PDF Solutions Inc (PDFS) 2015 Q2 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 second fiscal quarter ended Tuesday, March 31, 2015. At this time, all participants are in a listen-only mode. Later, we will conduct a question-and-answer session, for which instructions will be given at that time. (Operator Instructions) As a reminder, this conference is being recorded.

If you have not yet 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 sections entitled Risk Factors on pages 12 through 18 of PDF's annual report on Form 10-K for the fiscal year ended December 31, 2014, 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.

Thank you and welcome, everyone. Today I will start our discussion with a brief summary of our second-quarter results. Then I will provide some perspective on the environment and PDF Solutions' performance towards the strategic objective (technical difficulty) direction. Next, I will turn the call over to Greg, who will walk you through the financial results in detail. We will then take your questions.

In Q2, we had another strong bookings quarter for our solutions business. Solutions revenue increased by over 15% when compared to Q1, excluding the one-time impact of the resolution of the delayed customer contracts discussed on our earlier calls. We saw growth on this basis in both our Exensio sales and our IYR solutions.

Gainshare revenues, while up slightly quarter over quarter, reflect some muted volumes at some of our customers. We continue to see weakness at 28-nanometer across most of our customer base, which is consistent with the results reported for the quarter for both fabless and foundry customers.

We continue to remain positive on 28-nanometer over the long term and expect volumes to recover. However, we still don't have the visibility to say when that recovery will happen. We continue to plan on 28-nanometer volumes remaining depressed for this year. Our key customers have continued to make good progress on their 14-nanometer ramps and remain confident that we will start to recognize gainshare royalties on this node by no later than Q4 of this year.

Turning to a summary of the business for the second quarter, we experienced a broadening of our solution base in the areas of both yield ramp and big data analytics solutions, with particular strength showing in the Asian markets. In addition to the bookings for the quarter, when we look at the activities being performed by our Characterization Vehicle infrastructure deployment organization, we are driving significant expansion in the variety of customers and process nodes, including logic, image sensor, and memory.

These projects are being driven by both existing customers and many new customers in multiple end markets. This level of activity supports our confidence in the relevance of the technology that we are developing and deploying, and therefore drives our optimism about future solutions business.

The contracts signed in the quarter include a new contract at one of our major customers for technology development at the 22-nanometer FD-SOI node; a number of new solution engagements in the Asian market, including an engagement with an existing foundry customer for fabless-specific custom CD infrastructure deployment at 28-nanometer node; a new engagement for yield ramp services across multiple nodes at a new foundry customer -- this engagement covers both logic and DRAM; a new consulting agreement for a large automotive company -- this represents our first foray into the automotive vertical; an engagement to deploy CVs for a large image-sensor provider. This company already has Exensio for big data analytics and YieldAware process control. This demonstrates how you can apply CVs later in the node life to improve control, particularly when coupled to our analytics platform. And finally, multiple new software license and YieldAware FDC service agreements for our Exensio big data platform, including yield, FDC and test modules across a number of new and existing customers.

Given the success just mentioned and the leverage across our technology solutions that we continue to build, we remain confident that our customers will continue to increase their adoption of our solutions.

While we continue to drive adoption of our solutions at the leading-edge nodes, we are driving more adoption of our solution at fabless and fabs, at nodes typically described as more-than-more. Moreover, while our business has primarily been on the bring-up of nodes, we are driving growth -- driving a growing business tied to the ongoing control of a process, as well as the introduction of new products within an existing node. These activities are designed to broaden the applications from which we derive revenue.

There are three kinds of industries that are important to PDF. First, our history is moving to a position where mobile is maturing, and newer applications like automotive and IOT, while exciting, are not at the volumes that are significant relative to mobile. This is driving a proliferation of new derivative nodes in new applications. We are responding with our big data analytics platform, Exensio, and a Characterization Vehicles test chips targeted to process control and new product introduction.

We first developed Exensio for in-fab control challenges, but we are finding that there are a wealth of applications for the fabless customers. Our activities this past quarter with the foundries, fabless and automotive-system company all speak to this diversification. Our acquisition of Syntricity, which provides us a SaaS-based yield-management and product characterization system for our fabless customers, extends our leadership in this field and increases the volume of semiconductor manufacturing data we host.

The second trend is the shift of our industry to China and Asia. We have been in China for almost a decade now, and over a third of our engineering is centered there. This past year, we have been driving significant new business activity in Asia. We engaged in multiple new contract this past quarter in Asia and expect more in the second half of the year.

The third challenge we are responding to is the difficulty of inspecting and characterizing leading-edge processes -- i.e., those at 60-nanometer and below. There, as we have spoken before, we have been investing in design-for-inspection, or what we call DFI. For 2015, we have been clear that this is a year where we accelerate our investments but do not generate revenue.

That said, in this past quarter we taped out on two more fabless chips, bringing the total number of designs with on-chip inspectors to 10, with more chips in progress and expecting to tape out this quarter. We now have these on-chip inspectors running at three foundries on fabless customers' MPW or product chips, on 16-, 14-, and 10-nanometer. We expect to have design-for-inspection in place in 2016, and this is perfect timing for 10-nanometer, which we anticipate product tape-outs late in 2016 and the ramp about one year later.

In summary, we believe we have focused PDF Solutions on the opportunities and challenges in the industry that are significant and valuable. We are making great progress this year in building a foundation for future revenues and profits. Given the nature of our business model, in quarters where we branch into new areas, there's initially little change to our financials. But over time, these branches should grow and bear fruit.

Overall, we are very excited about the direction we are taking PDF Solutions. Now I will turn the call over to Greg to discuss in detail our financial results for the second 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 nonrecurring items, stock-based compensation expenses, and amortization of expenses related to acquired technology and other tangible assets and their related tax effects, as applicable.

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

Now let's turn to a review of the financial results. Total revenues for the quarter were $23.2 million, with a GAAP net income of $2.1 million. This resulted in GAAP EPS of $0.07 per fully diluted share. Net income on a non-GAAP basis totaled $5.9 million or $0.18 per fully diluted share.

Cost of sales and operating expenses together were $19.5 million on a GAAP basis and $16.7 million on a non-GAAP basis, which is a decrease in non-GAAP spending of approximately $300,000 from Q1.

Moving on to revenue details, total revenues of $23.2 million for the second quarter were $3.6 million lower than in the prior quarter. Total revenues were comprised of design-to-silicon yield solutions or solutions revenue of $14.2 million, and gainshare performance incentive or gainshare revenue of $9 million. Our top 10 customers represented 93% of total revenues in the current quarter. Two of these customers contributed revenues greater than 10% each for a total of 70% as compared to three customers and 82% in the prior quarter.

Looking at solutions revenue in more detail, 12 engagements contributed at least $100,000 of solutions revenue in the quarter, one less than in the previous quarter. Overall, solutions revenue at $14.2 million was $4 million lower than in the prior quarter. This decrease primarily reflects the recognition in Q1 of previously delayed solutions contract revenue of $6 million, offset by a new IYR customer engagement, higher software revenues, and higher percentage of completion revenue on existing customer contracts.

Gainshare revenue for the quarter was slightly over $9 million, an increase of approximately $400,000 from the prior quarter. The total number of node sites, which we define as an individual fab and process node combination, contributing to gainshare in the quarter was 17 compared to 18 in the previous quarter.

On a geographic basis, North America accounted for 49% of total revenues, which is up 10% over the prior quarter. Europe accounted for 32% of total revenues, an increase of 4% over the prior quarter. Asia accounted for the remaining 19% of total revenues, a decrease of 14% from the prior quarter, which was related to the previously mentioned Q1 recognition of the delayed solutions contract revenue of $6 million.

Moving to expenses, cost of sales for the quarter was $9.9 million on a GAAP basis, which is approximately $1.1 million higher than in the previous quarter. This increase in GAAP cost of sales was driven by a gain on sales of expensed equipment recognized in Q1 of approximately $500,000, in addition to the Q2 purchase of hardware related to a new software deal recognized in the quarter, annual merit increases taking effect in Q2, and a minor increase in travel expense.

GAAP gross margin was 57% compared to 67% in the prior quarter. On a non-GAAP basis, cost of sales was slightly over $9 million, which was approximately $800,000 lower than the previous quarter. This decrease in non-GAAP cost of sales was principally driven by the Q1 inclusion, on a pro forma basis, of $1.9 million of previously impaired project costs, in addition to the GAAP spending variances mentioned earlier.

Total GAAP operating expenses at $9.2 million were approximately $600,000 higher than last quarter and approximately 40% of total revenues, up 8% from last quarter. R&D expenses totaled $4.4 million, approximately $300,000 higher than in prior quarter. R&D expense as a percent of revenue was 19% in the quarter compared to 15% in Q1.

SG&A expenses totaled $4.8 million or 21% of revenues compared to $4.5 million and 17% of revenues in the prior quarter. Included in SG&A for Q2, on a GAAP basis, is approximately $500,000 of acquisition-related expenses.

The overall GAAP operating expense increase was primarily driven by headcount increases in our R&D organization, annual merit increases, increased patent filing activities, and the previously mentioned acquisition expenses.

On a non-GAAP basis, looking at operating expenses and cost of sales together, total spending was $16.7 million versus $17 million in the prior quarter. As stated earlier, this was principally due to the Q1 inclusion, on a pro forma basis, of $1.9 million of previously impaired project costs. This decrease was partially offset by the increase in spending I had previously mentioned.

The GAAP income tax provision for the quarter was $1.6 million, which reflects an effective tax rate of 42% compared to 37% in the prior quarter. This increase in rate is due to discrete credits recognized in Q1 related to expired statute of limitations. Overall, our full-year outlook for GAAP income tax provision remains the same.

Of the $1.6 million of tax provision in the quarter, approximately $700,000 represented cash tax liabilities. This represents an effective cash tax rate for the quarter of 19% of pretax GAAP income. This compares to a Q1 cash tax rate of 17%.

In regards to the tax rates for the remainder of the year, we continue to expect our GAAP tax provision rate to be in the range of 38% to 40% and our effective cash tax rate to increase during the year to the range of 20% to 22%.

GAAP net income of $2.1 million for the quarter resulted in GAAP EPS of $0.07 per fully diluted share compared to $6 million and $0.18 per share in the prior quarter. This reduction in GAAP EPS and net income is directly related to the recognition in Q1, on a GAAP basis, of $6 million once again of previously mentioned delayed solutions contract revenue, with no related cost of sales being recognized.

On a non-GAAP basis, net income was $5.9 million and non-GAAP EPS was $0.18 for the quarter compared to $8.2 million and $0.26 per share in the prior quarter. Once again, this decrease in non-GAAP income -- EPS and net income -- was directly related to the recognition on a non-GAAP basis of $6 million of previously mentioned delayed solutions contract revenue in Q1 and the recognition of $1.9 million of related cost of sales.

EBITDAR, which I defined earlier and is also defined in our press release, was $7.2 million in the quarter as compared to $10.4 million for the prior quarter. EBITDAR per fully diluted share was $0.22 compared to $0.32 in Q1.

Total cash at the end of the quarter was $131.7 million, representing a slight decrease of approximately $200,000 when compared to March 31. This decrease was primarily related to the delay in receiving customer payments of $1.8 million at the end of the quarter, which have subsequently been collected; stock repurchases of $2 million; and a $1.5 million prepayment for EDA software licenses, and also $1.2 million for the purchase of fixed assets primarily related to our design-for-inspection initiatives. These uses of cash were partially offset by $1.6 million in cash received related to stock option exercises during the quarter.

Cash generated from operations was $1.2 million. As mentioned above, the Company repurchased approximately 120,000 shares of its common stock for $2 million during the quarter. After these repurchases, up to $19.4 million remains available for use by the Company under its Board-approved stock repurchase program.

Trade accounts receivable or DSO was 81 days for the quarter compared to 69 days in the previous quarter. Trade accounts receivable balance at the end of the quarter was $20.7 million, an increase of approximately $300,000 over the previous quarter. The unbilled accounts receivable balance at the end of the quarter was $11.7 million, an increase over Q1 of approximately $2.5 million.

Of the $32.4 million of total receivables, $1.4 million or less than 5% was more than 60 days past due. Since the end of the quarter, $10.7 million has been collected, including the $1.4 million of the past-due balance I just mentioned. Total DSO for the quarter, including unbilled receivables, was 127 days compared to 101 days in the prior quarter.

Headcount at the end of Q2 was 360 compared to 351 at the end of Q1. This headcount increase reflects new hiring into our R&D organization, primarily targeted at our DFI initiatives.

In Q2, we continued to see strong solutions business activity in the Asia region, primarily in China and Taiwan. Additionally, we are seeing improvement in the sales of our Exensio big data platform. Overall, we remain positive regarding the solutions business for the remainder of the year.

In regards to gainshare revenue, we continue to expect 28-nanometer volumes to remain depressed for the rest of the year. As John indicated, the 14-nanometer ramp at our major customers is proceeding well, and we expect to see our first gainshare revenues from this node by Q4 of this year.

As you saw in our press release dated July 20, and as mentioned earlier by John, we have acquired Syntricity, a leading provider of a hosted solution for characterization and yield management. We expect that this acquisition will add, on a non-GAAP basis, approximately $2 million of revenue and about $1.9 million of expense during the remainder of 2015. Additionally, we expect Q3 cash balances to be negatively impacted by approximately $5 million related to the payment of initial proceeds for the acquisition.

Now I will turn the call back over to the operator for questions and answers. Operator?

Thank you, Mr. Walker. (Operator Instructions) Jon Tanwanteng, CJS Securities.

Thanks for taking my questions. Directionally, do you expect 28-nanometer gainshares to improve on a sequential basis, or is there still another leg down, given what we're seeing in various end markets here?

It's a good question, Jon. I think we have not great visibility at this point. We are kind of expecting to roughly be flat for the remainder of the year. Some people say it's going to get a little better, some say it's going to get a little bit worse; so we've been modeling it basically as flat.

Okay, got it, thank you. And then you mentioned a new foundry customer earlier. I'm wondering if you could provide a little bit more detail on that, the relative significance of that new customer to you guys, and from an industry perspective.

Yes, it's not the largest foundry in the world, but it's an interesting foundry because they've got a more diversified portfolio than a lot of them. It is in Asia. As you mentioned, they do have a DRAM portfolio, as well as -- the logic piece is really display drivers and things related to kind of more-than-more. And the infrastructure we provided them in the ramp is really targeted to a lot of derivative technologies, or what people call more-than-more technologies.

We think they themselves are a very fast-growing foundry, actually I think on a percentage basis the fastest-growing foundry in the world last year, albeit off a smaller base. And we see them as kind of an arbiter of other customers that have similar characterization and ramp needs and control needs.

Okay, great. Thanks for the color. And then, just going back to the customer you had an issue with earlier this year and late last year, any potential for new engagements at this point?

We remain in conversation with them, as well as a number of other companies that are not now our customers or not significant customers. And we can't comment on where we are in those discussions.

Okay, great. One more from me and then I'll jump back in the queue. Any update on potential cost-saving measures?

Yes, we've been going through a fairly detailed review of the trends we're seeing in the marketplace and how that's going to drive our business. We're going through the second half of this year and figuring out how to adjust our resource levels to match those demands properly.

I suspect that what we will see, more than an absolute reduction in costs on any kind of single, dramatic program, is probably more deployment changes to match where the growth is, more than anything else.

Okay, thanks.

Tim Diffely.

Quick question, I guess on the 28-nanometer softness. So, John, is it different today than it was a quarter ago, or after the initial impact, things have been pretty steady kind of at these lower levels?

Yes, it's a good question, Tom. So, actually Q2 was a little bit better than what we thought. We thought it wouldn't increase over Q1; it did a little bit.

We had anticipated it being flat for the remainder of the year off the Q1 level. Q2 definitely was a little bit better than we expected, I think by a few hundred grand. You know, we still expect it to bounce around at this level, no different than our expectation in April. You know, we expect -- it's kind of a refresh of products. These things are long in the tooth, is our perspective.

Okay. And then, when you look at the 14-nanometer FinFET capacity coming online, how much of that is new capacity versus basically 28-nanometer brought down to 14, reused equipment, if you will?

You know, there's two companies that drive this for us right now in terms of the volumes. One is purely new capacity or primarily new capacity, and they have kind of a more classic foundry characteristic, where each new node represents new capacity. The other is kind of a hybrid foundry IDM. And for them, we've seen in the past where the new capacity seems to cannibalize the old nodes to some percentage, not entirely, but to some percentage. And we anticipate that again in this case.

Okay. So for one test where you expect potentially to have all the 28 come back plus adding 14, the other only partial; 28 comes back and then 14 is the growth driver?

Directionally, that's correct.

Okay, okay. And then, Greg, you talked about a $5 million payment for the acquisition. It sounded like you said that was the initial payment. How much more is there?

Beyond the $5 million, a portion of which will go into escrow, there are earnouts based on downstream objectives that will be paid out over the next several years, into about -- I think it maxes at about $2.5 million.

Okay, and have you locked up the management team, or -- you know, some incentive packages to keep them on board?

Yes, we have.

Okay. And then, you're looking at the new foundry customer. You said memory and logic; sounds like it's older node stuff. Are you getting close to 28 at this point, or is it still kind of 40-ish range?

You know, if you look at it, the memory tends to be on the sub-30-nanometer feature size, but the remaining -- in a sense, the display driver and logic portions are much greater than 40, above 40-nanometer.

Okay. And that technology still benefits from your technology? (multiple speakers)

Yes, it's the point that I highlighted, actually, on the call. If you look at the vehicles we've taped out, some of which were on pilot, some of which were on these engagements, they went from 10 -- Q2 and Q1, we taped out on everything from 10-nanometer to 110-nanometer, logic vehicles, DRAM and image sensor. So it really speaks to the breadth of the application space for these Characterization Vehicles, especially when you tie them to the analytics platform.

Okay, okay. And then you talked about the 10 designs now with the chip inspector on them. How do you think that plays out as far as, at what point do you need to add the actual tools to do the inspecting portion of that?

Yes, a lot of these are on MPWs. So the volume of these chips, while there's a number of chips, there volumes are not -- the foundries tend to run tens of wafers. So, with the Series 1 machine that we have in development, which we will be shipping later -- be in the position to ship later this year -- it can hobble along for those chips in terms of being able to measure them. Hobble maybe is kind of too pejorative a term. It can measure them at reasonable throughputs throughout 2016. And we anticipate as we get through 2016, the Generation 2 machine, which is in development now, would be able to measure the chips that go to full volume production with really pretty amazing throughputs as we get into the second half of 2016.

Okay,. And so still the plan is that 2016, you get a few of these tools in place;. In 2017, if the tools work as expected, would be kind of the growth year for that new product?

That's correct.

Okay.

We do expect -- as we've been showing over the past year, Tom, we get more and more customer traction on these. The tape-outs are very important because you've got to put the on-chip inspectors on customers' products. And we're getting fabless customers quite comfortable with what -- this needs to include on-chip inspectors in their dummy area, and also what this means in terms of the information that's possible and the alignment between product and technology that's possible.

And the design community is quite excited about this. I'm very excited about this. And, frankly, from a technical standpoint, it's the most fun I've had in a decade. This is some cool stuff.

Yes. So, from a real estate point of view, you can do this all on the scribe lines, so there's no impact to the--?

No, if you look at -- that's actually, that's a great question, Tom. So if you look at a conventional SoC chip, somewhere between 5% and 10% of the chip area is fill, often what's called dummy fill. And the innovation PDF came up with back in 2012 or so was to instrument that fill with on-chip inspectors. The designer is going to waste the space because they need -- they can't put anything in it because of crowding on other layers. Why not use that space for an inspector to make the inspection problem easier?

And it's not all that different than built and self-test, except built and self-test, the designers were willing to waste area. Here, we're not wasting any area at all. We're just using the already-wasted area.

Okay. And is there a type of chip that this is the most efficient with, or a type of customer that seems to be running to this first?

Yes, we've been focusing on the leading edge, because we think that in the land of the blind, the one eye is king. And it's very hard to see what's going on in leading-edge chips.

We suspect there's applicability in other markets, potentially with the new, novel technologies that are coming out in memory and other areas. But we really haven't explored that yet. This is one where we have to select and focus. And we think we're focusing on important problems, as I said in my prepared remarks, and then we will look and see what more it can do.

Okay.

I mean, frankly, if we can make it work in that market, it's huge for us.

Yes. So, bottom line is, the optical inspection kind of runs out of steam when you get to these 1x levels, and you need to have another complementary inspection tool to give you more information about the chips?

Yes, I don't think -- you know, optical inspection has a place in the inspection flow and will on a going-forward basis, right? It's not that it's a replacement for optical inspection. But when you look at all of the methodologies for looking at features on these advanced nodes, the most accurate information is the electrical, right? The electrical tells you where you really -- a lot of our customers use our electrical CVs just to calibrate their inspection tools, right, to know where to go look and what sensitivities to apply, etc.

And what we're really doing is bringing electrical in line. You know, you can't fit a probe pad on a chip. Even though it's 5%, 10% of the chip area is fill, none of it's a very big area. It's all hundreds of nanometers, thousands of nanometers on a side. So you're bringing electrical in line. You're making in line now having electrical information, which is going to accelerate the use of many kinds of inspection and metrology capabilities because you're going to add more accuracy.

Okay, that sounds quite exciting, then.

Last question, then, for Greg: What are the incremental margins of both the solutions and the gainshare? Is gainshare 100%, or I assume close to it? And then what is it for solutions?

Yes, you know, as with the past, gainshare tends to be close to 100% margin. On the solutions business, I think the initial development of the Asia market will probably be at a slightly lower margin to start with. As we progress in that market and we spread our ability to sell on value, I think you will see that margin get back to what the rest of the world margin is. Overall, probably not a huge impact.

Okay, all right. Thanks for your time today.

Gus Richards, Northland.

Thanks for taking my questions. Just real quick on the design for inspection, you mentioned you're on 10 product die; how many customers does that represent?

Yes, so just maybe clear, Gus, it's on a mix of product and MPW wafers. So, there are early test chips that go -- something like 10-nanometer, we're on a number of tens. At least two customers, two customers' chips at tens, but those are MPWs at this point, and some of the 14s and 16s are MPWs.

Total number of customers -- you know, it's more than four and less than eight. I have to sit back and think for a little bit; maybe around five or six at this point?

Okay. And then how many of those do you expect to go into production next year?

So, most of the MPW -- of course, the MPWs, none of them will go into production. But (multiple speakers) some of them will, and some of the MPW content will be copied over into product chips. And the flows that are set up can be replicated.

You know, I think as we've talked before, we have one customer that, really, it's already embedded into their flow. And now we're working through that with a couple of other design customers.

Okay. Got it.

I say into their flow ;into their design flow, right? Stuff gets naturally incorporated.

Right. And then, it sounds like you're beginning to work a little bit more with memory. And there's been some announcements on some new, novel memory types coming out from Intel. Are you engaged with other customers as they try to find an alternative to Flash or DRAM, or--?

Yes, it's a great question. So we announced that we had one engagement this quarter, which was with a foundry that had a memory technology; that was a conventional DRAM. And that really is exploiting our conventional Characterization Vehicle infrastructure.

Tom alluded to this question, I believe, about applicability of design-for-inspection in the memory market. You've really kind of nailed right in on the 3D memory and the novel technologies, where an electrical measure would be very valuable from a control standpoint.

It seems like this would be a good, from a technology -- as a technologist, I believe this would be a good area for design-for-inspection. But at this point, Gus, we haven't really gone off and introduced that to customers. That is something we would do once we kind of get through the slog of the SSEs, right? We have a lot to get done on our plate right now. We do think technically it makes a lot of sense, there, though; we just haven't gotten to it.

Got it. And then, you're making a very strong push into big data analytics for the foundry. You did the acquisition. Clearly, your Characterization Vehicle is generating a lot of data, and DFI is going to be even more.

Right.

Can you sort of put all that together for me and give me an idea of the vision and the strategy over the next couple years, to bring these things together, and how it will all work, and how you get paid?

So, what we started -- as you said, we started in the foundry. The first thing we did back in 2009, when we started building this system, is we used Cassandra as the database. It's a very scalable database; it's the one Facebook used and put out there. And we used as our analytics Platform R. You probably saw this last quarter, that IBM announced it's supporting Spark, which is really the combination of Cassandra and R for machine learning and data mining. And this came out of Berkeley. We've been building on that platform since 2009, so we feel very well aligned with the direction the overall big data analytics is going.

You know, the chip industry used to be really the most innovative place on process control. And to be candid, analytics in the chip industry is in the Stone Age relative to what you see in the Internet companies and community. So we really drafted on that, starting back in 2009.

Our vision is, when you look at the industry, it is disaggregated in companies. But the technical problem is re-aggregating even more tightly. And that's why this automotive customer is kind of important. The automotive customers see a tremendous reliability and RMA problem because the electronics content is going way up. This is often due to changes in the consumables, way back upstream in the front end of the factory, of the wafer factories.

Tying that information together from a control standpoint -- not just a diagnostic standpoint -- really, it's first order of problem for the fabless and IDMs. The foundries play a role in this, too. And then understanding the microstructure that drives that variability is critical for them.

So, DFI gives you kind of the information on a microstructure basis. You know, these wafers we are measuring, a conventional SoC will measure something like 10 billion on-chip instruments per wafer. So, you know, a tremendous amount of information of many different layout styles.

So really this is why the SaaS part of this was very important. We believe there is value for the industry to host that data, from the most front end to all the way through to the back end and assembly, to allow the fabless customers to understand the relationship between all of those changes and that product, and eventual RMAs. So it's more than just yield.

And how we monetize that really is on the throughput of information and the volume that gets pushed through. So this will always be on volume-based business models, because really the application scales with volume.

Got it, got it. And the primary customer for the day is going to be the end OEM and the fabless guy?

We believe that there's -- you know, of course we see a lot of application in fab. And the fabs want to be able to do more and more to control or deliver value to the customers. But ultimately, when you look at the automotive companies, when they had a problem, they turned back around to the chip supplier, right? So the fabless entities need to manage this.

And if you look at the mergers that are going on in the fabless companies these days, these are all about manufacturing scale, right? And they've got now tremendous scale. And they need to have more updated ways of managing the volumes of data that they're going to push through, particularly if you believe in the Internet of Things and the volumes that these kind of represent.

Got it. All right, very helpful. Thank you so much.

At this time, there are no more questions.

Thank you, everyone, for attending today's meeting. We look forward to talking to you again.

Ladies and gentlemen, this concludes the program. Thank you.