Okay, we'll go ahead and get started. Welcome everybody to the third day of Needham & Company's 26th Annual Growth Conference. My name is Quinn Bolton. I am the semiconductor analyst for Needham & Company. It's my pleasure to host this fireside chat with SkyWater. SkyWater is a U.S. based technology foundry offering both advanced semiconductor process technology development, as well as wafer manufacturing services. Joining me for the com pany for the fireside chat is CEO Thomas Sonderman. We also have Steve Manko, CFO, and Claire McAdams for investor relations here in the audience. Tom, thank you for joining us.
Thank you very much, Quinn.
Tom, since there may be some in the audience who are less familiar with SkyWater, maybe spend a minute to give us an overview of the company, and I think especially the advanced technology manufacturing services part of the business, because I think that is unique. It's certainly for the, you know, U.S. foundry company.
Yeah, absolutely. So, SkyWater is a pure-play semiconductor contract manufacturer, so otherwise known as a foundry. We're literally America's foundry. We were created back in 2017, spun out of Cypress Semiconductor. This is before CHIPS and when everybody realized we needed to have computer CHIPS made here in the U.S. So we've been able to, you know, capture some of that momentum, have built a very strong, as you said, it what we call advanced technology services business. Think of this as a monetized R&D capability that allows us to work with customers very early in the product development stage, with the assurance that once their products move into production, we can also make them in volume. So we do both embedded R&D as well as high-volume manufacturing in the same manufacturing facility.
And the last comment is, we have two facilities. One is in Bloomington, Minnesota, focused on 200 mm CMOS, 90, and 130 nm and many derivatives. And then we have an advanced packaging, heterogeneous integration facility in Florida that is focused on bringing that technology back to the U.S. Only 3% of those type of solutions are made here in the U.S., which is even more dramatic than the 12%-
Right
... on front-end wafer manufacturing.
Great. You kind of mentioned a little bit of a lead into my next question, but if you just look out at the foundry or really the semiconductor industry over the next, you know, say, three to five years, clearly, there are some trends out there. I think reshoring of manufacturing in the U.S. may be one. But what longer-term trends do you see in the manufacturing side of semiconductors that you're trying to take advantage of, and position SkyWater for?
Yeah, great question, and it's really one of the main drivers of SkyWater is, you know, we believe we're leaving the smartphone era, moving into the era of AI. I think we hear a lot about AI. AI requires a lot of new solutions, photonics, various types of sensing solutions, the Himax presentation we just heard from, just tons of efforts around integrating, we'll call it traditional CMOS with other sensing capabilities. I think you're gonna see a lot more of that. That's what we call CMOS+. Of course, there's a lot of things happening in space. SkyWater is very instrumental in that, Rad-Hard solutions, uncooled infrared imaging solutions, and then, of course, superconducting and quantum computing. These are foundational areas that really, I think, when coupled with heterogeneous integration, is gonna...
You know, we keep talking about Moore's Law, but this is really what's gonna extend kind of the intent of Moore's Law as we act, you know, not only go through this decade, but really go into the thirties. And so, as someone who spent my entire career here, I can say we're just getting started.
Yeah, and-
The driving force, for sure, is gonna be AI.
I've got a number of questions about the advanced packaging facility in Florida, but that sounds like it. Yeah, clearly, the trend towards chiplets, heterogeneous-
Oh
... integration is only gonna get stronger over the next five years, and I think you guys are positioned as the only domestic, you know, advanced chip, or packaging facility in the States. So it sounds-
Yeah, I mean,
like a unique position.
Yeah, I mean, I'll hit it now. It's so exciting because just last week, we announced a $120 million initial investment, up to $190 million investment from the DoD to stand up the only wafer fan-out packaging solution using what's called Deca technology. I don't know how familiar people are with Deca, but it's a fan-out solution, only exists in Asia today. We'll be standing that up here in the U.S. It's highly competitive with Intel's EMIB process and TSMC's CoWoS process. So we will not only have a solution here in the U.S., but we believe highly competitive.
In our announcement, AMD came in as a commercial endorser of bringing this solution to the U.S., and we expect this to be our entry into the, we'll call it high, you know, performance computing market. We're not focused on, you know, extreme node, 3 nm, but we are focused on getting this capability in the U.S.
Yeah.
There's a lot of appetite for it.
Yeah. Since we're on the award, maybe spend just a minute talking about the partnership with Deca Technologies. I think they have a Adaptive Patterning technology, which is also somewhat unique, I think.
Yeah, absolutely. It's. So Deca Technologies uses a direct write solution. So just like e-beam is a direct write, it doesn't require a photomask. As a result, when you compare it to solutions from, say, Onto Innovation or Veeco, where they use traditional optical lithography, you have pattern movement issues. This cause yield issues. Obviously, when you're doing advanced packaging, the wafers are fully fabricated. They're all known good die, so you wanna have as high yield as possible because all the cost is into the part. So we believe having that solution in the U.S., it'll be a 300 mm solution as well, so it's our first movement into 300 mm.
We'll provide a lot of companies that wanna source here domestically, of which there are numerous, DoD being one, but also, you know, companies like AMD that wanna take advantage of having this type of solution available in the U.S. And, you know, there's two points to that. One is, obviously, having advantage or having access to the technology, but also, it's just more efficient in many ways because a lot of these fabless companies have packaging design teams that are all based here in the U.S., and so it's, you know, just much more efficient. You don't have to travel to Taiwan or overseas, and it makes the cycles of learning go that much faster.
The award, I think, was about $120 million, with an option for an additional 70-
Yeah
... over a five-year period. What does the $120 million pay for?
Yeah, so-
Potentially $190.
Yeah. Essentially, we're getting funded for all the equipment to stand up the line. This is very similar to 2018 when the DoD through a similar contracting agency awarded us our RadHard program. Now you fast-forward, you know, to 2024, that technology is in productization qualification, so it's preparing to go into production next year. I see a similar kind of situation here. They're buying the equipment. The facility is already there. We did have to expand our facility for the RadHard program, but we're also getting a lot of NRE to bring the technology, transfer the technology. By the way, this isn't, like, a new technology that's never been done before. It is working. Multiple fabs, ASE, Amkor, others, are using this technology on what's called gen one.
We'll be bringing the gen 2 technology into the U.S. And the other thing to keep in mind is we won't just be doing wafer fabrication or even advanced packaging, we'll be doing final assembly and test. So that's how fan-out works, is you take fully fabricated wafers, you singulate them, you take multiple products, recombine them on a new substrate, complete the fabrication, and then singulate again, put it in a package die.
Is this interposer-based, where you may actually have the opportunity to manufacture the interposer, or is it substrate-based, where you would buy the substrate, and then-
Yeah, we would basically buy the substrate.
Okay.
In this case, it's an epoxy substrate, but we do have an interposer technology, and we have a hybrid bonding technology. So really, as a U.S.-based foundry, there's no one that offers that comprehensive suite of solutions. The IMEC technology is the foundation of our interposer solution. Xperi, which is now called Adeia, is our hybrid bonding technology, and of course, we have Deca, which is the fan-out technology. And all that will really become, you know, really go into, we'll call it, production availability over the next several years. And we believe, you know, by working with companies like AMD, who have a very short timeline, AMD is an absolute leader in chiplet technology.
I think they are even more so than Intel-
I-
... just in terms of getting to production.
As a former AMD-
Yeah
... I'll say, yeah, they're for sure.
Okay.
For sure. But, I just think that... And it's something that, you know, through a lot of effort, the Commerce Department is paying a lot of attention to. Because you can make the wafers here, but if I have to send them over to Taiwan to get packaged, I've kind of defeated the purpose.
Right.
The goal is to have, for critical infrastructure, a complete fabrication process from initial wafer, you know, procurement, all the way through a packaged die here, sourced in the U.S., and that's kind of what we're going after.
Right. Okay. If I put my financial hat on for a moment, the $120 million, can you sort of walk us through, sort of the timing, receive-
Yeah
... those funds? Does it come in effectively as a grant, then maybe offsets CapEx? Does it come in through your income statement as tool revenue? And, and, you know, we, we can kind of circle back for those that may not understand what the tool revenue is.
Yeah.
But you know, is it an income statement impact, or is it really more of a cash flow balance sheet impact?
Yeah, I... And Steve's back there, so he can correct me if I get it wrong. But the basic idea is, and it's very, again, similar to what we did with our RadHard program. So, it will involve, you know, the purchasing of equipment. Sometimes that equipment flows to the top line, sometimes it doesn't, depending upon, you know, ultimate ownership of that equipment.
Okay.
The other piece is this NRE, so that would kind of go into our ATS bucket, so the ATS revenue. But the way it works is, every year, when Congress passes a budget, they put an allocation of funding.
Okay.
The initial allocation we got from last year's budget was $3.5 million. This year, of course, the new budget, hopefully getting passed as we speak, will then provide another tranche of funding. It's over a five-year period.
Okay.
Front-end loaded, of course, 'cause we wanna go as fast as possible. But, it's essentially a program that gets scoped, and then as we consume the program, you know, it, it kind of flows through the income statement. If I'm not wrong, Steve, it all flows through as some form of revenue. It's just, obviously, the engineering services revenue has much better margins than the tool revenue.
Right.
The tool revenue are... You know, given our expertise, our customers typically say, "We want you to buy the tools. We want you to enable the capability." But when we're, you know, doing a tool purchase as it relates to a program, you know, there's a little bit of margin there, but there's not a lot that you can-
Yeah
... go after, 'cause tool prices are tool prices. And so, you know, what you'll see, again, is the program will move through multiple phases. First, you'll get the equipment installed. We'll be working with Deca. We've already been working with Deca for the last couple of years to begin to do what we call hybrid processing. So as tools get ordered, we'll start creating, you know, kind of flows that where all the equipment doesn't reside in Florida, but may reside at a supplier's facility, so that we can get prototypes out by the end of this coming year.
Okay.
So basically, by the end of 2024, we expect to be giving prototypes to potential customers. We also see this, you know, the Florida opportunity as a, you know, we put a statement of intent to expand the facility down there tied to the CHIPS Act.
Mm-hmm.
And so we believe this you know award and this strong move into advanced packaging gives us a lot of momentum to go after you know some additional scale up. You know, when you talk about advanced packaging and specifically fan-out, the long-term goal is to do it on panels, not wafers.
Right. Right.
So this is where you move from a 300 mm wafer to a 600 mm² panel, and part of our thinking is that would, you know, envision putting a panel-like facility in the U.S. that could run these at scale.
Can DECA's patterning do panels, or do you have to then switch?
Absolutely.
Thank you.
Yeah, that-
Okay
... that's kind of, you know, I forgot to mention, but Qualcomm owns 18%, and this is all public, of Deca.
Okay.
So they've been a longtime user. All the Snapdragon and everything that they're making with any kind of advanced capability is using this technology. So we believe getting it in the U.S., getting it established... Again, the DoD sees multiple applications for their community, and we were excited to see AMD, you know, you know, verbally-
Yeah
... support, bringing this capability to the U.S. as well.
Right. No, makes sense. As a leader, probably the leader, as-
Yeah
... you said, in advanced chiplets. You know, I don't wanna try to get too close to financial reserves, certainly for the near term, but it sounds like, you know, in the third quarter conference call, you talked about increasing tool revenue-
Mm-hmm
... for 2024. I think that was more associated likely with the Rad-Hard program. Was this award sort of anticipated in the guidance and your outlook for 2024 back in the third quarter, or should investors sort of see this RESHAPE Award as potentially upside, a new opportunity that wasn't perhaps contemplated in that, you know, discussion from last quarter?
Yeah. So we were definitely anticipating it.
Okay.
We did not hear about it until Q4.
Got it. Okay.
You know, as we talked in the call, one of the things that we believe is going to happen as this year unfolds is we're gonna have more tool revenue than we ever have had as a company since 2017. And it's an indication that our partners are investing heavily. We talk about, you know, the Rad-Hard program, the focal plane array, our infrared imaging program. We have other programs also that we do with the DoD, and collectively those are causing us to, you know, have a period of time where we're gonna be investing and bringing in a lot of tools. And so you're gonna see a big jump this coming year. Now, we knew and we're confident that we would get this award.
Okay.
Of course, you never... And so, let's just say that this award now coming into play as we begin to look at when tools will be delivered, we may even see a higher tool level of revenue than what we were anticipating.
Okay.
So that would be my takeaway. The fact that we are planning on the program, we've built it into, you know, our plan this year. We're ready to execute on it.
Got it.
It's not like we have to ramp up. We were literally in Florida last week, kicking off the program with the DoD, some of our other partners. So, we expect it to be immediately accretive to 2024, and it will definitely add to the breadth of tools, 'cause now we're buying tools not just for Minnesota, but also for Florida.
For Florida. Yeah.
And you know, one thing to keep in mind is that we are in an industry downturn. There's a lot of other companies that are not ordering tools. So one of the things that's happened is, as others have pushed out, slots have opened up. We've been able to secure, because we have committed customers that wanna move as fast as possible, to securing slots in 2024. And one other point is, ATS revenue is great for us, but when you're installing tools, those are sometimes the same resources that are developing new technologies. And so, that's the other kind of margin dynamic that we have to navigate our way through.
Now, I get asked a lot, so I'll just say it before you ask is, "Why, why is tool investments from a supplier a good thing?" It's literally an investment in the future. It's us not having to invest in the tools. You hear about many semiconductor companies who are constantly having to take whatever money they're making and buy more stuff-
Right
... to stay competitive. Our model is we get our customers to do that for us. The other is a tool investment in 2024 will result in new revenue. Think of that two to three years down the road, and then that's kind of what you're gonna see now with the Rad-Hard program. We're... You know, I call it irreversible momentum. We have enough of these programs coming in play now, and the momentum is picking up on enough of them, that even if one slows down, others are moving faster. And that's why we've kind of continued to see record ATS growth, record revenue for the company, is we've really taken this ATS model and kind of went all in on it. And that... You know, again, an ATS program of today is a future wafer services program, too.
Right. Right.
So what we've done is taken a legacy facility, we've converted it with a bunch of new technologies. You know, Infineon is our main volume supplier today. As they went through their dynamics, of course, the slowdown is now beginning to affect automotive and industrial. And as those things occur, we just lean further in on ATS-
Yeah
... because that's not just building-
Building the future.
Yeah, it's building the future, and these are all technologies. Not all of them will make them to volume production-
Right
... but some percent will, and they'll all be new technologies that we own the IP for and that we're the sole source manufacturer of.
So you get tool revenue or, or revenue through the income statement, admittedly, maybe not at the same margins. The customer in that case then would own the tool, but you have sort of the rights to, to the use of the tool.
Mm
... to the extent that the customer doesn't need it 100% of the time, then you can use it for other customers or other technology development. Is that right?
Generally.
Okay.
In most cases, we own the tool.
Oh, you? Okay.
Uh, yeah.
Okay. Got it.
There are some specific cases where the end party would want to own the tool, and we usually have some timeline, like after a period of time, we would own it anyway. But that's actually the best part of the model, is that, you know, our customers buy the tool. The tool is ours. We can own it and operate it. Of course, your other part was correct, they get first rights.
Yeah
... to use the tool, and sometimes it makes buying a tool more complex, 'cause we may have a customer that wants a third of a tool. We may have another customer that wants a third, and another customer. Now, we have to go get a tool that can do all three things.
Right.
Right? And so sometimes, we're buying a tool and then working with the supplier to make it do things it wasn't originally designed to do. This is especially true in superconducting, which happens at very low temperatures. And so, that's something we've gotten really good at as well. And it's also something that is very valuable to the person that's, you know, you know, acquiring the tool with us because they know the tool is gonna run. Like, if I only need a third of the tool, that means that if I'm buying it myself, it's only running a third of the time. And in our business, you want everything running all the time.
Right. Right.
That's how you get-
Certainly, the best way to get nice margins-
Yeah, nice margins, nice.
... to be utilized.
Yep, absolutely. So, we think, you know... We constantly hear about semiconductor companies and the billions they're gonna have to invest in the future. That's why we call it a CapEx-light model. We do invest in R&D, we do invest in maintenance, you know, CapEx, but for big initiatives, we get our customers to pay for it.
Yeah.
Frankly, it forces them to have skin in the game, too.
One last question on tool revenue, and then I'll come back to a couple final questions about Florida. If the customer pays for the tool-
Mm
...does depreciation sit on your balance sheet or is that, does that sit on the customer's balance sheet?
It would still sit on our balance sheet.
Still. Okay.
Yeah.
Okay.
Correct, Steve?
Depends. If we own the tool, it sits on the balance sheet. If the customer owns the tool...
Then it's on there.
Okay.
Yeah.
So they pay for it sometimes and maybe even incur the depreciation. It's pretty capital efficient.
Yeah.
Got it. Okay. Back to Florida. Does the award accelerate the timeline for first revenue out of the Florida facility? And, and can you remind us?
Yeah
... expectations when you think that first revenue is generated?
Yeah. So, we have been generating revenue from Florida-
Okay
... albeit not very much. We had the program, the IBAS program, which we also announced moved to phase four. This was the Interposer program-
Interposer, yeah
... with, with IMEC. So, that program was existing, and there was a handful of others. We have some other hybrid bonding programs going, but... So let's just say we're in the early stages of revenue. This will certainly start driving more revenue, this year-
Yeah
... into Florida, and then that'll continue to grow and scale. But the, the other thing that, you know, we are a products or a services company, so we don't make our own products, but we do make platforms. So what we'll be doing in Florida is creating a heterogeneous integration platform that will incorporate, you know, not only DoD requirements... And the DoD community is interesting 'cause you're playing with all the same guys.
Okay.
So, whether it's our Rad-Hard program, our ROIC program, which is the focal plane arrays, or this program, it's all the same community, so you're interacting with the same people, and they're creating a more consolidated supply chain around SkyWater. That said, those programs tend to move slower. We see the opportunity to have this technology in the U.S. That's why the partnership with AMD was critical because that will allow us to move that much faster. So I believe getting the technology stood up, getting prototypes in the hands of customers, puts us on a path by... You know, as you get into 2025 and, you know, the second half of this decade, this is gonna be a huge driver for us because, you know, TSMC, Intel, Samsung, they get a lot of attention because of extreme node.
But to make those extreme node devices work, you need heterogeneous integration, right? And having an advanced solution based in the U.S. that can be done at scale, which is our ultimate goal, will really differentiate SkyWater and also move us out of just being 200 mm into 300 mm, and do it in a way where we don't have... You know, there's a lot of expertise, a lot of capacity at 300 mm. It's a way to move us in an area where we think we can be very, very successful.
Got it. Got it. I wanted to turn to the CHIPS Act. You submitted-
Yeah
... your application recently. You probably spent a good amount of 2023 sort of preparing that application. What was sort of most important, would you say, about your-
Mm
... your CHIPS Act application? And sort of, what would your expectations be for next steps, and when you might hear more about the application?
Yeah. So, you know, the process kind of kicked off in early 2023 through the NOFO, which is Notice of Funding Opportunity. Of course, the government's first focus was advanced nodes, right? So that was, you know, all the applications were focused there. We submitted letters of intent for both Minnesota, Florida, and our potential fab with Purdue-
Okay
... up in Indiana. So letters of intent kind of say, "Yeah, I'm gonna be doing something." And then, we prioritized Minnesota, focused around modernization and automation. You know, the fab was built, you know, back in late 1980s, early 1990s, you know, so it's 30+ years old. So it needs some upgrades, and a lot of existing older facilities are doing that. And so we got into the process. As we were doing that, some of these other government programs were coming into play, and so one of the reasons it, you know, arguably took us maybe a little longer than others, is because we had to, you know, calibrate all that. When you're getting money from the government for certain programs, you don't wanna be asking for money, you know... through another program.
So we, we spend a lot of time getting that calibrated, learned a lot about, the CHIPS program office. Now we have very good relationships with them, and, you know, once you submit it, once you submit your final application, you, you get asked to submit a final application. So you do a pre-app, they look at it, and if they think there's an opportunity, they say, "Do a final app.
Okay.
Once you do the final app, now it's kind of in the hands of the government, the what's called the Chief Investment Office-
Okay
... the CHIPS Investment Office, which has a Chief Investment Officer, Todd Fisher is the gentleman's name, and then he will go through and vet the investment. And so we... You know, that's probably a three-month process. So we expect some time, you know, certainly in the first half of this year-
First half you may hear something about talking.
... to be hearing what's going on. And now that that one's behind us, we'll be working with, you know, the Florida facility, like I said.
Florida, yeah
... to look at what can be done there. You know, Florida is already a public-private partnership. There's tons of enthusiasm from Osceola County and the state of Florida to expand what they call the Space Coast and build that out. They have a mechanism already. They built the facility that we have in Florida. We just lease it for $1 a year, and after 25 years-
Great lease!
... we own it outright. Yeah, exactly. That's the kind of deals that we like.
Yeah.
So, you know, we, we are, you know, probably, you know, gonna be moving and defining what that looks like next, and then in parallel, we're working with Purdue. For those who didn't see, they announced the CEO of what's called the National Semiconductor Technology Center. This is Deirdre Hanford. She came from Synopsys. She is going to be standing up, think of the $11 billion. So there was the $39 billion to build fabs.
Right, this is the more research-
There was $11 billion... Yes.
Yeah.
That's picking up steam, and a lot of our engagement with Purdue is kind of centered around that, creating that innovation engine for non-advanced node.
Okay.
you know, she's been very involved with Purdue and all that, so we just think that's-
Will Purdue try to become one of the areas of excellence? I think they're trying to set up, you know, you know-
Centers of excellence? Yeah.
Centers of excellence.
Yeah.
That was the term I was looking for, yeah.
Yeah. And we also, and this didn't get a lot of fanfare, but there's what was called Microelectronics Commons. That was basically the $2 billion that was part of the CHIPS that the DoD was allocated, and they've already identified what's called the hubs. They're called regional hubs, and now those hubs are all submitting proposals for new, you know, programs, new development programs, and SkyWater is a what's called a core fabricator, you know, for seven of eight of those hubs.
Wow! Okay.
So, we have engagements with those. So, you know, part of the questions that always gets asked is, you have the strong ATS business today, how do you keep that running? Well, the CHIPS Act, a lot of, the engagements we have with universities, national labs, that becomes kind of the funnel that keeps our ATS engine running, and then over time, we'll continue to transition programs. We have several biomed programs that we're gonna be moving to volume production this year. And then you'll start seeing the Rad- Hard programs, and that's when you kind of see the engine really working from ideation to productization, all within the same ecosystem.
I wanted to spend a couple more minutes back on the ATS business. You know, the business was $91 million in 2021. I think we have it at about $207 million in 2023, so more than doubling in two years.
Yeah.
You know, much, much stronger growth than the semi industry, which obviously went-
Mm
... through a downturn during that period. But maybe talk about a couple of the maybe bigger platforms or programs that are driving that ATS growth. And I'll jump to the biggest question we probably get from investors, and I'm sure you hear it, too-
Yeah
... is: How do you convert that pretty strong ATS revenue stream into wafer services?
Yeah. Yeah, great question. So, you know, we certainly have enjoyed a close relationship with the DoD, and that's driven some of that growth that you mentioned. But we also have multiple programs, a company called PsiQuantum, who is currently privately held. They're doing quantum-based computation using photonics-
Yep, yeah
... and superconducting. A very kind of unique combination. You know, great partnership there, moving very aggressively. So what, what we kind of saw happen as the industry slowed down is there was those that couldn't raise money for, you know, call it the start-up realm, and they went away. The others that had strong solutions got even more, got more investment. They wanted to move faster, so that also is accelerating the growth. And then, I think, you know, over time, what we, what we see happening, is these programs, and we've talked about the funnel, you start out doing initial kind of proof of concept all the way through, you know, qualifying a process to go to volume production, and we now have a lot of different programs throughout that entire funnel.
What you'll start seeing, and what's really required, there's two things required to get something to volume. One is the technology works. That's kind of what we're here to do, prove the technology works. Then the people that we're providing the services to, they have to go create a market for it, and the ability to, you know, grow and scale that market is highly dependent on them. Our goal is to have the product ready. And so that's one of the dynamics that we're dealing with, and one of the things we've learned is, in many cases, for these new emerging products, we have to know the markets as well as our customers do, because we're prioritizing things as well.
So, we see, you know, certainly in the med device area, that'll be what you start to see the first kind of transition occurring, and that'll be happening, you know, this year, the early part of this year. And then as we get into you know, later parts of this year, going into next year, you'll start hearing about the rad hard and the ROICs also moving into volume production. Then we, you know, we have other CMOS derivatives, like we have CMOS with our embedded memory today, which came from Sandia Labs. This is old Cypress technology. We're looking at Weebit, who is an Israeli-based company, who has ReRAM.
ReRAM, yeah.
Yeah, and they have a-
I think they present tomorrow.
Yeah, and we have, you know. So we are the U.S. solution provider for embedding their their memory. So this provides a path for, call it, legacy 19 nm-90 nm to get into a whole new group of solutions-
Okay
... because ReRAM has just got a lot of properties.
So you get some ReRAM stuff, you get some MRAM-
Yep
... opportunity, yeah.
Yeah, there, there's another one, you know, right now, LiDAR is done through a combination of, you know, a MEMS device, a sensing device, coupled with a CMOS device. There's a company we're working on that is moving towards an integrated solid-state device. So it's basically solid-state LiDAR, single solution. We see, you know, us, you know, really becoming a leader in this space. It's very, you know, complementary to core capabilities we have. But this could be transformational in terms of, you know, as LiDAR becomes more pervasive, especially night-based LiDAR, we, we just see this being an area for us. So our whole goal is to be in markets that are just starting, right?
Yeah.
You know, if there is anything that's frustrating, it's hearing that people call us legacy node, legacy fab, but we don't have any legacy, you know, products in our fab. I mean, we do have some-
Yeah, specialty-
... some, yeah
... is, yeah.
Yeah. But it's all emerging, and it's all exciting, and, you know, it's one of the things that I think makes SkyWater unique, is that there's a lot of innovation that occurs in our industry that isn't just on extreme node. And so we feel like having a solution that companies can come to is part of the reason you've seen that strong ATS growth. There's a tremendous appetite, especially, you know, there's thousands of fabless companies. There's about 10-20 that can get a lot of attention out of, we'll call it the Taiwanese foundry ecosystem.
Right, right.
We can offer that here in the U.S. in a differentiated way, and we think it's definitely catching on.
Yeah. I know A&D has been, I think you said on the recent call, about half perhaps, of that ATS business.
Mm-hmm.
Maybe spend a minute talking about the Rad-Hard program, because I know that is one of the bigger programs.
Absolutely
... you have. I think you're supposed to get to sort of a product qualification around the end of 2024, and so it starts to potentially generate-
Yeah
... maybe starts to transition into wafer services in 2025. But you know, walk through that, the sort of what you can say about that program and sort of how do you see it transitioning from, like, an ATS revenue stream to potentially a wafer services?
Yeah, great question. So back in 2018, 2019, the DoD decided they wanted a second source and a more capable source for their strategic Rad-Hard. This is tied to what's called the nuclear refresh, and it was also tied to a lot of new programs, new devices that, you know, are being envisioned for space-based applications. So they came to SkyWater and said, "You know, you guys have a 90 nm process. We want you to stand up a 90 nm technology. We want it to transfer from MIT Lincoln Labs. We want it to use a special substrate that MIT Lincoln Labs..." This is an SOI, silicon on insulator substrate-
Okay
... that they had fabricated. So think of MIT Lincoln Labs had done some prototyping work, but it would've all been on aluminum. The current solution was 150 nm on aluminum from Honeywell. They said, "We want a 90 nm solution using this rad hard substrate, based on the MIT technology, but we also want copper interconnects on it as well.
Okay.
So our solution has many technological advantages. It obviously has the Rad-Hard substrate, but it also has copper interconnects, which at 90 is kind of like where the industry went.
Right.
It just hadn't happened on Rad-Hard applications. So we're bringing a more powerful solution set. That said, you've got Honeywell as the incumbent. So, Honeywell, of course, was also working on a 90 aluminum-based solution.
Okay.
Different substrate starting material. So, you know, they're gonna be our competitor as we go, not only win new designs, but then, as we go through kind of the refresh, you know, the government's going to point the primes to SkyWater because they invested $270 million to stand up this technology. So, where a lot of our competition is, is for new programs.
Okay.
So there's a new design that someone's working on. You've got an incumbent technology that's maybe inferior, but it's got flight heritage and all these things, and then you've got a new one. So there's a lot of effort focused on Rad-Hard. We think, you know, it's moving the way we want it to move. To put it in perspective, though, that, you know, this is a four or five year program to kind of take this technology and get it ready.
Okay.
Once it's in production, it'll be in production for seven to 10 years.
Yeah.
So it becomes a very strong recurring revenue stream. Those wafers sell for... You know, TSMC gets a lot of fanfare. They sell $20,000 wafers.
Okay.
We'll be able to sell $20,000 wafers, too, with a fraction of the investment.
Right.
And then the other part of A&D is what's called focal plane array or uncooled infrared imaging. So this is the sensing that allows, whether it's a satellite or an F-35, to you know, sense to see things, right?
Right.
And then you have the sensing application, and then you have what's called the readout IC or a focal plane array that essentially translates that into a digital signal, and then that gets incorporated into the rest of the solution. These are two areas where we have many of the same customers. Many of our solutions are going into the same end state solution, and so we have a vested interest, as do our customers, as does our country, to get all these things ready to go in 2025.
On the $270 million award across, I think, two phases for Rad-Hard, where are we in that funding? Is a lot of that still to come? Obviously, you talked about, you know, once it's qualified, the primes will come to you for-
Yeah
... specific products, and that sounds like that's a whole another revenue stream that would be outside of that $270.
Absolutely.
Okay.
Yeah, so the $270 is to stand up the technology.
Got it.
We're getting towards the end. We got a final tranche of $99 million to move it into what's called productization qualification.
Okay.
That's what we're in right now. As we exit that, then it'll go from standing up the technology to using the technology, and that's where we'll start getting design wins. That's where you get to $20,000 wafers.
Got it.
And again, you don't have to... You know, if I make $1,000 wafers, $20,000 a wafer, that's $20 million. So you don't need a whole lot of those to have a significant impact. Again, very, very high margins.
I understand that some of this technology was developed at Lincoln Laboratory, MIT Lincoln Laboratory. You're taking this technology to market. Honeywell, I don't know if it's it sounds like it may be a slightly different process, but does anyone else have access to this Rad-Hard? Or, you know, my guess is the primes may wanna have a couple of suppliers for supplier diversity-
Yeah
... and so-
That's where we come in.
... if it's only you and Honeywell-
Yeah
... it still sounds like it could be a good position.
Yeah. Yeah, it's basically us and Honeywell are the two main providers. We're, again, the new guys, but we have a better technology. They're the incumbents. The other thing is we're a pure-play foundry. A lot of companies just like working with pure-play foundries because they're not in competition. That's frankly Intel's problem. When you have a product business and a services business, there's always a potential conflict.
Right.
We try to avoid that.
Well, it looks like we're coming up against the end of the session here, so we'll wrap.