PDFS - NASDAQ

Good day, everyone, and welcome to the PDF Solutions Inc. Conference Call to Discuss its Financial Results for the Second Quarter and Year-End 2023 Conference Call Ending Friday, June 30, 2023. At this time, all participants are in a listen-only mode. After the speakers' presentation, there will be a question-and-answer session. [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 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 15 through 29 of PDF's annual report on Form 10-K for the fiscal year ended December 31, 2022, 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 Adnan Raza, PDF's Chief Financial Officer. Mr. Kibarian, please go ahead.

Thank you for joining us on today's call. If you've not already seen our earnings press release and management report for the first quarter, please go to the Investors section of our website where each has been posted. The second quarter was very similar to our first quarter. Revenue remained strong as we experienced continued adoption of our end-to-end analytics by our customers. Before Adnan discusses the financials in detail, I have some comments about our observations from the second quarter and our perceptions of the market for the remainder of the year. Bookings in the second quarter was similar in magnitude to the first quarter. Significant contracts that closed in the quarter included customers deploying process control and front end fabs, a cloud customer deploying analytics for an internal chip design team and a chip company deploying analytics for complex 3D packaging. With continued strength in Asia, game share improved in Q2 versus Q1 as shipments improved. Finally, bookings for symmetric connectivity runtime license shows modest improvements in Q2 versus Q1 as our customers' equipment shipments, particularly in China, started to increase. Overall, given our strong backlog and business model, where most of our revenue is typically ratably recognized, we continue to deliver strong results in revenue and earnings. Beyond the business that we booked, we have experienced significant customer interest in our analytics and a number of pilots are underway with customers. We were pleased with the business results in the quarter as it demonstrates the strength of our business model. Subsequent to the second quarter, we announced an acquisition of Lantern Machinery Analytics. Over the past year-and-a-half, we've been evaluating with some battery manufacturers the ability to apply PDF's analytics platform to lithium ion battery manufacturing. Today, lithium ion battery manufacturing is under $100 million in revenue, and over the next 12 years is forecasted to be over $400 billion. So there's going to be a tremendous increase in capacity. Moreover, controlling variability of manufacturing to improve yields and product quality is becoming of interest. Through our work with manufacturers, we recognize that there is a need to collect and process images. This data, when combined with upstream equipment sensitive data, has the potential to be used to improve yields and product quality. Machinery Analytics have developed an ML pipeline for battery image processing. This capability may also have applications for our IC customers as well, particularly assembly processes where many images are collected. Machinery Analytics had some small revenue with early customers, but for the most part was a pre-revenue company. We will incorporate their ML pipeline in our products as we develop applications for battery and IC manufacturing. We're excited to have them join the team. Turning to DFI and our eProbe machine. The customer we previously talked about has qualified our new eProbe 350 and now has two machines running at their facility. After a lengthy evaluation in our fab, we anticipate shipping another eProbe 350 by the end of the year for an onsite evaluation with a different customer. We are very excited about the progress we are making with the DFI program. As our customers develop 3D processes like gate-all-around and backside power, we believe electrical inspection will increasingly be important to ramp and control yields. Now, let me turn to discuss our view of the environment and our perspective of the second half of the year. Midway into 2023, we are more cautious in our short term view. Our game share customers in China are reporting decreased wafer volumes, which will reduce game share in the second half of the year. We anticipate continued increases in Cimetrix runtime licenses, but a lower rate of improvement than we initially anticipated, as equipment suppliers, particularly outside of China, remain conservative on the increase in equipment shipments. Finally, for Exensio, in Q2, we saw customers delay some expected bookings to the second half of the year. Although some of those bookings already closed in July, we remain cautious about the timing of others. Early this year, we anticipated revenue growth for the year approaching mid-teens. Our expectation now is year-over-year growth will be in the low double-digit percentage. The cloud analytics growth is expected to exceed the overall growth, but it will be offset partially by year-over-year decline in integrated yield ramp. While the short term environment is unsettled, the long term drivers for our customers, which include increased use of AI, ML, cloud, smart devices and the electrification of the energy economy, remain in place. These drivers are being amplified by the various government investments in semiconductors we are seeing around the world and the increased diversification of the supply chains that many of our customers are embracing. As a result, our pipeline of business is strong and remain confident on our customers' continued success and growth. We would also like to remind everyone that, on October 24 to the 26th, we will have the PDF Users Conference Meeting at the Santa Clara Marriott. As with our pre-COVID event, we will combine one of those days with an Analyst Day, which will be on October 24. This gives our customers, strategic partners, analysts and stockholders a chance to see the latest capabilities from PDF and to learn from each other. The theme this year is applying AI/ML to transform manufacturing and technology R&D. The list of speakers is turning out to be the strongest we have ever assembled. Sanjay Nagarajan, SVP and co-GM of the Intel Logic Technology Development, will talk about the transformation they have made, which has enabled them to deliver five nodes in four years. Other speakers include executives from SAP, Siemens, Advantest, Global Foundries, and Analog Devices. We also expect additional executives and engineers will commit to speak at the event. Our attendees usually include executives and engineers from system companies, fabless, IDMs, OSATs, foundries, and equipment companies who all share a passion for analytics and ML that drives R&D in manufacturing. We are looking forward to the event this year after a multi-year hiatus after COVID. I want to thank all PDF employees and contractors for their efforts during the first half of the year. Now I will turn the call over to Adnan, who will review the financials and provide his perspective on our results. Adnan?

Thank you, John. Good afternoon, everyone. And good to speak with you all again today. We are pleased to review the financial results of the second quarter and to bring you up to date on the progress of the business. Our Form 10-Q has also been filed with the SEC today. Please note that all of the financial results we discussed in today's call will be on a non-GAAP basis and a reconciliation to GAAP financials is provided in the materials on our website. Financial results for the second quarter of 2023 continued to be strong, coming after a solid first quarter. Q2 total Revenue was $41.6 million, up 20% versus the comparable quarter last year. Analytics revenue was up 19% to $37.1 million in Q2 of this year versus $31.1 million in the second quarter of 2022 and represented 89% of total revenue this quarter. The growth in our analytics revenues came from growth in Exensio and leading edge products, offset by decline in Cimetrix runtime licenses. On a quarter-over-quarter basis, our analytics revenue was up $0.8 million. During the second quarter, revenue contribution from integrated yield ramp was $4.5 million, up 26% from last year's comparable quarter, primarily due to increased level of game share from higher volumes at some of our Asian customers. We're very pleased with the various engagements we have currently ongoing, the deal sizes of bookings we're working to close and the strategic conversations we're involved in with our customers, strategic partners, and major semiconductor governmental initiatives around the world. All of these factors evidence progress towards a goal to be the go-to manufacturing data analytics platform for the global semiconductor and electronics ecosystems. Our ending backlog at the end of Q2 of this year was $244.9 million, which is 33% higher than our prior year Q2 ending backlog. We reported gross margins of 74% for the quarter, up meaningfully versus 69% for Q2 of prior year. As we have said before, on a quarter-over-quarter basis, we may see some variations on this metric as we modulate the spend for our customer engagements and grow our cloud and people spend to support the growth of recurring revenues. We remain committed to our non-GAAP gross margin target model of greater than 70%. On the operating expense spending side, our R&D spend was down $0.6 million versus the prior quarter as we continue to take advantage of our leverage and shift our resources to pre sales and new business initiatives. Our SG&A was up $0.9 million versus the prior quarter, primarily driven by increased spend in pre-sales and marketing efforts. Overall, within SG&A, we have invested faster into sales and marketing, while ensuring that on the G&A side, we can take advantage of our scale and have brought down G&A as a percentage of our revenue slightly versus both prior quarter as well as Q2 of prior year. For EPS, we reported a profit of $0.19 for the quarter, similar to last quarter level, but meaningfully higher than $0.11 for the same quarter a year ago. We are pleased about our year-over-year $0.08 positive swing in EPS compared to the same quarter of last year. We ended the quarter with cash and cash equivalents of $124 million compared to $117 million at the end of same quarter a year ago and $134 million in the prior quarter, with the change versus the prior quarter primarily driven by an increase in our accounts receivables at the end of the quarter due to timing of billings. Since the end of the quarter through today, we have already collected the majority of our quarter-end billed receivables. During the quarter, we also spent approximately $1.9 million of cash to close the acquisition of Machinery Analytics that John mentioned, with an excellent team based out of Canada and Poland to expand our analytics platform for EV battery manufacturing. We continue to believe that the strength of our balance sheet positions as well to consider strategic investments in acquisitions as they become available. Like John mentioned, as we look forward, we expect to grow our revenue for this year on a year-over-year basis at lower double-digit percentage level instead of the approaching mid-teens level we had previously guided. We're being careful based on three key observations of the rest of the year. First, our game share from some Asian customers is expected to slow down for the rest of the year as they face their own economic and demand challenges and volumes decreased. Second, our Cimetrix runtimes licenses, while up versus prior quarter, are still facing a mix of increased demand from some regions, coupled with muted demand from other regions. Third, consistent with John's comments, the timing of bookings is less clear, given the industry dynamics. As we become strategically important for our customers and partners, the size of our bookings is growing, with many approaching the high-single digit or double-digit millions of dollars. Overall, when we look at the longer term, we feel emboldened by two factors, the demand for our products that we can see in our sales pipeline and the strategic relevance of our analytics platform, which is taking hold with all three constituents – customers, strategic partners, and the various government initiatives. We also believe that, in our analytic platform, we have three strong elements that complement each other, the Exensio analytics platform, the unique data collection capabilities of our leading edge products, and our Cimetrix conductivity products. We're excited about the future and the growth ahead for PDF. With that, I'll turn the call over to the operator to commence the question-and-answer session. Operator?

[Operator Instructions]. Our first question will come from Blair Abernethy from Rosenblatt Securities.

Just wanted to ask you about the acquisition of Lantern Machinery, just maybe give a sense of what needs to be done to take this – I guess, package this product and take it to market and what sort of are you thinking in terms of timelines?

A lot of the work early on in R&D is done with SEM images, scanning electron microscope images. We wanted to use optical images. The pilot we did with them before acquisition was to test that pipeline on optical images coming from some of the manufacturers we had been working with. We feel very good about that. We expect about six months of development to package this appropriately, and then begin pilots with customers. So, probably in early 2024, we'd expect to start seeing pilots begin with customers and revenue to follow from there.

It looks like from your 10-Q, you paid just under $2 million? Do you have any sense of how much money will be required to get it to market? Will we see that in the R&D or CapEx or how's the investment going to show up?

There's really two elements to that. One is the people spend, of course. The team is just under 10 people. And like I mentioned in my prepared remarks, it's based in Canada, as well as in Poland. So there's a bit of balancing what the costs were able to do there as well. The second piece of the cost is, of course, what we spend, to your point, on system improvements. We will, of course, go through the appropriate accounting treatment as far as P&L versus capitalizing, but given the small nature, it probably will hit our P&L. It's going to be small. We'll keep it manageable. Couple of 100k is kind of what we're thinking on this acquisition when you look at for this year. So, we'll continue to keep it manageable, is kind of our perspective. I'll give you one other context. The machines themselves, the equipment that's used, it's basically off-the-shelf camera, and then the structure that's used with it. So a lot of the value is in the AI/ML pipeline and the software that we'll be coupling with it. So, hopefully, once we get to the point that John is mentioning, the reason for the acquisition is it hopefully is very accretive to our margins.

Just shifting over, can you just highlight a couple of things on the partnership side that stand out in Q2, whether it's with Advantest or some of the other partners you've been working with?

I think in my previous remarks on Q2, I had already talked about anticipating the Advantest user conference. We presented there, had tremendous turnout jointly with customers, announced a number of new product capabilities in partnership with Advantest. At SEMICON, we were in the AWS booth to highlight Exensio cloud on AWS. I think that generated quite a few leads, as I saw the statistics. I think somewhere between 50 and 100 leads generated off of that activity. And then, also with SAP, we continued a number of selling activities in early demonstrations with customers beyond the first customer on the PDF SAP pilot. Nothing to describe from a booking standpoint there. But, again, I think we find that they're a very effective way of reaching at it. A lot of the customer base – it's the same company that we already knew, at a different entry point in the company, typically through the finance organization or C suite. I would say that partnership activity with those partners was quite heavy throughout the quarter. You might have noticed at SEMICON, we were quoted in a Teradyne press release of support of their edge box with Exensio applications. And we have a long standing relationship with Teradyne and supported their testers for close to a decade now. But this enhances our ability to provide the same machine learning capability on their edge box that we can provide on others. And that also kicked off a whole set of customer dialogues as well, as there's many customers on the Teradyne platform. We do believe that's an important economic relationship.

Our next question will come from line of Christian Schwab from Craig-Hallum Capital.

Given the substantial backlog and kind of a lot of long term positive growth drivers, despite bringing guidance down for the back half of the year, what should we be paying most close attention to as we think about what your revenue prospects for 2024 and beyond are? What would have to happen to return to the 20% plus growth rate and what would have to occur to kind of remain here in the low-double digits?

I think that's a great question, Christian. So continued weakness on equipment will create a headwind for us. As I said in my prepared remarks, if you take out the equipment weakness and IYR weakness, the core analytics business actually – the remaining pieces that are in control continues to grow pretty reasonably. And so, we'd like that. Ending that headwind would get us back up pretty close to the 20% range. Then I think it comes down to bookings momentum on core analytics. I think that would get you back over – we've been growing more than 30 plus range for a while. That will get us back into that range. Adnan said in his prepared remarks, we have an awful lot of pilots going on. I think a large number of things that are in the double-digit millions in terms of bookings value. In the past, it's been a small number that drove a lot of the growth. I'd say that the breadth of them has increased quite substantially. I think with those bookings, you start seeing, obviously, return to more robust growth. And, of course, that's on top of a larger base than it was in the past, right? Obviously, 30% on a small number is harder – much hard to do 30% on a larger number. But we do see that activity out there, that customer interest activity out there. So I kind of went on a ramble. Let me see if I can summarize that for you well, Christian. Let's get rid of the headwinds, stop declining on those things, growth on the equipment shipments again, would really help in terms of not creating a headwind and then the bookings activity on the larger eight figure deals, both on the leading edge and on Exensio.

Just a quick follow-up on that. Let's just say that wafer front-end equipment sees a muted recovery next year. Do we have enough double-digit million pilot businesses to take us to 15% to 20% growth in a flat WFE market?

Yeah.

Our next question comes from Tom Diffely from D.A. Davidson & Co.

This is Linda [ph] on behalf of Tom Diffely. First of all, congratulations on a good quarter. So I guess my first question is, I'd love to better understand the trends you're seeing across different customer types. Maybe you could walk us through what you're observing in customer behavior across memory, automotive, data center and markets.

If you kind of double click where we saw the bookings happen in the quarter, the two strong areas, silicon carbide in the high voltage related area, that was one of the process control bookings that we saw, some of the analytics bookings that we saw and actually just usage statistics coming off our cloud sites for customers. We track within our customers what parts of organizations are driving the most. And that definitely continues to be a strong part of the market. The Exensio analytics deployment for a cloud customer, obviously, it's data center related. The analytics for an IC company that does complex system and package, again, high performance computing and data center related. So the second area that we've seen quite a bit of strength has been in the data center area. That's definitely we see the second thing. We started seeing bookings for customers, I didn't talk about it my prepared remarks, that were on the test floor for our escalated customers that we were encouraged by that because that tends to be driven by mobile and cellular. We did see positive results in that area. So I think those customers are starting to feel a little bit more confident. They have been, I think, weaker in the past. And then lastly, on the memory side, when we look at our pre-sales activity, we do see a fair amount of activity on the memory piece. That was a little bit larger than we have in the past. And so, I feel like they are also starting to see the end of the tunnel for them. When you look geographically, as I said in my prepared remarks, we saw a pretty steep follow-up in the third quarter on wafer volumes because we measure the amount of wafers we test for our IYR customers. And we did see that drop off in our royalties. And that's how we became cautious on the game share royalties out of China. We saw that drop off in in July relatively meaningfully compared to the previous months. And when you look at our runtime licenses, though, we see factory activity, new equipment shipments for companies shipping into China, Chinese equipment companies, as well as companies that are primarily in Europe, but shipping to China, also being very strong. And when you look at what category of technologies, they tend to be more mature nodes with an emphasis on high voltage again, silicon carbide and IGBT and things like that. So, overall, what we see, I think, is very consistent with what you see reported in the market when we look at the segments that we track, equipment shipments, wafer volumes, and buying characteristics from our customers.

With respect to the forward pipeline, to what degree has it been influenced by the announcement of your seven figure S&P booking last quarter? And can you see any tangible effect on the developing interest from other customers?

The partners continue to be a meaningful part of our pre-sales activity with customers. It's very hard whether to say, okay, will they come in faster than the others, we have found that their selling cycles are different than ours. Some are able to give us pretty strong predictability. So we do anticipate them impacting our second half of the year. But the bulk of the selling in the second half of the year will come off pilots that are run directly in a two-way relationship between an end customer and PDF, with most of the bookings being driven by substantial contracts for Exensio and leading edge.

A follow up on that. Have you noticed any uptick maybe in conversations with other partners like SAP, where you can pair Exensio with other platforms and drive more of those top floor to shop floor linkages?

I mentioned the new relationship with Teradyne. That was one that they announced during SEMICON. We are also working on other things. Some are engineering related and some are more software related – engineering software related. We're not in a position to announce them yet, Linda, but I think, obviously, we have a user conference coming up. So you can anticipate us being in a position to make more announcements at our user conference around things that we're working on.

We'll be looking forward to that. And then a last question on the model, Adnan, on a year-over-year basis, gross margins have improved and they continue to progress nicely. So, aside from volume leverage on revenue growth, what other drivers are contributing to that improvement from this level?

I think a bunch of things. As the business scales, we've been able to focus a little bit more into all of the different dimensions, right? So whether it was on the Cimetrix side, taking a look at our prices and where they are positioned on the cloud side, okay, how do we charge versus markup on the AWS, what do we pay actually ourselves with our scale getting bigger, a combination of all of those things. And then, on leading edge, we've talked about the enterprise level deals that we have done. So a combination of those factors is what has allowed us, frankly, to be above the long term margin of 70%. And we hope to maintain that even in face of some of these are revenue comments as we have discussed on the call today for the next few quarters and the rest of the year.

And our next question will come from the line of Gus Richard from Northland Capital.

On PFI, I think you said you've shipped a system in the quarter and expect to ship another one in the fourth quarter. How many would you have in the field at this point?

What I said in my prepared remarks, we actually said in our last call that we had shipped in the second quarter a machine that was qualified, actually, in the first quarter, recently, at an existing customer. And then we would anticipate being able to ship again in the fourth quarter. That would put three in the field today.

The new systems that are going into the field, are they going to be used in lab or are people sort of thinking about trying them out for prefab?

If you look at how an e-beam has probably been used conventionally really sucks, e-beam inspection tool, most of the customers tell us, once you're kind of done with R&D, the number of defects you can see in a two hour recipe isn't enough to continue to run the machine. So, they typically are not used in manufacturing. I think the good customer base overall came into evaluating the e-probe, assuming it was a similar situation, and what they're finding with the 350, given the machine's ability to scan a few billion pieces per hour, so in a two-hour recipe measuring pretty close to 10 billion features. It's actually finding defects even as the node is maturing. Now I think people are just starting to scratch their heads saying, well, maybe this can actually be used in manufacturing. That would be a thirst. And as you know, for things like 3D defects, via voids, shorts, and gate-all-around structures, there's really no inspection technique other than an electrical test that can see them. So, usually, eventually, people give up on inspection and just try to control things with metrology. We believe the e-probe has the potential of giving them a way of controlling it with an inspection technique. The roadmap will continue to improve the throughput of the machine over the next years. And so, we believe we can keep up with improving defect densities at our customers. I think the customers are encouraged by that.

The revenue model here is equipment as a service?

As of today, yes.

Switching over to the acquisition, I just want to make sure I understand what you're doing here. It sounds like you've got – are looking at batteries. And if you're looking at the battery, you're looking at wells. And if you're going to move it into the back end, maybe it's wire bonding. And you're going to collect massive amounts of images and just use ML/AI against those images to sus out field failures.

The assembly on the backend [indiscernible] I think you mean for the IC production. There's a number [Multiple Speakers] are captured in the back end, not just bonds, but chip finishing and other things like that. So there's many places to apply an image recognition pipeline and assembly. For the battery area, the information is really about the grain information on the cathode and anode as the different materials are applied. Customers want to take very high speed images because these things are on a roller that move that many meters per second. And they want to capture images of those grains that predicts the features [ph] in the grains and different things give you a prediction about – what's this battery's life, like, going to be like, how stable is this battery, is it going to perform [indiscernible]. So we have a partnership with Voltaic. They have end-of-line extraction of electrical behavior. This gives you an inline measurement. And then our process control products gives you the ability to collect data on the equipment that's rolling the material, the rollers, the pressure being pressed when you're applying the pace, the alignment of the film as it's running over rolls, et cetera. That data we already collect. So it's really tying those three pieces together to create an end to end like situation for battery as we've done for our IC manufacturing. And you can think of the image as kind of like your inline inspection data collection point. What we're focusing on there [indiscernible] software.

Right, no. Got it. Absolutely makes sense. Just back to DFI, have you kept one inhouse for demos? And are you building a couple more? What's the pipeline of assembly look like?

We look to have a few machines inhouse, even after this to be able to do demos, incremental software and hardware development. We have placed some purchase orders for long lead items on further machines. And we've been doing that over the first half of the year, as we've increased our confidence. You probably saw that in our capital spending in the first half of the year. We do feel like the industry has wandered away [indiscernible] as long as I've been in the industry. And it's only exacerbated by the fact that you have 3D structures now. We feel like it's a very useful capability if we can prove you can do this in very high feature count. You can be measuring 20 billion on an inspection time. You can see very small deviations [indiscernible].

[Operator Instructions]. Our next question will come from the line of Andrew Wiener from Samjo Capital.

I just wanted to follow-up on DFI. And you did say the third tool that you're hoping to ship by year-end is a new DFI customer, correct?

Correct.

And is that for an advanced logic application as well?

Correct.

Maybe if we look out into 2024, how you're thinking about – you just expressed some increased optimism around DFI. We've obviously been waiting for a while for this to become a real commercial opportunity. So, you now have two with your lead customer. You're adding a second customer. Sounds like, as you said to Gus, you're ordering some more advanced parts. How do you look at the growth opportunity as you get out into next year? What has to happen? And is it around penetrating these two lead customers deeper? Or are there other, if you want to call them, pilots where you're running wafers inhouse, where you think there's an opportunity to ship new tools to additional customers in 2024?

Great way to frame the question. Let me see if I can give you an as well organized answer. So, first, you know because I've said this many times, to me the single most important thing was to get a second tool into the lead customer first. Because every company on the leading edge will always try something to see if it does something. When you can get two, that means you've filled up the capacity of one, and now they need to have a second. And so, this was a really super big milestone for us that – I think it's the first important thing that emboldened us. Along the way, we started doing some pilots, as you referred to them, with customers sending us wafers onsite. We did that only with one other customer at the time. Again, because we felt like 80% of our effort should be establishing that the lead customer would need two. There's no point in going and spreading yourself thinly across many companies, just to keep on doing the same thing. We've got to prove that we have a winning solution. We feel much more emboldened after the lead customer digesting a second. Now, we're in a position that we feel like, okay, we can ship a third machine to that new customer. What that lets us also demonstrate is the uptimes and utilizations have been really remarkable for an e-beam tool on the first machines. I think people were quite surprised by how strong they were. We want to be able to demonstrate that that really does work broadly across a couple of customers, while also increasing the aperture of taking on pilots from other customers and potentially other parts of the chip market. So we haven't really done much with memory yet. We will go back and look at that. There are additional opportunities that we see. So, in 2024, we would like to start seeing more success at the first customer. That's the most important thing. Secondly, starting to drive revenue at that second customer and getting more growth there. And then increasing the aperture on onsite pilots on our site – I'm sorry, in our lab pilots for additional customers. Some of that will expand the aperture onsite of leading edge logic.

Maybe then switching gears for a second. You called out silicon carbide as an area of strength. I think, historically, you've talked about it as a significant opportunity longer term, but given the volumes relative to other parts of the semiconductor industry despite the excitement around it, it was not material, are you starting to see that become material? And we've seen obviously some very large commitments made to sort of traditional Exensio PDF customers who provide silicon carbide solutions. Would you expect that to grow materially as you look out into 2024 and 2025? And how would you think about that trajectory relative to sort of the trajectory of silicon carbide industry as a whole?

A couple of data points there for you, Andrew. So, when I was meeting with some of our early cloud site customers and reviewing with them cloud statistics – I do occasionally drop in on some of our customer meetings – the customers have been telling me that they are – a lot of the driver for the existing cloud sites has been their silicon carbide business. So we are seeing that. We started seeing customers in the process control side start to deploy for these new silicon carbide in other IGBT-like technologies for our capability in this first half of the year, Q2 in particular. So, we do anticipate silicon carbide being a driver for our business as we look out into 2024 and 2025. And part of that is because a number of the companies that are in the silicon carbide business happened to be historic PDF customers. So, they're already natural – it's natural for them to extend their Exensio deployment into these new facilities as they stand them up and start tooling them up. We do anticipate that. And we do also anticipate, this gives us an opportunity to go to companies that have not historically been a PDF customer who are making a big investment in silicon carbide because, in comparison to IGBTs and other high voltage transistors, the manufacturing challenges are very substantial, as have been reported in the press. And so, using an analytics platform all the way down to the equipment control and equipment connectivity is more important for these customers than it has been for the conventional high voltage transistor production. It is a much, much more complex production as it's now becoming realized. We expect there to be four meaningful processing elements to these working flows, as opposed to really just wafer flow package and test. You have to grow the ingot, you need to use something like the Soitec Smart Cut technology, you have the front end processing, and then you have packaging and system packaging. So end-to-end analytics, we believe, will be important for that customer base. And so, yeah, we do anticipate that being a growing part of our business in 2024 and 2025.

I don't want to trump maybe something you're going to present at the Analyst Day, but if I think about like a dollar of investment in the ground or a dollar invested in capacity expansion across, let's say, like advance logic, silicon carbide and then maybe either like mixed signal or mature nodes, how do you think about what the revenue opportunity for PDF is across the different buckets? And if, one, that's even quantifiable, and, two, sort of maybe even on a relative basis?

Andrew, we've been trying to understand that a little bit ourselves. It's hard because some of our customers are multi end market customers. So, for example, we have IDM customers, they make silicon carbide transistors, they also do embedded controllers, they do mixed signal analog parts, they do a lot of different things. And so, for some of the enterprise class customers, sometimes it's hard to peel that out unless we know what sites they're logging in from specific sites and which sites those are, data is getting loaded from certain sites, so we can kind of get some understanding of kind of how they're using the system and what parts of their business are driving growth. So we don't have a really easy way to quantify it just yet. What I can tell you is, for Exensio, we've always found there's as much business on the trailing edge and high voltage and silicon carbide as there is on the leading edge. In other words, a company like an Onsemi or ADI or an ST, they drive as much analytics as our leading edge customers drive. And that's in part because the number of products tend to be a lot more. So while a leading edge company may have 100 or a couple of hundred different products, or maybe up to 1,000, our customers on the trailing edge may have 70,000 products they need to keep track of in Exensio. So, the characteristics are different. We haven't figured out a way to kind of model it the way you're describing. I find it intriguing the way you've broken it out. But it's not what we've done so far. We've looked at kind of just wallet share in customers and say, okay, how much are they spending on manufacturing, what should the analytics spend beyond that, and frankly, it's still tens of percents, is how we've modelled it. That may be a little conservative.

Maybe last, I just wanted to follow-up on Christian's question and put like a little bit more of a fine point on it. You made comments about the ability to grow the analytics business. And putting IYR aside, I think you talked about even the potential to return to sort of the 30-ish percent growth rates that you had experienced over a number of years. If wafer equipment sales were to return flat to modestly growing, were you saying that the pipeline is robust enough that if you guys can execute against closing deals, that you'd see a path to returning to that sort of 30-ish percent growth in analytics over sort of a couple of year period?

I think you're maybe stealing the thunder for what we'll talk about on the user conference. And we're still, I think, working that through. But, in general, what we feel is the following, Andrew. We do believe there is a very significant universe for our customer base. That is a big opportunity for Exensio and our overall platform for analytics. As I said to Christian, if we can get a return of equipment sales, growing a little bit, at least not decreasing, like it has started to grow a little bit more meaningfully, getting back to the levels that were before, and having the IVR return to the volumes they were at before, we do believe we can achieve over the 20% growth. To get back to the 30% growth, we're going to go back and do – sharpen our pencils and say, what needs to return in the overall environment? Because as I said in my prepared remarks, we did see a couple of things slide out from Q2, one of which already – a couple which already closed in the early part of Q3. We want to see that that robustness and bookings return on the overall analytics business. As I said, there's an awful lot of activity out there. So we do believe that demand is there. We'll just see how customer confidence happens in the second half of the year.

And my last question, both you and Adnan called out as one of the sort of constituents where you're seeing strong interest or activity is obviously around government, either subsidized or sponsored investment in new semiconductor capacity and sort of enhanced supply chains, are you seeing direct opportunities there? Or is that all going to be indirect? And has that shown up in pipeline yet? Or is that more still to come?

That's starting to show up in pipeline, Andrew. It's mostly through the companies the governments are funding, some of which are existing companies, some of which are new companies. But you can also imagine, given PDF's pretty unique perspective in the industry, our footprint across the world and our experience with everybody else that's tried to get into this industry over the last decade or so – the governments themselves, from time to time, do reach out to us. And we do have dialogues with them to understand what they should be thoughtful about. We have not really put in our pipeline business directly with them, and we really don't focus on it. But it does help us understand what their intention and motivations are. And since in effect, in many of these cases, they are the owners, right? They're the ones writing the checks. It's good to understand what ownership wants. So we do see those as an opportunity for us to be educated on their vision and their drive. And then for them to understand our perspective on what it's taken in the past to be effective at introducing a – creating a new entrant in semiconductors. It's not an easy activity at all.