PDF Solutions Inc (PDFS) 2013 Q1 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 ended December 31, 2012. (Operator Instructions.)

As a reminder, this conference is being recorded.

If you have not yet received a copy of the corresponding press releases, they have been posted at 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 16 of PDF's Annual Report on Form 10-K for the fiscal year ended December 31, 2011, 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 no obligation to update them.

Now, I'd like to introduce John Kibarian, PDF President and CEO, and Greg Walker, PDF's CFO. Mr. Kibarian, please go ahead.

Thank you, and welcome, everyone. PDF Solutions started off 2013 with another positive step towards our goal of being pervasively applied to leading-edge logic manufacturing and generating gain share from our clients' successful yield execution.

In the first quarter of 2013, we achieved revenue of $20.4 million, non-GAAP EBITDA profit of $0.26 per share, and GAAP profit of $0.15 per share. In a moment, Greg will go into more details of our first quarter financial results.

On the new business side, we have the following engagements to report, all with existing clients - our 20-nanometer leading-edge logic yield ramp with a foundry client, a 16-nanometer DFM engagement with a fabless client, an extension to a YieldAware FDC pilot based on initial positive results, and an expansion of a 28-nanometer DFM engagement with another fabless client to include an additional IC product. These engagements continue the trend of both fabless and foundry customers leveraging PDF Solutions' characterization vehicle technology and services across the IP process lifecycle to improve the yields and manufacturability of their designs and processes.

With this most recent business activity, PDF Solutions has now developed CV test chips on four different FinFET process technologies across four different manufacturers. As the results and feedback from this characterization vehicle infrastructure are generated in multiple fabs around the world, we expect to have collected more data on FinFET technologies than about any other company in the world by year-end. This will significantly increase our value to our fabless DFM clients, as well as our foundry clients, as we provide superior design optimizations and device characterizations to enable faster time-to-market. We refer these optimizations as virtual IDM enablement.

Our new 20-nanometer ramp engagement that I mentioned earlier is another sign of yield ramps at 20-nanometer are in full swing across multiple foundries. TSMC commented on its call that they expect their volume ramp for 20-nanometer will start in the second half of 2014. This indicates that multiple foundries will ramp volume production at approximately the same time.

Over the last four years, the foundry industry has changed, and the long time lags between the largest foundry and others are evaporating. Some of this closing of the gap in time is a result of PDF engagements with its foundry clients.

We continue to make progress with our YieldAware FDC solution. We're able to demonstrate both strong system performance of the software platform as well as powerful diagnosis and control solutions with our clients. Due to the nature of our business model, these will typically have impact for PDF Solutions over the long-term. Overall, the first quarter was another demonstration of the role PDF Solutions' technology is playing as a virtual IDM layer improving the productivity of foundry and fabless companies.

Gain share results in Q1 were strong. The primary driver for this improvement was increased 32-nanometer volumes and yields. Two factories previously not providing gain share at 32-nanometer had significant Q4 volumes, which resulted in strong Q1 gain share. This was partially offset by lower gain share at other facilities.

We continue to expect gain share to grow year-over-year as leading-edge volumes increase. Of course, there is always the possibility of variability quarter-to-quarter as foundry ramp volumes and improve yields and older volumes drop off.

Our perspective on the industry we serve continues to be generally positive, and we continue to believe that we will outgrow the semiconductor industry. As logical characterization becomes essential to ramping new products and processes, we continue to see strong interest from the fabs and logic designers for our technology and solutions.

However, we are mindful of the cyclical nature of the chip industry. So, while we see opportunities to continue to grow both our solutions and gain share revenue, we'll be cautious as we make investments and continue to seek to disproportionately grow earnings with increased revenue.

Thank you for your time and attention. Now I'll turn the call over to Greg. He will discuss in detail our financial results for the first quarter. Greg?

Thanks, John.

As a reminder, in addition to using GAAP results when evaluating PDF's business, we believe it is also useful to consider our results using other non-GAAP measures.

For internal purposes, the Company focuses on non-GAAP net income and EBITDAR. Non-GAAP net income excludes stock-based compensation expenses, amortization of expenses related to acquired technology and other intangible assets, restructuring charges and their related tax effects as applicable.

Additionally, income tax provision has been adjusted in our non-GAAP net income to reflect cash tax liabilities only. EBITDAR is equal to earnings before income tax, adjusted to exclude depreciation, amortization, restructuring and stock-based compensation. You can access the earnings press release that contains the reconciliation of EBITDAR and non-GAAP net income to GAAP results in the Investor section of our website located at pdf.com.

Now, let's turn to a review of the financial results. As John commented, during the quarter, we added a 20-nanometer yield ramp engagement, a 16-nanometer DFM engagement, an extension to an existing YieldAware FDC contract, and an expansion to a 28-nanometer DFM engagement.

Total revenues for the quarter were $24.1 million with a GAAP net income of $4.7 million. This resulted in an EPS of $0.15 per share, per fully diluted share. Net income on a non-GAAP basis totaled $6.7 million or $0.22 per fully diluted share.

Cash increased by $1.5 million during the quarter. Cost of sales and operating expenses were $17.8 million on a GAAP basis and $16.5 million on a non-GAAP basis, which is an increase in non-GAAP spending of $438,000 over Q4. Overall, we are pleased with the continuing strength in revenues, earnings and cash for the quarter.

Moving on to revenue details, total revenues at $24.1 million for the first quarter were up $295,000 as compared to $23.8 million in the prior quarter. Total revenues were comprised of design-to-silicon-yield solutions, or solutions revenue of $14.8 million and gain share performance incentive, or gain share revenue, of $9.3 million.

Our top 10 customers, on a consolidated basis, represented 91% of total revenues in the current quarter, the same as in Q4. Three of these customers contributed revenues greater than 10% each for a total of 74%, as compared to 73% in the prior quarter.

On a geographic basis, in the quarter, North America accounted for 41% of total revenues, which is up 3% from the prior quarter. Europe represented 23%, which is down 7% from the prior quarter. And Asia accounted for the remaining 36% of total revenues, which is up 4% from the prior quarter.

Looking at solutions revenue in more detail, nine engagements with a total of eight different customers each contributed at least $150,000 of solutions revenue. Overall, solutions revenue at $14.8 million was down from $16.6 million in Q4. In addition to our normal fixed fee revenues in -- Q4 revenue included revenue recognized according to certain contract terms for completed deliverables by the end of the year. Comparatively, Q1 solutions revenue primarily consisted of normal fixed fee revenues.

Gain share revenue at $9.3 million, as stated earlier, represented a growth of $2.0 million over Q4. The growth in gain share revenue was primarily driven by expanding volumes in our 32- and 28-nanometer engagements. This effect was enhanced by the addition of two new factories moving into volume production of these nodes. The total number of customers contributing to gain share revenue on the quarter was 10, an increase in one over the previous quarter.

Looking at expenses, cost to sales for the quarter was $9.7 million on a GAAP basis, which was $250,000 higher than the prior quarter. Total cost of sales reflected increases of $850,000 in expense, which was comprised of the recognition of previously deferred project costs related to the 20-nanometer yield ramp engagement that John mentioned, increases to depreciation and amortization expense related to the tester and software tool deployment, and other minor miscellaneous items. This increase was partially offset by reductions in variable compensation, travel and other miscellaneous expenses of approximately $600,000. GAAP gross margin was approximately 60% compared to 61% in the prior quarter.

Our total GAAP operating expenses for the quarter were $8.1 million, or approximately 34% of total revenues compared to $9.9 million, or 42% of total revenues in the prior quarter. The Q4 number included approximately $1.8 million of restructuring charges as compared to a credit of $52,000 in Q1. R&D expenses totaled $3.4 million, or 14% of total revenues for the quarter compared to $3.6 million, or 15% of total revenues in the prior quarter. The decrease in R&D expenses is primarily due to timing [base] reductions and variable compensation expense as compared to Q4.

SG&A expenses totaled $4.8 million for the quarter, or 20% of total revenues compared to $4.5 million, or 19% of total revenues in the prior quarter. On a non-GAAP basis, looking at operating expenses and cost of sales together, total spending for the quarter was $16.5 million versus $16.1 million in the prior quarter.

As explained above, cost of sales increased by $250,000 in the quarter. Additionally, non-GAAP operating expenses reflected a $400,000 increase primarily related to the 2012 year-end audit fees and one-time expenses related to the review of the Q4 release of our deferred tax asset allowance, offset by $200,000 reductions in R&D expenses which -- referred to earlier.

Other income and expense for the quarter was $250,000 compared to $82,000 in expense in the previous quarter as a result of foreign exchange gain. EBITDAR, which is stated above, is equal to earnings before income tax, adjusted to exclude depreciation, amortization, restructuring and stock-based compensation, was $8.1 million for the quarter compared to $7.9 million in the prior quarter.

EBITDAR per fully diluted share for the quarter was $0.26, which was flat compared to the prior quarter. The GAAP income tax provision for the quarter was $1.8 million, which reflects an estimated tax provision rate of 28.1%. Of this $1.8 million, $1.1 million represented cash tax liabilities. This represents an effective cash tax rate for the quarter of 17%.

Our cash tax liability increased from the prior quarter by approximately $566,000. This was related primarily to a large revenue shift from Europe to Asia for the quarter, which drove our foreign withholding tax liability up significantly.

Quarterly GAAP net income of $4.7 million resulted in earnings per share of $0.15 per fully diluted share. On a non-GAAP basis, EPS was $0.22 per fully diluted share for the quarter.

Total cash for the quarter was $63.2 million, an increase of $1.5 million as compared to year-end. This increase was primarily driven by accounts receivables collections and proceeds from stock option exercises offset by the purchases of fixed assets. Cash from operations were slightly positive in the quarter despite Q1 being our highest disbursements quarter within the past several years. These large disbursements primarily related to the payment of year-end variable compensation and a significant amount of prepaid software licenses.

DSO for the quarter, including unbilled receivables, was 122 days compared to 130 days in the prior quarter. Of the total accounts receivable of $32.7 million at the end of the quarter, [six point -- $6 million], or 18%, was more than 30 days past due. As of today, more than 33% of this past due balance has been collected. Trade accounts receivables DSO for the quarter was 99 days compared to 101 days in the prior quarter. Headcount at the end of Q1 was 342 employees worldwide compared to 345 at the end of Q4.

As with last quarter and last year, the overall financial results for the quarter reflect continued strength in our core business and ongoing attention to spending levels. The Company has effectively manage its spending while significantly growing revenues, providing increased income leverage and stronger balance sheet position.

This concludes the review of the financial results for the quarter. Now, I will turn the call over to the operator for Q&A. Michelle?

Thank you, Mr. Walker. (Operator Instructions.) Tom Diffely.

Yes, good afternoon. First question I guess, Greg, on the taxes, so look like the cash taxes basically doubled from where we were running at last year, and you said the move from Europe to Asia was customers. So, what kind of trend do you see, going forward? Are we kind of at the normal tax rate now that you foresee, going forward, or is this just going to bounce back and forth?

I think, on the cash tax rate, from what we can project based on the mix shifts we see in revenue by country, because a lot of the cash tax is driven by withholding taxes, we think we will be at approximately 17% to 18% for the year.

Okay. And is that driven by where the fab is, or just where the customer's home base is?

It's where the customer's home base, or where the contract says we can invoice to. So, for certain customers, one in particular, the fabs are spread out among multiple sites, but all of the invoicing goes to the home country, which does have a fairly high withholding tax rate.

Okay. Okay, good. And then, John, it -- sound like you were talking about the increase in the royalties coming from 32 nanometers, what's the percentage of 28 in the book at this point? And is 28 still tracking to be ultimately your largest node for royalties?

We expect 28 to be our largest node for royalties. The percentage, that is, today is quite small. It's basically 32 that's in a higher volume right now.

Okay. Is there any way to project when you think 28 might start to layer in?

We expect to layer as we go out throughout this year. I think that's why, in general, we said our comments were we expect volumes [on the] leading edge to continue to grow throughout the year.

Okay. And then, 20-nanometer, the activity you're doing today, that's royalties a year or two down the road?

Yes. Yes. In our comments, we really do think that when you look at the 28 -- the 20-nanometer -- the 28-nanometer node, because some factories brought up 32 first then brought up 28, there was kind of a time lag between when -- so the leader brought up 28, and when the others are bringing up 28, because they brought up pretty substantial lines at 32, when you look at the 20-node, we just don't think that's going to be there. When you look at the timing across the difference of 5s, you think they're all going to be pretty much in lockstep.

Okay.

Within a couple of quarters of each other.

Okay. And if you look at the royalties over the last year or two, there are periods of time when certain fabs came offline to move to the next node. Has -- is the negative impact of those largely over at this point and we're now just waiting to see the ramps of the newer nodes?

We put on the comment on, hey, there could be quarter-over-quarter variability exactly for that reason. We never really fully know how they plan their capacity, right, and there can always be hiccups. So, we never want to say we're over that. We do feel pretty good about the remainder of this year, though.

Okay. And then, it look like the design-to-silicon line item dropped a little bit more than just when it went up in the fourth quarter, so now you're below where you were in the third quarter. Is there a seasonality in that, or is it strictly just projects coming on and going off and just timing?

Yes, this is Greg. I think when you look at the detail which we spend a lot of time analyzing, it's more just a timing of when projects are coming on versus when they're dropping off. And when they're dropping off, they're on probably their highest percentage of completion revenue recognition, and as they start up, they're at their lowest. So, you do -- as these things move, you do tend to get some bumpy gaps, but when we look at the overall engagements, we see we're right on track.

Okay. And the four new projects you talked about this time, it seems that it's a little bit less than -- from a number point of view, anyway, from what you've had the last few quarters, does that keep us down at these lower levels for a while?

No, we don't believe so.

Okay. And then, there's also been kind of talk in the industry about how 20-nanometer is -- I guess there's questions whether or not these foundries are going to move to the 3D transistor at 20 or below 20. And so, some companies are actually saying that they're seeing a hesitancy or a pause in the activity level until the foundries figure out exactly what type of transistor they want to make and when. But, it sounds like from what you're saying though, you're seeing some pretty robust activity at 20 nanometers. I was just wondering if there's anything more you can add to that.

Sure, yes. But, for point of clarification, the 20-nanometer is basically the last planar node for most companies in the world, and what people refer to as either 16- or 14-nanometer is really the 20-nanometer technology from a dimensionality standpoint, with the transistor replaced with a FinFET transistor, or what sometimes referred to as a three-dimensional transistor.

So, ground rules why they seem like they're roughly the same, when you really peel the onion and look at the details and the rules, actually the other Co-Founder, Kimon, refers to the FinFET node is the first negative Moore's Law node, because the actual size of the standard cell elements at the 16-14 node may end up being slightly bigger than the 20-nanometer node just due to the extra manufacturability that the foundries are putting into the node to account for the three-dimensional structures.

So, as designers, I think last year, when everyone was very excited about the FinFET technology, there were -- I always joke with the software developers. You compare software in the field and all of its warts with a bunch of PowerPoint, and how beautiful your software is going to be in the future, which will also have warts.

So, last year, everyone wanted, "Oh, look the 20-nanometer planar node's not great, the FinFET node's going to be so wonderful. Everyone is going to use FinFET node. No one's going to use the planar node." Our perspective on this has been, when you really dig into the details on the design rules and try to create a library set on the FinFET technologies, they have their -- they have a lot of issues when it comes to getting to great area scaling, even though they do have wonderful performance characteristics or reasonable performance characteristics.

And the performance characteristics, as wonderful as they are, not that much better than the planar, and they are better. So, when you net it all out, both nodes have advantages. And we've spoken with VPs of technology and heads of divisions of fabless and fab-like companies. They really are picking different products for each of these. They're treating these as one node, and they're putting different products and different flavors, depending on the particular requirements of the products they're building.

And that's very similar to what happened at 28-32, where some people went to 32 for the high K, some waited and went 28 with the poly-silicon to get the cost savings. And others waited to get 28 with the cost savings of the area shrink as well as the high K, the 28 high Ks. And that's kind of bifurcation of the nodes we -- something we think is basically here to stay.

Okay. So, most of the 60-nanometer DFM activity you have is for the FinFET type activity?

Yes. Whenever we say 16 or 14, it's by and large, with almost all but one or two exceptions, FinFETs.

Okay. And Greg, maybe just one more thing on the taxes. So, how was Europe lower tax rate than Asia?

Withholding tax. So, if you look at our cash tax liability, the bulk of it is withholding tax-related. And what we see is the Asian countries, particularly one -- a couple of actually, have fairly substantial withholding tax rates, and we don't have a tax treaty with -- directly with them, whereas in most of the European-driven revenue streams, we don't have significant withholding tax.

Okay. If you--.

--So, any time we move revenue between Europe and Asia, our cash tax liability is going to go up.

Okay. All right. Well, thanks for your time today.

Andrew Weiner.

Hi, good afternoon, guys.

Hi, Andrew.

Hi. I was wondering if you could elaborate a little bit on the YieldAware FDC solution opportunity, in particular how we are going as far as commercializing with other customers, and in particular with our existing sort of logic foundry customers versus like the -- I believe the original customer was in an image sensor application?

Yes. Okay, yes. So just -- I know we have a lot of jargon in our business, so YieldAware FDC, really what that technology -- our systems are about is large software systems that use the models often built by running characterization vehicles to use the data that's coming off the equipment that's manufacturing the wafer, so data that's coming off the etchers or lithography systems to identify suspect wafers, right?

And the way factories are controlled today, by and large, they use metrology and inspection with what's typically referred to as a static inspection plan. In other words, they look at two wafers per lot, one out of every four lots manufactured. Our belief had been that you should somehow use the information from the equipment that's processing the wafer to identify a wafer that's a suspect wafer and spend your control dollars on the wafers that are more likely to be questionably produced.

That seems like a really simple idea. It's really hard to put in practice if you spent the last seven years or so building systems, acquiring companies at one time, and making capability that we felt was robust. Yes, you're right, the first beta customer was an image sensor customer, and we continue to build business with them.

We then went back -- and the system could be very useful to diagnose variability just from CV data and product data, and started doing pilots with our Logic customers over the last year and change or so. And now, some of them are converting from pilots into -- basically pay for extended pilots, and we expect in this year deployments. The reason why that's important to us is it extends the value or the life of our technology in the factory, increasing the revenue that we can generate from that factory, as well as the length of time that we can increase revenue of that factory.

So, for us, it's a revenue creation and extension. And for the customer, it's obviously improved variability in the production line, reduced variability, reduce excursions, and improve use of their metrology and inspection capital.

So, the extension that was referred to earlier in the prepared remarks, was that with a Logic customer?

That was with a Logic customer.

Okay. And so, if I think about it from a timeline, most of you had talked about sort of matching the long-term business model characteristics of the gain share business, is that you'd hope to secure a number of engagements this year with our traditional Logic foundries and-or manufacturers with the idea of that, starting in '14, it could have a material revenue impact.

Yes, we expect some revenue impact this year, but that kind of a variable production impact would be in '14 and beyond. But, we feel more comfortable about that now, as we're starting to see customers say yes, we understand why this is really valuable, and we understand how it fits to your usual business model. So, we're starting to get the kinds of nods that we had been hoping to get for quite a while.

Great. Greg, on the gross margin side, the cost of revenue, and I know you sort of mentioned a bunch of gives and takes on the call in absolute dollars was up sequentially, or the cost of revenue, I should say, was up sequentially from Q4 despite the dip. And as -- or should design-to-silicon ramp through the year, should we expect the cost of revenue to remain relatively constant with this quarter net of those [puts and takes]?

No, the design-to-silicon ramp will generate some increases in cost of sales due to the -- number one, the people employed in supporting those ramps. And then, also, as more and more of our testers get deployed, as you're aware, we don't sell those testers, but we do deploy them to facilitate our use, and we allow collaborative use with the customer while we're working. And we amortize those costs over a number of years, but as you start to deploy large numbers of them, you do see some cost increases.

So, while we don't think we'll see a tremendous amount of margin variability, well, on average, there will be some increase in cost of sales.

Okay. (Inaudible.) All right, thank you.

(Operator Instructions.) At this time, there are no more questions. Ladies and gentlemen, this concludes the program. Thank you.

Thank you. Take care.