GLOBALFOUNDRIES Inc. (GFS) Earnings Call Transcript & Summary

Good afternoon, and welcome to the first day of JPMorgan's 52nd Annual Technology, Media and Communications Conference. My name is Harlan Sur. I'm the semiconductor and semiconductor capital equipment analyst for the firm. Very pleased to have John Hollister, Chief Financial Officer at GlobalFoundries, third largest semiconductor foundry in the world, leader in specialty, mature manufacturing technologies, targeted at segments like analog, power management, RF, wireless, wired networking connectivity, targeting the comm infrastructure, mobile, IoT, automotive and industrial markets. Team reported solid results, constructive guidance in a tough macro demand environment just about a little bit over 2 weeks ago. So John, thanks for joining us today.

Thank you, Harlan. It's great to be here.

Now you've been with GlobalFoundries now for about 4 months. You came to the team with a great track record, 20 years of both finance and operations expertise. Chief Financial Officer at Silicon Labs for over 10 years. Wanted to get your thoughts of what you've observed so far, right, the team, the execution, the growth opportunities in front of GlobalFoundries.

Yes, Harlan, you bet. So I've been really struck by the quality of the team at GlobalFoundries. I mean, that's part of what attracted me to the opportunity in the first place. But it just really has been interesting to observe the culture that the company's created and how it really perminates throughout the organization of very hands-on problem-solving culture that's just also rooted in the deep experience of the folks that we have on board and also from a diverse set of backgrounds, both in terms of corporate backgrounds as well as geographical perspective. So it's really been interesting to get to know the team, and it's been a real pleasure. I'm very excited about the growth opportunities ahead for the company if you look at our position in: the smart mobile devices arena; what we're doing in automotive, which has been a tremendous growth driver for the company in the last couple of years; the opportunity we have in IoT, which is really remarkable over time here; added comm infrastructure and data center, where we've seen some amount of drop in the business, but I think we've reached a point of stability in the business and see growth ahead in the area of high-speed communications and power delivery. So very excited about the growth ahead and look forward to continuing to progress here.

Perfect. Well, again, thanks for joining us today. The semiconductor industry down cycle started in second half of 2022. Total industry revenues were down about 8% last year. The good news is that your customers are pulling out of the current down cycle. Many of them saw their businesses bottoming second half of last year. More of them are seeing their businesses bottom first half of this year. Second half of this year is looking more constructive across-the-board, including for the GlobalFoundries team. Could you just share your thoughts on the current state of the industry and impacts on GlobalFoundries' recent earnings update?

Sure, yes. We had a tremendous spike in business in 2021-2022 time horizon. And that led to an accumulation of inventory, actually, across end markets and customers and then, of course, the fed hikes rates 10x or so in a row in an unprecedented move to try to curb inflation. And that's had some impact on the demand environment and has led to an accumulation of inventory out in the channel and at our customers, and that's led to business being down as -- throughout the industry. I think the good news here is we're starting to see rationalization between demand and inventory levels, and we're starting to see inventory come in. In smart mobile devices, for example, we're seeing some pockets of improvement in inventory levels. It's been a bit slower than one might've anticipated and hoped for from the beginning of this year, but it is starting to happen. And we think that the end market demand is stronger than what the shipments would indicate due to the inventory levels. And that's even more true in IoT, where we see significantly high levels of inventory. And again, companies are undershipping the end market demand as the inventory draws down. And the good news there is that the demand signal's actually stronger than what the underlying shipments would indicate as the inventory comes in. In the automotive space, you've seen some inventory there as well. But the good news for GlobalFoundries is we have secular drivers in automotive that have kind of bucked the trends on that and allowed us to grow that business, even in light of a more challenging inventory environment. So that speaks to our market share gains, which speaks to our differentiation and the work that we've done over the past few years to gain design win momentum.

The team has done a really good job of navigating through the current industry downturn, right? Your peak to trough revenue declines this down cycle. Revenue-wise was about 25%. Your mobile RF customers were down 35% to 40% peak to trough. Industrial customers, down about the same. IoT and comm infrastructure, data center customers, down about 50% to 60%, right, revenues peak to trough. So when we think about the foundry industry, we always tend to think about the foundry being sort of at the back of the value chain, right? So it's -- the weakness tends to percolate sort of slowly all the way down and then slowly sort of all the way back. And your -- historically, foundry revenues tend to be magnified in terms of peak to trough relative to your customers. But again, you guys have done significantly better relative to your customer base. So how has the team been able to better weather this down cycle significantly better than the shipment profile of your customers?

Yes. I think it's 2 things I would point to. One is, again, the secular growth in some of these end markets that have allowed us to power through the downturn dynamics, and I'm thinking of automotive in particular. I mean, if you really think about automotive, we did over $1 billion in the automotive end market in 2023. We did around $300 million in 2022. So just as you mentioned at the top, this downturn for the industry overall began in late 2022. And for the year of 2023, we added $700 million of new revenue in automotive. So that's an example. I think our long-term agreement frameworks have also helped us in this regard by providing for more stability in the supply-demand equation and also contractual terms that have allowed us to somewhat offset some of the capacity shortfalls that otherwise rose through the downturn cycle. And that's going to remain an important part of how we operate moving forward in terms of having long-term agreements with customers. We -- as of the end of 2023, we continue to have around $20 billion of lifetime revenue captured in our LTA frameworks, which will help us going forward. And as we move ahead, we're entering into new long-term agreements with customers as evidenced with a major customer earlier this year.

With the -- on that note, with the long-term agreements or LTAs in place, it does seem like customers are more responsive, enabling you guys to reallocate capacity and wafer starts more efficiently, which was key to effectively, as you mentioned, managing your business during this downturn. As the business improves, LTAs provide your customers in key end markets, like auto, industrial, aerospace and defense, with confidence on assurance of supply, especially as 90% of your design wins continue to be sole-sourced business, right? And this was the rationale, I believe, for the LTA extension agreement you signed, for example, with the Infineon team right at the beginning of this year, primarily to support their leadership and growth outlook in the auto semiconductor market. On the flip side, you did have one of your RF smartphone customers terminate their LTA agreement with you at the end of last year. They're still a very strong customer for you. So is the team seeing a more, call it, diversified profile of business engagements now, more balanced between LTA commitments and sort of normal sort of book-and-ship type arrangements?

Yes, somewhat, Harlan. That's a fair assessment. And I think you have it right on the fundamental drivers of the LTA framework. It's really about providing surety to both ourselves and our customers around the supply-demand equation and that when we're investing in capacity, we're getting support for that from our customer base. And that -- likewise, when they're entering into long-term arrangements with us, they have confidence that our supply will be there for them. And during the shortages that we saw a couple of years ago, it really spoke to how important that is. If you think about automotive customers unable to ship a $70,000 vehicle because they can't get their hands on a $5 semiconductor, this is very important that they have assurance of supply around that. We're getting -- learning along the way. And as we proceed through the course of time here, I think you'll see LTAs not necessarily be as homogenous as they were before and oriented more around commitments that are more closely track the life cycle of the underlying products. And just as you say, just as a customer may decide to hold on an LTA, it doesn't mean that the relationship is over. We continue to have robust relationships with customers even where they may have decided to exit the LTA framework altogether.

We cover -- I cover over 20 semiconductor companies, many of them who are your customers. And we always hear quite constructive things about the technology portfolio, right, whether it's RF SOI, fully depleted SOI, your nonvolatile microcontroller products, which is targeted for IoT, for automotive, for a number of different applications and some new emerging technologies like, for example, some of your silicon photonics work and compound semiconductor work that you guys are doing. It's always been, I think, the big differentiator as it relates to -- and we haven't gotten an update on this recently, so I'd like to get your thoughts on it. But I think if I rewind back 12, 18 months ago, I think the percentage of your shipments that were sole-sourced were in that sort of 2/3 percent of your total shipments were sole-sourced business to GlobalFoundries, but your design win capture was more like 90% sole-sourced business. Is that still the right metric?

Yes. Yes. That's roughly right. Yes.

In terms of your capacity ramp, team is on track to increase capacity to about 3 million annual wafers per year exiting this year. Assuming that the industry is back on a growth trajectory next year, driving mid-single digits type of growth CAGR, mid to longer term, right, walk us through the capacity expansion across your fab networks, timelines to get you to your targets. And has the recent end demand weakness pushed out your CapEx or capacity expansion plans?

Yes, certainly. So that's right. So our current footprint is roughly 2.8 million wafers per year. As we exit this year, that will grow to roughly 3 million wafers per year. And if you think about our rough ASP, that's about $9 billion in wafer revenue potential. You add in the non-wafer revenue portion of our revenue mix, we're -- by the end of this year, we're built out to be roughly a $10 billion top line company. So we've got ample capacity to grow into. And as we can increase the utilization rates in production, accordingly, that will strongly benefit our gross margin outcome as well as we move forward. Now as we continue to grow the business, and we are strong believers in the growth of the industry, obviously, we think that the forecast of this becoming a $1 trillion industry in the next decade or so is right. And as that demand comes back, there's going to be the need for more capacity over time. The good news there is we've got the ability to grow our capacity throughout our jurisdictions of operation in the United States, in Germany, in Singapore, potentially even in Crolles in France. And the thing to keep in mind is we will navigate that and build out that capacity expansion in line with demand. We're not going to get ahead of ourselves. We're going to follow the demand signal, work with our customers, work with our government partners to activate that capacity as it's needed over time. And again, the good news is a lot of the frameworks that are necessary for all of that to happen are already in place or well on their way to being finalized.

Actually, you brought up a good point, which is the geographical diversity, right, of your manufacturing footprint. So if you take your 2.8 million, 3 million total capacity exiting this year, how does that break out roughly by geography, Asia, Europe and North America roughly?

Yes. It's roughly equivalent across our geographies.

Got it. And then you did mention the team has strong government -- you've got a strong government relations team, and we saw that the EU and French government recently approved the grant funding in relation to your partnership with STMicroelectronics in Crolles, right? And then when you couple that with the $1.5 billion of grant funding from the U.S. CHIPS Act, how is the team going to deploy these dollars and roughly over sort of what period of time?

Yes. We're really talking about a decade kind of time horizon, Harlan. As you look at the full activation of our capacity expansion opportunities and precisely how we go about that, TBD. We'll have to monitor the demand signal and see exactly how that activates over time. But suffice to say, we have good opportunities to build out capacity as demand recovers.

The one question -- it's sort of a more tactical question, but we might have talked about this before. But on the CHIPS Act dollars, Biden signed that into law -- into a bill, I think it was middle of '22, right? Obviously, dollars haven't even been appropriated yet, but you guys know at least what you're going to get. But Biden signed it in middle of '22, you guys obviously spent CapEx in '23 and part of it in '24. And some of it, I would assume is on some of your U.S. manufacturing capabilities. So can some of the grant dollars be appropriated like retroactively so you could get a boost in potential margins like right away?

Yes, not necessarily. I mean, the CHIPS framework that we have in place is really oriented around 3 programs. You've got what we refer to as [ 8.auto ], which is diversifying the multi-fab and the capabilities that we have there beyond FinFET technology into 22 nanometer, 28, 40, 45, et cetera. It's also around expanding the multi-fab over time. That's referred to as 8.2. And it's also around modernizing and enhancing the Burlington fab. Those are really the 3 programs. And it's really prospective, Harlan, looking ahead as demand recovers and we can activate dollars to service forward demand.

Got it. Any questions from the audience before I carry on? If you do, please feel free to raise your hand, and we'll get a mic over to you. We've got one up here.

Very sorry for the short-term question, but you talked about this sort of recovery panning out and inventory getting burned down. I mean, how quickly could that ramp accelerate? I mean, if demand comes back, it could happen pretty quickly. And what would you have to see to kind of get that sense of, hey, stuff's really back and -- it's going to come. I'm just kind of curious like how quickly could that happen for you.

Yes. I think if you break it down by end market, we see smart mobile actually having the opportunity to grow modestly this year already, call it, low single digit. In automotive, we see the ability to grow that business mid- to high single digit in 2024. And we think in comms infrastructure and IoT, we're at a stable point of revenue with Q1 as a base, and we'll see how the demand can recover. In IoT, in particular, it's really going to be around the inventory drawdown. And that really is the key across the industry, across the markets that we serve is seeing the rate and pace of inventory stabilization and seeing what can happen on the demand front.

Any other questions? So one of the things that investors have been concerned around has been some of your compute and data center customers moving away from your 12-nanometer FinFET technology in Fab 8 in Malta, right, to single-digit nanometer technologies, which GlobalFoundries obviously does not do. Tom has been articulating some of the dynamics around diversifying the technology mix for the multi-Fab 8. It's a smart strategy, right? You talked about 40-nanometer embedded memory MCU, RF SOI, FDX and, on the mainstream, 12-nanometer FinFET capacity mixing in new customers and applications as older programs phase out. So John, maybe you can expand on this, timelines on phasing in new technologies and the success that the team has been in just backfilling the plain vanilla 12-nanometer FinFET programs.

Sure. I think, first, it's important to understand and acknowledge that the FinFET corridor that we have is already more diverse than, I think, folks may realize. So yes, some of that technology is serving the comms infrastructure and data center market, but we also have FinFET servicing automotive applications as well as IoT applications. And actually, by far, the largest end market that is served by our FinFET technology is in smart mobile devices, actually, by far. So first point is that the FinFET is already more diversified than just serving data centers, is the first point. Next is actually enhancing and diversifying the FinFET offering itself to add new features to FinFET to address even more market opportunities. And then finally, as you said, is looking at the Malta site, the Malta infrastructure that we have and diversifying the full range of GlobalFoundries technologies into the Malta site, and that's underway. We have actually begun those technology transfers and, they're well underway as we progress through fiscal '24.

On the rewinding back to sort of navigating to the environment, capturing new opportunities, especially in automotive, right, and I keep thinking about how your automotive business went -- in '22, went from like $300 million to over $1 billion last year, right? And you said it's going to grow another mid- to high single digits. Is that also 12-nanometer FinFET-based technology on automotive? Or is that BCD? Or what is that?