Wolfspeed, Inc. (WOLF) Earnings Call Transcript & Summary

All right. Thank you, everybody, for joining us. If I could ask everybody to please find a seat, we're going to go ahead and start the next presentation. So I'm Craig Irwin. I cover sustainability or clean tech or whatever we're calling it these days at ROTH. Wolfspeed, we are very lucky to have the founder of the company -- co-founder of the company, John Palmour, with us. Tyler Gronbach, who runs the Investor Relations program, is also a key person for Wall Street. And this is one of the most exciting situations I've encountered as my 20 years as an equity analyst, as a sell-side analyst. So I believe that silicon carbide chips and modules are going to be a $30 billion, $40 billion, $50 billion market in my lifetime. And what that is, is it's a conversion of the silicon power semiconductor module and chip market, where 1/3, 1/4 of the slots get converted to silicon carbide, 5x the price, pricing on value, and then absolute growth because of the new applications allowed. So I'm widely bullish on the long term. On the medium term or medium short, what matters is capacity, right? It looks like these guys might actually be sold out or close to sold out. And with the GM, strategic procurement, and several other contracts, including one, which was a couple of years ago, $500 million MOSFET contract, biggest ever in -- for chips in power semiconductors, I believe at the time. How are you guys looking at capacity, right? What are we planning for '24, '25 because you're going to get new customers. There's a lot of them out there. How do you -- what's the decision process on the next fab?

Well, so yes, capacity is something we spent pretty much, all day every day, talking about both near term and long term. Because we set out plans and had certain assumptions about the adoption of battery electric vehicles and what that would do for the silicon carbide market, looking at 2030 and what amount of penetration BEVs would have in the total market as well as what penetration silicon carbide would have into that market. And every time we go back and look, the time line is getting pulled in, the volumes are going up, the people are looking at more aggressive penetration of BEVs as a total percentage of automobiles sold and are looking at more rapid adoption of silicon carbide into that. So we're devoting a lot of our time to capacity discussions. We've been very open about our Mohawk Valley fab. We've cut the purchase orders to fill that about half up with equipment. So now the very immediate topic is, okay, when do we pull the trigger to go ahead and populate the rest of that fab? And then with the upward pressure in those outer years, yes, there is a discussion of when do we pull the trigger for the second fab as well. So it's really looking at what our backlog is, what our customers are saying and what they're committing and then weighing that against the amount of CapEx that has to be done, but it's a constant ongoing discussion.

Craig, and what I'll also add to that, Craig knows this, but the design-in pipeline or the things that we talk about every quarter in terms of what have we won or been awarded, we've totaled, in the last 10 quarters, it's been about $6 billion worth of design-ins. So that kind of gives you a sense of what we're winning. And then as Craig points out, that will give you some idea of how quickly things are going to fill up. And we started talking about design-ins about 2 years ago. At that time, we were talking about $0.5 billion worth of awards in a particular quarter. In the most recent quarter, we announced $1.6 billion. So I think the leading indicator on capacity for you is going to be when we talk about on a quarterly basis the rate of design-ins and what's happening there, and then also the increase in the opportunity pipeline.

So talking about the opportunity pipeline, right? Just big picture, 6-inch to 8-inch, you announced your 8-inch reactor guide today, so congratulations, Felix. First press release ever for an equipment supplier like this. So this is a big deal.

Yes. We've never discussed our equipment suppliers. We've always held that very proprietary. But in this case, we did do an announcement with AIXTRON. We've worked with them for actually a number of years on the development of this new tool with an eye on 200-millimeter because we knew we wanted to ramp up Mohawk Valley at that 200-millimeter diameter and have worked very closely to make this tool a reality and very cost effective. And so that was our -- well, AIXTRON's big announcement today. But that was the first epi tool we've ever acknowledged.

I know. It's a big deal. So when I look at the mechanics of going 6-inch to 8-inch, there's a 75% improvement just based on area. I know there's yield sometimes positive, sometimes negative, other issues. The MOSFET manufacturing is pretty well established. And from what I understand, you're going to invest some of that in pricing for your customers, and the strategic agreement with General Motors was interesting. Obviously, there's always a little price in a strategic agreement like that. It kind of puts a lot of pressure on the other guys out there that don't have access to 8-inch, right? And I'm wondering what the capability is if there's an early adopter in the market that's already a good customer of yours, maybe on the different side of the house, what's the capability to serve them with the existing footprint? You really need another fab to actually -- to become the primary supplier on that account, don't you?

Well, let me get to that. I do want to say that in terms of cost, yes, there are 70% more die on a wafer, but you end up going to, like, some single wafer process tools, et cetera. But generally, when you go up from 6- to 8-inch, if you look at the history of silicon, there's generally been about a 30%, 35% cost reduction. All of that is based on the assumption that the dollars per square centimeter of the substrate is the same. And we've been open. We're not there yet. So -- but we're close enough to where we can go ahead and start pushing the 200-millimeter through. So I just want to make that clear. But we also know that over time, we will get that cost per square centimeter down, the same as what we're at on 150-millimeter. The other big cost advantage we think we'll have is this is a very automated fab. Today, in our North Carolina fab, it's a 6-inch fab, it's very manual. It looks very much like a silicon fab you would have seen maybe 25, 30 years ago. And literally, wafers get touched thousands and thousands of times throughout the process. When Mohawk Valley comes in, it's completely automated. The number of touches per wafer will be effectively 0. So we expect to get some yield improvement just by going to this much more highly automated fab. So we're going to see where we're at in terms of throughput, in terms of the yields, the throughput, the cycle times, et cetera. But yes, we're evaluating the need for another one because even under pretty rosy projections for yields and things that we expect to see, we're looking and trying to calculate when is that fab basically fall.

I think, Craig, there's 2 key decisions within the next 12 months we have to make is the build-out of the second half of Mohawk Valley. As you know, 50% of that clean space has tools in it, and there's another 50% to go. And I think the second decision that we're going to have to make is, what John just talked about, inside of the next 12 months, we're probably going to need a decision on a second fab. Where do we locate that. There are lots of things. Our CEO and CFO are currently in Europe right now doing 2 things: meeting with customers, which is something that we like to do to get a read on demand, and then we're also going to be talking with European officials about possible incentives. So it's something that we're going to continue to watch closely, but I think you're right, Craig, it's something that we really have to give key consideration to inside of the next 12 months.

So one of the questions I'm still getting, which I'm kind of a little bit surprised by, but I'm a techie, I'm a nerd, right? People are asking, is the whole EV fleet going to be silicon carbide? Are all the converters going to be silicon carbide? What do you think? What do you guys say?

I think a lot of them will be. We talk a lot about, let's say, the standard bus voltage for an EV today, by and large, is a 400-volt bus. And silicon carbide is winning in that bus voltage. But there's a big push to double that, to go to 800-volt bus. And one of the big reasons for that is for fast charging. And there's -- the idea of being stranded out on the road and having to wait an hour to charge your car is a detractor for EVs. The realization is that with an 800-volt bus, you can charge a car twice as fast. And you now have the ability to maybe -- now, you're talking about 100 miles in 5 minutes or something like that, enough time to go in and buy a cup of coffee and come out and leave. When you go to that 800-volt bus, it really makes silicon carbide almost the only choice. There is 1 800-volt bus car out there using silicon. But I would say the very large majority of people looking at 800-volt systems are looking at silicon carbide because it has a strong advantage over silicon IGBTs at 400 volts. It has an incredible advantage at 800 volts. So I think the penetration is going to be pretty good. There will be applications where silicon carbide may not play such as like peaking inverters, where silicon carbide might be in the main inverter for driving the car, but then if you need burst power, then it's harder to get the value out of silicon carbide if that inverters only on 5% or 10% of the time. That might stay silicon. But I think silicon carbide is going to have a very strong play in it.

I would agree. So Wolfspeed or Cree, as it was formally known, right, always behaved a lot like a semiconductor stock, right? It is a semiconductor company at its core, highly differentiated versus most of the industry. So margins and the margins in the margin cycle end up being of tantamount importance. And it seems like we've moved past a margin trough. I'm not saying there's not going to be volatility as we bring on new equipment and change processes a little bit. But what should we be thinking about margins and the factors that impact margins over the next couple of quarters? Are these factors going to be higher visibility to you? And maybe if you could list the most important ones and just speak generally about the puts and takes.

No, Craig, it's a great question. And where we sit today, we're in the mid-30s, and with an expectation that we'll get to 50% gross margins by the end of our fiscal 2024. And when you think about the bridge and how that happens, and Craig's right, we're being measured like a traditional semiconductor company. There's a couple of things that are happening. There's some efficiency gains that are going to be made within the business in the Durham print. So before we even get to Mohawk Valley, just think Durham, there's efficiency gains that are going to be made there. Missy Stigall, who ran one of the world's biggest bay and chase fabs for Texas Instruments, she joined just back in August, and she put in a lot of, what I would call, blocking and tackling work. We went from looking at performance reporting on a 30-day basis to a weekly basis to a daily basis, and Missy wants to eventually get to a per shift view, which allows us to gain great efficiency there. So we've seen some improvements in that regard. And we've done some investment in that fab. But it's never going to get to anything but think high 30s in terms of a gross margin. The real rev on the engine is the move to Mohawk Valley. Because as John talked about, we do thousands of wafer touches a day in the bay and chase setup. And we're going to a ballroom set up in Mohawk Valley, where the equipment stacked right against each other. We've got 8 miles of track built into the ceiling that are going to be running Smith cartridges all around. But let me just give you some rough numbers. I mean, the move to Mohawk Valley is going to drive a 50% improvement in cycle time. So the time it takes for our wafers to get through the fab is going to be cut in half. We're going to probably see 20 to 30 points improvement on yield, and that's just because we're not going to be touching wafers any longer, and some other improvements there. So Craig, that's what -- when you really feel the move on the gross margin is going to be the turn up of Mohawk Valley.

And the move from 6-inch to 8-inch wafers, to 200-millimeter. There will be more die per wafer, and we should get margin improvement there.

Excellent. Excellent. So another question I wanted to ask. So there's a conversation out there about growing silicon carbide bulls and the merits of inductive versus resistive. So Cree has been a strong, staunch defender of resistive, right? You've very successfully led the industry by a hundred miles. But there are others that are making progress with inductive, faster cycle times, higher defect rates. What is the -- so you have done inductive in the past. Cengiz actually was with INTRINSIC Semiconductor was using inductive when you bought them for a Zero-Micropipe technology. What's your opportunity to use these other furnace technologies for silicon carbide bull growth? Is this something that there is a portfolio of options, and you're using the best one because it delivers the highest quality wafers that are adequate for your needs? I mean, does this evolve in the future?

Well, your question is interesting. We've never been a staunch supporter of anything. We've never ever discussed how we do our commercial growth systems, ever.

I didn't know that. I'm sorry.

So -- no, we owe you a lot, literally.

I talk to a lot of people.

Well, people may think they know what we do, and they may be right, they may be wrong. We literally have never ever said. What I can say is we've been building semi -- or silicon carbide crystal growth systems for 34 years, and we've gotten really damn good at it. And people can do what they want. There's not much in the way of technology we haven't looked at over the years, but we've continually reinvented and refined the process, but we do not talk about how we do it.

And Craig, I think for the benefit of those of you that don't know us really well. We believe culturally -- somebody said to us earlier today, "You guys are like a 30-year-old startup." And the truth be told, that is the truth. I mean, last year, we had revenues low $500 million. We spent $575 million in CapEx. I mean, we're a growth company. And what we're trying to do is preserve the culture a little bit. And so while at times we might be shy or not forthcoming about our approach to the market, we feel that, that's a competitive advantage for us. And it's something that we really like where we are today. And Craig, in the competitive landscape, we kind of -- we view everyone as rivals. And the reason why I say that is because anything that the silicon carbide win at this juncture, it's still so early stage. So we get excited when we see things happen when -- maybe we didn't win something, but somebody else got someone to move from silicon to silicon carbide. And as we think about the market, that's the way we're going to see it because we are still early-stage transition from silicon to silicon carbide. And we think that there's still a lot of upside for all of us.

Yes. And whoever won, we're very happy to sell them wafers.

I remember when wafers were -- actually, we had the margins on those, and those are 50% gross margins. Man, what a business. So what do you see as the biggest challenge for the next year? What's -- there's a lot of moving parts. I think the panic in certain members of the investment communities has been washed out. I don't think it was justified. But you have a lot to execute on in the next year. I mean, what would you say that we should watch for as sort of key milestones other than obviously the big announcement today?

Well, I mean, I think it's exactly what you said. It's execution. I mean, everything is lined up for us. We have to execute. We've got to get Mohawk Valley up, get it going. We got to show good yields, good quality, get the products qualified. And then we have to get our customers to qualify them. And that's a fairly long process. So we got to execute and execute and execute on that. Ramping up the production of the 200-millimeter material to meet that demand, a big execution issue there for the crystal growth, the polishing, the epi as we talked about, et cetera. So it's a lot about execution. And then the other side, as we already touched on, is planning for that next tranche of capacity, which we already discussed. I'd say all of those things are really the big metrics for us.

So it's the next fab.

Capacity. Capacity. It's all about capacity, baby.

It is, literally.

A double and then a double again, right, 4x?

Yes. I mean, we're in a multidecade opportunity here. And we're really at the very, very beginning. Electric vehicles is still a tiny percentage of the market. Now you're talking about 10 years from now, it could be 30%, or some people say 40% of the market.

So investors actually asked me, so I'm very bullish on the electric vehicle opportunity. But when someone is a long-term oriented investor, a buyer of stock and holder for decade, that's what they do. I point out the utility market and the power conversion opportunity there, right? And in power supply sizes, cost is very simple. If you got to pour concrete, put something on a pad, and this is a giant [ Hulka Steel ] and you've got all sorts of other cooling and equipment in there. If you can do something with silicon carbide, it's going to cost you a tiny fraction for a lot of those pieces, except for the silicon carbide. So it brings a lot of the value added to a supplier like Wolfspeed. If you were to take sort of a big global view. What do you think that the high power markets need to start adoption?

I'm going to take an even bigger view and just talk about the industrial market in general, because we do a lot of business in industrial. We talk a lot about the EV market because it is going to be so big. But today, big markets for silicon carbide are server power supplies. If you own a data farm, your single biggest cost is energy. We make that more efficient. So server power supplies, solar inverters, chargers, fast chargers. Kind of automotive, but by and large, actually an industrial customer base. We do a lot of business there. Energy storage systems is a big market. And that's going to be a big growing market as we move forward in terms of how we manage the grid and distributed power with a combination of solar, wind and fossil fuels. So we're very active in that. The next level of that is the grid level, what I'll call grid-tied inverters and things, which I think is more of what you meant. There, we are working on higher voltage products. So basically what you need there is something up in the 3.3 kV range, 6.5 kV and even 10 kilovolt, which is what we're working on. Today, the very highest power silicon device, you can get is 6.5 kV, and they're very rarely used because they're so lossy. The bulk of it is really 3.3 kV. So once we can get those products out, then we can start doing a lot of grid-tied solar inverters, grid-tied energy storage systems and stat VARs for controlling power quality on the grid. Those are all future markets. We just don't actually have the products out yet to address those, but we've been working with numerous companies on, I'll call it, the transmission and power distribution aspects of the market.

So can you -- I mean, it's been a long time since I reviewed this, but if I remember, there was a sort of rule of thumb, 20 microns per 100 kV?

It's about 10 microns -- or excuse me, 1,000 volts per 10 microns.

1,000 volts per 10 microns. Does that -- so when you were focused on making LEDs in the past, you were doing much thicker epi than what you're doing now, right? That was much more chunky epi.

No, that was not terribly thick. I think those layers were probably in the range of a few microns or something like that. So the bulk of what we make today is actually thicker epi than what we did for the LEDs. If it's 600 volts, it's going to be 6 or 7 microns. If it's 1,200 volts, it's more. But we've also worked on 10 kilovolt devices. That's 100 microns or so of epi, very low doped epi.

Yes. No, I mean what's coming is so exciting. And I'm so pleased you're able to join us at this conference. And I just want to say thank you, John. Thank you, Tyler. I'm obviously excited about the opportunity, and I appreciate everybody joining us. Thank you all. And you should see them. Go to the one-on-one desk.