GLOBALFOUNDRIES Inc. (GFS) Earnings Call Transcript & Summary

Welcome back. I'm Vivek Arya. I cover semis and semi cap equipment at BofA. I'm really delighted to have the team from GlobalFoundries join us this afternoon, Dave Reeder, Chief Financial Officer; and Sukhi Nagesh, the Head of Investor Relations. And I plan to run through a few of my questions. But if you have anything you want to bring up, please feel free to raise your hand at any time. So with that, welcome, Dave, welcome, Sukhi.

Thanks for having us, Vivek. Great venue, great event. Thank you.

Thank you very much. So maybe, Dave, since GlobalFoundries just kind of came public in the last year, I think it will help to kind of give a quick overview, what really differentiates GlobalFoundries from there is this notion of trailing edge fab, but what really differentiates GlobalFoundries, and then we can talk about all the fun things about the cycle and so forth.

Sure. So let me address that a couple of ways. I think the first thing that differentiates us is that we take a real end market approach to the foundry market. So what that means, practically speaking, is that we go out, we hire engineers that work at fabless companies that work at IDMs. We bring them in-house to GF, and we put them in charge of our R&D process. And we say, all those technologies that you wish you had when you were designing those chips for those end markets, design that foundational and that specific IP here at GF, so that those companies can then win in end markets. So we take an outside-in view, a very end market approach. We've identified -- we split the market into about 150-ish individual battlegrounds. We then broke those down into about 30 that we specifically target that aggregate up into 5 end markets. And so I think that's one of the big differentiators for us is we don't talk in terms of just commodity platform performance. We don't talk about nodes. We talk about end markets and we talk about unique differentiated technologies that satisfy those end markets.

I see. So what percentage of what you do, Dave, at GlobalFoundries would you say is proprietary versus, let's say, what the Asian competitors can do?

Sure. Well, I always like to talk in terms of proof points first. So rather than talking in broad generic terms, if you say you're differentiated, well, how do you define that? Well, I think one of the key indicators of differentiation is what do you charge ASP per wafer? Well, if you took our first quarter numbers and you took our wafer revenue and just divided by our wafer shipments all on a 300-millimeter equivalent basis, you would see that our ASPs are about $2,800 per wafer, which is about $1,000 more than our closest competitor. So I think we command a premium ASP in the marketplace. I think that's one point of differentiation. Now in terms of our technology platforms, so we have differentiation on our CMOS platform. What do we mean by that? We mean embedded nonvolatile memory, and we mean low power, high-voltage. We mean bi-CMOS, BCD. So these are all technologies that when we look at an end market, we think about how can we be differentiated in that end market. How can we add unique technologies to it to help our customers win, which then manifests itself single-source design wins. About 80% of our design wins are single source. About 2/3 of our revenue is single source. And so I think those are a couple of ways that you can talk about differentiation, you can also not only talk about the differentiation but provide proof points. So you've got CMOS, and then you also have FinFET, and you also have SOI, both FD-SOI as well as RF SOI. You also have SiGe, and then finally, silicon photonics. And those are our technology platforms, all of which have varying degrees of differentiation.

Got it. So it's more specialized processes as opposed to just nanometer, right? That type of definition, which I think is very important to realize.

Absolutely.

How hard is it for your competitors to build that capability?

Well, I think there's various degrees of whether the competitors that you're talking about and then specific technologies. So let's talk about -- let's say there's competitors that are trying to get into the space. So if you're talking about competitors that are trying to become a foundry, well, then the first thing that you have to have besides a fab and some equipment is you have to have a full ecosystem. And this is an ecosystem that's built upon a stable process with specific IPs that have been proven on that process. So for example, if you want to have an automotive-grade process, that takes time. That takes years of testing, development and proof points. If you want to have nonvolatile memory on that process, that takes time. If you want to be able to satisfy power management on the same process node as RF capability, then that's a proof point that has to be tested and designed over time. And so I think if you're new to the space, it takes a long time to develop the ecosystem. If you're talking about existing competitors, then you have to go kind of a step deeper and you have to say, well, where do maybe some of the single-digit nanometer competitors that we have that also play in the greater than 10-nanometer space, where do they put their best and brightest? We put our best and brightest on satisfying 12-nanometer and above, and specifically, satisfying franchises in connectivity and power management. And so that's where we put our best and brightest. That's where we put our engineering talent. And that's where we put our engineering dollars, as well as our capacity dollars. And not everyone does that. Sukhi, anything you'd add to that?

Pretty well said.

One of those emerging competitors that you mentioned, announced the decision to buy a tower semi. And do you think that gives them enough by way of IP, customer relationships, right? And then they can bring their scale and resources and funding from EU and U.S. chips [ activator ].

I had a feeling that's where that question was headed. Look, I think it's no surprise to me that tower was available on the open market. I think that's been known for a while. I think the challenge there is that their technology roadmap essentially ended at 200-millimeter, which that technology platform, 200-millimeter, kind of ends at 90-nanometer. They had a little bit of 300-millimeter capacity, but not a lot. And so you then have to develop all of these PDKs. And the PDKs, which is what a customer uses to design their products, it's a set of design rules that they use to then take their design and port it onto your process, all those have to be developed for below 90-nanometer. And they have to be developed from a foundry perspective, not on designing a product with a process at the same time, but I'm developing a truly commercial hardened process and I've got to take that and develop 65-nanometer, 55-nanometer, 40-nanometer, 28, 22, 12. And then I've got to add nonvolatile memory, I've got to add automotive nonvolatility. By the way, there's like 6 different types of nonvolatile memory. And then I've got to add bi-CMOS, and I've got to add BCD. And you start putting that all together, and that takes a lot of time, right?

We looked at the product portfolio that tower has. And really, I mean, we've been taking share organically for a while, right? And so at least in 2 areas that there were a little bit of overlap. So we feel pretty comfortable with our product roadmap and our traction with customers in the areas that we overlap with them.

Got it. The next thing, Dave, maybe digging into kind of some of the real-time trends. Your business has very big exposure to the consumer, right, especially mobility is half of your business. And that's the end market that seems to be under -- facing the most macro headwinds. How do you think you can offset those headwinds? Because every day, we hear about more news about smartphone weakness, especially in the Android side, for example.

Right, so about half of our business in Q1 was smart mobile devices. And that business actually grew quite nicely for us on a year-over-year basis. And really, that was driven by the 5G transition. And so we have a real franchise in the front-end module. And we've had it for years, and we continue to extend our leadership position in that franchise. And so we're quite pleased with the demand that we see in smart mobile devices. We think that even if there is a compression in total handset units, that the transition from 4G to 5G, which is significantly silicon accretive for us, as that transition happens, it still averages us up from a growth perspective on a year-over-year basis. And so we think that business continues to grow for us this year. It won't grow as fast as the other businesses. But we think it continues to grow for us this year. Sukhi, you want to maybe give them a proof point? Maybe China as an example would be a good proof point of that type of transition.

So let's just look at the funded module business, right? It's about 40% of our smart mobile business. And we have a who's who there, customer list for Sprint and modules. I think you know the list of customers. And if you just take that part and look at -- and what's happening in China because a lot of people talk about China weakness, the weakness is really on the low end 4G area. That portion declined about mid-teens in Q1 year-over-year, right? But at the same time, our 5G business in China grew over 30% year-over-year.

And the net result was a growth rate of about mid-single digits, even within a market. We use China as an example. Our China exposure is, I don't know, 8% or 9% of our total business. But I use that as an example because here's a market that had, by and large, a pretty large contraction, particularly on consumers with having a large metropolitan area shutdown. And yet even within an example market like this, we still had growth year-over-year and the growth was driven by that 5G transition. It was not driven by the low end of the market, which is an area where we're not very exposed.

Is there a rough number like content-wise, 5G versus 4G or percentage growth-wise that we can think of?

I don't think we've...

Yes, I don't think we've disclosed any...

Suffice to say, most of your business is now 5G-related, right? So it's not as exposed to the 4G.

It's fastest-growing business within our smart mobile devices, right? And so the other thing to notice is, one shouldn't just assume that it's just over handsets. We have a very nice growing WiFi 60 connectivity business in the smart mobile device. We have an audio processor business. We have power management devices. We have an ISP business, image sensor business, right? I mean, it's a pretty broad portfolio of business lines within smart mobile devices, which is with a very broad customer base, which gives us more confidence that it's going to grow nicely this year.

All right. So pricing has been an important lever for growth, right, across the industry, right? And it's obviously the foundry side. But your customers have been very successful in passing along, right, and getting value.

They passed some of that pricing along before we did.

Yes, [indiscernible] So do you think that that's a sustainable trend? Like I know you have given view of roughly 10% or so, right, price increase this year. How should we think about pricing dynamics for the next 2, 3 years?

One before I get to exactly that point, I think it's important to think about how the silicon awareness of the market has changed over the last 24 months. The number of OEMs that have discovered where their silicon comes from over the last 24 months has been pretty incredible. It's in every industry. It's not just in automotive. It's not just in consumer. It's in industrial. It's in data center. It's in NPC even where some suppliers or some OEMs, I should say, didn't really understand where the underlying chips came from beyond maybe their fabless company or beyond their IDM or beyond their Tier 1 supplier if you're talking about automotive. And some of that realization has led them to the conclusion that if you're going to sell a $100,000 car, as an example, if we take auto, that it probably doesn't make sense to have a factory shutdown for what would be a $2 in device. And so for that part of the market, it's pretty price insensitive. So whether that chip was $2 or $4, it's still worth it to have that supply so that you can then sell $100,000 device as an [indiscernible]. And so I think you've seen that awareness of maybe supply chain security and maybe continuity of supply is more important to me than the absolute product to an extent. And so I think that silicon awareness is important. I think it's kind of become pervasive throughout the industry. And people are just much more sensitive to continuity of supply more than the absolute lowest dollar cost. And so with that as a backdrop, let me actually address your question. In first quarter, we had, and I'm going to lump ASP and mix together, we had 19% year-over-year increase in ASP. If you looked at that on a like-by-like basis, ASP was probably about 12 points of that 19% with the other 6 or 7-ish percent being really mix related as we accrete up in our end market mix. We did communicate that we would have about 10% ASP growth at the enterprise level, at the company level for the full year on a year-over-year basis. And the way that we think about '23 and beyond, '23, you're going to get the full year benefit of what you got in 2022 because not everything ramps on January 1. The pricing actually ramps as those LTAs ramp, which are largely associated with capacity coming online. So we'll get the full year benefit of that in '23. And then we've modeled it flat thereafter. Obviously, that was before we started to see some of the impacts of inflation. So we do have the ability to pass along those inflationary costs. We have built that into many of our contracts. And so I'm not quite sure what inflation will drive in '23 and beyond. But think of pricing this year, up 10%-plus on a year-over-year basis. Next year in '23, you get the full year impact of that and then plus whatever inflation is.

What does full year impact mean? That it can grow up? If you take a similar-ish level or...

Take a midyear convention, right? So if you had a full year 10% improvement for the full year, if instead of you got everything January 1, you got everything midyear, then that means you get half of that again next year in the subsequent year.

Right. Which would sort of align with what we have heard from other foundries, right, somewhere in the mid- to high single-digit kind of price increases for next year. Are there other levers to also expand content, right, instead of price? I think you call it mix, but in that, are there other things you are adding from a technology perspective that you charge more per wafer?

Yes. I think the -- when you think about being differentiated as a foundry, then the ultimate kind of quantifier is how much of your business single source, right? So today, 2/3 of our business is single source. 80% of our design wins are single source. Eventually, the revenue catches up to the design wins over a 2- to 4-year period, depending on whether you're talking consumer products or more long-lived automotive products. And so I think when you think about accretion and the first thing you have to ask yourself is how much of your business is single sourced and how are you making progress on your goal or objective to have more single source business. So I think that's number one. I think the second portion that you think about is are there end markets that are just naturally more accretive. And I think there are. I think by and large, most in the industry would think data center is a more accretive market. I think most would think that automotive is a more accretive market. If you took technology, I think most would think something like a silicon photonics would be a more accretive technology in the market. I think same thing for SiGe. So I think you can look at it from a single source perspective, and you could say growing single source means that you're growing in differentiation, which has a correlation the value and margin accretion. I think you could say the same thing about certain end markets and then I think you could say the same thing again about certain technologies. So there's kind of this matrix of 3x3 different things intersecting that are ultimately driving up the mix of our business.

Dave, one interesting trend we have noticed is that many semiconductor companies, right, including some of your customers, when they are asked how they will behave in a downturn, they say, well, we will keep internal back utilization very high, but so our foundries will help us with the flexibility. But then you have long-term supply agreements, which says your demand is kind of set in stone to some extent. Who should we believe in this?

Have I mentioned single-source business yet?

Yes.

So internal, external.

Got it.

Okay. Look, I think -- and I'm going to make a broad statement here. And Sukhi, if I misstate anything here, then keep me on track. Almost all of the business that we do with IDMs, it's where their roadmap ended, right? So for example, we do business with certain IDMs that have basically said, we're not going to go below 200 millimeter, which is essentially 90 nanometer, right? And so where we're engaged with them is we're engaged with them on 65, we're engaged with them on 40, we're engaged with them on some 22 FDX. And I could go down the list. But my point is that this idea of I'm an IDM, and I'm taking my overflow to foundry, that's not a very accurate statement when it comes to GlobalFoundries. It may be true for others. I think it's less true for others, quite frankly, but it's certainly not a very accurate statement when it comes to GlobalFoundries. And so I think from that IBM overflow perspective, kind of take that one off the table a little bit. Well, then you look at, okay, well, maybe your second source. Well, I mentioned single source before. I won't log that point again. So we're becoming increasingly single source, which removes that dynamic. So now you're kind of back to this dynamic where you're having a conversation with the customer, not because you're being played against your either internal capacity or against the external capacity vis-a-vis another supplier. It's actually the end market that's suffering, right? And so if a customer comes to us, even if we have a signed LTA, remember, our long-term agreements are fixed price, fixed volume, fixed duration, take-or-pay contracts. We've signed more than 30 customers to those LTAs. We've received more than $3.5 billion of customer prepayments and access fees. And so those are the LTAs. And when a customer with an LTA, if they were to come to us and say, hey, I don't quite need that much capacity. Well, that creates an opportunity for us to sit down and have a conversation. And it creates a framework for us to have a negotiation. We have the contract itself. It's backed by $3.5 billion of prepayments, which are predicated upon the successful completion of that contract. So obviously, you have the prepayments on the table. And then, of course, you're having a conversation around if your end market is suffering, then what can we do together? Because you bet on me as a single-source supplier, and I bet on you as a single-source customer that I committed capacity to. So I think that framework is a healthy model. I think an LTA engagement is a healthy model. And Sukhi, I think, we've had that engagement in place for a long time, right?

Yes.

It's to a lesser extent.

If you look at prior to that, right, I mean, one of the big reasons we wanted to pivot the company also was because we didn't want to be in a situation where we build capacity and we didn't have customers. Prior to 2018, we had one customer on the contract. I think everybody knows who that is. Now fast forward, we have 30 customers, right? So the negotiations and the conversations, real partnership conversations with these customers. And then we have their money, right? And so it's really a different business model that GF has instituted.

The one thing, Dave, that I think investors struggle with is that if semiconductors are so cyclical, how come none of the semiconductor companies are feeling the heat of the cycle at all? Like where is that fundamental disconnect? And they get even more nervous when they see companies with large consumer exposure, right? So I absolutely get the content. But have any of your customers come and spoken with you that, look, things may not be as great, we want to reach -- like have you seen any sign from any customer wanting to talk? Has anyone called you and said, let's talk?

Yes. So let me -- I'm going to regurgitate a little bit of what our CEO said in our earnings call in mid-May. We entered the year with about 25% more demand than we could satisfy. And when we had our call in mid-May, when you look at our book-to-bill, it was closer to 1.5 than it was to 1. In fact, it rounded to 1.5, I believe, was the actual phrase that was used. And so we're not demand limited right now. We're capacity limited. In fact, we're capacity limited for the next couple of years, '22 and '23. And so, I mean, we read the same press you do. And as CFO, I get paid to be a little bit paranoid. And I'm looking out trying to figure out where these headwinds are. And so I don't want to discount them. But I also -- I'm in an environment where if there's a little bit of weakness, it actually gives me an opportunity to mix up. The analogy I like to use is you may go out to your car tonight, and your car may be able to drive 150 miles an hour. You probably don't want to drive at 150 miles per hour all the time, right? And we're driving the industry -- by the way, the industry is driving utilization at the highest extent possible. So GlobalFoundries is fully utilized, and we're fully utilized for the next couple of years. And it's not -- from a financial perspective, it's great. From an accretion perspective, it's great. But you don't want to run at maximum all the time. And so to the extent that we had any weakness, if we were to see any weakness, then we would mix that up with some of the available unfulfilled demand that we have. But by and large, it still looks like a pretty robust market to us.

Got it. Next thing, Dave, on gross margins, right? So fairly or unfairly, people will compare you to the best in the business, right? And they would say, well, those guys have gross margin in the low 50s, GF has a gross margin in the low 20s. Obviously, it's improving consistently every quarter. Why is there such a large discrepancy? And when will you start closing that?

Sukhi, do you want to take this one, and then I'll chime in?

Yes. Look, I mean, I think the artifact of our gross merchant being where it is today is because of mostly legacy reasons. We built a facility in Malta, that was kind of underutilized, right? I mean it's been underutilized. We had to live with the legacy depreciation rolling off there. So a big part of our gross margin improvement last year was because we saw some of that depreciation roll off. And also, we had good fixed cost absorption, right? This year, we should see a nice tailwind to our gross margin from ASPs and fixed costs for absorption. And we're mixing up the business, right? And so all of that, I think, is going to help our gross margin. But I think the biggest reason why our gross margins were where we were mainly because of our legacy spend, capital spend.

Yes. Let me talk a little bit about where we're going as well and maybe thematically give you some data points. So 2021 was really the year of normalization of depreciation and fixed cost absorption. 2022. 2022 is the year of ASP increases and continued fixed cost absorption. 2023, there's a little bit of ASP, there's a little bit of fixed cost absorption and now you're starting to mix up the business. And all of which leads us to kind of our longer-term financial model, which kicks in 24-ish, it's that 40% gross margin level. So that's one way to look at the business. Another way to look at the business, which again, I always ask for proof points, another way to look at the business is let's take our manufacturing footprint and look at it that way. So we generate about 40% of our capacity from Singapore. Singapore is already above our long-term financial model. Let's look at Dresden. So David, you told us at IPO road show, that Dresden as it went from 325,000-ish wafers per year to about 850,000 per year as we tool that facility. Again, the facility already exists. But as you tool it, you get fixed cost absorption. Well, when we exit this year, we're essentially going to be at our long-term financial model in Dresden. That's 30% of our capacity. So that's 70% of the total. Well, then you say, okay, well, what's happening here in the U.S.? Okay, will [indiscernible] is transacting with on, I think you had them earlier. I saw [indiscernible] and team earlier. And so take that one off the table for a minute, then you say, well, how do you feel about your Burlington facility? Very good about our Burlington facility. It's largely in line with our financial model. In fact, as it transitions to SiGe, it's looking quite good, actually. So now you're left with Malta. Malta, that facility, when we built it in 2015, it's got 2 challenges. One is it's not fully utilized, so you don't get the full fixed cost absorption. It can actually produce about 600,000 wafers. I think it's about 575,000. But not fully tooled, and the tooling is ordered, but it's not in yet. And so there's some fixed cost absorption to go. That's one challenge. And then the second challenge is that it's carrying some legacy depreciation that was originally built as a single-digit nanometer fab. If we were going to build it today, it's probably $7 billion of CapEx instead of, call it, $15 billion. And so we're carrying an extra $7 billion or $8 billion of depreciation, and that will be with us through 2024 until the middle part of '25.

$7 billion to $8 billion in depreciation.

Noncash.

Noncash.

Right. But is was...

[ But it is a hard ] fact in our P&L that will stay with us until the first half of 2025.

Understood. The next thing is on getting better funding, right, from all these EU and U.S. chip sector. Interestingly, we had a chance to host [ Ingrid ] yesterday, and one thing they mentioned is that they could actually see euros before dollars in terms of just the funding that Europe seems to be interestingly moving faster, right, to support the semiconductor.

Did you ever think he'd say that?

Yes, it's amazing. So maybe politics works differently there. But the point is that, so far, all the benefits that you have really seen tangibly has really been more in Singapore, right, as opposed to U.S. and Europe. What's the plan to take advantage of these facilities? When do they start showing up because -- and I imagine these will be incremental to the norm.

Yes. So let me speak a bit more broadly. We are seeing a great desire, both on the European continent as well as domestically here in the U.S. for local production and supply chain security. So I think we actually have customers that probably, for the first time ever, in an action-oriented way, not from a marketing way, but from an action-oriented way saying, we want supply chain security where we manufacture matters, and it may even matter more than the last penny that I'm trying to negotiate out. So for the first time ever, I think you're seeing that, right? There's still a kind of awareness that I spoke about earlier. So I think great desire for local manufacturing, both on the European continent and then on the U.S. continent. So then you look at well, what are you doing about it? Well, we have developed our plans to expand capacity in Europe as well as to expand capacity in the U.S. We've developed them jointly, not together, but concurrently, I should say, we develop them. And we've also developed the customer list and some MOUs, nonbinding MOUs, such that when funding becomes available that we will be there with customers, with prepayments with long-term agreements that will take the output from any incremental facilities or incremental capacity that's built, and it will be available to be then submit to the government to work in partnership GlobalFoundries with customers with governments in both Europe and the U.S. Now in terms of timing, I actually agree. I think since Russia invaded Ukraine, the EU has really accelerated their efforts. I think they have the potential to get home with some deals this year, in 2022. That's not something that I would have even thought that I could say 6 months ago. But I think that's kind of the speed and pace at which they're operating. It's actually pretty impressive. And then on the U.S. side, who knows? I think if you had asked me 6 months ago, I probably would have said that potential to get a deal done in the first quarter of 2022, now we're where we are, we're [indiscernible] and competes act coming together, trying to get something negotiated before recess in July. And if we don't get something before end of July and before recess, then you're going into lame-duck session and not a lot gets done in lame-duck session. So TBD on the U.S. But that said, we are ready with customers and we are ready with plans on expansion both in the EU, as well as the U.S.

Right. And I think the Euro point is interesting because in the last few months, you have made announcements with automotive customers, right? And that's especially important, right? So talk to us about what those -- I know you've made it with U.S. customers, but also I believe you made the BMW as well. So what does those relationships mean? What kind of -- what are they getting and what are you getting in return?

Sukhi, you want to chime in on this, and then I'll fill in.

Sure. We're extremely excited about these negotiations and partnerships that we're starting to have with the auto OEMs, right, both in Europe and here, as you mentioned. It's really built on 3 pillars, if you will, our conversation with them. First and foremost is to give them the supply assurance that they want, right? I mean they've gone through a horrendous time the last couple of years where they didn't see product and missed a lot of revenue. So just having that supply assurance to them is very important. That's one of the pillars that we're working on. The second pillar is really working with them for co-development of technology. They all want to have unique technology, differentiated technology so that they can actually sell product in the market that can make the run, right? So developing product technology with them. RADAR and LiDAR would come to mind. And the last pillar there would be long-term partnerships where we can co-invest for expansion, right? So those are the 3 pillars that we're working on with many of these auto OEMs. They are very receptive to that. And I think stay tuned, I think you'll hear a lot more along these lines from us. Dave, do you want to add anything?

Well said. Well said. I know Sukhi mentioned co-development. That doesn't necessarily mean that the OEM is designing their own chip. It could just mean that they're in -- you've aligned roadmaps, and they've been involved in more of the architectural decisions than maybe what was historically performed at those companies. It doesn't always mean necessarily that they are doing the design themselves.

You're not removing the chip designer from...

That's correct. I think there's various degrees of maturity depending on which OEM you're speaking to, right? On one end of the spectrum, you've got an Apple that can design their own chips. On the other end of the spectrum, I think, is where historically auto manufacturers have sat, where they've kind of taken not only the design of the chip, but maybe even the entire subsystem and farm that out to a Tier 1. And I think what you see the OEMs becoming more aware of is the customer experience. They want to make sure they control more of it. And so I think that puts them somewhere between those 2 end points on the spectrum.

Got it. Dave, just again, what's waning on people's mind is potential for slowdown, et cetera. And I know you said next year, you have, what, about 10% additional, right, capacity comes around online.

We're going from roughly 2.6 million wafers this year to about 2.8 million wafers of capacity next year.

Got it. So if we look at that and let's say, we get into a period where semiconductor industry sales are flat or down next year, will that mismatch worry you that you have 10% more capacity coming at a time when semiconductor -- I know it's not perfectly 1 to 1 aligned, but with that dynamic worry you?

Worry is such a loose word, I'm going to go to maybe proof point.

Bother you.

Let's go to a proof point. So what would 10% underutilization, what would it mean to us? And then what would actually have to happen for that to occur? Maybe that's how I can talk to it. So about every 5 points of utilization is about 2 points of gross margin. And that's without -- that's just a pure mathematical calculation. That's without taking any action, right? So obviously, if there was a significant reduction in utilization, then we would actually take actions as a company so that you don't get impacted 2 points for every 5. Maybe for every 5, you get impacted 1 or 1.5. So that's just a mathematical calculation. So then you think, well, what would have to happen? You use 10%, I'll use 10%, if utilization was to decline 10% in '23, what would have to happen for that to occur? So I think the first thing that would have to happen is this 25% more demand that we have today than we can satisfy, that would have to completely disappear. So that's the first thing that would have to happen. Then I think the second thing that would have to happen is that for the commitments we already have for 2023, they would have to then decline by 10%. And that feels pretty significant to me. So 25% has to go away first because that mismatch [indiscernible]

Essentially...

And then 10% additional has to go away, and that's a 10% that signed LTAs with prepayments and access fees that are already committed. And so I don't want to forecast macroeconomically what could or couldn't happen to the market. But there's a lot that has to happen to then get us to not from 100% utilization to 90%. So we feel pretty good about -- I'm paranoid. I get paid to be paranoid, it's in the title. So we do a lot of work around it. What would we do? How would we do it? What would our process be? What would your plan be? But right now from kind of where I sit, I don't see that scenario.

Do you have flexibility in kind of delaying or slowing down additional capacity for next year?

Yes. So what would you do, right? What would you do to take 5% of utilization instead of 2 points? What would you do, right? Well, the first thing you do is you start to push out equipment, and you'd also start to push out other materials that are coming in, right? So just like we would get pushed out, we would then push out below us, our supply base, we would push those things out, right? That stated, we don't see -- if you look at the foundry space, by the way, I know everyone talks about semiconductors being cyclical. I've grown up semi and TI and [ broad com ]and then went to Cisco systems company. But if you actually look at what's happened with foundry over time, it's actually a great chart to go and look at. And when the industry, semiconductor large industry has had a little bit of a wiggle, foundry would kind of flatten off, but then it would almost immediately go back up because the IDMs want to make less investment in their own internal manufacturing capacity, which means more goes to foundry. And so I think the long-term trend is very good for us. And so we don't particularly think about, hey, what could happen for a 6-month period or even a year period because we're making investments for 5-, 6-, 7-year type periods. And so we believe that any near-term perturbations will weather through. We'll do the right things to be fiscally responsible. And that ultimately, the foundry industry comes out stronger on the other side.

All right. Last question in the 15 seconds left.

It's so fast.

Delay in tools, right, we have heard a number of cap equipment companies talk about their supply challenges. Are you getting enough tools to meet your growth targets for this year? Or are those issues for next year? Or it doesn't change?

So the good news is we got our orders in early. So we placed orders really mostly in the kind of the first half or around the first half of last year. So from a FICO perspective, we're in a good spot in the Q. That stated, we are seeing tool delivery start to slip a little bit. If I had to characterize that slip, I'd say something like somewhere between 2 to 4 weeks type of slippage is what we're seeing today. We built more than that type of slippage into our plans. And when I say plans, I mean both LTA commitments as well as financial plans that we have communicated. So at this level, call it at 30-ish days at the high end, I'm not particularly concerned from a commitment perspective that we've made. I think if it started to get double that, then I would start to maybe have a different opinion. But right now, those guys are working very, very hard for their customers. Very impressed with the WFE vendors and really appreciative of their support. It's a tough environment.

Right. Great. Thank you so much, Dave. Thank you for your time. Really appreciate it.

Thank you.