PDF Solutions Inc (PDFS) 2017 Q4 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the PDF Solutions, Inc. conference call to discuss its financial results for the fourth quarter and full year ended Sunday, December 31, 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 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 ended 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 obligations 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. If you've not already seen our earnings press release, please go to the Investor section of our website where it is posted. The management report presentation for the quarter and full year will be posted later today.

Today, we will discuss the fourth quarter of 2017, our performance across the entire year, our perceptions about the market we serve and the implications for 2018.

First, let's discuss the contracts closed in the fourth quarter and other significant events in the quarter. It was a busy quarter, particularly for the Exensio team. Contracts to highlight that closed in the quarter are a process development engagement that utilizes our CV infrastructure to characterize 28-nanometer embedded nonvolatile memories, a new enterprisewide deployment of both Exensio-Yield and Exensio-Control at an existing memory client in China, a contract for renewal and expanded deployment across multiple factories for Exensio-Control at a large Japanese IDM and a large engagement -- an engagement for engineering services related to our new Exensio-ALPS product at an existing customer of ALPS, CV infrastructure and Exensio.

Some of you may remember that in the third quarter last year, we acquired ALPS, a software tool that keeps track of chips as they're removed from wafers and inserted into packages. This kind of die level traceability is increasingly important for situations like sensors, which often have multiple chips placed in a complex package.

This is our first contract that combines the service capability from our Exensio team with the technical capabilities of ALPS. These ALPS-related services provide expanded opportunities for this customer to employ new control techniques across their [manufacturing] flow. In addition to the business just mentioned, we closed many other contracts for Exensio platform and modules, including Exensio-Test and related services in the quarter.

We had 4 strategic objectives for the year: first, expand the use of our CV infrastructure beyond leading-edge process ramps to include development and deployment in derivative nodes, such as 28-nanometer embedded memory; second, expand the adoption of our solution -- of all of our solutions in China; third, expand the market for Exensio both vertically and horizontally across the electronics supply chain. The contracts I highlighted earlier show the progress towards these 3 strategic objectives. The fourth strategic objective is to develop and deploy a new inspection solution based on electrical characterization, a common language between design and fab. We call this our Design for Inspection solution or DFI for short. This allows fabs to inspect 3D structures to provide a link between fabless layout patterns and in-fab process control.

DFI has 3 components codesigned for optimal results. The first component is test structures that are placed on test chips, scribe lines or in dummy silicon products. The second component consists of E-beam tools specifically designed to measure these test structures. The final component is a complete analytics platform optimized for the application.

Today, there are over 100 chips taped out with DFI content on them. These include full production designs as well as R&D test chips. While R&D wafers have tens of millions of these test structures on them, full product wafers have billions. The first generation of our E-beam tool, the eProbe 150, is being used to measure test chips in R&D and ramp. The eProbe 150 is now 6x faster than it -- when it was first shipped to customers, capable of measuring 30 million test structures per hour.

For the last few years, we have been developing the next generation of this tool, the eProbe 250, with the goal of measuring billions of test structures in under 2 hours. While the eProbe 150 is well suited for R&D and ramp, the eProbe 250 is being designed for production. While the key differentiation for the production -- the differentiation for production versus R&D is speed, through the last half of 2017, we discovered that optimizing the 250 to have improved electrical sensitivity increased the potential value at our customers. We made the decision to delay the customer demos from the third quarter to enhance the capabilities of the 250. Since our call in October of last year, we have now reached critical milestones in the development of the 250 and have started processing test wafers at a clean room facility in Milpitas.

Now that we're able to demonstrate improved electrical sensitivity, we have scheduled additional customer demos. We anticipate being semi certified by Q2, in the position to ship the first in the first half of this year. We are very encouraged by the demonstrated performance of the eProbe 250, which has already demonstrated about an order of magnitude faster in speed than the 150 in the same application.

Overall, in Q4, the company made significant progress towards all 4 of our strategic objectives. As we look back on the entire year of 2017, we had a mixed performance. Our Yield Ramp business, both solutions delivery and gainshare, did not perform well. Wafer volumes at 28-nanometer and below at the second-tier foundries have not kept up with the foundry leader. Their lower volumes have negatively impacted our business, both in terms of gainshare and these companies' development and capital investments at these nodes. We believe that collectively, they have not delivered competitive node derivatives. Electrical characterization is critical to developing and ramping these derivative processes, and we are transitioning our CV infrastructure to improve our customers' ability to develop derivatives. The embedded nonvolatile memory R&D engagement that I mentioned earlier today is one example of this transition.

While the Integrated Yield Ramp business did not perform well in 2017, Exensio did. By the second half of 2017, Exensio and DFI together represent about 50% of our design-to-silicon solutions revenue. Exensio now has 130 customers, including many fabs, IDMs, fabless, system companies, equipment companies and OSATs. Customers tell us they select Exensio because it links data from wafer fab, wafer sort, assembly and final test, it delivers machine-learning capability customized for semiconductor manufacturing. And because many in the industry use Exensio, it is easier to collaborate with your OSAT foundry design partner if your team also uses Exensio. Overall, we are very pleased with the progress we have made with Exensio and believe we have much more we can do with this platform.

We achieved many technical milestones for DFI in 2017. We have 4 machines running in 3 fabs. In Q3, our first customer signed our first renewal contract to continue the use of the eProbe 150. Although we were late on some of our development milestones for the eProbe 250 last year, we have achieved significant performance above the 150's capability. In summary, while we made good progress on Exensio and DFI, we have work to transition our CV business.

As we look to 2018, we will make adjustments based on our learnings from 2017. By the second half of the year, as we complete initial development of eProbe 250, we will reduce our R&D spending with third-party suppliers. By the middle of 2018, incremental spending on DFI will track more with revenues as the fundamental development should be complete. For Exensio, we anticipate expanding investment in the field as we continue to drive adoption of our solution. For the Yield Ramp business and the general Characterization Vehicle capability, we will reduce spending in our deployment until we see improvements in the overall situation. We will innovate in new characterization techniques and explore modifications in our business model to improve return on investment. We believe that our core Characterization Vehicle technology is more vital than in the past, but we need to be more creative in our delivery and business model.

Finally, I would like to end by providing some perspective of the journey we've been on these past few years. In 2014, we concluded that our business, which primarily was with foundries as they brought up leading-edge processes, needed to be transformed. Decreased capacity in the leading-edge, long-node transitions and a concentration of customers prevented us from growing the business further. While we had those headwinds, we took advantage of opportunities in Design for Inspection, Exensio and the expansion of semiconductors in China. This has enabled us to expand across the leading-edge and mature nodes from fabs to fabless, from front-end wafer fabs to back-end assembly and test.

Today, PDF has over 130 customers. Our revenue concentration has come down from the top 3 customers contributing 80% of our revenue in 2014 to 56% in 2017. China contributed virtually 0 revenue in 2014, and in 2017, is contributing over 15%. Many leading Chinese foundries, memory and fabless companies are customers of PDF Solutions today.

Design for Inspection has brought electrical test in line to leading foundries and, equally important, has engaged the design community in inspection. This transition has not been without risk or investment. We are very mindful of the latitude our stockholders have given us to make this transition. We're also thankful to the dedication and support of PDF's employees.

The changes in investment we have made have not been trivial, and while we are proud of what we've managed to -- managed the company to continue profitability through the past few years, we haven't achieved our goal yet. As we steward the company through this important inflection, we remain committed to capturing the value from our investments over the next years to come.

With this, I'll turn the call over to Greg.

Thank you, John. As John mentioned, our management report with comments regarding the financial results of PDF for the quarter and the year will be posted on our website later today. Given that, I'm going to focus my verbal comments for the quarter and the year on a few key highlights reflected in those results.

First, on revenue, total revenues at $26.8 million for the quarter were up $259,000 as compared to Q3 2017. Gainshare revenues at $7.8 million increased by $478,000 from Q3, while solutions revenue at $19 million decreased by $218,000. Gainshare revenues improved during the quarter, primarily driven by revenue increases from one major customer at 28 nanometers. During the quarter, the decrease in solutions revenue was primarily related to lower hours worked across multiple IYR projects and customers, and that's the Integrated Yield Ramp business, which was substantially offset by increases in Exensio software revenues that were driven by strong Q4 bookings.

For the year, solution revenues were $2.7 million lower than in 2016 as a result of weakness in our IYR business when compared to 2016, partially offset by growth in both Exensio and DFI sales. Gainshare revenues in 2017 were $2.9 million lower than in 2016, primarily driven by continued weakness at the 28-nanometer node across all of our customers, partially offset by the continued ramp-up of 14-nanometer revenues at one of our major customers.

Turning to expenses. On a GAAP basis, total expenses for the quarter were $25.9 million, approximately $150,000 lower than in the previous quarter. This decrease in expense was primarily due to nonrecurring expenses recognized in Q3 related to cost reduction actions implemented during that quarter. This decrease was partially offset by higher SG&A expenses in the quarter due to variable compensation earned on strong Exensio bookings during the quarter. On a non-GAAP basis, total expenses for the quarter were $22.6 million, slightly higher than in the previous quarter.

On a GAAP basis, cost of sales in Q4 was $12.3 million, approximately $100,000 lower than the previous quarter. On a non-GAAP basis, cost of sales for the quarter was $11 million, the same as in Q3. On a GAAP basis, R&D expenses in Q4 were $7.6 million or approximately $200,000 lower than the previous quarter. On a non-GAAP basis, R&D expenses for the quarter were $6.7 million, also approximately $200,000 lower than in previous quarter. On a GAAP and a non-GAAP basis, SG&A expenses were -- in Q4 were $5.9 million and $4.9 million, respectively, with each increasing approximately $200,000 over Q3, primarily due to the previously mentioned variable compensation earned on strong Exensio bookings during the quarter.

For the year, total GAAP spending at $101.7 million increased by $7.2 million as compared to 2016. This was primarily related to increases in depreciation and development expenses for our DFI solution; salary expenses related to our annual merit performance program; additional R&D investment in Exensio; increased audit and tax fees related to accounting policy and tax law changes; and finally, legal expenses related to an acquisition completed during the year.

GAAP net loss for the quarter was $2.6 million compared to GAAP net income of $590,000 in Q3. Non-GAAP net income for the quarter was $4.2 million, up $600,000 from Q3. GAAP net loss for the year was $1.3 million compared to GAAP net income of $9.1 million in 2016. Non-GAAP net income for the year was $13 million, down $8.5 million from 2016.

Looking at the balance sheet. Total cash at $101.3 million increased by approximately $500,000 during the quarter. This increase was primarily the result of cash generated from operations of $4.3 million and $500,000 in proceeds from stock option exercises, being partially offset by PP&E purchases of $3.1 million -- excuse me, $3.3 million, most of which were related to the development of our DFI solution; and $1.1 million for the purchase of company stock related to the settlement of employee tax obligations and RSU grants.

Accounts receivable, including total unbilled receivables, was $66.2 million as of December 31, 2017, which is an increase of $4.2 million during the quarter. This increase is primarily related to slow payments by several of our customers in China. DSO for total accounts receivable increased from 213 days in Q3 to 225 days or a 6% increase in Q4. The total current and long-term unbilled accounts receivable balance of $30.9 million increased during the quarter by $1.1 million. Of the $30.9 million of unbilled AR balance, we expect to bill $22.2 million over the next 12 months, of which $12.1 million will be billed during Q1 in 2018. Since the end of the quarter, we have collected approximately $11.1 million of the $35.4 million trade accounts receivable outstanding.

Now turning to taxes. Our GAAP tax provision for the quarter was $3.5 million. The provision was primarily driven by a reduction in the value of the company's deferred tax assets related to the adoption of the tax rates as prescribed by the new tax law. This tax rate was reduced from 35% to 21%. Cash taxes incurred for the quarter were $64,000 or a cash tax rate as a percentage of GAAP pretax income of 7%.

Now I'd like to turn for a moment to the outlook for 2018. On an apples-to-apples basis, we expect 2018 total revenue to grow at a rate in the high single digits as compared to 2017, and gainshare, DFI and Exensio revenue growth rates should be above that single-digit level. However, we expect to see continued downward pressure on our fixed-fee IYR revenues. In each case, because of the accounting changes driven by the adoption of ASC 606, it remains to be seen how 2018 results will compare to 2017 results on a revised basis under 606.

Based on continuing weakness in our IYR business, however, we will reduce our spending amounts related to that business. These reductions, however, will be partially offset by increased investments in Exensio and DFI field organizations. Overall, our total spending on a non-GAAP basis is expected to be down moderately as compared to 2017. In conclusion, given moderate growth in revenues and reduced spending levels, our non-GAAP net income should increase year-over-year.

And with -- for a moment, I'll talk about the accounting and tax and reporting changes. With the enactment of the new tax law, we expect our GAAP tax provision rate for 2018 to be at 18%. However, we may modify this rate as we get further into the year and get more information on the impact of the various tax changes. Cash taxes, expected -- are expected to be at 12% of pretax non-GAAP income. For modeling purposes, non-GAAP pretax income should be taxed at our GAAP tax provision rates.

In regards to the new revenue accounting standards, we are currently finishing our estimates of the impact of these changes on our expected financial results for 2018 and '17. We will discuss this impact in detail on our Q1 earnings call and update our guidance accordingly.

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

(Operator Instructions) And our first question is from Jon Tanwanteng with CJS Securities.

Just a quick update on the demand picture for both your gen 1 and gen 2 DFI machine. Are you seeing additional interest in there for more series 150? And then just how do you see the series 250 playing out in terms of shipments and orders as you get that actually demonstrated on the floor at your clients?

Sure. This is -- Jon, this is John. So we have a number of demos and pilots going on with existing customers as well as new customers. With the systems, some of those pilots are for applications -- the majority of those are for pilots and applications the 150 can serve, and we still anticipate the 150 being useful in ramp and R&D. As we get through the year, we start shipping the 250s. We do believe that for production, the 250 will be a much better solution. If you have a tremendous amount of R&D or ramp need for customers running well over 100% utilization on the 150, the 250 is a better alternative because on a cost per unit measurement, it's much lower. But for customers, we have some R&D customers that keep the machines only running at 50% or 60% utilized. Being 10x faster doesn't really buy you much if you're not at 100% utilized anyway. So for those customers, we foresee that the 150 will continue to be a useful machine in the foreseeable future.

Got it. And then just relative to how you positioned yourselves in the Investor Day, call it a little bit over a year ago now, you'd told us that the potential from Exensio and DFI could drive a doubling or even a tripling in your revenue power. How would you update that given the situation now with IYR's low engagement, where it is?

Yes. That's a great question, Jon. We actually still believe that opportunity is there. A couple of points on the IYR business. We've been working on this for quite a while. We know that our customers have made much bigger investments than we have. So they're as motivated or much more motivated than us to figure out how to improve their utilization, improve their amount of capacity in their factories. So we believe there is an opportunity there. We do still believe very heavily in the opportunity in China. However, if you look at the spend in China over the last year, while our investment in China has been primarily with the local native Chinese companies, most of the manufacturing that's come online in China has been at the multinationals, primarily in memory, right? So be it Samsung and Hynix's memory investments in China, which we do not participate in.

So we still believe the China opportunity plays out the way that we thought it would. It will take -- it's taken longer than we would have liked. And we see -- still see the significant opportunity on DFI that we saw, let's say, 1.25 years ago when we had that Analyst Day. I think the positive part has been we believe in a larger opportunity on Exensio than we saw even at that time. The more and more I spend time with SVPs in operations at the fabless and system companies, I keep on seeing more opportunity for us to expand Exensio. And I do believe that, that is a more substantial opportunity than what be represented at the users -- at the Investors Conference in November.

Okay, great. And just on the gen 2 machine development, are you up to the speed that you want to be in terms of being able to inspect structures and wafers for a unit of time? And when do you expect that?

Yes. So we are expecting to be -- while we -- when we ship at the performance that we've targeted for customers, just like the 150, where once it was in the field it sped up by a factor of 6x, there's a number of firmware upgrades that we anticipate on the 250 that will incrementally speed up its performance once it's out in the field again. Again, by a similar amount as what we saw in the 150. So it'll ship at where we expected it to be, and it'll have a road map for basically getting faster than where we expect it to ship.

Got it. And then, Greg, just a quick question on taxes. What was your cash tax rate in '17? I don't know if you mentioned it before, but...

Yes. If you look at total cash taxes for '17, give me a second, they were actually -- let's see here. So in the quarter, they were $64,000. And for the year, it's a little complicated to calculate out, but they were actually about $1.2 million. When you look at that, it had a lot of influences in the year. We had some windfall tax gains, both on the book tax rate and the cash taxes. So it was lower than we expected. Remember, we were expecting it to be about 22% to 25%, I think, of the pretax GAAP net income.

When we look at 2018, the relationship that we've had in the past between GAAP and non-GAAP income is changing. So we actually -- for 2018, we think it will be 12% of non-GAAP pretax income. And we're not going to bother trying to relate that as a percentage of the GAAP pretax. The book tax provision on the GAAP pretax is 18%, down from what we were projecting to be 38% to 40%. Now we may see that come up 1 point or 2 as we get into the year and start to see some of the impacts and do some more measuring. It's not that we can complete everything with our tax advisers at once. And so -- but we think 18% is a good starting point. So there's a significant drop as we move into the year. And we start to approach book and cash tax, starting to close that gap to where we're down to about a 6% differential right now.

Great. Last quick question. Just the days outstanding and the receivables, any outlook of -- on that improving?

Yes. If you remember, last quarter, we improved it, but I said that, that's a temporary thing because it is kind of like playing whack-a-mole with the customers overseas, that you put your focus on one customer, get them caught up, but in the meantime, the other customers may get worse. And that's exactly what we saw. We're placing a more and more emphasis of resources both in the finance team and of the people out in the field, responsibility for getting those numbers down. But I would not anticipate a great deal of change rapidly there. We'll try to keep pressure on it such that it comes down year-over-year, but I wouldn't look for it to go down by 50 or 60 days all of a sudden. China just doesn't change that rapidly.

Your next question is from Tom Diffely with D.A. Davidson.

This is Franco in for Tom. So my first question is mainly on the DFI tool. And so I was hoping you would give me some additional color in terms of the timing of the release and how maybe that's -- how that's affecting the interest for the 150s. So maybe some customers are pulling out their purchases for the 150 and are waiting for the 250.

Okay. Yes, so this is John. So as we said, we expect to be in a position to ship by the end of Q2. We feel pretty comfortable about that. In terms of customers holding up purchases, I think as some of our customers come into renewals and also new potential customers, they are curious about seeing demos on the 250. Because we don't actually sell the machine, right, we provide the machine, we provide a certain capacity of measurements per hour, we will price or make it at the customers' option to increase -- for increased fees to increase the amount of measurements they can measure per hour, i.e., switch from one hardware platform to another if they so choose. So we're trying to make the individual machines a nonissue effectively for customer decision making. Because as they need more capacity, for some customers, it will make very much sense to trade in a 150 for a 250. And we'll use those 250 -- the old 150s in other customer settings. By the way, if you look at our Yield Ramp business, we've done the same with testers for years, where we provide a certain capacity of test, and over the life of the contract, we'll take out some testers and put in other testers to give the customer a better mix of capacity and precision. And we'll do the same in this case, too. We're never turning over -- we're not planning on turning over title to the test -- the machines themselves.

All right. And I guess beyond the existing customers that you've been doing demos with, what's the interest like at other customers? Do you think that you'd be able to supply them all in a timely manner? Or how is this holding up?

Yes, it's funny. I think the -- a lot of times the most significant interest we get is from the design community. The design community is quite concerned about electrical performance on these advanced nodes as they relate to product performance and product reliability. And as more of the designs that are going into the leading edge are very large chip sizes with applications in high-performance computing and automotive, they become really concerned about parametric reliability and performance. That's why we held up the 250 a little bit because we knew that that's what the fabless community really cared about. So our marketing effort is as much to the designers as it is to the fabs today. And I'd say we see very good interest at both, but probably a little bit more interest with the product teams than with the factories themselves. And that's super important, by the way, just because it allows you to go across multiple factories. So the work we do with a fabless company at foundry A immediately can be part of a discussion to foundry B or C because many of those larger fabless entities use multiple foundries.

Okay. Yes, that make sense. And switching over to the financial side. As it becomes a larger part of your business, when should we expect to see a more -- like a fully realized margin benefit, if you will? And I'm not sure if you've disclosed what like the margins are like for Exensio, but if you could give any additional color on that, it'd be helpful.

Yes. So at our Analyst Day, we were expecting to return to our kind of our pre-investment levels of margins, which means, on a non-GAAP basis, operating margins approaching 40%. We were hoping to be back at that. I think it was in -- sometime in '19. I think we're maybe more towards the end of '19, early '20 to actually achieve that. And 2 of the ways we get there is by increasing participation in the revenue stream from DFI and Exensio. And we expect DFI and Exensio in the long run to have very similar margin profiles, in that they would be at the gross margin line in this 75% to 85%, which will look very much like a technical software term license or an IP license.

Right now, Exensio is approaching that because it is scaling up and is getting very close to that. We probably will obtain that by the time we exit '18. DFI, because it's at an earlier stage in its life cycle, is not there yet, but on the 1 or 2 transactions that we've had, we see that we can get there very easily. We just have to scale up and build up field organizations and so forth. So it will be -- the margin improvement at the bottom line will be driven by DFI and Exensio.

Okay, that make sense. And then lastly, on the IYR business, just trying to get a sense there (technical difficulty) expanding, any chance that this business (technical difficulty) in deployments?

Yes. What we believe, as I said in my prepared remarks, we see -- like this embedded nonvolatile memory, we see many other applications where electrical characterization is important, is important for development, important for control. As I said, we need to be a little bit more creative on delivery and on business model. So what did I mean by that? Well, when you're bringing up a single node where it's being measured on one product, and it's a 2-year R&D cycle, we tend to have a very large team with the vehicles and systems to help the customer get up that ramp quickly. Many of our customers for these embedded and derivative programs, they have 4 or 5 programs going simultaneously. We are -- we've been piloting throughout 2017 on a lighter platform where you're using less support to help customers with a broader range of derivatives. And we believe that, over time, we'll drive more benefit to them and to us. And we may twist the business model in terms of what we get paid for and already have been doing that in some of these contracts on the delivery portion versus what we make on the back end. And how we charge for the back end, we will look at, too, we -- how much risk we take on that back-end piece.

So there's -- we believe it will come back, but it will come back looking different than it did when it was a new node every couple of years, and there's 3 to 4 customers, and you throw a huge slug of resources at that node. It will look very different than it does today, at least, my -- if I were going to handicap how it evolves. And there's some other elements we're looking at that -- I mean, I'm not -- I don't think I want to announce as being done yet, but we're also looking at what else with the Yield Ramp makes sense -- with the Characterization Vehicle technology, make sense to help the customers get over their challenge. A lot of the customers really value the vehicle data as a marketing tool to their fabless customers, and we're looking at how we can help them do that in more efficient ways than we do today.

And your next question is from Christian Schwab.

I just want to understand better the difference between the high single-digits top line growth you expected now versus kind of the solid double-digit previously. Is that all -- is that a combination of a slowdown at 28-nanometer and below at the Tier 2 customer base plus a pushout of DFI? Or is that just predominantly IYR?

Yes, it's predominantly conservativeness around the IYR business. What we've seen all year was the weakness in both the gainshare revenues driven by the customer volumes but also whenever you see -- talk to these customers or read their press releases, they're cutting back on their capital expenditures, and we think they're cutting back somewhat on their development programs, which causes us to be pretty conservative on the near-term outlook. So as far as the rest of the business, the Exensio and the DFI, there's been no change in our outlook on those. In fact, as John said on the Exensio, we actually have an expanded view of the opportunity there. But in the near term, it's really being cautious around the IYR business.

Okay. And then on non-GAAP operating expenses, is that going to be held steady in '19? Or is there a chance that, that could go down given some of the reduction in R&D third-party expense?

Chris, this is John. I think it's a little bit hard for us to forecast what we do with the R&D expense in '19. I think if we -- if the universe of opportunities for DFI is leading-edge logic limited, then we would reduce the -- and this is kind of off the cuff of my head, but as I think about it, we would reduce the overall R&D programs and focus on commercialization on leading-edge logic. We've been doing some early demonstration into other areas, non-leading edge and embedded and nonvolatile memory applications for DFI. If those pan out, we may have a different R&D profile as we get out into that time period. It's a little bit hard for us to say. I mean, frankly, if we're spending more on this it's because we see a bigger opportunity, and we've made some traction with DFI. So we should have -- accelerate our top line growth before we make increased R&D spend. And if we don't have accelerated top line growth, we'll have reduced R&D spend.

If you were going to sell a DFI box, the 150 or the 250 as a stand-alone box, what would be the -- what should we assume would be the expected ASP?

I haven't really given it any thought, Christian, because it's not been something that we've been asked or thought about doing with the system. Because the box without the software and without the test vehicles is -- it's not that useful. So I don't really know. What -- yes, I mean, what we're trying to do, and I think as Greg said in the question with the previous caller, trying to make sure we're getting to good gross margins on the solution. And I think if we can get to the gross margins north of 75%, it would be very reasonable for a hardware business. And the reason why it's, I think, north of most hardware businesses in the capital equipment world at that margin is because the customers are valuing the software and IP that comes with the solution. So then go and peel off one part of it, it's hard to say what that piece of it itself would be worth.

Okay. The reason I ask is some of the industry people that I talk to about kind of the challenges potentially facing your ramping that technology is that large-scale purchasers don't like sharing money, of success or pain, kind of prorated for outcomes. They would rather buy the box and the software and the test vehicles and then figure out how to use their knowledge process know-how to use that tool or that box better than the competition, similar to the way that they buy almost every other equipment box that they use inside of the fab.

Yes. So a couple of things there. I think people have complicated -- the people that you spoke to complicated a couple of things. So the -- separate how folks pay, and we -- it's just an economic discussion. If a customer came to us and said, "I don't want to pay ratably, I want to pay for it in a different way. And I want to pay for it upfront, and I want to own it some number of years, if not in perpetuity." That's just an economic discussion, and we can sit down on that discussion all the time. I think the part that folks are complicating is, what we've done in the Yield Ramp business and what we're doing in DFI. For our lead customers, we offer them a design, what we call a design -- what's the acronym for it? But basically a design development kit. So customers can design their own test structures that go in DFI.

In fact, the customer that we have in Asia, they design their own structures all the time. Those get put in the system. It's a software configuration. They run the analysis on what they do with it. So from a using standpoint, it's no different than using any other piece of equipment in a factory or any other piece of software or even any of the tests. So the, "Oh, I don't want to share with PDF part of it," or whatever you heard, that's not actually something that they need to do or is required to do with respect to the way customers work with us on DFI. Some choose to do that; some do not. Our customer in Asia says, "What we'd like is we'd like to get you guys to put down your content, and then we'll put down our content, too. And we get the best of what PDF already provides, and we differentiate on top of what we think you provide, too." I suspect that for the leading customers, that is a very leading company out there in the world, that's how they'll all actually get at this capability. The second part you bring up is just an economic discussion. And frankly, we're happy to sit down and have that conversation with customers. We can -- it's all just a way of modeling out good returns for us and a good value to them.

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

Thank you.