabless, system companies and OSATs. At the conference, customers presented their applications of Exensio, spanning factory control, yield improvement, product test and product quality. The conference reinforced our belief in the accelerating need and opportunity for analytics, leveraging integrated data from the fab through final package test data and that PDF's uniquely positioned to meet this need.

The second recent event was the acquisition of the significant assets of Kinesis Software. We acquired Kinesis due to their leading ability to track individual die from wafer through final package. This capability greatly enhances the analytics resulting from integrating fab through final test data. Kinesis has a strong position in assembly fabs, with its software deployed to thousands of assembly tools around the world. Combined with Exensio's connection to over 20,000 tools in wafer factories and thousands of connections in test floor, we believe Exensio is the only commercial system with full integration of front and assembly and test equipment. While Kinesis was generating revenue, it was small relative to our Exensio revenue. We believe the acquisition enables PDF to develop high-value solutions addressing chip quality and reliability. These applications depend -- are dependent on and require data from all aspects of chip manufacturing and vital to growing markets such as automotive.

As we've done every quarter, we are providing an update about Design for Inspection or DFI. As you know, DFI's our innovation for wafer inspection metrology. By placing small on-chip instruments on the wafer in empty space, and typically 5% to 10% of chip area is empty, we can provide a proprietary inspection tool that can read those instruments. Coupled with our Exensio analytics platform, we've been able to demonstrate that customers can see electrical defects that are typically not possible to see with in-line inspection. As more integrated circuits employ 3D structures, and electrical failures become more critical than geometric failures, DFI has the potential to be a new source of process control for fabs and greatly expand the available market for PDF Solutions. In the second quarter, an eProbe 150 was placed in 24-hour production mode at a customer's factory. This demonstrates the capability and stability of the system. Moreover, the DFI program at one customer was extended from 10-nanometer to 7-nanometer, which provides further validity of the capability. In the first 5 months of this addition to the contract, we are taping out many new DFI on-chip instruments. These are being included on test chips as well as product chips. Overall, the tape-out velocity of DFI has increased across all the customers in the last quarter and we're seeing that continuing this quarter as well.

In Q2, we shipped our fourth system as part of an overall 14-nanometer contract that was signed in the prior quarter and are bringing it up this quarter. We also released new eProbe 150 capability, providing higher throughput and sensitivity as well as demonstrating the ability to catch partial opens and shorts. This ability to identify and continuously provide new capability is the result of the business model of partnering with our clients and selling unique data versus selling conventional CapEx business model. Our customers tell us that identifying partial opens and shorts defects are particularly hard to find with conventional systems. Since soft failures can result in quality and reliability issues, fabs and fabless see them as critical to inspect during production. For example, at our Exensio users conference, John Chen, the VP of Technology at Nvidia, gave the keynote speech, talking about the importance of seeing those invisible defects. His focus of the talk was that for chips in automotive, high-performance computing and AI, these soft failures are key to success. His challenge to the industry for a number of years has been to detect the undetectable. DFI's demonstrating that capability.

The eProbe 250 development plan remains consistent with our comments from last quarter. The first system's built, and it's -- and we anticipate wafer testing in our lab this quarter. Depending on customer demand, beta shipment to customer can be as early as the end of the year. These achievements in Q2 mean that our DFI program is on target to have more impact on our business this year with a significant impact next year.

In summary, while gainshare recovered in Q2, the first half of the first -- the first half of the year was a difficult period from a revenue standpoint. However, we are pleased with the progress we have made with DFI and the success we've had with Exensio. Initial 28-nanometer gainshare from China, and our first significant contract in 3D NAND were important milestones for the company this quarter. We have confidence that in Q3 and Q4, we will return to quarterly growth and set up a strong future.

I will now turn the call over to Greg.

--------------------------------------------------------------------------------

Gregory C. Walker, PDF Solutions, Inc. - CFO and VP of Finance [3]

--------------------------------------------------------------------------------

Thank you, John. As you may have seen in our earnings release. We have posted in the Investor Relations section of our website a management report with detailed comments regarding the financial results of PDF for the quarter. Given that, I'm going to focus my verbal comments for the quarter on a few key highlights and issues reflected in those results.

Turning to revenue. As John stated, total revenues of $24.3 million for the quarter were flat as compared to Q1 2017. Gainshare revenues at $7.8 million increased by $3.2 million over Q1 while solutions revenue decreased by $3.2 million. To reiterate some of John's comments, gainshare volumes and revenues recovered during the quarter primarily driven by a 14-nanometer ramp at one of our major customers, the achievement of R&D milestones at other major customer and a new 28-nanometer node ramp at a foundry customer in China. During the quarter, solutions revenue was negatively impacted by a slower-than-expected close rate on Integrated Yield Ramp or IYR contracts in Asia.

GAAP net income for the quarter was approximately $200,000, lower than Q1 by $300,000. Non-GAAP net income for the quarter was $2.7 million, up slightly from Q1. Looking at expenses, on a GAAP and non-GAAP basis, total expenses for the quarter were approximately $200,000 higher than the previous quarter. This increase in expense was primarily due to merit-based salary increases, which were effective at the beginning of this quarter. On a GAAP and non-GAAP basis also, cost of sales and R&D expenses were approximately flat compared to the previous quarter. Sales and marketing expense increased by approximately $100,000 on both the GAAP and non-GAAP basis. GAAP G&A expense increased during the quarter by approximately $100,000 and was flat on a non-GAAP basis as compared to the prior quarter.

Looking at the balance sheet. Total cash at $109 million declined by $5.9 million during the quarter. This reduction was primarily the result of stock repurchases totaling $4.8 million; plant, property and equipment purchases of $2.7 million, most of which was related to development of our DFI solution; and -- let's see, and that's it. The total of which was partially offset by cash generated from operations of $1.4 million.

Accounts receivable, including long-term unbilled, increased by approximately $1.8 million during the quarter. The total unbilled AR balance of $30.7 million decreased during the quarter by $3.8 million. Of the $30.7 million in unbilled AR balance, we expect to bill $21 million over the next 12 months, of which more than 45% will be billed during Q3. Since the end of the quarter, we have collected approximately $6.5 million of the $33 million of trade accounts receivable outstanding at the end of the quarter. Looking at taxes, our GAAP tax benefit for the quarter was $800,000, driven by our Q2 pretax GAAP loss. Cash taxes incurred for the quarter were $200,000. For the remainder of the year, we expect our GAAP tax provision rate to be in the range of 37% to 39% of pretax GAAP income. However, changes in tax regulations, as we implement them, may result in large variances in our actual rates. Cash taxes incurred for the remainder of the year are expected to be in the range of 27% to 30% of pretax GAAP income.

In summary, while total solutions revenues was below our expectations for the quarter, our Big Data, Exensio solution and our new DFI solution met expectations for the quarter. However, as I mentioned earlier, delays in the contract closures primarily drove the quarter-over-quarter reduction in our IYR revenues.

Driven by recovering gainshare revenues and continued strength in our Exensio and DFI businesses, we expect total revenues to be stronger in the second half of the year when compared to the first half and approximately flat compared to the second half of 2016.

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

================================================================================

Questions and Answers

--------------------------------------------------------------------------------

Operator [1]

--------------------------------------------------------------------------------

(Operator Instructions) Our first question comes from Jonathan Tanwanteng.

--------------------------------------------------------------------------------

Jonathan E. Tanwanteng, CJS Securities, Inc. - Research Analyst [2]

--------------------------------------------------------------------------------

Should we view the delay in the engagements in the solutions business as pushing any business out of this year in to 2018 though?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [3]

--------------------------------------------------------------------------------

I think from a booking standpoint, as we said in my prepared remarks, we expect them to close this quarter or next. From a revenue recognition standpoint, because they're 10% complete, they will slide as a result.

--------------------------------------------------------------------------------

Jonathan E. Tanwanteng, CJS Securities, Inc. - Research Analyst [4]

--------------------------------------------------------------------------------

Okay, got it. And then gainshare was, obviously, up nicely. Was there any catch-up payments here? Or was this all organic strength in the quarter?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [5]

--------------------------------------------------------------------------------

Primarily, all organic strength in the quarter.

--------------------------------------------------------------------------------

Jonathan E. Tanwanteng, CJS Securities, Inc. - Research Analyst [6]

--------------------------------------------------------------------------------

Okay. Got it. And how should we think of the prospects for the older 28-nanometer nodes and their contribution to gainshare this year? Do see a recovery there at all? Or is it going to be growth from new foundries?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [7]

--------------------------------------------------------------------------------

We anticipate some new contracts starting gainshare up in the second half of this year. It like -- when we had such a soft Q1 on gainshare and we commented that we felt it was going to recover in 28-nanometer and we got a lot of questions around that, part of that was because we know contracts were starting to contribute. So that drives some upside even though we expect, by and large, for a lot of these foundries, they're going to be soft on demand. If you look at some of the prepared remarks from the foundries in Asia that came out last night, they expect (inaudible) on demand for the second part of the year. So we're a little bit insulated from that because of the new contracts starting, and that's how it works.

--------------------------------------------------------------------------------

Jonathan E. Tanwanteng, CJS Securities, Inc. - Research Analyst [8]

--------------------------------------------------------------------------------

Got you. And then switching over to the topic of DFI. Do have a backlog or order book for DFI? How do you gauge the demand from your end?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [9]

--------------------------------------------------------------------------------

Jon, we are still figuring that out. Our demand, the way that we model it, is basically on tape-outs. We know tape-outs, in all situations, drives demand for measurement, particularly as those tape-outs are, in part, combinations of fabless content and fabless involvement in the foundries themselves. And we're seeing more and more of that across all the customers, where fabless are partnering with foundries to put on the content they want to monitor things that they feel drive performance and yield and reliability for their parts. So we usually look at that as early indicators and where we're going to then get customers signed up. As I said, the velocity has been quite good this year with tape-outs. And not just on the exist -- the places where we have machines today, but in a number of places where we don't have machines installed yet. And that's kind of how we've been monitoring it. But we don't really know -- we don't have a good model beyond that at this point.

--------------------------------------------------------------------------------

Jonathan E. Tanwanteng, CJS Securities, Inc. - Research Analyst [10]

--------------------------------------------------------------------------------

Okay, fair enough. Is there any way that you can say the first Gen 2 machine is spoken for at this point, or first 1 or 2 beyond that? How do kind of think about that kind of demand?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [11]

--------------------------------------------------------------------------------

Yes, for Gen 1 machine, we expect to demonstrate capability to customers in this quarter and next. We do not know where that first machine will go as of yet. We're still figuring that out.

--------------------------------------------------------------------------------

Jonathan E. Tanwanteng, CJS Securities, Inc. - Research Analyst [12]

--------------------------------------------------------------------------------

Got it. And just a question on the business that you acquired, Kinesis. What were the trailing revenues for that? And then kind of what are the expectations for it going forward?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [13]

--------------------------------------------------------------------------------

Yes, it was pretty small. So I don't -- and a lot of it is going to change with the way that we bring it forward. As I said in my prepared remarks, we didn't really buy it for revenue. Although it did have some revenue associated with it. But we look at it -- in fact, how we -- I don't know how to say, stumbled on this or recognized this, is we had customers that were deploying our yield and test solutions that were using Kinesis to keep track of MEMS elements, as they were included in their package, with other silicon elements to build system and package capability. And the end customer for that, the consumer product that these things went in, was demanding traceability of all elements. So if a system came back that failed, they wanted to be know -- to know exactly where the MEMS element came from and exactly where the SOC came from. So you could understand how large -- how many other parts they should be looking to call back, right? If this was on a certain -- assembled with a certain machine. And that means that all -- on that day, maybe other ones are at risk, et cetera. And it was really that work that we did together with them over last year that got us to recognize that this was a general need for a lot of these applications like automotive. Even things like microphones and cell phones, you have systems and packages these days. So we are really breaking out a series of new solutions based on their data collection and our analytics. And we'll price those separately than the way that Kinesis sold its products in the past.

--------------------------------------------------------------------------------

Jonathan E. Tanwanteng, CJS Securities, Inc. - Research Analyst [14]

--------------------------------------------------------------------------------

Okay. That was helpful. And finally, just -- first of all, congratulations on breaking into the 3D NAND business. Is there any way you can size that market opportunity for us and what you bring to the table there?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [15]

--------------------------------------------------------------------------------

Yes, thank you, Jon. We're very excited about that. We are -- this is -- as I said, it is a gainshare contract so it has solutions revenue as well as gainshare tied to the -- basically, paid per wafer. And so it has the potential to be substantial because, as you know, a lot of the volume capacity build-out right now has been in 3D NAND. We are -- I think it's premature for us to say how large that market opportunity is for us because we need to demonstrate what we're able to achieve, how much yield we're able to attain, what the volume profile looks like for these programs once you hit volume. So today, I wouldn't be able to put a number to that yet.

--------------------------------------------------------------------------------

Operator [16]

--------------------------------------------------------------------------------

And your next question comes from the line of Tom Diffely with D.A. Davidson.

--------------------------------------------------------------------------------

Thomas Robert Diffely, D.A. Davidson & Co., Research Division - MD & Senior Research Analyst [17]

--------------------------------------------------------------------------------

I guess, just following up on that last question. Is the new 3D NAND customer a new customer or new player in the space then? Someone who hasn't done 3D NAND in the past?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [18]

--------------------------------------------------------------------------------

It is a new customer for us. I believe they've been in the memory market before, but this may be a new entry for them.

--------------------------------------------------------------------------------

Thomas Robert Diffely, D.A. Davidson & Co., Research Division - MD & Senior Research Analyst [19]

--------------------------------------------------------------------------------

Okay. Great. And then when you look at the DFI, has the success of the 250 being built on schedule with all the plans still on schedule, has that changed your customers' view of what the demand for the 150s -- additional 150s going forward?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [20]

--------------------------------------------------------------------------------

Not really, the 150 now, as we've said in our prepared remarks, it's been qualified for mass production. It's running in mass production. And as we said in my prepared remarks, we continue to bring out new capability on it with firmware upgrades, which is -- and new structures and new analytics. So we're able to demonstrate that we can see things in the 150 that we couldn't see, let's say, a quarter ago. And so right now, I think the customers are quite pleased with our capability. As the 250 demonstrates its performance, we know that within chip, when you are inside -- looking for DFI structures inside a chip, it will be substantially faster than the 150. In the scribe line application, where a lot of our customers are using the 150 today, and in R&D, it will be also substantially faster, but there will be a cost-benefit analysis there that we're going to have to run for the customers. The 150 is quite productive for the customer at the cost that we're able to charge.

--------------------------------------------------------------------------------

Thomas Robert Diffely, D.A. Davidson & Co., Research Division - MD & Senior Research Analyst [21]

--------------------------------------------------------------------------------

So that being said, would you expect to get more 150 business before the 250 is ready to go out the door?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [22]

--------------------------------------------------------------------------------

We've built a number of 150s or are in the process of building a number of 150s for where we expect to see business from customers over the remainder of this year and into next year, and we anticipate shipping those as we kind of went into this year thinking that was going to happen. Originally, as I think we -- if you look at our -- go back and look at our prepared remarks in 2016, we thought we'd build 1 or 2 150s and then move on to the 250s. But what surprised us was when -- for example, the customer that we talked about put the 150 online for manufacturing and ran it 24 hours a day and ran inspecting scribe line on product wafers. We didn't anticipate that, given the amount of area a 150 can test. Most of our customers will allow us to inspect a wafer for under 2 hours when they're in high-volume manufacturing and then the wafer needs to move on. And typically, E-beam tools was 2 hours of inspection time. Once you get to good yields, you can't see anything. So we had thought we would get to a point where at -- with 2-hour budget, you wouldn't be able to see anything. However, we're still able to correlate what we see on the 150 with customers' product yields, even as they get to manufacturing defect densities, so competitive defect densities. So at this point, we see -- we start -- as we've said last -- I think it was Q4 of last year, we started seeing that there was a manufacturing application for the 150, and we scrambled in the first half of this year to make that possible, and Q2 was when we demonstrated. I mean, the customer has tested literally hundreds of wafers with this thing now. And so now, we believe there is an ongoing application in the scribe line for the 150.

--------------------------------------------------------------------------------

Thomas Robert Diffely, D.A. Davidson & Co., Research Division - MD & Senior Research Analyst [23]

--------------------------------------------------------------------------------

Okay. So when you talked about a significant -- potentially significant impact to the model next year, is that more 150-driven? Or is that the 250 ramping up?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [24]

--------------------------------------------------------------------------------

We built the model assuming it would be on the 150 itself and it would be significant. We believe there's even more we could do with the 250. But we've got to demonstrate to customers. Well, one of the customers said, "If you can hit those performance specs, I'll be blown away," right? So we need to show them that that's actually achievable. We work with manufacturers, right? Not designers. They like to see how stuff actually really works, right? So people have asked us to send us wafers this quarter and we will do that.

--------------------------------------------------------------------------------

Thomas Robert Diffely, D.A. Davidson & Co., Research Division - MD & Senior Research Analyst [25]

--------------------------------------------------------------------------------

And I did move over to the model. When you talked about ramping business in the third and fourth quarter, was that just a royalty stream increase or -- of the gainshare? Or did you expect actually an increase in the design-to-silicon as well?

--------------------------------------------------------------------------------

Gregory C. Walker, PDF Solutions, Inc. - CFO and VP of Finance [26]

--------------------------------------------------------------------------------

Yes. We expect an increase in both that will bring us back to levels similar to the second half of last year.

--------------------------------------------------------------------------------

Thomas Robert Diffely, D.A. Davidson & Co., Research Division - MD & Senior Research Analyst [27]

--------------------------------------------------------------------------------

Okay. And so -- I know for a couple of the contracts you said that you're probably going to sign them this year and then revenue really hits next year. What is the typical rollout of revenues for kind of a standard contract for you?

--------------------------------------------------------------------------------

Gregory C. Walker, PDF Solutions, Inc. - CFO and VP of Finance [28]

--------------------------------------------------------------------------------

Yes. In these particular contracts, the reason they had an impact on revenue this quarter by not signing is that these are contracts where we've been working with the customer already. And in fact, in some of the cases, we've done quite a bit of extensive work with them. So that when they -- in fact, they do sign, which we're fairly confident they will, it's just a matter of timing, we'll have catch-up revenue in whatever quarter that happens. And then they'll go forward from there.

--------------------------------------------------------------------------------

Thomas Robert Diffely, D.A. Davidson & Co., Research Division - MD & Senior Research Analyst [29]

--------------------------------------------------------------------------------

Okay. And then finally, when you look at the ramp of 250 as it matures over next few quarters, what is going to be the impact you think to the CapEx and OpEx?

--------------------------------------------------------------------------------

Gregory C. Walker, PDF Solutions, Inc. - CFO and VP of Finance [30]

--------------------------------------------------------------------------------

Yes. That's a good question. We're actually going through our modeling as we come out of this quarter into next quarter right now. But the major thing is that the 250, like the 150, will be sold on a usage model with fairly ratable revenues over extended period of times. The cash to do the actual BOM is incurred upfront. So it will have some impact on cash flows as you initially build up to the ramp, but once you get past a certain number of machines, the cash being generated by the existing machine base overwhelms the cash you have to spend to build additional machines. So we have a lot of variance on that model right now.

--------------------------------------------------------------------------------

Thomas Robert Diffely, D.A. Davidson & Co., Research Division - MD & Senior Research Analyst [31]

--------------------------------------------------------------------------------

Okay. What about on the OpEx side?

--------------------------------------------------------------------------------

Gregory C. Walker, PDF Solutions, Inc. - CFO and VP of Finance [32]

--------------------------------------------------------------------------------

On the OpEx side -- basically, if you -- what we said -- and it's on our website, we expect R&D expense to basically flatten at the level it's at now. And then as we get into 2018, we expect to see R&D come down related to these programs as we go into production on the 250, and you'll see -- you'll start to see the cost of sales line pick up, which will be in depreciation expense.

--------------------------------------------------------------------------------

Operator [33]

--------------------------------------------------------------------------------

And your next question's come from the line of Christian Schwab with Craig-Hallum Capital Group.

--------------------------------------------------------------------------------

Christian David Schwab, Craig-Hallum Capital Group LLC, Research Division - Senior Research Analyst [34]

--------------------------------------------------------------------------------

On the design-to-silicon yield solutions, when did you know this was a problem?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [35]

--------------------------------------------------------------------------------

As we -- pretty close to the end of the quarter. So as Greg said, a lot of these were extensions of pilots we had done or actually early R&D engagements that converted over or anticipated to convert over. As I said in my prepared remarks, we had some customers where there were changes in leadership in the quarter. And as we got to the end of the quarter, we found the signing process was a little bit different than we anticipated.

--------------------------------------------------------------------------------

Christian David Schwab, Craig-Hallum Capital Group LLC, Research Division - Senior Research Analyst [36]

--------------------------------------------------------------------------------

Okay. Is -- and the guidance that you gave for total revenue in the second half of '17 to be, in aggregate, a similar dollar amount to the second half of '16. Does that include an assumption of a significant catch-up in that? Or are you assuming that if all these contracts do get signed then that would be on top of that?

--------------------------------------------------------------------------------

Gregory C. Walker, PDF Solutions, Inc. - CFO and VP of Finance [37]

--------------------------------------------------------------------------------

I would say it's a partial catch-up.

--------------------------------------------------------------------------------

Christian David Schwab, Craig-Hallum Capital Group LLC, Research Division - Senior Research Analyst [38]

--------------------------------------------------------------------------------

A partial catch-up. I had to jump on the call late. Did you quantify exactly -- the exact dollar amounts that these contracts not signed cost you in the June quarter?

--------------------------------------------------------------------------------

Gregory C. Walker, PDF Solutions, Inc. - CFO and VP of Finance [39]

--------------------------------------------------------------------------------

No, we did not. But it was material to the results of the quarter. Let's put it that way.

--------------------------------------------------------------------------------

Christian David Schwab, Craig-Hallum Capital Group LLC, Research Division - Senior Research Analyst [40]

--------------------------------------------------------------------------------

Right. Is it safe to assume it was probably roughly $4 million?

--------------------------------------------------------------------------------

Gregory C. Walker, PDF Solutions, Inc. - CFO and VP of Finance [41]

--------------------------------------------------------------------------------

I would not say that, but it's in the millions.

--------------------------------------------------------------------------------

Christian David Schwab, Craig-Hallum Capital Group LLC, Research Division - Senior Research Analyst [42]

--------------------------------------------------------------------------------

Okay. And then, following up on an earlier question. Would you anticipate that gainshare revenue continues to improve sequentially from June to September, and then potentially September to December, depending upon your largest customers' 14-nanometer volume?

--------------------------------------------------------------------------------

Gregory C. Walker, PDF Solutions, Inc. - CFO and VP of Finance [43]

--------------------------------------------------------------------------------

I would hate to try to predict the individual quarter to say, are they going to be in a nice linear ramp. But we're fairly confident that the second half in aggregate will be significantly higher than the first half.

--------------------------------------------------------------------------------

Christian David Schwab, Craig-Hallum Capital Group LLC, Research Division - Senior Research Analyst [44]

--------------------------------------------------------------------------------

Okay. I'm struggling to remember a company, and I've been doing this a long time, who constantly cannot predict their business with a degree of accuracy. Are we putting in checks and balances? What are we doing? Or is it just as big of a black box for you as it is for us?

--------------------------------------------------------------------------------

John K. Kibarian, PDF Solutions, Inc. - Co-Founder, CEO, President and Director [45]

--------------------------------------------------------------------------------

We understand your disappointment, Christian. And certainly, we were, too, about the lack of predictability. I think we've actually been pretty good over the previous few years. This first half of the year has been a surprise to us. I think there've been a number of factors. There's been an awful lot of start-up of new business activity with our customers in Asia. They've had some leadership changes, too. That's certainly caused some timing issues on the signings. With respect to the volumes, they've also, I think, struggled a little bit with their volumes over there that have been more substantial than we anticipated. We will as we -- if you look at when -- if you go back '14, '15 and '16, as we had times where we understood the customers pretty well, we will get a pretty good model on what's going on with the customer business and customer behavior.

--------------------------------------------------------------------------------

Operator [46]

--------------------------------------------------------------------------------

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