GLOBALFOUNDRIES Inc. (GFS) Earnings Call Transcript & Summary

Welcome to the 5th of 5 straight semiconductor fireside chats. So hopefully, you're finding value in seeing these companies lined up one right after the other. I'm very excited to have the -- I believe it's the newest semiconductor company at the conference, and that's GlobalFoundries. So today, we have Dr. Thomas Caulfield. He is the President and CEO of GlobalFoundries, joined in 2014. Prior to being named the CEO, he was the SVP and GM of The Fab 8, which is the Saratoga County one. And he led operations, expansion, and manufacturing ramp at that Fab. And before GlobalFoundries, he was the President and COO at a gallium nitride company, a solar power solutions provider. He was also at Novelis and IBM for 17 years. We also have Sukhi Nagesh. He's the VP of Corporate Development and Investor Relations for GlobalFoundries. Prior to GF, in 2018, he was the CFO and VP of Finance for Benitec; Interim CFO and Head of Investor Relations at Marvell. Gentlemen, welcome.

Thank you, Mark.

Thank you, Mark.

Great. Thank you. So since you guys are the newest semiconductor company here, Tom, why don't you just give investors a little description of what's GF? What are you guys about?

Okay, 3 minutes or less? So newest in the public markets, but we've been around. I would put GF into 3 phases. I'll call it the creation phase, 2009 to 2018. I became the CEO in March of 2018. And the creation was, remember, the very beginning of GF was a spin-out from AMD. And AMD had to decide what's the best version of themselves, a product company or a technology manufacturing company. They didn't have the scale to be the latter, so they spun out GlobalFoundries and went on their way as a product company. Now just because you spin out a subscale fab doesn't mean you have scale. And for the better part of almost a decade, GF collected the assets to create global scale. They bought the #2 foundry at that time, Chartered Semiconductor. They acquired the IBM assets and a check that came to GF. I always seem to get a chuckle when I say that. And then of course, they did their own organic investment in this Fab 8 facility you mentioned. And began, in 2018, as a company of $6.5 billion, global scale, global footprint. But fundamentally, our business model didn't work. More money left the building than came into the company, and there was real problem with that. And so we made a pivot. Because what had happened since the inception of GF in 2009 to 2018 was the semiconductor industry fundamentally changed from a compute-centric industry to the pervasive deployment of semiconductors, where chasing Moore's Law represented 25% to maybe 30% of the market opportunity. But where we play, 12-nanometer and above, was 70%. So it gave us an opportunity to focus where we can have real differentiation, create real value, and get the, what I call, Phase 2 of GF, which was the transformation to go from a scale, globally competitive foundry, to one that could do that and make money. And then with the advent of joining the public markets, I think we've entered Phase 3. And I call it the realization of GF, where we become the very best at everything we do. Our financial model, our execution of differentiated offerings to our customers, everything we do, we should strive to be excellent and realize the full potential of our company. So that's GF in a nutshell.

Great. That's perfect. And I think you guys are a semiconductor foundry on the front end. And I think when most investors think about foundries, they think about TSMC, which has been a phenomenal success story. How should investors think about GlobalFoundries relative to TSMC? What's the same? What's different?

So I think what's the same is we're manufacturing services company. We don't make products, right? We make our customers' products. So the intimacy in our relationships has to be strong. Given the scale of TSMC, remember, TSMC is 9x bigger than GlobalFoundries; twice as big as the next 4 foundries: GlobalFoundries, Samsung, UMC, and SMIC; you add up all their revenue, TSMC is twice their size. So another difference is their scale. And a little bit of their go-to-market strategy is more menu-driven. They create a menu, and customers design into it. And they have the scale of manufacturing. And by the way, UMC, just in that vein, they practice a second source strategy where they align to that TSMC menu. GF plays it a little bit different, maybe a lot a bit different. We look for differentiation. We look to take our different nodes; 40, 55, 28, 14, and create rich platforms between the IP we enable, the different device types. We look for what we call single-sole source business, where our differentiation creates value for our customers. They use the catch-up value for themselves and we use the catch-up values for ourselves. To give you an idea of that single-source nature of our business, in 2021, 80% of our design wins, future sockets we'd win, was single or sole-source business. In Q1, we just reported that year-over-year, our revenue from single-source business grew 47% and today represents 66% or so. And so you can imagine the design wins, all our future design wins are more and more single source, more and more accretive to our economic model. As our business moves forward, we'll be more single source, more differentiated business and better value capture.

Great. Let's shift gears to the operating model. In the first quarter, you just released, you reported 25% gross margins, 14% EBIT margins, which is up from 2 years ago where you were negative 13% gross margins and negative 30% EBIT margins. And so when you took over as CEO in 2018, what were the -- and that's an impressive turnaround in a short bit period of time when you took over. What were the key ingredients of this turnaround you engineered?

Yes. You have to get conviction. It's easy to decide what you're not going to do anymore, right? And then you wake up at the crack of noon every day and make sure you're not doing it. It's a build conviction on what you're going to do and be successful. And for us is we had a lot of proof points. We understood with customers where we had, where our differentiated matter. We had a site in Singapore, our Singapore Fab 7, which was already at our economic model and thriving. And the key was to make the rest of the company look like Singapore. So we repositioned our portfolio. We stopped talking in business units and technology and rather made them end market focused: IoT, automotive, data centers and communication, personal compute, and then smart mobile devices, which by the way is 50% of our revenue. And we used that capability to make sure if we were going to be differentiated. We knew what we needed to spend our R&D dollars to create that differentiation. And then with that came how do you improve the profitability for the company? And it's never one thing. One is making sure where we were value selling, that we were capturing the value of our products. Two, we had a global footprint, but it was half used in a sense. Not utilization the way you think of tool utilization, but the fabs didn't have all tools in them. And if you run a manufacturing operation, you know that you have this high fixed cash cost, whether you make one wafer or a million wafers, that's there. And so part of our improving profitability trajectory comes with filling our global footprint, doing it with better mix, so a higher ASP. Our depreciation, that first 9 years of our growth, we did a lot of hyper investment, I would argue not the best investment, so high level of depreciation. And so that combination is coming into play as we see -- we're probably maybe newly minted IPO, but also probably have the highest profitability trajectory. We see our business growing top line 8% to 10%. We'll get to 40% plus gross margin, 25% operating income, 45% EBITDA, and then be balanced in our capital deployment at 20% of revenue. Now you say, okay, what do you got to believe to make that happen? Well, for me, it's always the proof point. Let's go back to Singapore. Today, our Singapore facility is at our long-term model. Next year, when we finish this year's investments in our Dresden facility in Germany, we'll be at our operating model. And then our U.S. facility in Fab 8, where we have a high level of depreciation, that gets there after 2024. So we see the existence there or the existence of where it's already at our model and how we drive there. And the key for us is, how do we accelerate that journey to get to that model?

Got you. And...

Hold on, are we missing anything?

Yes. So to talk about the LTAs and how is the visibility of the LTAs.

Okay.

Actually, can I set that one up? For me, this is a really important idea. And what we had learned from you guys and also from the industry, we're learning a whole new lexicon in semiconductors, like things like LTSAs and manufacturing corridors. So maybe we could just start out in LTSA. And I think during the pre-IPO process, one of the most impressive things that I saw was that list of 10 to 15 customers with over $20 billion and signed up for LTSAs and $2 billion of prepayments. And this is something that I had never seen before in covering semis for 20 years, that kind of visibility. So how did you get there?

Let me put a little context around it.

Yes.

So I'd like to give this example. If you went to a builder of custom homes and said, you're going to build me a custom home. I'm going to live in it. I'm going to forecast I'm going to buy it. And when it's done, I'll come back and maybe I'll buy it, but I won't, and I'm certainly not going to buy it at the price you've built it for. That was essentially the fabless-foundry relationship. And that's why, for the better part of a decade, there was underinvestment in the area where pervasive semiconductors were needed, because the economics didn't work. The reason there's only 5 foundries of any scale left in the world is because the economics didn't work. And so this inflection point that we see today, where demand outstrips supply, in my world, it was always going to happen. And maybe it was going to be 2023, 2024, worldwide pandemic, digital transformation gets accelerated. And what was going to happen, happened sooner and much more exaggerated. And what it created was a new economic model. Why would GlobalFoundries invest with 10-year paybacks and 6% returns, right, to take the risk of adding manufacturing capability? And what we really needed to do was to lead the industry, to articulate the need that a better balance and economic model that matters, that works for both the fabless and the foundry. And we began that journey. And I would tell you, as much as you say it was a shock, it was a shock for our customers. But the other end was they had amazing demand to grow their business, but there wasn't enough supply. And maybe as an industry, we all did ourselves. We made it too easy as the foundry for our customers. And maybe they had it too easy. So we started to sign these long-term agreements. And with their vouch for us is 3 criteria: certainty, durability and profitability. The certainty is, we can't invest billions of dollars not knowing there's a use of that capacity. GF will never put capacity on for GF. We don't make products. We have no use for it. We will only put it for our customers with their commitment to it for their need. So certainty, they're signed up for that. These are take-or-pay contracts, reciprocal in nature. The second thing was durability. We don't like second-source business because you're less differentiated there. And there are some applications that drive so much demand, they need multiple sources. An image sensor processor, for example, needs millions of wafers a year of supply, right? But we look for durability, meaning it's a segment where we have real differentiation and it has long legs. And then profitability, as we can't make investments without the discipline around. And for us, that means less than a 3-year payback if we're putting capacity in an existing facility; less than a 5-year payback if we're putting a shovel in the ground, because you'll lose a whole year, right, in just building the facility. And that means that translates to ROICs of 18%, 22%, all depends. That's the model. We sit with our customers, we figure out prepayment, access fees, what's our capital contribution, what's theirs, what's the actual capacity they want that sets the corridor? We want 40 nanometer. Okay? But which version of it? You want it with embedded memory? You want it with RF? You want it with different features on it? And so these corridors become very defined for an application and customer. There's some fungibility that we can mix and match. And that's where corridors come from, customers signing up 4-year take-or-pay capacity corridors. And what it's created is a new economic model for an industry, that if you believe the industry is going to double in the next 8 to 10 years, and that will take more than $1 trillion, you can't have 5 companies or a dozen companies foot that bill. It has to be a broader participation. And I think that's where as one of the few foundries with the ability to do this, we created that model. In the beginning, a lot of shock. By July, our customers wanted long-term agreements more than we did. They needed the certainty of supply to convince their shareholders that they would have the ability to grow their business. And the way they were able to do that is they said, I signed up a long-term agreement, I used my balance sheet to create capacity at GlobalFoundries to meet my product. Sukhi, I missed anything there?

You got it right.

Great.

He got my back.

Do all the LTSAs, do they all include an increase in pricing? And do they all include a prepayment element from your customers?

So what they all include is the return on those investments, all hit that threshold or we just don't do it. Some of them have prepayments. Some have an access fee. But pretty much every one of them has cash upfront, because it helps us with free cash flow. And if it's not our cash, it's someone else, it helps that return. And so a lot of the modeling around that was to achieve a different objective, the certainty and the profitability. They're pretty much all 4 year. But more importantly, they all have, after the first year, or after every year, requires the parties to sit down and say, what are we going to do for that fourth year out? And so customers can say, hey, I'd like to mix it differently. I'd like even more. I'd like less. If they want less, I have 4 years to go find another taker for the capacity. And also introduce the opportunity to say, okay, let's revisit pricing in that next year, maybe some of that gets pulled in. But pretty much they're fixed duration, fixed volume, fixed pricing, other than inflationary clauses that allow us to be balanced in sharing some of those unfortunate headwinds that we face.

Right. And I know that you have lots of different nodes. If someone were to come to you today and say, hey, I would like to start a new LTSA, what year would it start? Because you're sold out?

We're sold out this year. We're sold out next year. We're sold out on capacity. We're still adding. In fact, I'm leaving here to go to Dresden to meet with my team, who's done an amazing job building up the capacity in our site there. Next week, I'm in Singapore. We broke ground on the Phase 1 of a 3-phase expansion in Singapore last June. And next week, we're putting the actual first tool in there. Tim Archer, the CEO of Lam, and I are going to be there to, I think, actually push the tool, right, in the ceremony. And so a lot of this capacity is being put on at customers' needs. So it's broken for. What we're trying to do now is a next round of capacity adds with customers with a longer-term view. Probably it would come online late 2025 at best given the lead times for equipment. But more importantly, again, doing it on something that has legs and durability and in concert with our customers. I think you'll see something from GF in this year where we'll be adding new capacity somewhere in the world, where we will not be able to meet with our customers or ask us.

Got you. And I think, when you put up like the list of LTSAs, I think it's met with some level of skepticism that in a downturn, that they might be hard to enforce. And I think what I had learned from you guys just recently is that the LTSA is for a corridor, and there's actually flexibility. They're not signing up for a chip. They're signing up for capacity. Is that -- so your customers have some flexibility?

I think for us, we entered all of these long-term agreements, I'm not sure we get the S in there, but long-term agreements, with a partnership in mind. And the point there was, let's go really scrub the demands. Remember, in the past, our customers didn't have to spend much time really understanding the demand. They were nonbinding forecasts, right, we would need to respond to. They spent a lot of time. They had to go to their Boards for approval. They were signing things that were, in many ways, 8-K reportable, right, requirement. And so they gave a lot of thought to what they were signing up to. Now a contract is a contract, and I get that. But remember, 80% of our design wins are a single-source business. The point of these LTAs was not to say, well, in a downturn, use software I don't. It's to create more balance in the downturn. Where we had a partnership as we entered the contract, how do we work with each other to kind of share whatever pain there is going to be? Because we know one thing. If you think the downturn means this industry goes down from there, then contracts don't mean anything. We all know one thing is going to happen. When it comes back, it's going to come back strong. And the key is how do we get through the downturn in a more partnership manner, where we both share in the reasonableness, the contracts, the framework to have that conversation. Single-source business is another way to have that conversation. So that when we come out, we come out strong. I'll tell you, I wasted 3 months in my life, March of 2020 through say June of 2020, doing all kinds of scenario planning around what if this is 3x deeper than the great financial crisis and 4x longer? When would the balance sheet run out? By August, I had my sneakers on, we couldn't make enough wafers. And I know that whatever we go through now, it's just a question of time and depth, that the world will require more and more semiconductors over the next 8 to 10 years. And we're going to be better prepared to respond to that with our customers when that happens.

Got you. So you have excellent visibility. You're sold out this year. What is left to guess? Like why would there be any variance at all between the quarterly guidance and what you report? Isn't every quarter just a lay-up?

I wouldn't call it a lay-up. We perform miracles every day called semiconductor manufacturing. We have a 15,000-person team that's just amazing. But what we do is really, really hard, really complex. And every day, we have to be focused on it. At my heart, my best as CEO, I'm an engineer, I'm in operations. And we have to be disciplined. And there's better versions of ourselves, right? This is the period of GF where we become the best at everything we do, how we run our enterprise, our ERP, our manufacturing operations. So it's not a lay-up. I think we create certainty in our business that we know how to execute. We make commitments. We take great responsibility in delivering those commitments. But it's not easy work. It's easy for me; 15,000 others, very hard.

And what's the bigger risk to meeting your projections? Would you say is it a negative demand shock or some kind of a supply issue?

Yes. There's a lot of uncertainty. I think what plays to our favor is the very nature of GF is a global supply chain. We operate in 3 continents, so we have 3 supply chains. If one gets a little tight, we actually can easily leverage. I read a note last night, I was really -- some of the tools we're installing in Dresden was short on some stainless tube, high quality. Our Fab 7 in Singapore is sending some over, right, because we have this global supply chain. And so I think while there's risk in that, we're a little bit more insulated because we have a lot more touch points in our supply chain. But look, we live in a world today that's a lot of processes are set up like you live in a perfect world and you don't. And so those are the things that keep me up. How far does the Ukraine situation go? How high does energy translate into all our businesses, right? And does that create consumers and more social unrest around that? Those are the things that keep us all up at night. But execution of our business, prosecuting of our business, we spent a long time to get good at it. Now our job is to get great at it.

Got you. And so we talked about like the supply chain issues that have been impacting the industry. More recently, there's concerns around inflation, interest rates, potentially the risk of recession, higher energy costs. Is there a framework...

Other than that business link...

Right. Is there a framework that you have or you might share with investors to think about how GF operates in high inflation environment, high interest rate environment, higher energy costs, like how should we think about how you guys work in this newer environment?

So listen, I'll let you take the lead.

Yes. So look, I mean from an inflation standpoint, we think about it in 3 main buckets, right? There's direct materials that we use to manufacture. So that's mostly -- that's covered by supply contracts that we have with many of our suppliers. And then if -- most of our contracts...

That's wafers, chemicals, gases.

Yes. So we've signed a lot of -- when we signed the LTAs, we were also signing contracts with our suppliers. So we're covered there. We do have inflation riders in contracts with some of our customers. So it's customary that when you see some of this, we can pass it along. There's another big pillar of inflation that's more people related, labor-related across the different geographies that we have. What we have done specifically is to move more towards a variable comp structure where if the company performs, its people perform, they get paid. Everybody gets paid, right? So that's something that we've shifted more towards as a part of being a public company now. And finally, there is energy costs, which is something that we watch very carefully. On the energy costs, we are heavily hedged, so 95% hedged. And that's a 2-year hedging program that we have. And so we feel pretty comfortable with where we stand. So I think in the last earnings call, we said our inflation, we did the analysis earlier this year, what that could impact our total business this year. It's about between 1% to 1.5% of revenues is what the inflation impact could be. But with all these programs we have in place, we think we'll be in a good spot.

Got it. Got you. When you talk about capacity expansion, if you look at the equipment suppliers, semi-cap equipment suppliers, some of them have been guiding down or reporting misses for the out quarter, at least for the guidance. And they themselves have a challenge getting components, not all just semiconductors. How do you think about managing that risk given that your growth is predicated in part upon adding capacity?

Yes. So let's talk a little bit about our growth and then where we were in that cycle. We measure businesses on revenue, but we get the revenue by the number of things we sell times that in the ASP. So if you look at the -- in 2020, our revenue was based on 2 million wafers, 300-millimeter equivalents of chip. But we're done with this investment cycle that requires building out within our existing footprint in Fab 8 Upstate New York, filling in all the floor space in Dresden, and our new brick-and-mortar in Singapore. By the time we get to 2024, that will be 1.6x higher, like 3.1 million, 3.2 million units. That's what this investment goes in. Now the equipment guys are suffering, right? And I think this is a great example of supply chain balance or supply chain cost and efficiency. And so I think you'll see the world shifting a little bit more towards balance than single points of failure, because you see how much economic havoc you could do to your business by missing a couple of quarters without components. For us, we made these decisions in late 2020, early 2021 about all the investment we're going to make, locked in our tool purchases. Equipment companies were kind of first in, first out. You get a slot then -- and so we've been able to maintain our ramp, our schedule. We've built good manufacturing and operations teams. We've built contingencies into them. We still have contingency in our plans. But we've seen tools move out as much as 30 days or so, and we find ways to work around it. So, so far, I think because we got there early, we planned well, we're executing just fine. But every day is a new day. And we have to make sure that our equipment companies, we help them get the components they need. It's a little ironic that some of them can't build tools because they can't get the chips that their very tools help us create. The good news is they use such a small amount of them, once they can figure out where the chip comes from, usually someone can help out.

Yes. It's a Catch 22, right? So we have a question here, and I want to work that into the fireside chat. Isn't the lesson from the pandemic that semis are so important for nations like governments are going to participate in investment, and that ultimately, the outcome will be a lot more supply coming online, supported by government investment. And so can you just share with us? I believe your Singapore facility has some support from the local government. What kind of government support have you been getting in your capacity additions? What do you expect going forward?

Yes. So there's 2 questions here.

Yes.

One is the role of governments in the past, in the future. And then I think you alluded to, is there going to be too much capacity because of all this?

Right.

Well, let me take the first one. So that's not new. Governments have participated in investments to create semiconductor technology in manufacturing. By the way, those investments paid back in spades. GF AMD has a fab in Dresden, Germany, because when Western Germany needed to reintegrate Eastern Germany, and they wanted to create manufacturing jobs, then they chose Dresden. Upstate New York Fab 8 at Saratoga County you pointed to. New York State put $1 billion. By the way, GF was to invest $3.2 billion, 1,200 employees. GF invested $14 billion in 3,000 employees. Economic trajectory of that region has never been stronger or better. So the return on that $1 billion was amazing. Our Singapore build-out. The reason we're there in the first place is the Singaporean government. The EDB has always been very supportive in doing their part and helping the economic development. And the facility we're doing there has got that. So I think it's not new. I think what's new is more regions are recognizing how important semiconductors are to sovereign security, supply chain security and economic security. And that having manufacturing jobs in high-tech industries are important, and they want to make sure it happens there. With regards to overbuild, we track this really, really carefully. First of all, we play 12-nanometer and above. A lot of the CapEx you hear about is the single-digit nanometer stuff that is really capital intensive. So in 12-nanometer and above, we look at 2 types of build-outs. One, where there's actual work going on, either there's construction going on or tools are being put into a building that's already constructed, versus someone went to a farm somewhere, cut a ribbon and said, we'll be back in a couple of years to build capacity. So if we take the first one of those, and then by the way, we say what's putting announced in China doing that versus the rest of the world. We started this year as an industry, GF much higher than that, with about 8% to 9% more demand than capacity. We see that capacity coming on in the next 5 years, all the ones that are in, that will add capacity somewhere around the 5% range, while we think the industry is going to grow greater than that. So this notion that, well, all this is coming on. Pretty soon, we're going to be floating in semiconductors. I don't see it. I mean we've -- because the vantage point we see where things are coming. Now what's our insurance against we're wrong, we don't build it for us. There's nothing speculative in our investments. It's customers with their prepayments, ownership of this capacity. So we think ours get used before any speculative capacity would be used. But that's kind of our view of the world. I think governments want this. They're going to fund it where they can for those reasons. And we'll be very careful of how we make investments to create capacity, mission built, customer-centric built capacity.

Great. Well, we ran out of time already 5 minutes ago. So we could talk for a long time. Fascinating topics.

Because they want to listen for longer.

Right. Nobody left for lunch.

Right. Tom, Sukhi, thank you very much.

Thank you, Mark.

Thanks very much.