John Kibarian: Yes, I mean, off the top of my head, I don’t know that I could give a really great answer, Gus. I mean, as always is the case for us, on a dollar value basis, we live by the 80-20 rule, so, you know, there’s – or maybe even 90/10. There are probably 10% of the deals that represent a sizable 80-plus percent of the dollar value because these enterprise things tend to be quite large. There are some out there that we’re working on right now that are, as I said, substantially bigger than those first two, that are going to drive a meaningful piece of our bookings this year. And then there’s a number that are similar size to those, a handful of them, I don’t know off the top of my head, but there’s a fair number. Collectively overall, as I said in my prepared remarks, we do have a lot of confidence about the bookings this year.
It’s a – the level of activity of customers is quite meaningful given those three drivers that I spoke about, electrification of the energy economy, that’s really just driving high-voltage silicon battery technologies, the advanced process nodes really geared towards AI. That includes advanced packaging. So a ton of activity we have going on for the ML Ops piece there. And then all the DFI and leading-edge capabilities all around just advanced semiconductors driven by AI. And then the geographic diversification, customers really wanting to get to these more enterprise-wide control schemes because of the nature of their manufacturing. Two in Q4 are kind of somewhat representative of that, but we see a number of those. So overall, it’s probably, I don’t know, in the 10 range, but there’s probably two or three that drive a sizable fraction of that.
And those 10 probably are more than 80% of the bookings value for the year.
Gus Richard: Got it. Okay, that was very helpful. Thank you. And then just flipping to DFI, it sounds like you’ve got pretty decent visibility through this year in terms of what you’re building. Over the last 90 days have you – I know there’s some long-lead time items for DFI. Have you had to sort of go back to your vendors and order more material for potential deliveries in 2025?
John Kibarian: That’s a great point, Gus. Yes. If you look at our expense spend on capital, you’ll see it step up in 2024 versus 2023. And 2023 had a modest increase versus 2022, if my number memory is correct. And that step up is because of ordering for things, machines we expect to ship in 2025, more than what we expect to ship in 2024 per se. It does impact 2024 a little bit. And yes, we’ve been going back to our vendors to try to tighten up availability and delivery times. We would like to get, and we feel like we’re getting close to the point where we’re really going to be looking at how we can pull those in for our – because of the customer interest. So yes, you’ll see the spend go up this year. It’s anticipation of 2025.
We’re setting ourselves up to be able to ship significantly more in 2025 than what we’re shipping in 2024. And we are going back and even discussing what the vendors want. If we needed to pull in even more for that, how would we be able to, more than what we’re already planning, how would we be able to effect that? That would probably not affect our capital spend in 2024 very much, but it could affect our capital spend in the first part of 2025 and then shipments that would impact second half of 2025. We think we have the first half of 2025 mostly okay.
Gus Richard: Got it, got it. That’s super helpful. And then you mentioned in your prepared marks that you expected to add additional customers this year. Any color? Is that memory? Advanced logic? Is there any – is it one or two? Any color there is helpful.
John Kibarian: Yes. So we’ve – we continue to see opportunities in advanced logic, and we do expect advanced logic incrementally to contribute this year above what it contributed last year. We’ve also – we’ve been in the industry for quite a while. As you know, PDF has worked on yield ramps all around the world on – with virtually everybody. And I was chatting with an executive at one of the companies that’s just getting into what would be more advanced than a trailing edge, but nowhere near the two nanometers. But as you think about everything from 12 nanometer to 28, and we’ve started to see interest in those areas where people said, hey, if we had this, it could really accelerate our bringing up of new products. And one of the guys joked with me, hey John, we worked on yield ramps with you at my past company in these nodes, and this would have really helped.
So we think that the aperture is broader than just the leading edge. As you know, we were very focused on just what can you do on the leading edge in some way because it’s a talisman about where are you going for the industry. But now we’re going to double back and look at some of those other opportunities. And we’re also starting to get some early looks in the memory space where we’re starting to have some early dialogue. So part of our goal this year is to make sure we’ve got enough capacity in our own lab to be able to do demos and evals for customers, much like we did for that Asian fab in 2023 that enabled us to ship at the end of the year. We’d like to be able to do that this year for some additional customers in the trailing edge, quasi-trailing edge, in memory, while we continue to penetrate on the leading edge logic.
Gus Richard: Okay, and I promise this is my last question. When you talk to these not leading edge but more advanced geometry logic guys, DFI would be helpful for them because it would accelerate their learning over time the parametric tester you used to use, or is it a cost savings because you wouldn’t scrap as many wafers as you go through the development process?
