PDF Solutions Inc

NASDAQ:PDFS

Earnings Call Transcript 2024-Q1

[Operator]

Good day, everyone, and welcome to the PDF Solutions, Inc. conference call to discuss its financial results for the first quarter conference call ending Sunday, March 31, 2024. [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 16 through 36 of PDF's annual report on Form 10-K for the fiscal year ended December 31, 2023, 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 would 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.

[John Kibarian, executive]

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 first quarter was a strong start to our year. Business activity continues to be robust as our products and solutions are well aligned with the larger trends driving the IC industry. Before Adnan discusses the financials in detail, I have some comments to make about the events in the first quarter and our perceptions of the market in the second quarter and for the remainder of the year.

Building on a strong Q4, Q1 bookings were again up over the previous quarter, which continues to build backlog. When we look at the nature of the business in Q1, we see that bookings and presales activities were driven by the alignment between larger macro trends in the industry and our product development investments. I will describe 3 of these.

First, AI is driving strong demand for advanced logic processes that leverage 3D transistors like nanosheets and 3D interconnects such as backside power. In the quarter, our largest booking was with a new customer deploying our systems, including DFI, CVs and Exensio, to accelerate the development and manufacturing of advanced 2-nanometer logic.

Second, customers are driving digital transformations to achieve more efficient operations. Our systems from Exensio Cloud, Exensio Test Operations and Sapience Manufacturing Hub are gaining increased attention from our customers for this workload. Bookings in the quarter include cloud customer expanding Exensio usage, and presales activities include a number of customers evaluating our Sapience Manufacturing Hub as part of the SAP S/4HANA deployment.

Third, customers are accelerating their use of AI and machine learning to achieve their yield and operational targets. In Q4, we announced our MLOps system that enables engineers to build AI models and publish them in manufacturing, both at the factories they own as well as their partners' facilities via our DEX network. Customers report to us that they are mastering the ability to train AI models, but the challenges are to deploy them in the field, monitor their performance and act on the results.

Our MLOps system was designed for these challenges. It leverages our Exensio Cloud, DEX and test infrastructure. In the first quarter, we started running MLOps pilots with our lead customers. Overall, we believe the strong engagement we have with customers is due to our investments in our eProbe DFI systems, our SMH digital platform for manufacturing and our MLOps for testing of complex system and package products.

I know many of you are curious about the progress with our DFI and eProbe. Utilization of the tools in the field is extremely high, we believe because it's unique ability in seeing 3D yield issues. In February, we said we anticipated a third customer starting to deploy DFI and we would ship a third tool to our lead customer. The contract we signed this past quarter brings us a new customer for the eProbe. We also remain on track to ship our third tool to our lead customer this quarter. Given this acceleration of our 2024 goals for DFI, we are now looking to accelerate our manufacturing as we believe demand is quite strong for this solution.

Our results have occurred while the industry is at an unusual time. Government investments have resulted in outside capital spending in parts of the world. Demand for some chips has been very weak, while silicon for AI has been strong. That said, we believe the industry is starting to see a general improvement due to the metrics we track about its health.

As we look to our run-time license sales, which track our equipment customer shipments, we see improvements when compared to Q4 of last year. This, as well as our strong bookings in Q4 of last year and Q1 of this year, suggests customers expect stronger business as they go through the year. We are pleased with the business activity in the quarter as it demonstrates the strength of our strategy and investments.

Consistent with our comments in the February call, we expect the first half of the year to be flat year-over-year with the second half of the year getting back to our growth target of 20%. I want to thank all the PDF employees and contractors for their efforts during the call -- during the first quarter, sorry.

Now I'd like to turn the call over to Adnan who will review the finances and provide his perspective on our results.

[Adnan Raza, executive]

Thank you, John. Good afternoon, everyone. Good to speak with you again today, and I hope all of you and your families are well.

We are pleased to review the financial results for the first quarter of 2024. As mentioned, our earnings release and the management report are posted in the Investor Relations section of our website. Our Form 10-Q was also filed with the SEC today.

Please note that all of the financial results we discuss in today's call are on a non-GAAP basis, and a reconciliation to GAAP financials is provided in the materials on our website.

We are excited about the booking momentum during the quarter, especially the meaningful, multiyear, leading-edge booking with a new customer. Our backlog ending the first quarter was $262 million or approximately $32 million higher compared to our prior quarter ending backlog of $230 million.

Total revenues for the first quarter were $41.3 million, up slightly versus the prior year and up slightly on a sequential basis as well. Analytics revenue came in at $38.5 million, an increase of 6% year-over-year. On a year-over-year basis for the quarter, our IYR business was down driven by lower fixed fee and gain share. However, the growth in analytics allowed us to be slightly up for total revenue versus prior year period.

Like John said, for our Exensio products, we continue to engage with customers on opportunities we are seeing for the digital transformation initiatives and deploying the Exensio platform across their manufacturing operations.

For our leading-edge solutions, we are emboldened by the booking with a new customer, which encompasses our broader offerings, including CV infrastructure, DFI and Exensio software. We're starting to see opportunities for further expansion over the coming years with continued investment in the DFI system.

For our symmetrics products, we saw growth in run-time licenses versus the prior quarter and remain cautiously optimistic. Taken as a whole, we believe our portfolio of product offerings provides us the opportunities for strategic engagement with customers for our analytics solutions.

IYR revenue came in at $2.8 million for the quarter and was down compared to $4.4 million of prior year period, primarily driven by lower time spent on fixed-fee projects and lower gain share. We remain optimistic about the IYR business in the longer term as eventual customer product shipment volumes increase.

Our gross margin for the first quarter came in at 72% versus 75% for Q1 last year and flat versus 72% of Q4. On a year-over-year basis, our cost of sales was driven by the lease accounting treatment for the DFI hardware, increased personnel costs and increases in cloud costs. Our operating margin for the first quarter came in at 12% versus 19% for the year-ago same period and 15% for the prior sequential quarter.

On a year-over-year basis, we were able to better manage and reduce our R&D expenses, while our SG&A expenses increased as we utilize our technical resources and added sales and marketing headcount to support the presales and sales activities with our customers. Net income for the quarter totaled $5.7 million or $0.15 per share, both essentially similar to Q4, however, lower on a year-over-year basis.

Turning to the balance sheet. We ended the quarter with cash, cash equivalents and short-term investments of $122 million compared to $136 million at the end of the prior quarter, with the change primarily driven by $7 million of share buybacks completed during the first quarter, annual bonus payout for calendar year 2023 and CapEx expenses to support the build-out of our DFI systems that John talked about. We have also adopted a new, larger $40 million share buyback program.

Overall, we are pleased with the new leading-edge booking during the quarter, growth in our backlog and are excited by the opportunities we are seeing in our pipeline. As we look to the rest of the year, we remain committed to our prior guidance of flat revenues for the first half of this year, with revenue growth returning to our 20% long-term target for the second half of the year both compared to the respective prior year periods.

With that, let me turn the call over to the operator for Q&A.

[Operator]

[Operator Instructions] And our first question comes from the line of Blair Abernethy with Rosenblatt Securities.

[Blair Abernethy, analyst]

John, just wondering if you can expand a little bit on the DFI market traction you're seeing there. This new customer, I assume this is the second customer. It looks like one machine now. What sort of is the size and scope of that opportunity with that customer over the next few years?

And secondly, are there other pilots ongoing now? Or sort of what do you see in terms of activity with other potential customers for DFI?

[John Kibarian, executive]

Sure, Blair. So as you know, we talked about our lead customer, right, who had 2 machines, and now we will ship a third. We started in Q4 a manufacturing evaluation with the customer, a second customer. That was a [ nonpaying ] evaluation at this point. That is ongoing quite well. We feel very good about that one as well. And then this past quarter, we signed a contract with a third customer.

And for all of them, I don't speak about any one specific one, what we believe is the following: the yield loss for these advanced technologies is very much due to the conduction between layers or underneath the surface of the wafer. Conventional inspection is to shine a light and look at the surface of the wafer for a defect. So you obviously can't see something that is going down below the surface of the wafer, like a [ viad ] or a contact that's open or short in the gate all-around structure.

So this voltage contrast or using an electrical test method has always been desirous, but the issue has been you could never see enough on the wafer. The throughput was so slow. You can't see what's going on. And you don't know enough about the design to know what matters and what doesn't matter, what is it could be a defect and what's not.

What's unique about what we've done with DFI is the software first knows what's in the design. So it knows where to look and it has a simulation model of what to expect. It drives the machine to look there. It's measuring billions, if I thought we'll ship later on this quarter over 10 billion features in an hour. So it's seeing huge statistics.

And then it's an analytics problem on the other side of that, okay, [ 20 ] sales. That's still less than -- well, let's say, it's less than a failure in a billion and infractions of a failure in a billion. What's unique about that from a design perspective, that whole end-to-end capability is what is embedded in the DFI with the eProbe.

And I think the question for customers and what they've told me is, hey, if this can really do this, we would use these in production, right? And we hear that from multiple customers. And then the market would be significant for this, right? Right now, the market for e-beam inspection in and of itself is hundreds of millions of dollars, but it's been primarily used in the early stages of yield ramp because once you improve the process to a certain level, typically, the sensitivity of the machine isn't good enough.

What we've been able to demonstrate with the eProbe and the road map for how much it's evolved is that it's able to see quite a lot of silicon, and hence, it can see things as the process is maturing. And that's really the question mark, how far can we push this? How much can we prove you can use this in manufacturing? If that was the case, it would look a lot more like the optical inspection business than the e-beam inspection business today. But that's an unproven thing.

And then lastly, to answer your question around pilots, we do have a number of other requests from customers on pilot wafers and demonstration capabilities outside of the 3 that we've discussed. And we are slowly opening up the aperture to take on a little bit more of this work. But obviously, we really want to succeed at everywhere we go.

[Blair Abernethy, analyst]

That's great. John, I really appreciate it. And just second question, just over on the battery manufacturing side, the Lantern technology acquisition you made last year, any update on that in the market?

[John Kibarian, executive]

Yes. So we've been engaged with a couple of lead customers on some pilots. Those are ongoing. They look promising. We think we understand what our common issues. And again, it's one of those things, can you -- can the engineering achieve what the customer is looking for in terms of analytics? And the biggest thing there is around AI for extracting critical issues off their in-line data.

[Operator]

And one moment for our next question, and it comes from the line of Gus Richard with Northland.

[Auguste Richard, analyst]

Congratulations on the new customer. That was a rapid addition. I was just wondering, the relative size of that contract relative to the size of the one you booked a couple of years ago with the contract that included DFI and other services, could you give sort of an order of magnitude? Is it [ capital sized ] or what is it comparable?

[John Kibarian, executive]

Comparable. Yes.

[Auguste Richard, analyst]

Okay. Wow. And then can you talk a little bit about how quickly you can ramp DFI manufacturing to meet increased demand? I think the plan is for 3 or 4 systems this year, and how rapidly can you ramp that into next year?

[John Kibarian, executive]

Sure. Yes, our original plan was really to effectively commit to 2 more machines this year, and now we've actually committed those. So that's -- we were planning on building 3 or 4 machines, so they were to build -- completions would be into next year. We're looking to do now, as I said in my prepared remarks, Gus, is accelerate that, pulling them as much as we can. And we have, of course, preordered some long lead items and been somewhat thoughtful about in anticipation of success and then accelerate the build beyond that number.

It's too early for us to really say what we would be able to commit to because we're still working with our suppliers. But we do want to be able to meet customer demand if it materializes the way that we hope it will. And we still have a lot of work to do there, too.

So I think at this point, it's kind of a wait and see. Gus, we can't promise that we're going to build more than we originally committed. We think we have some flexibility there. We hope to at least at the end of this year or early next year to squeeze it a little bit more. And then we expect to have a more steeper manufacturing ramp in 2025. That, we have more flexibility on, obviously, because there's plenty of time there.

[Auguste Richard, analyst]

Right. And just are you, at this point, contemplating outsourcing the manufacturing of the equipment? And...

[John Kibarian, executive]

It is outsourced today. I mean we don't -- up until now, we've done final QA and assembly here in the States and we review it, but all of our production is offshore. And we have the ability to just if we want to not ever land them back here and complete final QA at the producers.

[Auguste Richard, analyst]

Yes, okay. So then what's -- is the long lead time items the limiter to a more rapid ramp in like the next year?

[John Kibarian, executive]

Correct. Correct.

[Auguste Richard, analyst]

Okay. And...

[John Kibarian, executive]

Obviously, the technology in this thing is not simplistic and it's not off the shelf, but it's pretty bespoke stuff, right?

[Auguste Richard, analyst]

So refresh my memory as to typical lead times for your longest lead time item for the tool.

[John Kibarian, executive]

Well, I think what we've been able to do this first quarter, I really applaud the team, is take items that did not have a deterministic lead time, right? In other words, they were -- the suppliers would build them, then we would test them and then we would send them back because they didn't meet spec or you would start more material than you needed just to get out a small amount with relatively poor yields.

And what do you know, a company that's kind of predicated on manufacturability, we went back and worked on a lot of that with our suppliers and got the manufacturing yields up to get the times to be deterministic. So you could put an order in and in fixed number of months get a subsystem out that actually meets requirements.

So I think we've been working through that in the first quarter. I think I feel really great about the results the team achieved on that. And we're really rolling that up to say, okay, now where can we take slack out of our overall build schedule? Because our build schedule had a lot of slack time, and it still accounts for some components that were basically undetermined. You could order and you didn't know when you're going to get them.

[Auguste Richard, analyst]

Got it. Got it, I understand. And then just switching to the software side of the world, the Exensio, I know you're working on some large enterprise deals and you had some potential for some rather 8-figure deals, I believe. And I was just wondering, are customers beginning to come out of the fog and start getting closer to committing to orders? And sort of where is that funnel look -- what does that funnel look like today?

[John Kibarian, executive]

Yes. So we did have some, I would say, 7-figures contracts last quarter. This quarter, I do see -- this quarter or the next quarter, it's always hard to pick timing whether they're going to be Q2 or Q3. We do have a number of larger opportunities out there. And they are -- I chose my words around Exensio Cloud, Exensio Test Operations and the Sapience Manufacturing Hub, a lot around customers' digital transformations. This seems to be one critical area for the customer. So the SAP connection is an important one for a lot of this work and a lot of these contracts.

And we're part of an overall larger engagement in those cases. So it's a little bit trickier for us to have clear visibility. But we feel pretty good over the next couple of quarters to see what will be larger contracts, what's the total size of them and how they break up. I think, of course, it's early, but we do feel engagement quite strong with the comps.

[Operator]

One moment for our next question, please, and it's from the line of William Jellison with D.A. Davidson.

[Will Jellison, analyst]

For the first one, I wanted to start out by asking you about the core Exensio business. What sort of trends are you seeing in customer behavior with respect to adopting incremental Exensio modules or their usage trends of the software overall? I'm curious what you're seeing there.

[John Kibarian, executive]

Sure. Yes, so I think there's basically 3 places where we see, I'd say, common assets: so the cloud customers evaluating and looking at test; cloud customers looking at the guided analytics, AI module, right, so automate -- more automation in the way they're doing analytics; and then I would say our larger customers really looking at the Exensio and Sapience connection between SAP and their factories.

And those are -- like kind of the way I answered Gus' question in my prepared remarks, Will, we're all around those, basically, those 3 capabilities. That's what we see. We look at that as, a, customers trying to kind of change their digital platform, digital transformation, some of them call that, that often starts with an S/4HANA deployment and then how they get kind of real-time feeds from the manufacturing tends to have an impact on test because -- and the DEX networks and things like that because a lot of their supply chain isn't in their control.

So the test operations is a piece of that. And of course, it starts with the cornerstone of the cloud. So like I said, after cloud, it ends up being guided analytics, test operations and the SMH-SAP connection. And those are all ongoing now. We've had -- we had stuff closed in the quarter in Q1 related to some of those. And we have stuff in the pipe in Q2 and Q3 on those activities as well.

[Will Jellison, analyst]

Okay, great. And then as a follow-up, to dive a little bit deeper into that, with respect to the ModelOps pilots currently going on, can you talk about how you expect customers to ultimately uptake ModelOps and how it affects PDF's monetization overall of that Exensio business?

[John Kibarian, executive]

Yes, that's a good question. We're learning as we go, Will. But what we think so far is the following: it does drive our TestOps, so as you noted in my prepared remarks, the ModelOps is primarily around testing for advanced packaging. So multiple test insertion points, data feed forward, predictive bidding, all the things around being more able to make sure the chip that you put in the package is going to really be there, be at the performance levels you need and then make sure that the testing is efficient because they've added so many test insertion points. They are trying to be as efficient as possible with their tests.

So this has been where a lot of the engagement on the pilots, almost all the pilots are in that area right now, all related to advanced, typically, advanced packaging and chiplets. That will drive ModelOps deployment, which basically is tied to the number of machines that they deploy ModelOps open. So it scales with the capacity of the deployment. And it also has kind of a follow-on effect because then you need more cloud, you need more analytics on the other side and more of the ML training capability in the cloud as well.

We think that we'll close first is the stuff around the moving data around MLOps as deployed. That's why in my prepared remarks, I talked about the challenges of being able to monitor, maintain and act on the models. And over time, as they get more and more sophisticated, we think it will also have an impact on our general cloud business as they drive more compute, roll out more models, manage larger data sets and do more complex things. I would say, overall, the industry is really relatively simplistic in what they're doing in AI models today.

[Operator]

[Operator Instructions] One moment for our next question, and it comes from Christian Schwab with Craig-Hallum Capital.

[Christian Schwab, analyst]

So if we use -- if you take the DFI eventually and move it in as a production tool, what would you charge for that versus the way that you're kind of charging customers to use it as what I guess I would call more of a research and development tool?

[John Kibarian, executive]

Yes. So that's a great question, Christian. I think we're still working and in active dialogue with the customers on this right now. The -- today, we charge the whole thing as a bundled subscription. And the customer gets quite a deal because we actually manage the uptime on the machine, we manage the spare parts, we provide the entire software stack for the application layer.

We said we're open to selling the machine piece itself, and that's on a capital purchase basis. And they would still be potentially on a subscription on all the analytics and systems that drive the machine and drive the data off and particularly for the product-level inspection. So there are ways where we could make it less capital-intensive for us and maybe more conventional to the customers. We've had dialogues with customers around that point. I don't think we have anything that is set in stone at this point.

And I think as we've gone through that, we've learned a lot about just our subscription business and how we've priced that. I think we did that in a way that made it very, very, very easy for the customer to get into the machines and see the value that's there, right? While they're relatively meaningful subscriptions, when you consider the total value, I think it's quite a good deal.

[Christian Schwab, analyst]

And should we assume that if you sold them as a piece of capital equipment that an ASP would be $4 million to $5 million, is that still fair?

[John Kibarian, executive]

No, I think that would probably be much lower than where you'd expect them to be on a capital basis.

[Christian Schwab, analyst]

Okay. Okay. And then on a capital -- I guess my last question is, if it was used in production, do you have an idea yet of how many tools would be needed based on wafer starts per month with every 10,000, 50,000, 100,000, whatever, have you done the work to know how many would be needed?

[John Kibarian, executive]

I think we don't really know that answer yet, Christian. What I can tell you is the following, right, when we first started doing this many years ago, right, even before -- for the last couple of years, all of the customers that I talk to, they said, you need e-beam a lot in the first 1 or 2 years of development. Then once you get to a certain defect density, you can't see anything anymore.

So everyone was very comfortable with the subscription because they didn't think you'd see it. You would see defects are very long. But the reality is we've proven with the capability that, especially as we've done further revs of the machine, right, so the first one we shipped, we called it 250, then we shipped last year the 350, this year, we'll ship the 450. These are getting incrementally faster and faster and faster really due to the software layer.

And we know overall what the market is in terms of defect densities, and we think we're very unique in being able to see these 3D problems even as customers are getting to pretty good yields. So that's why I think customers' early idea about how long you would need to be in the tool, and what's turning out to be the case is, I think, surprising the industry overall. And that was our original thesis. So it's nice to see it come true, but we didn't know when would we start it.

[Operator]

One moment for our next question, please, and it's from Gus Richard with Northland.

[Auguste Richard, analyst]

Just one more question. And I kind of want to be clear, there's 3 buckets for DFI: process bring-up, product bring-up and then there's in-fab in production. And the conversations you're having with at least one or more of your customers, does that include the third bucket where you're talking to them about production tools?

[John Kibarian, executive]

So I think people are asking about that, Gus. The people have used e-beam historically for test vehicles and process bring-up, SRAMs and very simple structures in the product where they know what's there where the "nuisance rate" is low. I think what customers have been surprised, and we see this already in the way we've engaged with the customers across -- at least a couple of them so far, the eProbe is very, very good at looking at product.

So now you can bring up each new chip, each new design and see via open issues, complex issues because it is very targeted on where it lands the beam. You can even sequence the landing of the beam. So turn on areas and look in other areas is very, very sophisticated in the way it can work. And so I think customers now themselves are getting on that. It can be very powerful for product bring-up.

They are asking us about the third application as well. And now, of course, because we've deployed this on very advanced nodes that are just getting to production now, now we're starting to begin those engagements. That pilot that we announced in Q4 of last year was really around the manufacturing question, okay, can you really run one of these things in manufacturing? That was their question.

And we've seen, of course, across the fleet, the uptime data and the consistency data, and we feel pretty good that, again, e-beam tools tend to be like Formula 1 cars, right? And customers will question, okay, can you really run one of these things? And we've seen our uptime data, and we feel pretty good about it.

So we'll have to go and answer that third question. People are asking it now. I think they're pretty comfortable now with the second question. You can do product inspection, even have ultimately mature yields as you bring in new products, and I think that's a very new application for e-beam.

[Operator]

One moment for our next question, and it's coming from the line of William Jellison with D.A. Davidson.

[Will Jellison, analyst]

Adnan, I wanted to ask you with respect to gross margin during the quarter, my impression was that much of the year-over-year drag was related to that lease accounting for the DFI that you mentioned. And so the question that springs from that is, if DFI accelerates and becomes a bigger part of PDF growth into the future, does that change at all the way you think about the long-term greater-than-75% target?

[Adnan Raza, executive]

Yes. Thanks very much for asking that question. So look, now with multiple machines shipped into the market, we are actually getting quite better. And the team has been doing a great job about managing the costs over the longer term, especially as we think of some of these advanced -- these early orders that we are starting to look at that we are placing with our vendors. So feeling pretty good about the machine side of it as well, being able to not hurt our 75% gross margin targets that we have said.

Well, yes, the last 2 quarters, some of the margin was lower compared to our gross margin targets. But as the revenue ramps through the rest of the second half of the year, we certainly expect to be getting back to our target gross margin model. Particularly by the time we are ending the year, I think we're starting to think that the 75% gross margin and the 20% operating margin targets that we had set for the year seem like they would be achievable ending the year.

[Operator]

One moment for our next question, please, and it's from Andrew Wiener with Samjo Management.

[Andrew Wiener, analyst]

I wanted to follow up on Christian and Gus' questions, I guess. The first is with respect to the potential, and I realize you're still evaluating it as a model, but the idea of selling the tool and then charging a license for the analytics and the design, et cetera, and the software, it sounds like you believe, and I don't want to give away anything from a pricing strategy perspective, but you believe relative to sort of other e-beam tools or inspection tools and then sort of the software that's required to manage them that perhaps our initial subscriptions were underpricing the overall value, in part, because we need to prove that it works and drive adoption. Is that a fair sort of interpretation of your earlier comments?

[John Kibarian, executive]

Yes, I mean -- and also the machine got a lot more capable. So from that first generation we shipped in 2020 to today, it's about 16x faster. Because of the software, because of the capability of getting smarter about the design, there's a lot on the software side and the control systems side that make that possible. There is changes to the machine, too, to basically take advantage of what the software does.

But that software -- improvements to the software, improvements to the machine to take advantage of the software have made quite a different experience than we think it continues to be. So yes, then you can go and look at where it is, performance-wise, competitively in the marketplace, and we feel pretty good there.

[Andrew Wiener, analyst]

Okay. And then second, I realize we're not ready to put a fine point yet on what the number of units or sort of market sizing. But if I take your earlier comments of e-beam being a market in the hundreds of millions of dollars and then I take your response to Gus' question about that we sort of prove it -- or we're in the process of proving that you can not only use it to bring up a process, but you could actually use it to bring up sort of each individual design or each individual product.

It's sort of -- it somewhat suggests that you believe at minimum, the market for eProbe or DFI is larger than the current e-beam market. And then the question obviously would be how much larger, and that would be somewhat dependent upon sort of the proof of getting into manufacturing. Is that a fair comment?

[John Kibarian, executive]

It's fair, Andrew, yes. I mean I think, look, the conventional uses of e-beam, there's more than just voltage contrast. They use it for fine feature inspection, right? And the number I was quoting was the inspection market, not the review market and not the CD market or the overlay markets, all the other ways that people use e-beam tools. They're just the e-beam inspection market.

And the -- we focus primarily so far on the logic producers, right? And a lot of e-beam goes into memory, right? So you've got to go and slice that all up a little bit and then say, okay, can you make the market for logic voltage contrast bigger if you can do product and not just early vehicles? Yes, I think that's true.

How much bigger that is in the regular, the whole market or is it being the whole market? Well, that depends a little bit more about how relevant to what we're doing to the memory customers, et cetera. So that's why I like some wiggle room there, Andrew, because I think we're still really digesting all that.

I think we are feeling pretty comfortable that we -- this will be an important part of PDF's business, and we think it can be an important part of the subscription for PDF on a subscription basis for PDF's business.

[Andrew Wiener, analyst]

Okay. And then lastly, I just wanted to sort of make sure I understand. So we're planning to ship this quarter a third tool to our leading customer. I think on prior calls or prior public comments, you talked about shipping sort of another tool in the fall. Is that -- was that always intended for the customer who signed the current contract? And is that sort of the timing of when you'd expect to ship that tool?

[John Kibarian, executive]

Yes. So that is that customer. We had said that now we can help them in this stage of being able to use the machines in our facility as they go through this year. And then at the end of this year, potentially [ good flight ] internally next year. If they're not ready, that machine would move from our site to their site as they transition their status, right? So -- but I think what we're able to get done here was the ability to start supporting them sooner than later this year, right?

[Andrew Wiener, analyst]

Okay. And then the comments about trying to pull forward potentially availability of tools into late '24, is that mostly for pilot sort of evaluations? Or are you seeing actual sort of customer interest in installing a commercial tool in their facilities?

[John Kibarian, executive]

Customers are asking us now, Andrew, about availability of additional capacity, which we really -- I wouldn't say they've come and said, we want to buy one now. They're saying, okay, if I wanted to buy one, what would it take? And as I think we would want to be able to take advantage of those opportunities, we are looking at what we can do.

Obviously, if you've got an instantaneous supply chain, it gives you maximum flexibility, right? And anyone that's taken control fairly knows the longer the feedback loop, the more likely you go out of control, right? So it eludes us to figure out how to shorten all of that for maximum business flexibility and also because customers are asking about it.

And we think the way we thought about the supply chain should give us some advantages. I think historically, people said, oh, PDF is not a manufacturer. There's some disadvantages to that. And go ask a fabless company, it's not having a side as an advantage or a disadvantage. I think most of the fabless companies would tell you that they're quite happy with being fabless. And I think similarly here, if we think there's ways that we can be -- we can leverage the way that we operate to maximize flexibility.

[Andrew Wiener, analyst]

And I guess the last question around that is the customer is doing the manufacturing evaluation. I think you thought you might know by the end of this year sort of if it's a go or no go. Have they given any indication whether if it's a go, they would require additional tools?

[John Kibarian, executive]

Yes.

[Andrew Wiener, analyst]

And I assume if the answer is yes, they would require additional tools?

[John Kibarian, executive]

Yes, they would. Yes. No, that's the purpose of the manufacturing evaluation, right?

[Andrew Wiener, analyst]

Yes. Okay. No, that's what I thought. I just wanted to confirm.

[Operator]

[Operator Instructions] Well, at this time, there are no more questions. Ladies and gentlemen, this concludes the program. Thank you for joining us on today's call.