All right, so good morning, and welcome to day two of the Evercore ISI Semi and Semi Equipment Conference. I'm Matthew Prisco, semiconductor analyst at Evercore ISI, and this morning we have the pleasure of hosting Gregg Lowe, CEO of Wolfspeed, and Tyler Gronbach, VP of External Affairs. Now, Wolfspeed is a company leading the way in silicon carbide materials and devices, and during Gregg's time at the helm for nearly the past six years, has undergone quite the transformation to make it the silicon carbide pure-play that it is today. So thank you, Gregg and Tyler, for joining us.
So I have a list of questions I'd like to run through and then save some time at the end for potential Q&A. With a lot of moving parts that the company discussed, we'd love to just jump right into it. So given its importance today, I think the 200mm material ramp is a great jumping-off point for the conversation, and would love any color you could offer on technical metrics of the materials they stand today: boule height, wafer thickness, defect rate, et cetera, maybe versus analyst day expectations or versus the more mature 150mm materials or whatever you think most helpful for investors to get comfortable around the company's material transition [crosstalk] and where we stand today.
Great. We demonstrated our first 200mm wafer actually way back in 2015, and when I jumped into the company in 2017, we began a very intensive program to get that technology into production. In fact, I hosted and chaired a monthly program review for 200mm, so tremendous amount of effort over the last six years at developing 200mm technology. Today, the technology is in production in what we call Building 10 on our campus. It is exactly where we want it to be from a defect density perspective, which is actually better than where we are at 150mm. So the, you know, the quality of crystal is exceptional. We had about a six-month delay in putting into production.
That delay was caused simply by a supply chain issue with a piece of equipment associated with an electrical substation. That equipment is now in order, working now, and we got certificate of occupancy on Building 10. Building 10 will have the amount of capacity for crystal growth that will allow us to get to 20% utilization in our Mohawk Valley Fab. It'll feed the Mohawk Valley Fab. As of our last investor update, 75% of all of the intended machines are in the building, are functioning, and are growing excellent crystals. It's actually more than that today, 'cause obviously we've had a little bit of time going between there and then and there. So we feel we're in great shape right now with the crystal growth.
Wafering is in good shape as well. Epi is in good shape, so we're actually feeding Mohawk Valley now at a pretty nice rate, you know, as we go to ramp the Mohawk Valley Fab. So, you know, six-month delay was not caused by any technical problem with the crystals or growing or anything like that. It was simply a substation that, you know, we couldn't get a part for. That's completed, and crystals are growing, and they're in excellent condition.
Perfect, very clear. So competitors in the materials market appear to be making progress towards their goals. So as you reflect on that competitive landscape, how have you seen Wolfspeed's positioning evolve over the past six, 12 months, and do advancements in that merchant market worry you at all, particularly as it relates to China and the clear strategic prioritization there, and capital being deployed?
Well, I think a couple of things. So first off, all of our device competitors are currently wafer customers of ours. So they buy wafers from us and some with epi and so forth, and all of them have stated some specific objective of having their own internal capability to do crystal growth and epi and so forth. Some have stated that they're gonna be 100% internal by the end of this calendar year. I think that's been modified to now more than 50% by the end of this calendar year, and I believe that was ON Semiconductor that has said that. ST has stated publicly, I believe, 40% of their supply is intended to be internal and 60% will be external. I think for companies to build an internal capability makes a lot of sense. I would do exactly the same thing if I were them.
So, we have in our plans, and as our, you know, our revenue outlooks and so forth, that all of them will be successful to the extent that they've publicly stated they would be. So we consider that in our plans, that they are gonna be successful. I think they'll likely find it more difficult than they originally anticipate. Silicon carbide is not an easy subject to master. It's one that we've spent 35 years on, and just in 200mm, you know, I can personally tell you, it's actually been six years, 'cause I've been at the company now six years. Basically, six years of intensive R&D to get us where we are at 200mm. You know, I can't imagine being the program manager of somebody trying to go from zero to, you know, full capability in, w ell, I guess by the next four months?
Yeah.
You know, so that's gonna be a bit of a challenge, I would imagine.
Okay, perfect. And I guess when you think about that 200mm ramp, given competitors' commentary and plans to you know, ramp their own 200mm in relatively short order. How do you think about your lead in that process specifically, and what have you learned during your ramp that maybe differentiates you from the pack today?
Well, we've spent six years perfecting, you know, the 200-mm technology. The capability that we have right now is exceptional. As I mentioned, it's better than our 150mm, and we're, we shipped $1 million, you know, last quarter. We'll ship $2 million this quarter and ramp up the wafer fab. I think the technology is, I think, substantially ahead of everybody, and it's one thing to be shipping wafers and ramping a fab, and it's another thing to plan to be shipping wafers. So I think it'll just have to play out.
All right, perfect. And then, you've previously talked about capacity constraints through the end of the decade, and given the continued advancement in material capacity build-outs and competitors working to increase output through processes such as Cold Split or Smart Cut, do you still believe that we will be in a shortage through the end of the decade? And how do you think this supply-demand imbalance normalizes over time?
Yes, I think that will, that remains the case. I think the growth of the industry is substantially greater than anyone would've imagined, certainly five years ago. And, you know, I've seen everything from 40%- 50%- 60% compound annual growth rate for silicon carbide over the, you know, through the end of this this decade, and it's just really gonna be really hard to catch up with that. It's a message that I tell all of our customers, that, you know, the, the supply is gonna be less than the demand for quite some time.
But what's gonna be different than what happened with silicon over the last couple of years, and the crisis of, you know, not being able to buy a car or buy, you know, a PlayStation or whatever, is that during the silicon crisis, the supply did not equal demand, but it was actually less supply this year than they got last year. What's gonna be different with silicon carbide is, every next year will be substantially more supply than the previous year. So while it won't be great because supply will be under demand, the supply will be substantially higher than the previous year, you know, every year through the end of this decade, because of the ramp of the various different capacity plans.
Okay. And then, you know, as we think just quickly about these wafer cut technologies, that Cold Split, that Smart Cut, what is your interest today in incorporating similar techniques to drive greater efficiencies? Or is there something else that Wolf is doing to drive an output advantage, making adoption of these types of processes less important?
Historically, silicon carbide was cut by diamond. Silicon carbide by the way, is the second hardest material on Earth, so, you know, cutting it isn't for the faint of heart, so. When I joined the company, 100% of our wafers were cut by diamond saws. As of a couple of years ago, none of them were. We've already incorporated many techniques that have eliminated kerf loss and some of the things that people are talking about implementing at some time in the future. We don't cut anything with diamond saws anymore, and we haven't for a number of years.
Okay, interesting.
And Matt, if I could add, I think, you know, it's still an emerging market silicon carbide and I think that there's gonna be a lot of interesting applications and things that come to market. You can rest assured that, you know, by being the leader silicon carbide we get a chance to kick the tires on everything. But I think, as Gregg kinda outlines here, we're very comfortable with our approach, but we're not, y ou know, we're kinda technology agnostic. We'll do what's best, you know, for the overall business, is the way to think about it.
I would maybe add to that a little bit. We're at the early phase of the adoption silicon carbide and that ramp is gonna be, you know, an amazing growth opportunity for us. I believe that there's gonna be tremendous amount of innovation on how the technology is used in the application. What applications can use silicon carbide? How you separate a wafer from a boule. There'll be more innovation that's happening there. How you grow the boules, and the height of the boules, and the defect density. There's gonna be a lot of innovation, you know, on that. Over the last five years, because we've been supplying wafers to most of the industry, we've had nearly all of the learning cycles in silicon carbide. Nearly all.
I think it's a pretty bold step to say, without any learning, we're gonna go from zero to 100%. That's just gonna. There's just gonna be a lot of challenges. I can tell you, you know, the expression that most people know is, you know, what can go wrong will go wrong. And silicon carbide what can't go wrong sometimes also goes wrong. It's just a strange technology. You know, it grows at 2,500 degrees Celsius. That is half the temperature of the sun, and a small variation in temperature has an impact over 2,500 degrees Celsius. It's quite astounding. So this is not an easy technology to master.
The only time silicon carbide was formed in nature was one time, and it was when a meteorite struck the Earth. You can imagine what that environment must have been like. So that's kind of what you're doing when you're silicon carbide and it's just not an easy technology.
All right. Perfect, so with those challenges, when we look at the JP Center ramp, what can investors look to in order to gain confidence in the progress being made? And maybe how has the delay at Durham potentially impacted Wolfspeed's strategic approach to the JP ramp to ensure a smoother process?
So I was at the JP on Friday last week. Construction is going exceptionally well. The building has gone vertical. You can go on campus and go on the campus and look, and you'll see, you know, the, you know, a building is now forming. The roofing will be complete, I think, maybe even this week. You know, something that phase one is gonna be two million sq ft.
So you can imagine, you know, this thing going up. You know, it's an amazing, you know, facility. Right now, everything is on schedule, and that means that we would be ready to install crystal growers in the first half of calendar 2024, and we would turn those growers on in the second half of 2024, most likely in the third quarter of calendar 2024, is, you know, the target for that. And, you know, right now, everything's on schedule. You know, I met with the construction crew, the operations guys, you know, and so forth, and that's all on plan. You know, I would say we're also preparing for unexpected problems, and so we're trying to put buffer in the plan.
Part of that buffer is doing some things at the JP faster, and part of it is taking our Durham facility and squeezing more than 20% utilization capability out of that. So taking it from 20%- 22% to 24%- 25%, or something like that, to give us a little bit of buffer. But right now, everything's on plan, everything's on schedule with the JP. We will be building crystal growers in parallel to actually building the building itself, and then when the building's ready for installation, we'll have them, and so it won't be, you know, a serial process. And I feel real good about that, too, because although we had a delay in turning on Building 10, when we turned it on, s orry, guys, I'm just gonna turn this thing off. It's giving me too many, t here we go.
We had the delay with the substation equipment. When we turned on the building and energized the crystal growers, it took us about a week to fine-tune. We had a week of, you know, some troubleshooting that we had to do. But after that week, crystal growers were humming. And so, you know, we've demonstrated that we're able to bring up another facility, and so doing the JP should be relatively straightforward. And I would add one other thing, which is really important. The JP is a 40-minute drive from our current campus, and that was very intentional. Same people, same engineers, same scientists, same installation crew that did Building 10 will be doing the JP. It's just another. It's 40 minutes from campus, but it's probably a 20-minute drive from where people live.
Okay, so a lot of learnings that could be leveraged-
Absolutely
[crosstalk] from Building 10 to JP. Okay.
Absolutely. Very purposeful on where we chose that location.
Perfect. So putting this whole conversation together thus far, materials business, obviously some interesting developments over the last 12 months including the delays to ramp on the negative side, but the signing of the Renesas agreement on the positive. So when I think about these moves within the context of your prior material revenue forecast of $700 by FY 2027, first, how is that tracking today? And second, given these moving pieces, what do you think is an appropriate, kinda upside, downside band, just to put around that $700, and what would be primary drivers of risk in either direction?
I'll say a couple of different things. So, first off, the Renesas deal, Renesas deal was an important thing, and I think it shouldn't be lost on people that this is, to my knowledge, the largest capacity reservation deposit in the history of semiconductors. $2 billion deposit for a capacity reservation. I'm unaware of any deposit that's higher than that. Maybe there is one that wasn't announced, I guess, but I'm unaware of one. So $2 billion capacity reservation deposit. And Renesas is arguably the most respected automotive semiconductor company in the world. It's certainly one of the top three. You know, I think, you know, Infineon, NXP, you know, they're pretty, ST, pretty good companies. They're in that league of, you know, very, very respected automotive semiconductor companies.
Presumably, they searched the world for determining who was gonna be their silicon carbide wafer supplier. Presumably, the Chinese companies, who some people say are ready for prime time today, weren't asking for a $2 billion capacity reservation deposit. In fact, I'm positive they weren't. So if, if all of this is gonna be easy, why is there a $2 billion capacity reservation deposit? I think that's a pretty legitimate question. And number two, it's a 10-year agreement.
So that says we're years ahead is my take. And then finally, so to answer specifically your call or your question, you know, obviously this gives us upside on the $700 million, because we weren't banking on that, you know, supply agreement. It's also our first opening for 200mm deployment. Now, we won't be shipping 200mm until 2027, so we're obviously going into production in 2023 ourselves, so it's about a four-year head start, you know, for anybody else to get into production on our materials for 200mm.
And it'll be a, y ou know, it's not gonna be a boatload of wafers in 2027 it'll be a, you know, a ramp like you would anticipate. And then finally, our $700 million plan had contemplated that all of our device competitors who are building their own internal capability are gonna be successful to the extent that they say they're gonna be successful. And I think it's, it's probably a pretty good assumption that not all of them are gonna be successful. Some might be, some probably will be. Our plan is that they all will be, but I think the probability that they're all successful to the plans that they have, it's probably a long putt.
Gotcha. And how are you thinking about appetite for further 200mm deals now that the Renesas one's in the books?
Well, what's really important is, we, you know, we have an appetite ourselves to fill the Mohawk Valley Fab. So we don't really have any capability to supply anything soon. Which is one of the nice things about the deal with Renesas is that it, it's 2027. That'll be well after we're, you know, nearly fully loaded. In fact, I think it is fully loaded out of Mohawk Valley. So, you know, it'll be a really nice amount of time for us to refine the process for building or growing crystals, and to refine the process for slicing them and epi and all that kind of stuff, and s o I think Renesas was a great first customer. I think the capacity reservation, you know, deposit obviously was a huge, you know, positive on that. But right now, it's I think we're just focused on filling our own capacity.
Makes sense. Now, transitioning over to the device side, you've been very clear on Mohawk Valley hitting 20% utilization in the June 2024 quarter with revenues flowing through it at that level, one, two quarters thereafter.
Yeah.
So to start, would love to quickly hit on your confidence in achieving that 20% utilization, and then revenues of $1 million in 4Q, and then a couple million more in the 1Q guide, suggest a fairly, you know, meaningful acceleration thereafter. So why is that growth so incremental early on, and how are you thinking about the steepness of the revenue ramp through calendar 2024 and potential challenges going?
When you start a fab, you have to dial in the process. Each of the machines has to be dialed in to make sure you're gonna yield what you're expecting to yield. That's a very normal process. You never start a fab, and you load it up with thousands and thousands of wafers, and you just press the button. You'll end up with a lot of low-yielding wafers, scrap wafers, et cetera. So every single piece of equipment has to be kind of dialed in and made sure that we understand what the corner cases are, and what the variability is, and how is that gonna affect yield, et cetera.
We're going through that process right now, and that process, you know, you take some amount of wafers, you run them through, you do characterization at each of the different points, and then once you get the process dialed in, then you can actually load the factory pretty substantially. Now, the hardest thing to do in ramping the Mohawk Valley Fab is getting the materials production going. That is finished, and so we are 75%, it's more than 75% now, of the capacity for Building 10 is operating and developing crystals or growing crystals and turning them into wafers, et cetera. And we are shipping those wafers up to Mohawk Valley, and they are in place. There's tons of storage. We have overhead tracks that, you know, help us with managing the supply and so forth.
So once it's ready, turning it on will be relatively straightforward. I will say a couple other things. We have now qualified a number of products out of Mohawk Valley, a number of MOSFET products. We have qualified our largest device, which is always a great sign. Typically, the largest device has, you know, challenges because of the area and contamination and so forth. And we've also qualified our most complicated automotive device, you know, that has plating and some metallization, you know, and so forth on it. And the results of those qualifications are actually very, very encouraging.
When you do a qualification, you characterize the part, and then you run it through a bunch of stress tests, very high temperatures, a cycling of high and low cold temperatures, you know, different kinds of stresses, and then you characterize the part after you've done all that. And what you're looking for is shifts in the parameters of the device. And if you look at our first couple of qualifications that we've successfully qualified on the parts, the variation that we're seeing out of the Mohawk Valley qualifications is smaller or tighter than the variation we saw on the original qualification out of Durham. That's very encouraging. Quite frankly, it's also expected because it's a substantially better, you know, wafer fab with, you know, cleaner characteristics and so forth, but we're very pleased with that.
That's perfect. And then I'd like to dig a bit into the ramp above and beyond that 20%. So, you mentioned already the R&D projects to drive better output and expansion into some other sites to drive capacity. Can you maybe break down your efforts into each of these two buckets and what that means for supply potential above the 20%, just from Durham alone, or?
It'll be, w ell, from Durham, you know, I think we've got a pretty very, very clear line of sight of getting to that 20% you know, number, and then we've got a number of projects to get better efficiency and better productivity out of those machines. And I think there's a pretty good chance that we're gonna be able to do that, you know? And so you can kinda think of that as something north of 20%, and probably in the zip code of 25% is probably where we're gonna end up. And it's not reinventing science or anything like that. It's really a productivity kind of, you know, effort that we're doing. And we're also doing a lot of things to make sure that the JP stays on track.
So, you know, we're building the crystal growers, you know, in parallel and so forth. So I think as we, as we, ramp the facility, I think going from 0%-20% is gonna take us a certain amount of time, and going from 20%-40% will likely take less time, just because, you know, it's, it's just simply increasing what we've already got stabilized.
Okay, so, you know, once we get past this, say, 25% utilization, JP is at the epicenter of all growth from there, right?
Yeah.
So, we've already talked about your efforts to ensure the smoother ramp there, but how are you thinking about the slope of that ramp in terms of servicing Mohawk Valley? Should we be thinking there's a few quarter pause in utilization rate expansion before moving higher, maybe fiscal 2026 or, or something? Is it, is it reasonable to think that the ramp could add a few points of utilization rate, even in the, the back half of 2024 as you begin production?
You know, so first off, when you turn on these crystal growers, you don't line them all up and then flip the circuit breaker. These things ramp up to 2,500 degrees Celsius, so there's a process for making sure you don't blow the substation.
That's good.
So, now, we did that in Building 10, and, and we're now utilizing 75%. So it's not a long period of time, but you don't just turn them all on at one time. You turn them on, turn one on, you know, a few hours later, turn on another, y ou know, you're doing a, a kind of a sequential process, if you will. So I think that process will be sequential, but it won't be a long delay. So don't think of it as it's gonna be some kind of, you know, very, very slow process for ramping those. It'll be the normal process that we have.
Okay, that's helpful. And based on the ramp plans today, is it still reasonable to think that Mohawk will be running at that full $2 billion run rate in fiscal 2027? And then on the demand side of it, how are you thinking about the ability to fill this fab, today, based on your current design pipeline and conversion rate expectations?
We are still on track for the fiscal 2027 on the full fit-out and production and utilization of Mohawk Valley. Nothing's changed there. And I think the demand that we have is well served by that full output. We have plenty of demand, quite frankly, that's above that. We do know that not every company that's planning to ramp every single car is gonna happen exactly the way that they had planned.
So there's, you know, when we look at our design into, anticipated revenue, equation, there's a drop-off, a funnel kind of effect, if you will, assuming that not everyone's gonna be successful at, at doing everything in the exact timeframe that they anticipate. Maybe it's their cars aren't selling as well, maybe it's they, they have a supply chain issue with something else, you know, who knows? But, I think there's plenty of demand out there to satisfy Mohawk Valley and then some.
Okay, perfect. So then, you know, within this construct, you previously forecast $2.9 billion of power revenues in fiscal 2027. So is this still the right way to think about that split, you know, $400 million Durham, $2 billion Mohawk, and then $500 million from Saarland? And if so, can you offer an update on those Saarland build plans now that the capital's been secured, and when do we need to break ground, and what milestones should we be looking for to ensure that that $500 million baked into the forecast is on track?
No change in the 2.9, so that's staying. Saarland, we still have to get official approval on the funding. We're anticipating that that will happen relatively soon. Could be as early as this quarter. That's the signals that we're getting from Europe right now. We have a pretty good understanding of how much funding we're gonna get, and we're quite pleased with that as well. If that all goes as planned, then we would begin the permitting process. Now, in parallel, the government of Saarland, the state government of Saarland, has secured the land. They now own the land. We will be leasing the land from them for, I think, 99 years, I think.
Yeah.
So they have secured the land, and they have already begun work on the land, tearing down buildings, moving a creek, you know, preparing the land for construction. We've also hired the construction firm. They've done architectural designs. We're very far along in terms of what the building is gonna look like. Now, obviously, it's a, t here's a lot of learning that we had from New York that's gonna be applied to Mohawk Valley or to the Saarland fab. We've also hired the fab manager for the facility. It's a German national that's been working in the semiconductor industry for many, many years and understands fabs, et cetera. He'll actually start with us in January.
He'll come over to the United States, working with our teams in Mohawk Valley and Durham, you know, because he has background in silicon fab, so understanding semiconductor fabrication with silicon carbide is a little bit different. So we've secured that as well, which is, you know, obviously a positive thing. And then, we would anticipate that the permitting process would take about nine months, which would put us in kind of that July to October timeframe, where a shovel would go in the ground, and we'd begin construction. So all of that's on track to plan. And, you know, in terms of building the facility, I would think that since it's our second facility, we'll be able to do the construction faster because we're more experienced.
But we also, well, hopefully, we don't have another COVID lockdown, but you know, our, our Mohawk Valley Fab, construction started in March of 2020. That was week 1 of COVID in the United States. And so that entire facility was built during COVID, supply chain issues, all the shenanigans we had to deal with. So presumably, that's not gonna happen again, too. So I think the, the probability that we can get the thing, you know, up and running faster, I think is pretty logical.
And Matt, though, so to think about that, that means then for fiscal 2024, it's ramping Mohawk, ramping Siler, and then as Gregg outlined, then you kind of pivot your focus after you get those things going, you'll then look at Saarland as kind of the next opportunity.
All right, that's perfect. So next I'd like to hit on just customer conversations in light of some of the recent delays. Acknowledging that you guys have mentioned no loss at first chair position at any customers, have you seen any change in requested out-year volumes at all, as customers maybe look for second, third, you know, the dynamic, which is not unusual in the auto industry at all? Then, maybe just help us better understand the customer conversations you're having and the work you are doing to ensure that supply-demand alignment.
In some respects, we are seeing a change, and the only changes are positive and more. So that is what's happening. We're getting more traction across the auto industry. We're getting more traction across other modes of transportation, whether that's trains, earthmoving equipment, vertical take-off and landing, kind of Uber drones kind of things, and Jet Skis and, you know, personal watercraft. A lot of different applications there. So basically, the demand continues to be very, very high. And in fact, the demand, as everyone probably knows, the demand for silicon semiconductors right now is seeing a downturn. Memories, you know, in the doldrums right now. Industrial markets are down, you know, et cetera. The demand for silicon carbide with one exception, is all pulling in and going up.
The only exception to that is China industrials. China car companies, automobiles, still up and to the right, for silicon carbide, you know, technology. The demand for automotive generally across the world, still very, very strong. Someone asked me the other day about, what about an auto strike, and will that impact everything? You know, our production ramps are really, you know, in the early phase right now for the U.S. auto companies, and it's really gonna take off in the, in the coming years. So if a strike lasts a couple of months, it'll have no impact on us. So you know, basically, there's only one area of weakness, and that is, China industrials. Everything else is up and to the right right now. Then, for the longer period of time, like I said, the only, the only thing we're seeing is people just want more.
Okay. So I guess just digging in there a bit, can you cue up your thoughts on the primary driver of Wolfspeed's success, in particular, of securing these new wins? First, I guess, how are you thinking about the share of new device wins today, and then what are the most important factors that drive a customer to choose Wolfspeed, and what are potentially some areas of pushback that you may hear?
I think when a customer makes a decision, a device decision, it's never one thing. It's kind of a mosaic. They look at a number of different things. They look at the performance of the device in their application, and they compare us with others, you know, et cetera. We always land in the top category on performance. Sometimes we're there with other people, but we're always in that top category. The second thing they look at is your knowledge of the underlying technology, silicon carbide. We also land in the top tier in that one, and there's nobody else in that tier. We have more knowledge with 35 years of experience, and we're alone in that category. They look at your experience or their experience with you as a supplier, an automotive supplier.
And here we land in the second tier. We have substantially less experience in the automotive market than ST, Infineon, Rohm, you know, et cetera. So, you know, they have just ON Semi, they have more experience in automotive, so we're definitely in the second tier in that one. What we're doing, by the way, to rectify that, is we're hiring people from the semiconductor industry who are automotive people, and we've had hundreds of people join us. So you know, for instance, Lisa Fritz runs our quality organization. She reports to me as a Senior VP. She ran automotive quality for I don't know how many years, but she was 25 years at Texas Instruments. So when an automotive company comes to audit our facility, maybe it's the first time they've seen us, but it's not the first time they've seen her.
So there's a trust kind of thing that's been built, and we've been doing that ever since I've been at the company. So that's a positive thing. They, of course, look at pricing too, and I have no idea where we land on pricing because they never call you and say, "You're the lowest price." So you know, you always kind of take it with a grain of salt. I will say to that extent, though, we are no longer bidding on being a second source. Our capacity is very, very precious. We bid on programs to win as the lead source, and that's kind of 70%. We totally understand that there's going to be a second source. We don't have any issue with that. But we tell our customers, "If you choose us as a first source, you'll get our full attention.
But if you don't choose us as a first source, we are not gonna bid as being a second source." We don't have the capacity to do that. We're giving all of our capacity to people that have really banked on us. So it's a mosaic of things that customers look at. It's never really one thing. But I would say the other thing that kind of is in the back of their minds is we currently supply 60% of the world's silicon carbide substrates, and the next closest participant in that is Coherent at about 15%. So it's a pretty substantial difference. And we do that 60% out of our Durham campus, you know, that we're, we've been talking about.
The JP will increase our capacity for wafers by 10x. So, you know, if somebody's not building, you know, a scale to at least get to where we are today, which is 60%, you know, the probability they're gonna be able to get to 10x that is pretty low, too. So I think they look at that, and they say, you know, "Can I really bank on somebody learning this technology so quickly and ramping so quickly, versus somebody that already has 60% and has a building now that has gone vertical, that's gonna increase the capacity by 10x that, and it's gonna be functioning in one year from now?" That, I think, is in the back of a lot of people's minds. All told, you know, we've done $19 billion of design-ins over the last three years, I think it is?
Four.
Four years. You know, so, you know, we've got a pretty healthy backlog of book business.
Perfect. With that, I'll just pause for a moment, see if there are any questions. All right, perfect. So let's, let's keep running through. You talked about gross margins of 50%-54% [crosstalk] still being the right target to think about in fiscal 2027. Can you maybe walk us through the moving parts to get from the 13% today to that 50%+ over the next few years? Primary levers, timing, anything that can help us get comfortable. And then, just how should we think about the ramps of JP and Saarland impacting that trajectory?
Maybe I'll kick it off and you can give a little bit more color on it. But there's no change on the target for 50%-54% in the timeframe that we've talked about. So we're ramping that, you know, we're ramping the Mohawk Valley Fab, we're ramping the JP, you know, and so forth, and all of that is feeding into our plans for 50%-54%, gross margin. Now, we had originally non-GAAPed out the underutilization and the start-up costs, and we had targeted that at a 70% utilization. By the time we're, you know, in that timeframe where we said we'd be 50%-54%, we'll probably be north of 70% utilization.
Yeah. So the way to think about it is what we talked about on the call. And if you look at the midpoint, you know, we've kind of called the trough for non-GAAP gross margin. What you would hope to see, and what Neill talked about on the call a couple of weeks ago, is, you know, several points of improvement between now and when we get to 20% utilization. So you're probably looking, Matt, something in the low 20s by the time we get to the, you know, the June quarter of next year. Then, even though you have Siler coming on, and you're gonna have some start-up costs at Siler, the top line's still gonna be growing, so that still creates that kind of improvement engine for non-GAAP gross margin over time.
You're just gonna continue to watch for improved trajectory on non-GAAP gross margin as we continue to ramp Mohawk.
Perfect. Helpful. So capital concerns have largely been put to rest following the $5 billion raise over the last year. So maybe you can walk us through your debt optimization strategy from here, given the company's leverage. And how are you thinking about timing and magnitude of government funding at this point?
So, we've had engagement with government funding from Europe and Germany for quite some time. We've got an application going in for CHIPS funding. We feel very confident in the funding from a European perspective. Like I said, that could be coming out really any time, and as soon as the end of this month. It certainly will be, I think, before the end of this year. So that's all looking quite positive. I think we are a very, very solid candidate for the CHIPS Act funding. A lot of the things that are in the interest of the U.S. government we're doing. You know, we know, the part of it is about securing the supply chain for semiconductors in the United States.
And the U.S. currently has Wolfspeed and Coherent as the two leading silicon carbide companies. Between the two of us, it's over 80% of the world's silicon carbide is between these two companies. And the Chinese are obviously investing in all kinds of semiconductors. You know, they've been on a path to be dominant in memory, and logic, silicon carbide and GaN, and everything. They've had, you know, limited success, you know, so far, but they are determined. And, so what I think is, in the minds of, certainly in the minds of the CHIPS Act folks, is that we currently have a leadership position, and it's a hell of a lot easier to support, maintain, and encourage the retention of that leadership than it is to try to regain a leadership in, in semiconductors, in silicon semiconductors.
So that is an important aspect, and I think it's important that, that we do that as well. But investing in maintaining a lead, it really makes sense, I think, from, from that perspective. The second thing is that silicon carbide is important for the adoption of electric cars. It helps the car go further, 5%-15% further, and it helps it recharge faster. The two most dominant care abouts of anybody buying an, an electric car; I want it to go far, and I want it to refill fast, and that's what silicon carbide does. You know, when I joined the company six years ago, one car company was committed to using silicon carbide in the inverter, and that was Tesla. Today, I can't think of a car company that's not using silicon carbide.
Everybody's silicon carbide and some of them are going 100% silicon carbide. These are, you know, big companies like General Motors, you know, high-end companies like Jaguar and Mercedes, startup companies like Lucid, you know, they're all adopting silicon carbide. With the, certainly the current administration's, you know, drive for green energy initiatives, we're a unicorn. I mean, we're U.S.-based companies leading the industry in a growth pattern that's amazing, and it's gonna help the adoption of electric cars. I mean, so it's, it's kind of a no-brainer.
All right. So I guess just to round out the conversation a bit, what's your view for trajectory of operating cash flow and free cash flow from here?
Don't you, Don't do that.
When do these metrics turn positive, and how do you think about growth?
Yeah, I think, Matt, you know, we've talked about this before, but, you know, operating cash flow, you see that towards the end of fiscal 2025, right, a nd you see that free cash flow generation sometime, you know, probably early 2027. So, you know, the, the underlying economics have not changed. It's a timing question in terms of ramping Mohawk and bringing that up. That's the way to kind of think about operating cash flow and free cash flow.
That's fiscal.
Fiscal, yeah.
Yeah.
All right, perfect. So with under a minute left, any closing comments or any areas that you think we've, we've missed that investors should be focused on?
I think the hardest thing to do in creating a silicon carbide MOSFET is generating a high quality crystal, whether that's at 150mm or 200mm, and being able to separate it from the boule in a very efficient fashion, and to be able to lay a layer of epi on top of that. We've done that for 35 years at the lower diameter numbers. We're 60% of the market in doing just that, and we've now accomplished that in 200mm.
Wonderful. Well, with that, we've unfortunately run out of time. But Gregg, a pleasure chatting with you-[crosstalk]
Thank you, Matt.
Tyler, as well, and thank you very much for joining us.
Thank you.[crosstalk]
Thanks, Matt.
All right, thanks, everyone.