GLOBALFOUNDRIES Inc. (GFS) Earnings Call Transcript & Summary

All right. Good morning, and welcome to JPMorgan's 50th 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 Dave Reeder, Chief Financial Officer; and Sukhi Nagesh, VP of Corporate Development and IR of GlobalFoundries. They are the third largest semiconductor foundry in the world, leader in specialty and mature manufacturing technologies targeted at segments such as analog, power management, RF, wireless, wired networking connectivity, targeting the communications infrastructure, mobile, IoT, automotive and industrial markets. The team reported strong results a couple of weeks ago, and it's been a busy earnings season. So I thought it'd be great for Dave to sort of start us off with a summary of the March quarter and June quarter outlook. And then we can go ahead and kick off the Q&A. So with that, gentlemen, thank you for joining us. Let me turn it over to you.

Thanks for having us, Harlan. And just wanted to say great event and good attendance. Good to do this in person again. Look, we did report earnings just a couple of weeks ago. We did have a strong Q1. So we -- revenue was about $1.94 billion. That's up about 37% on a year-over-year basis. Volume was up about 14 points, and ASP's about 20%. We had good earnings. The high end of our guidance was about $0.27 earnings per share. On an adjusted basis, we delivered about $0.42. And then, of course, for our guidance for Q2, our midpoint of the range was about $1.97 billion on the revenue side. And then we had an EPS range of about $0.43 to $0.48. So great year-over-year performance from a delivery perspective in Q1, good guidance for Q2. And that's really underpinned by the demand that we see in the market. We've got long-term agreements that essentially have consumed all of our capacity. We're sold out for 2022 and for 2023. Our customers have contributed more than $3.5 billion in prepayments and access fees for us to grow the capacity to be able to support those long-term agreements. And we just see a very constructive demand environment on a go-forward basis.

I appreciate the opening commentary. And to your point, I mean, demand appears to be quite strong overall for the industry. Chip demand continues to outstrip supply across most end markets, right? But there's also a lot of cross currents, too, COVID, supply chain disruptions. There are some pockets of weakness in the end markets. But the team is very well diversified by end markets. And so you have the capability to reallocate capacity from weaker segments to stronger segments. So help us understand for the full year, right, relative to your high teens revenue growth outlook profile, what end market segments are growing above and below the corporate profile?

Sure. And you're certainly right about a very constructive environment. We entered the year with about 25% more demand than we could satisfy. And so we certainly have the ability to be able to remix some of that demand in the event that we see any changes in mix as determined by our customers. With respect to how the year is playing out so far and how we expect it to play out, we entered the year articulating that home and industrial IoT would be our fastest-growing segment. That remains true. It certainly looks to be the forecast for the year that, that will remain true. We talked about how comms -- or communications infrastructure and data center would be our fastest -- our second fastest-growing market. That certainly looks to be true as we look at our demand profile between now and the end of the year. And then we also communicated that automotive would grow quite nicely for us. Smart mobile devices, I would say, in line with enterprise average. And then personal compute, which is an area that we really pivoted away from a couple of years ago, personal compute, we expect to decline from about mid-single to high single digits of our business in '21 to sub-5% in '22. And that's playing out exactly as planned.

In the March quarter, your book-to-bill approached 1.5x. Your backlog grew sequentially. Your backlog grew year-over-year. Lead times continue to extend. Clearly, the demand profile environment is quite strong. Given your LTAs, your backlog, capacity expansion plans, when does the GF team start to see some closing of the supply-demand gap? Clearly, it doesn't appear to be this year, but is it 2023? Is it 2024? What's the team's view?

Well, as I mentioned, we're sold out for 2022. We're sold out for 2023. We're majority covered for demand even in '24. And this is in an environment where we're bringing on capacity. So even capacity that we haven't yet brought on is already allocated through a long-term agreement with our customers. We think demand and supply remain quite tight, certainly through the remainder of this year and probably well into 2023. Sukhi, anything you'd add?

No, you said it right. It's pretty filled up.

So there's been a lot of focus clearly for the past 24 months on manufacturing localization efforts, right? And so obviously, we're focused on -- you guys are focused on the U.S. CHIPS Act bill, right? House and Senate are finally coming together to try and craft a unified $52 billion CHIPS Act bill, strong bipartisan support. My understanding is that the target to get the unified bill signed off by Biden is before the end of this summer, before the August recess. You guys have a really strong government relations team. What are your views? And what are your views, and assuming it does get passed before the August recess, what's your sense on appropriations and funding timelines from there?

Sure. We agree with all those comments. We think good progress has been made. We think there's broad bipartisan support for the CHIPS Act. And so we think there's a real opportunity for us to bring back some domestic manufacturing here to the U.S. With respect to timing, I'm going to shake my magic 8 ball and give you whatever answer comes up. I'm not sure how quickly or slowly things will get done on The Hill, but they are in reconciliation. They do have 2 bills that they're working through to merge into one. And I do believe that there is broad bipartisan support and that something can get done this year. What's interesting to note is that Europe who was behind the U.S. has really accelerated their efforts. And so if you had asked me this question, Harlan, 3 months ago, I would say the U.S. is significantly ahead. I think there's the potential based on what I see now for Europe actually to get home on some deals before the United States gets their CHIPS Act in place. So the good news is we've got manufacturing in Asia, in Singapore. We've got manufacturing in -- on the continent there in Europe in Dresden, Germany. We've got manufacturing in the U.S. We've got a long list of MOUs with our customers signed. And as we work through those MOUs and we work with the state and local governments, we think we have the opportunity to expand our capacity in partnership with customers and governments in all 3 regions.

To give us a sense on the potential benefits to GlobalFoundries, I don't know if you can look out over the next, let's say, 3 years, look at your expansion plans and your CapEx profile, what percentage of your CapEx over the next 3 years is going to be U.S.-focused? Obviously, Malta. And obviously, you're investing in Dresden now. But over the next 3 years, what percentage of the CapEx is going to be Europe-focused and U.S.-focused?

So I'm going to take 3 years from '21, '22 and '23. We're going to spend about $10 billion over that period. Our customers are actually funding about 1/3 of that. So that $3.5 billion that I mentioned earlier, our customers are contributing to that capacity expansion. And so if you were to take that $10 billion and kind of parse it by region, you've got about, call it, $4 billion to $5 billion that's going into Singapore, the majority which is that new facility there in our Singapore campus, enabling about 450,000 wafers of incremental capacity. So that's $4 million to $5 billion. You've got about $2 billion to $3 million that's going into Dresden, Germany. That ramps that capacity from about 300,000 wafers up to about 850,000 wafers out of that existing facility such as tooling that existing facility and then of course, the remainder in the U.S.

Perfect. Any questions from the audience? If you have a question, raise your hand and wait for the mic. So we've got one up here. Just hold on for a second.

[indiscernible]

Want to repeat the question?

Yes. I think the question was gross margins versus peers. Is that correct?

Yes, and the plan to get them up and why they're lower.

Sure. Sure. So the single biggest challenge that GlobalFoundries has had, and we talked about this in some detail at the roadshow for our IPO. The single biggest challenge is that we had this large fixed cost footprint, but it wasn't fully tooled. And so what we had to do in 2021, I mean, there's kind of 3 years of gross margin expansion. 2021 was the year of normalization of depreciation and fixed cost absorption. And practically speaking, what does that mean? What that means is that we had a fixed cost footprint in Germany, as an example, where that facility was tooled to about 300,000 wafers per year, but it can actually support with fully tooled facility about 850,000 wafers. And so you have capacity. You've got the physical building, the process piping, the gas farm that can support 850,000, but you're only running 300,000. So once you tool that facility for about 15% incremental operating cost, you can take all those expenses and divide by almost 900,000 wafers versus 300,000 wafers. And as you do that, you get massive fixed cost absorption and fall-through. And that's actually what you're seeing play out between '21, which is one of the bigger years. '22 is also a significant year, and then a little bit of that will also happen in '23 and '24. And so similar to Dresden, you also see that phenomenon play out in Singapore on a fixed cost absorption basis and in the U.S. In addition to that, you also have some normalization of depreciation. So as that depreciation normalizes, you overbuilt the footprint initially, that starts to decline as a percentage of revenue as well as roll off. That adds some additional benefit to gross margin as well. So '21 really fixed cost absorption and depreciation. '22, a lift in ASPs. So '22, we've communicated you get about a 10% increase in enterprise ASPs. And that happens throughout the year. It didn't all happen January 1. As those LTA contracts ramp, it bleeds in throughout the year. So ASPs increased. You also get fixed cost absorption. That's really the story of 2022. You get a little bit of benefit of normalization of depreciation, but it's primarily driven by fixed cost absorption. And it's also driven by ASPs. By the time you roll into 2023, you're starting to hit on all the cylinders. We're mixing up the business. So during our pivot in 2018, we migrated to end markets that are accretive to GF. You're starting to see some of that play out in our ASPs, almost $2,800 per wafer. So our mix of the business is increasing, more accretive markets, automotive, good example. We're also going to continue to get a little bit of fixed cost absorption. And then, of course, you get a little bit of normalization of depreciation, and you also get the roll-off of our East Fishkill facility, which adds about 3 points to enterprise gross margin. So all those things come into play in 2023. So what we've communicated is that our long-term model is 40% gross margin, and we talked about long term being in that 2024 time frame.

And then, I guess, maybe as a follow-up to that. On top of that, I think in steady states, I think on the incremental revenue growth once things start to normalize, I think you guys have talked about 55%, 65% incremental gross margin fall-through. Is that kind of the way to think about it?

That's correct. So the -- once we reach steady state, in other words, once this footprint is built out, we're getting the fixed cost absorption and entitlement out of our model then every incremental dollar essentially falls through at kind of that 60% rate.

Perfect.

Okay. So do you envision the CHIPS Act going back to that topic and the $52 billion to be a competitive process, amongst the fabs out there? If that's the case, what's your proposition to be one of the major beneficiaries from that?

Sure. I think our proposition would be that we are currently the only pure-play foundry in the United States. And the portion of the market that we serve is greater than 12-nanometer portion with really an emphasis on connectivity and power management. I think that's a pretty differentiated position in the market. So if you think about 5G connectivity, we don't think the 5G transition is happening because of GF, but we think it doesn't happen without GF. We have a real franchise in the front-end module with that RF connectivity. We have some real franchises on the connectivity side and the power management side. We have a growing and burgeoning automotive business, where there's a lot of interest in domestic manufacturing. You've seen some of the agreements that we've signed there. And of course, we're a large manufacturer here in the U.S. If you look at this Northeast corridor between Malta, East Fishkill in Burlington, we're employing roughly 5,000 people, rounding up slightly. So we're a big contributor to the local economies here, and we've been a great partner to the federal government for a long time. Sukhi, anything you'd add to that?

Yes. We also have a pretty sound and long-standing defense business that's going to continue to go with us in the U.S., especially in the Northeast. That's going to be a big play for us when we continue to work with the Federal government here.

Question back there?

Can you just give us an update on gas? And if you're not impacted, who is?

Are you referring to Russia, Ukraine?

Yes.

Yes. So I wouldn't have thought this before the conflict in Ukraine. But one of the biggest benefits that you get, but -- besides just general supply chain security in being on 3 continents is that you have a lot of local suppliers that are fully qualified at a company like GlobalFoundries. So we manufacture in Singapore. We manufacture in Dresden, Germany. And we manufacture here in the U.S. And we actually have cross-qualified products across all of those facilities. And we also source from a lot of local suppliers in those regions. And so one outcome of having this type of diversification is that when you have a conflict in an area like Eastern Europe, we're able to take a supplier that's been fully qualified in a region like Singapore or a region like Malta, or Northeast U.S. And we're able to take that qualified supplier and through a cross qual with a product that's already qualified in both locations, use the output from that supplier in another region in a pretty effective and efficient way. And so by and large, we've been able to overcome all of the supply chain issues associated with the Ukraine-Russia conflict. Now there are some challenges in the region, right? I think the challenges in the region would be the laser gases. They -- Ukraine provides about 30% of the laser gases for the industry. I think some of the fluorites, they provide some of the fluorites for the industry. We've largely been immune to those effects. We have suppliers in other regions that are ramping up for the industry, not just for GF. We feel like we're largely mitigated. But the industry, if it lasts for a long period of time, the industry could suffer if those other suppliers could not ramp the volume required.

Let's talk about the differentiation because I consistently hear comparisons between the GF team and some of your, what I call, plain vanilla CMOS players. Well, like UMC, like SMIC in China. But your process technology is quite specialized. And I think a reflection of that is in the percent of sole-source business or products that customers run only on your process technologies and nowhere else, right? Good example is analog devices. We just found out recently that their market-leading battery management solutions for EV is sole-sourced to GlobalFoundries. In silicon photonics, for example, the leaders of the optical market, Marvell, Broadcom, NVIDIA all sole-sourced to you guys for their silicon photonic parts. So I think the team has continued to add enhancements to the portfolio. So what percentage of your revenues or what percentage of your design wins are sole-sourced to GlobalFoundries?

Yes. So in the first quarter, about 2/3 of our total revenue was sole-source revenue, and about 80% of our design wins are sole-source design wins. And that's been consistent now for, call it, the last 1.5 years or so at those types of levels. So over time, I think what you'll see is you'll see our revenue start to kind of asymptotically approach the design win numbers, the 80% design win. And I think what it really speaks to is it speaks to the differentiation that we have in our model. So we've got a real franchise on the connectivity side of the house. We've got a real franchise on some of the power management by CMOS, BCD side of the house. Sukhi, anything that you'd add to that?

Well, the other thing you would want to look at is in general, you pointed it out earlier, our wafer ASP is about 30% to 50% higher than some of the Taiwanese players, right? You don't get to that kind of ASP increase if you don't have differentiation. That really speaks to a lot of our differentiated technology that we have.

Another reflection of the differentiation, so post IPO, you guys had 25 significant long-term agreements under your belt. In 4Q of last year, you added 5 more LTAs, so 30 in total in 2021. Our estimate is well north of somewhere between $20 billion and $30 billion in future commitments, right? Supply-demand gap backlog for your customers has continued to increase through the first half of this year, and they want access to your technology, right? So given all of this, I mean, can you give us a sense on how many LTAs the team has signed through the first half of this year?

Sure. I think the most relevant statistic here after I'd already told you that we were sold out for '22 and '23 is how do customers see the longer-term supply and demand imbalance. And so what I'd point you to is some of the commentary that we had from our fourth quarter results and our first quarter results. So our fourth quarter results, we commented that we had more than $3.2 billion of customer prepayments and access fees. Most recently, just a couple of weeks ago, I highlighted that we had more than $3.5 billion of customer prepayments and access fees to bring online capacity. That's not capacity for '22 and '23. We're sold out. That's capacity that comes online in '24, '25 and '26. So the actual most relevant statistic at this stage is not could you be more sold out for '22 and '23? That's not the most relevant statistic. The most relevant statistic now is that are customers still giving you money contributing their balance sheet to grow your capacity that won't even ramp for them until '24, '25 and '26? And the answer is yes, they are. And I think what they're telling us is the same thing that we're seeing in the industry, which is there is a long-term structural supply and demand imbalance. And it looks like it's getting better, but it doesn't look like it's being solved over the next couple of years.

And that's a good lead-in to my next question because back in February, the team did some analysis on industry supply growth for mature and specialty manufacturing expansion, right, over the next few years. And the team's conclusion was that demand for mature and specialty was going to continue to trend above supply growth for the next few years. And interestingly enough, during the current earnings season, a few of your customers like NXP and Microchip also agreed with your view as they look at their internal and external supply expansion plans. Can you just walk us through this analysis because I believe that many are just looking at the total magnitude of total semi industry CapEx and coming to the conclusion that not only is leading edge going to potentially be an oversupply, but mature and specialty has the same potential issues.

Sure. So let me parse a little bit here and also provide some context. When you look at top line dollars being spent in the industry, the first thing you have to do is you have to take a step back and say, where is that being spent? Well, about 90% to 95% of all the headline dollars that you see being invested to expand capacity is actually in single-digit nanometer. So it's not in a space that we play. That's a space that's really dominated by digital compute power, and it's a growing space, but it's not a space where GlobalFoundries plays. GlobalFoundries plays in this 12-nanometer and above space. And that's a space that's really dominated by macro trends around connectivity, power management and the electrification of the automobile. And so when you look at that portion of the market, what we see is we see that, that market is growing capacity over the next 5 years on a compound annual growth rate of about 4% to 5%. And about 2 points of that is actually growth in China. And so the first question you have to ask yourself is, to domestic companies in the U.S. and Western Europe, how much of their new design wins do they want to put in China? I'll let them answer that question. But from our perspective, we see some hesitancy there, given some of the policies as well as some of the licensing challenges that have been had or have occurred over the last couple of years. So supply coming on 4% to 5% with China, ex China supply coming on 2% to 3%. Then if you look at the long-term growth of the industry, it's growing at about 5% to 10%. So in a best-case scenario, you see the rough supply-demand balance. At a worst-case scenario, you see something like a handful of points imbalance where demand outstrips supply. And that's the way we see it. I think you've cited some others who see it in a similar way. And I think there's an IDC report that's been published in the last month or so that has a similar perspective.

We've got a question from one of the clients that are online. So any views on Intel Tower acquisition implications for the competitive landscape?

I would say we don't see -- we didn't see a lot of Tower in our customers' kind of virtual lobbies over the last couple of years. They don't have a significant 300-millimeter footprint. I think similar statements on the Intel side of the house. They don't have a significant presence in the spaces where we compete today. We see Intel largely competing with TSMC and Samsung on the leading edge. We think the addition of Tower for them still requires some build-out of the ecosystem. It's not sufficient to have a fab. You actually have to also have the ecosystem, the hardened commercial platform, the IP blocks, both foundational as well as specific to be able to deliver a full solution to the market. And we don't see them in a lot of the virtual lobbies of our customers today. Sukhi, anything you'd add to that?

Yes. From a technology standpoint, there's probably just a couple of areas we can have overlap with Tower. And in both of those areas, we have been taking share from them organically, right? So that would tell you that we were not really seeing a whole lot of competition from Tower prior to -- prior to this [indiscernible] acquisition.

Any questions from the audience? Hold on for a second. Right up here.

Just wondering if you could speak to the announcements that were made that -- since you made announcements in Japan in terms of building a new fab, it's kind of geographic angle with Sony as a partner. Intel made an announcement in Germany because they are building out. So I was wondering, in the down payment you've taken, assuming you have all these LTAs spread out to many customers. Could we think of maybe capacity had been more geographic? And what was specific angle for a lot of these broad partners or is it something that's more concentrated with the geographic angle in terms of the capacity.

Sure. Let me answer that question a couple of ways. I think the first thing I'd do is I'd point you to the expansion that we're doing in Singapore. We're not putting on capacity in Singapore for GF. We're putting on capacity because our customers have committed long-term agreements, and they've actually committed funding to it. So what that Singapore expansion looks like is it's a roughly $4 billion project that enables about 450,000 wafers per year of capacity. Our customers contributed more than $700 million to that incremental capacity. The government in Singapore there contributed about $0.5 billion in grants. They also committed a little bit north of $1 billion loan, 20-year duration less than 2% interest. And so what you see in Singapore is similar to what TSMC announced after we had started that deal, which was a partnership between government, customer, manufacturer. And we think that's the type of partnership that we can do and we can utilize to expand in Europe. We think we can also expand in the U.S. in a very similar way where you have the foundry or GlobalFoundries combined with customers, combined with government, expanding domestic manufacturing to secure the supply chain.

Any other questions?

Can you get to all the tools that you need? And if not, what would delays mean? And also if you have to allocate particular tools through particular processes, how do you make that decision?

Sure. So we placed our orders for tooling really -- probably centered about the first half of 2021. So here we are, we're sitting essentially in the first half or the middle of 2022. And I would say, by and large, the wafer equipment industry has delivered as planned, by and large. Now what we are seeing is we are seeing some slipouts of tooling probably in the magnitude of, call it, 15 to 40 days. If you had to pick an average, call it, 25, 30 days. And so the implications for a company like GlobalFoundries, where we -- our intentions were to receive about $1.5 billion worth of equipment in the fourth quarter based upon the POs that we placed about a year ago from today. They're probably running about 30 days late. So practically speaking, there's about $500 million of CapEx that will slip out of the fourth quarter of '22 into January of '23. Now in terms of our ramp plans, we had included contingency in our plans to be able to accommodate that type of slippage in delivery. It's a challenge. I mean, I'll be honest with you. The -- it's a hand-to-hand combat every day on making sure that we get the equipment that we need, the spare parts that we need to ensure that we can hit all of our ramp plans. But I would say, by and large, they've delivered very well for us. I think we got our POs early. We're in the queue from a FIFO perspective in the right spot. And from an expansion of capacity perspective, all of our externally communicated plans to grow capacity are well intact, and we even have a little bit of space.

We're seeing a trend in the market. We're seeing this resurgence of end customers, OEMs, cloud titans, right? They're starting to design their own chips. And we're also seeing these same customers taking more ownership of their supply chains to prevent similar supply-demand dislocations like the one that we're currently going through today. Your partnerships with Ford and BMW encompasses this, right? They're working directly with you to secure supply for chips -- for their chip suppliers, but they've also talked about potentially doing some of their own in-house chips as well. You also have end customers like Cisco, Microsoft and Amazon as direct customers. Do you see that more cloud titans, auto OEMs, consumer device OEMs coming to the table to talk to GlobalFoundries? And how many of these types of discussions are you having? And how do you see this benefiting GlobalFoundries and the industry sort of longer term?

Yes. I think coming out of this really large the last couple of years of supply-demand imbalance in the semiconductor space, everyone's just become more silicon aware, right? You used to not think about where your chips actually came from, especially if you worked with maybe a fabless company or perhaps on the automotive side on a -- with a Tier 1 company. And I think the one thing from this supply-demand imbalance that you can take away is that people are much more aware now of where their silicon actually originates. And I think that sensitivity, that silicon awareness has led a lot of companies to want to control their own destiny. And so I think what you're going to see in the market is you're going to see a lot of different models. I think maybe more mature silicon-aware companies that have the capability will try to do -- some of them we'll try to design their own chips for areas in which they can be differentiated. I think you'll also see models where some of the OEMs want to reserve a capacity corridor and perhaps direct some of their designers or their fabless companies to use that corridor. And then I think you'll continue to see some of the traditional model as well. So I think the outcome of this supply-demand imbalance is that you'll see companies that have been negatively impacted in their end markets, they'll want to act to control their own destiny. And I think that will take several different forms.

Part of the team's gross margin improvement strategy, you mentioned this in answer to one of the questions was -- is the sale of the East Fishkill fab to ON Semiconductor. You're targeted to hand over the fab at the beginning of next year. You've got good visibility on the mix of their products, right, and the remaining products that you'll continue to run in that facility as well. So first, can you just give us a sense on the positive impact to gross margins once the handover occurs? And then how long do you anticipate running products through East Fishkill post the handover?

Sukhi, why don't you address this one.

Yes, so as we mentioned, the positive impact on gross margin from the Fishkill transition is roughly about 300 basis points next year. We'll see about 150 basis points of that in the first half of next year. And as far as products running from ON and Fishkill, we will still see a significant amount of wafers from Fishkill coming to us in terms of revenue next year, and then it will taper off. So think about it like a 2- to 3-year transition from our products tapering off and their products coming in.

And to give you a little bit of color on that. So as ON has been ramping into that East Fishkill facility, we've been slowly ramping out. So they're ramping in. The transaction, we'll hand over that facility at the end of this year. We'll slowly ramp out. In that way, we didn't have any underutilization from when we were using it. They won't have any underutilization as they ramp up, and we ramp out. It keeps the facility a little bit more fully loaded, and it's more capital efficient for both companies.

Dave and Sukhi, thanks. Great insights. Thank you for the participation.

Thanks.

Appreciate it. Thank you.