GLOBALFOUNDRIES Inc. (GFS) Earnings Call Transcript & Summary

It varies from case to case. But the way I think about it is using those longer durations, it's typically speaking to the development phase, the ramping phase and then getting to some steady state run rate. And then honestly, the -- once you've solved for that steady-state run rate, you're in pretty good shape to sort of see those volumes continue well beyond the end of those agreements. And so what it sets up is as you get a couple of years into these LTAs, and this is sort of true across all LTAs. It's easier to go tack on another incremental year or 2 as that -- as you grow into that volume. So I think these are evergreen LTAs in most cases where you set a duration to get you to a certain capacity level with certainty and durability. And then each year, you decide how much more you want to tack on. So these are very much evergreen arrangements. And if you're in a market like auto, these can be 10 to 15 years of total lifetime demand easily for which you cover the first half with the first instantiation of the contract in the back half with extensions. So that's what we would expect. And again, it speaks to the attractiveness of those automotive markets to GF.

Makes sense. So more recently, GF established the first of its kind, at least that I'm aware of direct-to-source LTA with General Motors, where you are going to have dedicated capacity corridor and Malta to support them. How do you think about this trend as it relates to kind of pricing margins, duration of contract and in terms of working directly together, how do you think about the potential of securing more and more market share?

Yes. It's a very unique, and again, I think not -- first of a kind, not last of a kind contract. You have to take yourself back to the beginnings of the automotive shortages. And frankly, most OEMs would tell you the back then that they were -- they were not very aware of where their semiconductor capacity was coming from or what choices were being made on their behalf. And what they thought was a multi-sourcing strategy through multiple Tier 1s and Tier 2s, which was -- they thought was creating resiliency in their supply chain was actually just creating an enormous amount of possible defects when it turned out that the supply-demand situation turned against them. So deals like the one we announced with General Motors are acknowledging that it's far better for the OEM to secure capacity and then allocate that capacity to their authorized purchasers and to drive a strategy that ensures supply. The second thing that is really unique about that opportunity is deciding where you want to have capacity built and deciding what technology you want sort of implies that you've got to put in place new capital to deliver those products. And the economic model that is sort of -- in this first-of-a-kind contract is to acknowledge that if that capital expenditure generates depreciation and amortization cost to the foundry, that winds up getting marked up multiple times before it reaches the OEM that it's very much in the interest of the OEM to intercept that expense much earlier in the process. So that's how we get to a different economic model that allows an automotive OEM to get the 3 things they want. The technology we want, where they want it. The economic model that makes it affordable. And frankly, the security of supply, knowing that if I built a wafer in 2021, we might not know immediately if it was going to go to an automotive application or an industrial or consumer. Even if it was going to automotive, there will be a question of which fabless company it might go to. There might be a question on their side as to which Tier 1, that chip wound up in and then ultimately, which OEM. So circumventing all of that is securing the wafers and making sure that they can only wind up in that OEM's vehicle is part of the game-changing part, and it gives us the certainty and durability to make those capital investments with great confidence. And that's the thing I think you'll see more of as we go forward.

And from just securing the capacity perspective, is that kind of actual wafers or more virtual in the sense that a customer an OEM could decide to toggle between Malta and Dresden and Singapore? Or would it be a specific factory?

So it's sort of a -- that's a mix of sort of 2 good points. In this case, it's very specific. It's very tangible. It's capacity at a certain level and a certain process that -- to your other point, we are able to transfer in. It's an existing process that we can reshore back to America from Singapore to give not only the certainty of the aggregate supply, but also make it domestically supplied, which was attractive in this case. But it also does highlight the utility of having not only a breadth of product, but a breadth of interchangeable product across different regions. So we can build in many cases, as you roll out this strategy, you could build the same process technology and at least 2 possibly 3 different geographies. So you're creating resilience, not only in terms of avoiding some of the more geopolitical hotspots in the world, but you're actually creating the ability to multisource. So specifically, this agreement, a specific capacity in a specific place but the benefits of being able to run the same process in Dresden that you run in America or in Singapore and in Dresden, those are benefits to everybody, the OEMs and the Tier 1s because they get the diversity. And for us, we get the flexibility that when one market is up or one customer is up and another one is down, we have the flexibility to flex across the footprint of facilities that we have worldwide.

Excellent. So maybe we could focus on capacity and obvious focus for basically the entire world. But clearly, for automakers over the last handful of years. So I guess, can you speak to what your plans are in terms of capacity additions specific to auto this year, next year, whatever time frame you deem appropriate?

Yes. So I guess the way you can think about it is with the CapEx investments that we've sort of already made or have said we intend to make, we probably are getting close to the ability to go from sort of the 2.4 million wafers that we talked about at IPO up to around 3 million. Now the rate and pace at which that shows up in production is based on a lot of different factors, but it gives us the flexibility to meet the automotive demand that we're seeing, and that's been -- there's 2 parts you need to get automotive growth, right? You need to have the designs and then you need to have the capacity. So we feel really good that we've got enough CapEx and tooling that we can grow up to 3 million wafers. And with the demand that we see with automotive, automotive can get more than their fair share of that expansion. So that's sort of short to midterm. And then long term, we just -- we look for deals where we can get optionality on larger chunks of future capacity. And a great example is what we're doing in [ Crolles ] with the French government and SD -- the EU has actually come through with not just the notion of doing funding of chips, but has actually approved that funding. So now here we are sort of spring loaded where we have got the ability to expand capacity a little longer term on good economic terms, but always predicated on actual customer demand. So nothing speculative, build it and they will come. Just having the flexibility to say, you know where we build it, it's all ready to go. And given the backlog of automotive of interest and demand, really likely combination. And then you'd probably see the same thing, you will see the same thing in the U.S. where we have ability to expand the capacity in the existing facility in Malta. GM is a great example of using some of that headroom. And then ultimately, via things like the CHIPS Act, the ability over the long term double the size of the wafer output that we have coming out of the U.S. So sort of a short, medium and long strategy on being able to meet upticks in demand and do it in a very capital efficient way.

Excellent. So maybe a simple question next. Why does GF win in the auto world? And I guess, can we apply sort of the same drivers that drive success elsewhere at GF, namely, the importance of connectivity, power management and feature-rich analog and/or anything specific that we should be taking outside of that within the automotive bucket?

Yes. So look, why we win in automotive is the same reason we would win in any market. You have to bring value to capture value. It also turns out that automotive is not a market you can dabble and you have to be very committed across all your various processes and all your different factories to deliver the kinds of features, the kinds of services, the kinds of ruggedness that don't happen with fleeting interest or fleeting investments. So as I said before, automotive is an elemental part of our strategy. It's not accidental. We don't build for the commercial market and then hope there's some automotive applications. We really focus a disproportionate amount of our R&D, again, punching above the current revenue run rate to automotive applications, automotive innovation, automotive process technology. So we're putting our money where our mouth is and delivering value. And then as we discussed, because automotive is very dynamic, it's been through a pretty traumatic experience and awakening as to the strategic importance of semiconductors in that end market to deliver the user experiences that they'll need to that it's created a great opportunity for us to get just much closer because of the vital nature of semiconductor technology to that market. And so think about it, 3 years ago, I think automotive was 4% of our revenue. This last quarter, it was 10% of our revenue. We've told you it's going to continue to grow. So you can almost imagine what that would be as a year '23 number, clearly solidly in the double digits. So this -- on one hand very simple, you focus and you add value. And if you do that, you have you win. And I think in foundry, you look at the strategies of other competitors. And they've got other very valid strategic interest in different areas, but I think we're the only ones that can raise our hand and say, we doubled and tripled down in these specific end markets and auto is a great example of that.

Very helpful. So maybe we could pivot to your future road map. Given the seat that you're in, you obviously have to take a longer-term view. So what gets you most excited in terms of your tech road map thinking out over the next 5 to 10 years? Is it kind of silicon germanium, gallium nitride, silicon photonics, what gets you excited?

Yes. I guess, you know. If you just sort of walk through on -- in the data center and the AI world, we think there's massive opportunities for us in power delivery, managing the power that those massive data centers are requiring and also realizing that there's going to be a big transition from electrons to photons, from carrying everything in wires to carrying it on fibers and making that transition. In automotive, we talked about the 3 big ones. We're putting more and more radar and LiDAR technology into cars more imaging, just playing automotive cameras. We've got a -- we've got a strong incumbent position on the microcontrollers that still has multiple nodes left to carry it out through the next 2, 3 decades. The advantages we see in the IoT world with 22FDX will be directly applicable to the MCUs today that are just ramping in 40. And beyond that, similar plans for 12FinFET. In the electrification space, we've got a very strong battery management portfolio, but we are looking at wide bandgap materials like GaN, which we think has a lot of promise, not only in automotive but in RF, in micro LED. So it's an exciting area as well. And then just some of the core markets that we've always served, continued innovation on the RF side as we go from FR2 5G to FR3. There's constant innovation to sort of continue to deliver better wireless performance across both wireless handsets and connected low-power IoT devices. So all the engines of growth are growth engines because of the R&D that we're doing and the road map that we see out in front of us. And I think we're actually just getting started.

I guess we would love to add on 2 points there, if I could. Can you elaborate on your exposure on the electrification side, you talked about strong BMS? And then secondarily, with the shift to more domain controllers, oral architecture and vehicles? Can you speak to your positioning there as well as what kind of increased content you could see there?

Yes. So in electrification, it's interesting, you can focus on the specific thing that's being electrified as being going from an internal combustion engine to a traction inverter. But probably the bigger story is that now every single element in the car becomes power sensitive. That wasn't the case before. When you had an alternator being run on a battery, right? You had a sort of -- as long as you were driving it at an endless source of power, you can air condition, you can heat, you could play videos on the screens. You can have as many screens as you wanted. You had a sort of a renewable source. The opportunities in electrification for us are really the management of all that power, all those different voltage domains, all the different conversions and the conditioning and the preservation of energy across, now every single system in the car. That is probably the biggest area for the electrification gives us is not only the direct control of the different voltage domains in the car, but the overall premium on power, power, power. And that leads directly into your second question, which is the technologies that were chosen for the current crop of microcontrollers. Power was on the list of important things, but in the next generation, power becomes primary. So the benefits of things like 22FDX that we've seen in the IoT world, and they're beneficial because they run off of batteries, right? Well, guess what? It's a bigger battery, but it's still a battery. So in automotive, the transition to zonal control will still benefit us in terms of combining that with a real focus on power and moving things from 40 to 22. And if those zonal controllers are [ be fewer ] -- and maybe fewer to our 12FinFET product line with embedded nonvolatile as well. So that's sort of how zonal thing fits in. But what we always have to remember is the car is -- it's like an IoT element and an IoT network rolling around. And in that IoT network model, there's real-world sensing and control and actuation that is never going to go away in a car. And so there may be a different architecture for the zones that are controlled, but they're all controlling very specific functions that require actual real-world sensing, high-voltage actuation, the microcontroller that's going to control the suspension or the braking or the movement of the seat will still be something that has to drive a motor. So you'll always have a very rich mix of different MCUs, whether that's 90, 40, 22 or 12. And I think we cover really the next several decades of however that architecture evolves.

Excellent. So we've got 2 minutes left. I was going to turn it to you and maybe ask, is there anything that we've missed that we should be thinking about to truly understand GF's auto business and/or if you had kind of concluding thoughts?

Yes. So look, I think the proof is in the results. We started focusing on auto about 4 years ago. We've stuck to it. And now we're the beneficiaries of a fast-growing business in a market that's valuing semiconductor technology in a very different way than it had been historically. It's a good proxy for what we aim to do in all of the end markets we serve. The differences for automotive for us are that it's sort of unique in its durability and certainty. So it's a very good representative of the GF strategy when applied to a market that has those attributes, you get the best possible outcome. So look to us to replicate that in other areas. Everything from what we do in smart mobile, IoT, data center and even things like aerospace and defense, they all fit our business model. And remember, as an $8 billion company with a $120 billion worth of SAM out in 2030, we can be very selective to make sure we're doing it in the most profitable capital efficient way. So that's, I think, where I'd put a pin it and thank you for the opportunity.

Well, Mike, thank you very much, Really appreciate you taking the time to spend with us this afternoon and best of luck, and look forward to talking again soon.

Thanks, C.J.