GLOBALFOUNDRIES Inc. (GFS) Earnings Call Transcript & Summary

Just sort of went under everybody's radar screen, but we had the approval of the National Defense Authorization Act, right, the NDAA, I think it was in mid-December. And people tend to think about the NDAA as, yes, it's the defense budget authorization, right, for the following fiscal year. But it actually did place further restrictions on China-based semiconductor content in U.S. government applications and equipment. And it seems that this combined with all of the supply chain disruptions that we've had over the past 3 years, post the COVID reopening, is really motivating your large U.S. auto OEMs, industrial OEMs and so on, right, to diversify their manufacturing footprint away from Asia-based suppliers. So given the announcements you made with large-scale customers such as General Motors, Qualcomm, I mean, how are these dynamics playing into the decision-making amongst the customers to drive more and newer engagements with these constituents?

Yes. In a phrase, I'd say we're in the early innings of that dynamic. It's really inconvenient for a supply chain that's been developed and honed and made efficient over the last 3 decades to immediately disappear overnight. But there is a recognition that for security reasons, for the inability to maybe operate that more balance is required. So here's what we're seeing. Western companies are still thinking through how they want to change that dynamic. They're certainly not going to go back and redesign old part numbers into a different global footprint like GF's. They're certainly thinking about future designs. But none of that is really baked into our model going forward. It doesn't mean we don't think it's going to happen, but until we see it. Now the other end of this is there's this worry that maybe China is overinvesting and it would hurt the foundry players like GF because there'll be overcapacity. I'll tell you 1 thing I am seeing. Our customers may not be designing fast out of China. They're certainly not putting more into China. So there will be capacity in China and may be excessive, but it's not going to hurt a foundry like GF because our customers are not going to do more volume there. The other thing we're seeing that, that surprised me at first, but then when I thought about it, it made a lot of sense, the customers that are the most anxious to move out of Taiwan and China are the Taiwanese fabless and the China fabless companies. Taiwanese fabless say, look, if there's some kind of geopolitical dimension that blockades our ability to use Taiwanese foundries, we could service our customers in the Western world through GF's global footprint. And so they're moving. And we'll start to see some design wins just based on moving products from Taiwan into the U.S., same thing with some of our Chinese customers. And I think you'll see more of the Western companies figuring out how they're going to go deal with this inconvenience of the East versus West. And look, it's not going to happen overnight. It doesn't matter the investments that China is putting in. For the real endgame of this industry, it could be there is a Western and Eastern supply chain for semiconductors, 2 different ecosystems, highly inefficient, but it probably is what it's going to be, but it's going to take a decade. It's not going to happen overnight.

Let's talk about some of your end markets and some of the diversification strategies, customer-wise and market-wise. And I'd like to focus on the automotive business because you touched upon it in some of your answers here. The team more than doubled the automotive business year-over-year in the first quarter. You're on track to drive $1 billion in revenues this year in automotive from $375 million, right? 2.5x growth, as you mentioned. GlobalFoundries is broadly engaged with most of the auto OEMs, tier 1 auto suppliers, auto semiconductor companies, so the entire value chain, right? You secured $2 billion of long-term agreements last year in automotive. So based on the current design win, ramp-up profile, I mean, how should we think about the revenue trajectory for automotive? And what do you think the revenue mix in auto looks like over the next 3 to 5 years?

Yes. So beauty of the tiering is small numbers, you could look like you have really high percentage growth rates, but you're starting in 2020 with like $75 million of revenue. But it's substantial this year, $375 million last year. We're going to bump our head on $1 billion. We think it's an important market for us, and I'll give you the answer to where I think it's going in a moment because it's a market that requires where we play. It doesn't need a lot of high-speed digital computation. It needs a lot of analog, mixed signal, high frequency, whether it's connecting the car or autonomous driving the car. This transition from internal combustion to what we call ACE, autonomous connected electrified car, plays to all the technology nodes that GF invest in. So it's an important market because it's accretive to our business. It typically commands higher price because it has higher reliability because of the mission-critical applications. And so if we stopped any activity today in automotive and just harvest it for the rest of the momentum we have going into it. That business easily doubles from the $1 billion today. But we're not stopping there. We're going to try to double down and continue to build out that platform because the technology requirements in that space really align well with what GF does, whether it's RF for radar, whether it's low power for electrification of cars, precision battery management or microcontrollers for intelligent and secure functionality. That's the place we play in.

So help us understand, though, because when the announcement about the signing of the GM long-term agreement. We got a lot of questions from investors, what is GM doing securing capacity? Are they designing their own semiconductor? They're already -- GF is already engaged with the top automotive semiconductor suppliers in the world. So help investors sort of understand the dynamics around why is it that your signing and inking these deals with the auto OEMs directly, but yet you also have pretty strong partnerships with all the semiconductor suppliers.

Yes. First, it's a great question and what you understand what's it foot there. It makes sense for why, it might have been a first-of-a-kind deal but certainly not a last-of-a-kind deal. So what is it on a company first need to do? It's a little like electronics snuck up on them. They started putting more and more electronics in cars. So they wake up 1 day and find out, my God, we have 80 different SKUs of a microcontroller when we need maybe 4 different types. And so the first thing all the automotive companies are doing right now, and this is the end of the realization through 2021 is, hey, how do we harmonize and rearchitect the electronics of the car so we can be more balanced and have a more efficient use of semiconductors. And then once you have that, you say, okay, how do I go out create the supply that I need on the technology nodes that have the durability and legs and get it at the best economics. And the GM deal does all 3 of that. It's our 40-nanometer embedded memory for microcontroller technology. It's a winning platform. It will be here for a long time. They wanted capacity in the U.S. They could have asked for it in Germany or Singapore, they wanted in the U.S., and they wanted the best economics. So to create that capacity in the U.S., we'd have to make an investment. To make that investment, we need a return on that investment. So that will be an extra cost on their wafers. Let's take an imaginary number. Just say that cost to get a return is $1,000 in every wafer. If we took that $1,000 and passed it over to a fabless company who does the design, they don't just pass through cost, they margin it up. They'll pass it on to a tier 1 who creates some functional anti-lock brake, airbag or some function and they'll mark it up. And so something that should cost GM $1,000 for that investment on top of the manufacturing service to create the wafer is now $2,500. So how do you get the best economics? You work a 3-way deal. You work with the fabless company with that product is going to cost. You work a deal with GF on how much we want to every wafer you ship to return on investment. So you pay us directly for that investment, you pay the fabless company the value they've created in design and GF charges the fabless company for the wafer without that investment. Best economics, capacity where they want it on the technology they want. The reason why it's not the last-of-a-kind is because it makes economic sense. I think you'll see more of those come along in different variations. By the way, GM could have easily said, you know what, I'll use my own capital. I'll make the investment, then no one has to mark up anything. They chose a better use of their capital was to build their automobiles and their capital investments and chose to use GF's. But you can imagine a world where other companies would say, no, I have a big balance sheet, I might as well use my balance sheet to help my gross margins.

Perfect. On the last earnings call, you called out strength in your aerospace and defense business through this weak period in our industry. And it's the first time I've heard the team mentioned aerospace and defense as a part of your mix. But I do know that GlobalFoundries has what we call ITAR and trusted fab certification by the U.S. Department of Defense. What types of programs are you engaged in? And did the NDAA that was signed that we just talked about in December, did that motivate more of your aerospace and defense customers to redirect programs to GlobalFoundries?

Yes. So ITAR trusted foundry, this enables GF to make the most sensitive components for U.S. military applications. The NDAA kind of created awareness. There wasn't really a policy that talked about where parts can be sourced other than when they were super sensitive, and it required trusted foundry. So a lot of the defense industrial base who serves the Department of Defense and the military took note. We need to get a better understanding of what are the uses of semiconductors in our systems. Many of them give a specification and have a box made to them as kind of a black box that delivers a certain function. I do think it was a surprise to find how much content was not U.S.-based or even Western based. And so it's creating a sense of awareness. I think it was no different than the auto companies in 2021 when semiconductors as an afterthought until they left about $100 billion of auto sales on the sidelines because they didn't have enough semiconductors to create awareness, they started to think about the industry differently. I think we're in the early innings of the aerospace and defense to think about, hey, where are we getting our chips from. There's real risk to them. There's reputational damage if something goes wrong in that supply chain, there's risk that the U.S. government will want more proof points on what they're sourcing. And it creates opportunity for us to work in partnership with these -- with our existing customers and maybe new ones in the defense industrial base to create opportunities for them to leverage our technologies.

Well, talk a little bit about differentiated technology because the metrics that you guys throw out, 90% of your design wins are sole sourced to GlobalFoundries, 2/3 plus of your revenue, so it will gradually approach that number. 2/3 of your current book of business is sole sourced to GlobalFoundries. And we cover 20 semiconductor companies. And one of the biggest trends in the market, especially high-performance networking, is all of the big guys are moving to co-package electro optics. And more often than not, when we ask them, who is your manufacturing partner, they'll say, GlobalFoundries. And to that point, I mean, you were one of the first to market in terms of developing a manufacturable and the key term here is manufacturable silicon photonics solution for next-generation data center connectivity. I think you guys said that you have 16 customer tape-outs on your new 45-nanometer CLO silicon photonics capability. So I guess the question there is, how much revenue is the team driving today? Have you seen any customer pull-ins on some of the technology roadmaps? Because of all of a sudden, there's this big focus on accelerated compute, artificial intelligence. And this is really driving some of your customers to really step it up in terms of higher data rates. Silicon photonics is going to help them achieve that.

Yes. I think the catch phrase, it's certainly like around [ iGPT or ChatiGPT ]. These are the technologies that are kind of overhyped in the near term and underhyped in the long term. Silicon photonics is an important play for us. It's -- we're the first to monolithically integrate CMOS, RF SOI and photonics on a single chip. So it creates great functionality at the best cost of ownership and integration. But it's a technology that takes a while to deploy because of how special it is. And we see these design wins and these tape-outs are really important. They're really more teaching our customers how to best leverage it. This is a story that's in early innings for us. It's -- I think it really ramps in the 2026 time frame. But it could be a market that grows to over $1 billion, and there's no reason we shouldn't have 50% share in that. It's a very good margin business for us because it's so special. And differentiating creates value. But this is going to be something that's going to take time to get that deployed into the industry and with our customers.

Well, we're just about out of time, but I did want you to maybe hit on several of the key factors influencing the growth opportunities ahead for GlobalFoundries, particularly in the face of some of the near-term headwinds that we're seeing across the industry? And then against this backdrop, are there any other remarks that you'd like to maybe leave us with today?

Yes. I think kind of the new news is, given the investment profile we made and how it can grow our company from where we are today, we're probably running ahead of schedule to our free cash flow positive for our business. We needed to invest to get to that $10 billion and now we have that investment in place, so we can be free cash flow positive. I think from an industry perspective, the way I look at semiconductors, it's always been a series of golden ages, some call the different epochs of computing, right? It started with the mainframe, PC, networking, handsets and now IoT. And each generation of technology, it wasn't replacement, it was additive. So we still sell 300 million of PCs a year on top of 1.2 million handsets. Now we're going to have 15 billion of IoT devices. Additive. The thing about IoT, it's the first time in the history of this industry that the technologies that service that next generation of compute is not leading edge, Moore's law. It's where GF plays because it's about compute at the edge. It's about taking data to the edge using it for inference in action, parsing it, compressing it and then send it to the cloud because it's impractical to turn every piece of data collected to the cloud for a decision and come back. So sensing devices at IoT now have to become intelligent devices, and we have the technology for that because we have the 3 important features that make that a winning play. The first one is like real estate. It's not location, location, location. It's about power, power and power. The second one is best RF connectivity wins. And the third one is secure intelligence, which means digital processing with embedded memory. Those are the key features that we bet in our platform, and we win in the space. Give an important application, think about a camera -- a surveillance camera you want to be able to work on the edge for you. You want it to be pasted up in a remote spot 2 to 3 years on 2 AA batteries. You're not going to get that if you don't use GF's 22FDX technology, so that it delivers the performance in power that lets the battery life. And how does it do that? Any kind of device that's not tethered spends most of its time in the idle mode. In this particular case, there's 2 microcontrollers in the design. The first one wakes up 10 milliseconds every 1,000 milliseconds to look around to see if anything is going on. If it does, it wakes up the big microcontroller and the data and action takes place. So what's really important is the minimization of power when the device is kind of in its sleep mode that fits to our technology. We've developed to be not only power-efficient when our devices are running, but super efficient when our devices are sitting in the idle state. And so we talk a lot about our automotive growth. A huge opportunity for GF is to be doing at the surface of the edge, the kind of AI and compute through the inference. Let the edge do the inference, let the data centers do the machine learning and then teach one and other. I think that's the big play for GF as we participate in this new area of AI.

Great insights, Tom. Thank you for your participation today.

Appreciate it.

Yes. Thank you.