Welcome, everybody. Good morning. My name is Quinn Bolton. I'm a semiconductor analyst with Needham & Company. Thank you for joining us at the 27th Annual Growth Conference. It's my pleasure to host this fireside chat with SkyWater. SkyWater is a semiconductor manufacturer and a DMEA-accredited Category 1A trusted supplier that serves the growing markets of aerospace and defense, automotive, biomedical, industrial, and quantum computing. Joining me from the company this morning are Thomas Sonderman, CEO, Steve Manko, CFO, and we have Claire McAdams in the audience, who heads up investor relations and strategic initiatives for the company. Tom, Steve, thank you for joining us.
Thanks.
Thank you.
Since, Tom, there may be some investors in the audience that are new to SkyWater, can you give us a brief overview of the company and touch on the company's ATS business, which makes you unique in the domestic foundry space?
Yeah, thanks. And thanks for having us at the conference. So SkyWater Technology was created in 2017. We spun out of Cypress Semiconductor. When we were created, one of the things we wanted to do is differentiate the capabilities we were bringing to market versus, say, traditional foundry, which was either advanced node like a TSMC or specialty. So we refer to ourselves as a technology foundry. And a big part of that strategy was monetizing R&D and a high-volume manufacturing facility. And that's really what we've proven over the last seven years. We have over 50 active engagements spread across kind of half and half across DOD customers and commercial customers to innovate inside our volume manufacturing facility, creating new emerging capabilities. The fab, foundationally, is a CMOS facility with capabilities down to 90 nm.
With that foundation, we've been able to leverage emerging technologies like MEMS devices, superconducting films, photonics devices, bringing those and integrating them into our CMOS foundation, which creates a lot of differentiation. So what you've seen over the last seven years is an acceleration in our growth of ATS. It grew around 35% through the last three years. Through 2023, we haven't put out our 2024 numbers yet, but very solid growth in ATS. And that's been against the backdrop of a declining wafer services environment because of the correction that's going on in the semi industry. So we've really been able to accelerate many of those programs. We've also announced last year multiple conversions of companies that started in ATS that are now moving to volume production.
Today, we announced our ThermaView platform, which is a 90 nm platform for thermal imaging devices, thermal imaging cameras, and a great example of a program that started in ATS, multiple individual customer engagements. We created a platform. Now, while we still have a lot of those one-on-one customer engagements that are going to actually be going to production later this year, we now have a kind of traditional foundry offering that others can design and use to bring their products to market.
You touched on a lot of points here in my next question, but you guys have grown nine quarters consecutively in a pretty tough market for analog mixed signal and specialty foundries. How much of that is your ATS business and this technology service, but maybe also then touch just at a high level on the tools side of the business, which has also been very strong and sort of the idea of customer-funded investments in the business?
Yeah, one of the things that also makes SkyWater unique is we're CapEx light. And so what we've really been able to do is, again, against the backdrop of a soft, broader semiconductor market, we've leaned into ATS. A lot of the growth and the top-line revenue is definitely coming from ATS. Wafer services has actually been declining. But the strategy, when we took the fab over from Cypress, we had legacy technologies that were not very profitable. And so the way to think of it is, over the next three years, we're taking the same volume in the fab, lots of tool investment, $200 million between 2024 to 2026. And all that new capability is going to allow us to think of it as create a 5x multiple in terms of the average selling price of the technologies that are coming out of the fab.
We're replacing legacy volume with new volume that's on technologies that we created while we also, in parallel, had this very strong ATS business. The reason the ATS business is strong and we believe will only get stronger is because of what you alluded to, this tool investment, which is literally our customers buying tools that we don't have to procure. We think projecting the $200 million through 2026, coupled with the $120 million we had leading into 2024, and then we announced about $35 million of state and federal matching for our CHIPS award. That's $350 million of investment into SkyWater that we did not have to go raise money for.
If you go back to the time of the IPO, we had around $60 million in debt. We now have all this new investment, and we still have about $60 million in debt. And I think it's an important testament that, not only to SkyWater, but when you ask the customers to pay for the R&D, we found they're willing to do it because we allow them to innovate. And in the end, it's all about velocity. We go from ideation to commercialization a lot faster with our model compared to a traditional foundry.
Thinking just sort of over the intermediate term, next three to five years, what are some of the biggest trends you see affecting your business and maybe touch specifically on the reshoring trend and how you may benefit from that?
Yep, absolutely. And Steve Manko, if I leave anything, but I mean, when SkyWater was created, we were like salmon swimming upstream. It was 2017. Everything was in Asia. And so that's why we decided we had to create something unique, which was the technology foundry, technology as a service concept. Then COVID happened, and everyone said, "Hey, we need semiconductors made in the U.S." And so we think, as really the only 100% U.S. investor-owned pure-play foundry in America, certainly on 200 mm for front-end manufacturing, we provide a very unique capability. One of the other things, and the thermal imaging platform I mentioned is a good example, is by offering differentiated technology, you can draw people from one foundry to another. Most of the consumers of ROICs were getting their silicon from Tower and on 150 nm. We offered 90 nm. Now we have a leading position.
We think by bringing advanced solutions, you can begin to draw customers that get their silicon fabricated in Asia foundries into the U.S. We also think the new administration will be leaning in hard into Made in America. We get lots of questions about tariffs and all that. Who knows what's going to happen? But clearly, there's a desire within the customer base to have things made in the U.S. And we think we can be a large participant in that, especially given our unique capabilities.
Perfect. Let's go to the unique aspect of the story, the Advanced Technology Services. That continues to be sort of the bulk of the revenue stream today. It's been about $60 million a quarter the last couple of quarters. Maybe talk about some of the specific applications, whether they're aerospace and defense or commercial programs that you have in ATS that are sort of driving the growth of this business.
Yeah, so one area is advanced compute. So think of that as quantum computing, which is getting a lot of attention these days. We've been working with D-Wave back to 2013, long history making quantum annealing QPUs, quantum processing units. So we have a long history there. PsiQuantum is another company we talk publicly about doing photonics-based quantum. So there's a lot of ATS investment going into that space. The area of photonics is another emerging area. Photonics, again, crosses over into quantum, crosses over into the data center, so crosses over into moving into solid-state LiDAR versus having to use other MEMS-type devices to detect motion thermal imaging. So we see a lot of the emerging technologies being drivers of our ATS, but it's also the fact that they can integrate those new capabilities with traditional CMOS. And I think there are MEMS fabs, and there's power fabs.
One thing about SkyWater is we're kind of this combination of CMOS coupled with a lot of other intricate technologies that we have the ability to integrate into that CMOS platform. That's where advanced packaging comes in. A lot of advanced packaging is getting a lot of attention, but many of the solutions we've been making for years are basically advanced packaging. If you look at a thermal camera, you've got a microbolometer that we also make, and you've got a ROIC, and you put that together, and now you have a thermal imaging camera. We think that that combination, that's almost like a Lego strategy where we have a bunch of Lego blocks, and customers can come in and combine them in different ways. The customization is what makes this unique. Most foundries do not like customization. We embrace it.
And so we think the ATS business will continue to be strong and grow. But the other equally exciting part is the conversion of ATS customers to volume manufacturing. And again, think of it as a 5x multiplier from what we would sell a commercial wafer on legacy technology versus what we can sell with these emerging technologies. And that allows us to reach, we'll call it similar revenue levels to when we acquired the business, but with much better margins and technologies that we're the originator of. And we own all the manufacturing IP for and for products that are just beginning to hit the market.
Obviously, much better margins.
Yep.
If I look at ATS in 2024, you saw pretty good growth for the full year. That growth was a little bit front half loaded. What drove the sort of relative slowdown in the second half, or at least in the third quarter, and you got it for a little bit of a recovery in Q4, but probably stronger first half than second half.
Yeah, I'll start, and Steve, you can because he has to live with this. One of the things with R&D programs is you typically have an R&D budget, and success and innovation sometimes can consume the budget faster than anticipated. I think that was part of what we saw happening, so you kind of get through the budget faster than you expect. We were very successful in the front end in the second half. The other dynamic that we deal with is these continuing resolutions and the Congress and the President not being able to agree on spending levels, so we always kind of, whenever a continuing resolution just means you're going to stay at your prior year spending levels.
But if you're not approving the New Year's budget until you're six months into the current year, then you're kind of at half the spend levels of the prior year. So a lot of those dynamics come into play. I would say that we feel like the overall ATS business has got a robust set of customers to where we're less sensitive to one or two customers slowing down. And that'll only get better, but that's part of it. It's just a matter of how fast you consume the budgets and then when the new budgets get renewed. We were hoping that in Q4 we'd go into the new 2025 budget, but even as of now, that budget still hasn't been approved.
Yeah, that's going to be kind of where I was going with this is to the extent that some of this is government funded and there can be puts and takes. How much risk do you see just one around the new administration coming in next week? Could that be just a source of volatility? And second, I think the government, as you just mentioned, is sort of only funded through March 14th. And so we probably have a budget battle over the next couple of months. How are you feeling about those different events?
Yeah, well, I think the new administration has actually given, again, the focus on made in the USA, but also the focus on efficiency. The whole, I would say, DOD community is going through kind of what we saw happen in the commercial space when we went from laptops to smartphones. That's really when the fabless foundry model took off. And it's because we had to make billions of smartphones versus millions of PCs and laptops. And then the DOD community, especially the products we make, and thermal imaging being a good example, you're going from making 1,000 F-35s to making a million drones. And the drones are advancing very rapidly.
The desire to have a centralized foundry model where someone like SkyWater can be the fabricator for thermal imaging, RadHard, advanced packaging, and all the primes essentially move away from their inefficient fabs and they kind of come to SkyWater, that's kind of what's underway. I think that'll only be amplified with the new administration because we're already proving it with three of the programs that won't be announced today, RadHard and advanced packaging. The desire to, there's a lot of CHIPS that are made overseas that go into mission-critical systems. SkyWater has the capabilities to make some of those CHIPS. Most DOD CHIPS are 90 nm-130 nm, believe it or not. We think that will only help things. Then the threat of tariffs also is making people that maybe would be looking to move something to a China fab because they have a lower price. They'll think twice about that because many customers want to have it made in the U.S. or it may not be economical to not have it made in the U.S. going forward.
Yeah, so it sounds like the change in administration probably, if anything, maybe a net positive. Could be some volatility around the budget, but obviously that's a very short-term thing.
Yeah, I'm optimistic that it's going to be a net positive. I mean, there's still some uncertainty. We talked about the slowness of the budget may reduce spend over the course of Q1 or how long it goes out. But if we think back some of these programs, some of these programs were actually started in the 2020 timeframe. So we don't think that these programs are just automatically going to go away. There's going to be a difference in direction. A lot of these were started during the previous Trump administration. I think they would continue with the same force. Once there's been four years of investment that have gone in, we've moved the technology forward. So I don't think there's a risk to us losing programs. I think there's just some uncertainty and some slowness over the first part of the year.
Yeah. What's the typical life, if there is a typical life of an ATS program? Is it a couple of years? Can it go longer? And how do you think about the transition from ATS into wafer services at the end of that ATS program?
The ThermaView program, our platform we announced today, that began back in 2020 as a series of one-on-one engagements with many of the DOD primes, and then in 2021, late 2021, we kind of said, well, let's create a platform around this. We collected kind of all the requirements, and in parallel, we've been still working with each of those primes. We expect to see wafer services revenue on our ROIC and microbolometer platforms this year. It's not like way in the future, but think of that as five years from initial engagement. We were able to take the Cypress IP and manufacturing capabilities we inherited, do some adaptation, and now have the best solution. Again, took customers away from Tower and onsemi at 150 nm because we had a better solution at 90 nm. In many ways, that's one extreme.
You could look at our RadHard program. That'll even be longer because you have to go through a bunch of qualifications, space heritage, which means you got to put your chip in space, and so that program began in 2019. It'll probably become meaningful from a wafer services revenue 10 years later, 10 years later, 2028, 2029. So that's a very long program, but then we're committed to making it for another 7- 10 years once it's qualified. And then you have on the biosensing side, we announced NanoDx back in 2022. They licensed some technology from IBM. We brought it into SkyWater, and then in Q4, we announced our manufacturing supply agreement with them. So think of that as like three years, two and a half to three years.
Depending upon the technology and how complicated it is, but also what's required to kind of freeze the process for qualification, I would say two to 10 years. But all of it is the thing I always like to point out is you always have to do innovation and R&D in our world. But our difference is we get our customers to pay for it. And a normal foundry, it's the foundry who pays for it and recoups that investment when it goes to production. And we get paid whether it ever goes to production or not. Of course, we want it to go to production, but we still make money on the front end. And we own all the manufacturing IP. So if they want to leave, they're going to have to pay us either way.
I know you've mentioned the ThermaView Solutions announcement from this morning a couple of times. I think you hit some of the highlights. But is there anything else you'd like to say about the ThermaView Solutions business? It sounds like it may contribute to wafer services this year, which is great to see.
Yeah, it's a good example. And why we're so excited at SkyWater is because it's our first platform. So it's a technology that was created solely by SkyWater, again, derived through our ATS business. It's the one-to-many approach where we were able to, t his is somewhat different, like the RadHard program was, the government invested $270 million to go stand up a platform. This same community we were working with on RadHard, we were working with in the ROICs and again, same approach. The difference is it was a bunch of ATS. It wasn't like a big funded program. We had a bunch of ATS engagements, and we created a common platform. So now you can download the PDK, you can go design CHIPS, and we expect many, basically anyone making ROICs to take advantage of this.
It's a 15% improvement in terms of the amount of information that you can fit in a pixel, which is kind of the language. So the pixel density is much higher because of the technology approach we use. But that said, the six customers that we have been working on that helped define the platform are much further along in their development. That's why while we're launching the platform today, we expect to see wafer services revenue later this year. And again, ROICs is, so in a digital camera, you have a microbolometer, which senses the image on an uncooled application. And then the ROIC translates that to a digital image. And so the thing we announced today is both the microbolometer technology coupled with the ROIC technology. We've been in production on microbolometers for several years. And again, we just continue to shrink those devices. They're also for us targeted towards DOD, but they're also applicable to automotive and medical imaging.
Yeah, it was going to be sort of my next question. I think most of the six customers that you helped define the platform were probably DOD.
Yep, exactly.
Aerospace and defense customers, are there commercial applications? You'd mentioned sort of F-35s going to drones. Could you see commercial drone applications? Medical, I mean, where could you see this technology going?
Yeah, I mean, again, it's a digital camera that uses cooled and uncooled thermal sensing to drive the image, and so automotive, the microbolometer technology is for uncooled infrared imaging. So that's like night vision goggles, security systems. These are all applications, and that's another technology we've been developing for many years with a key customer, and so I think in the automotive space, as we go to autonomous vehicles, the car's got to work as well in day as night. So a lot of the solutions, ROIC-based solutions are going to be really critical for not just LiDAR, but moving towards autonomous vehicles. So a lot of interest and even some of the customers that we deal with on the DOD side are starting to say, how can we get in the commercial market?
But because we have a good presence and a very strong solution with DOD, a lot of the commercial customers are also starting to come to us. In fact, we have several engagements underway in that space. So I think automotive imaging cameras for med devices. And then again, going back to an F-35 versus a drone, you have to miniaturize and you have to make it a lot more cost-effective. And you have to make it work in varying environments. And so that's driving a lot of innovation and the pace of innovation. And we believe that having the solution available will just accelerate that. And so I think a lot of the near-term interest will still be in DOD. And then the drone community will take advantage of it as well, the commercial drone community, once we figure out exactly what all that means, right?
Maybe you have Steve just looking at the tool revenue here. You guys saw a very strong uptick in tool revenue in 2024. I think you've sort of guided to cumulative tool revenue of about $200 million from 2024 to 2026. Just talk to us about sort of the pattern you see in that tool revenue over 2025 and 2026.
Yeah, well, it's an interesting pattern. Coming to the organization back in 2020, there's been a lot of surprises since I've been here dealing with so many R&D programs. I'll say one of the pleasant surprises I had that we really didn't have a lot of visibility when I came in was how much continued investment there would be in the organization. I'm quite thrilled that we're having this $200 million coming through. Before my time, the RadHard program was already in process that we saw investment coming through there. And I think our first quarter as a public company, we saw around $17 million of tool revenue. Really after that in 2021, there was really a stop to tool revenue being there.
So to see this amount of investment continuously coming back into the organization is something that really was a surprise to me and I'm very thankful for, of course, but it's going to be lumpy along the way. There's two easy ways to think about tool revenue. One would be a customer comes to us because we own a fab. Fabs are costly to own. They'd rather come to a fab as opposed to running and owning one on their own. And because we have the CMOS platform and a number of tools in there, we can do a number of things for them. They may do a little bit of customization that we welcome, but we don't have that specific tool that would open up that customization.
So our toolset may get them 90% of the way through the process, but we need one or two additional tools specifically for them. That's where a customer would come in and fund that additional one or two tools to complete the process flow. On the flip side, what we've seen in some of these larger dollars that have come through in 2024 and even towards the back half of 2025, these tool investments of this size are coming through just because it's not a one or two tools coming in for a process flow, but it's a platform being established. And a platform is going to be an immense flow coming through. Yes, maybe it can utilize some of what's already in the facility, but there's much more specialization. And that's what we're seeing.
So we saw a lot of that coming through for some of our A&D programs in 2024. The one that we're expecting in 2025 is the second half of the year. And that would relate to the $120 million award we received about a year ago for the fan out packaging in our Florida facility. So again, you're standing up a new platform. That's why there's a significant amount of tools coming in. So with that, if we quantify that, we're talking about maybe $300 million of investment that has gone into SkyWater. I mean, that's 3x the size of our initial public offering. So a lot of free capital coming in that I don't have to raise debt for, I don't have to raise equity for that's coming in. So we're thrilled with that. It shows that our customers are committed.
It shows that it's very difficult to leave our company once they're there. People would question, are these science fair projects? Well, if you're investing hundreds of millions of dollars, very likely not. So the one downside is it's a lumpy part of the business, right? So you're going to see ups and downs across quarters. And there's been some quarters where it's zero. I think as long as people understand the nature of what the tool revenue is, you need to take it out. We show that tool revenue specifically so you can really understand where the core revenue of the business is coming from. Evaluating us on our ATS and wafer services business and just seeing this tool revenue not as like something drastic is happening because tool revenue is dropping. It's just our outside basically free investment is down that year. But the amount of investment that we're getting through tools is nothing but positive, in my opinion.
Just to size it for the audience, you did sort of $75-$76 million or so in 2024. I think you expect a little bit of a pause in the first half. And then as the Florida investment starts to pick up, a sort of stronger second half and probably a pretty strong 2026 to get to that $200 million.
That's right. So clearly very lumpy to be going from a $30 million tool revenue quarter to something like potentially zero or $2 million in the first and second quarter of 2025. But again, once you understand really what that is, it's not a portion of our business that's not performing. It's just outside free investment that's coming in and it comes in kind of lumpy depending on where the program is.
Got it. Wanted to spend a couple of minutes on wafer services. Obviously, this business was, I think, affected by the inventory downturn. Where do you think we stand in terms of the inventory correction demand for your legacy part of wafer services? Does that recover at all? Do you think that continues to kind of grind lower? And then you talk about, I think, four or five customers transitioning from ATS to wafer services in 2024. Sounds like they're more transitioning in 2025, especially with ThermaView. What's the outlook for wafer services through the year, kind of that mix between legacy and new wafer service programs?
Yeah, so I'd say first off, we're projecting flat kind of Infineon type wafer services revenue this year.
Sorry, on a quarterly.
Yeah, quarterly, yeah. But just we're not projecting any growth in legacy this year just because that's kind of what the market's saying. And what we are seeing is if you look at Q4, it's kind of the trough as we begin to grow off that base, it'll be with all these new programs that we're talking about. So it's kind of flat with where we exited the year with Infineon and the legacy businesses. And then the growth will all be coming from the new converts, again, conversions from last year that will start ramping this year. And then the ThermaView customers that will start producing volume products in Q4, Q3 of this year, second half. And then the other thing that we're kind of, there's two buckets. So when we look at the ROICs and that, that's fairly predictable for us.
When you look at bio devices, these rapid diagnostic sensors, these are new products that have yet to hit the market, so it's always unclear exactly. We have some very aggressive forecasts, but we're modulating those. We want to have room for wafer services growth from Infineon if the market recovers and the demand comes back. Because right now, we're using ATS to absorb a lot of the fixed costs of the fab. The quicker we get wafer services growth, then that allows more of the ATS margin to flow through, so we're eager to do that, but we also want to rebuild back on the new stuff, and so we still have a strategic partnership with Infineon. They're a good customer, but we'll make the things that we need to make there, but only as much as we need without sacrificing ATS or the new wafer services business. We're kind of waiting to see. I guess the clear message is we're not dependent on the broader market recovery with the legacy business. A lot of this is all new stuff that we're really excited about. Our customers are also excited about.
Great. Let's talk about the advanced packaging business. Tell us a little bit about your Florida fab. I think it's going to be the first advanced packaging fab in the U.S. You also received a $120 million DOD award for the Deca fan-out wafer-level packaging technology. So just sort of tell us about the award and the facility.
Yeah, so Steve alluded to the $120 million that we were awarded a year ago. So in 2024, we've been ordering equipment. We're going to have around 50, 60 new tools in the Florida facility. This is an example of the DOD funding to stand up an entire flow. So this isn't buying one or two tools. This is standing up the complete Gen 2 flow that we've licensed from Deca Technologies. Deca's Gen 1 technology is basically designed by Qualcomm into smartphones that we all have in our pockets today. If you have an Apple smartphone, there's two CHIPS using the Deca technology, Gen 1 technology. We'll be bringing in Gen 2, and the goal is to stand up a full flow for not only prototyping, but early stage volume production. We have a partnership with ASU. ASU also licensed the Deca technology.
They're standing up a similar flow, and so as their capabilities come up, we move our tools in the second half. We expect to have prototypes available this time next year, early stage prototypes, and then over time, of course, example of the DOD having all the primes come to us to leverage advanced packaging, and again, advanced packaging, while it gets a lot of attention in the commercial side, it's equally important with the DOD, and a lot of it ties back to what we just talked about with drones is how do you make it more efficient to create computation environments where you don't have $150 million per plane to spend and ridiculous numbers.
So I think we feel good about that t he fact that Qualcomm's already using the technology, AMD is not a partner, but they are paying a lot of attention to what we're doing and are very enthusiastic about having, again, an open chiplet foundry here in the U.S. Our goal is to stand up the technology, obviously deliver it within the DOD sector, but also really think of it as create a business within a business. When I refer to an open chiplet foundry, it's open because if you go to TSMC or Intel or even Samsung, you have to use their front-end services to be able to leverage the advanced packaging capability. Deca's technology is also very competitive with TSMC. It allows very tight packing densities and multiple levels of fan-out interconnect, which again just allows you to pack more computational capabilities into a smaller footprint.
So it's kind of another way to think of Moore's Law, extending Moore's Law. And that really hasn't been driving any of our ATS business. If anything, it's been a drag. And so when you look at ATS growth, it's not only think of the $200 million of investment that's coming into the business. That's going to drive a lot of growth, continued growth in Minnesota, but also now Florida contributing. And we're really excited about both. We hired an individual from Intel who's coming in and again setting up as its own business unit. And it's very different than front-end ATS. And so we're kind of taking a fresh approach to how to maximize the impact for SkyWater.
Perfect. You guys in December, I think it was announced CHIPS Act funding and funding from the state of Minnesota Forward Fund. Just talk about the funding amounts and the planned use of those proceeds.
Good. You want me to get it? Go ahead. Yeah, so we have. It's a $35 million total grant, $16 million from the federal CHIPS program and $19 million from the Minnesota Forward Fund. And so that $35 million is part of a $127 million program for modernization and automation of our Minnesota facility. Obviously, with all the tool investment we've been talking about, we didn't have to go buy a lot of tools. And so we're focused on upgrading the infrastructure, making it more sustainable, putting in full automation. We're converting a test floor to more production capacity, upgrading our manufacturing systems. And a few one-off tool buys. And the way to think of this is these are investments we would have made anyway without CHIPS. But by having CHIPS, it's an accelerant.
So when coupled with the investment tax credit and the $35 million, we feel like we can bring in, we'll call it a seven-year investment program that we are funding through our annual maintenance CapEx into a five-year timeframe. And so it's just an accelerant, but it will allow all this other tool investment. So we believe we can really drive growth, not only continue growth in ATS, but in wafer services without having to go put shovels in the ground and talk about new fab for quite a while.
Perfect. Last question I wanted to end with is just for folks trying to model the business.
Good luck.
Talk to us about how we should be thinking about modeling gross margin. It seems like you guys have sort of recently encouraged analysts to look at your combined wafer services and ATS business and then model an incremental gross margin on the growth of that combined revenue stream. So tell us, is that the right way to think about it? And what kind of incremental gross margin should folks be using in their models?
Yep. Great question and you're spot on. That's right. The one thing that with tools, I guess, the one negative is it confused people on the way it was really flowing through what the business was doing on the margins and the flow through. Again, with the tool investments, we wanted to keep that CapEx light. We weren't marking up these tools. We're not building these tools. We are procuring them from the other big tool providers that are out there. We're not marking these up for significant profit and margin on what we're doing. So when you remove that and you put but you have that flow through our revenue profile, our gross profit doesn't change, but our margin percentage points change. And I think that's what was really confusing people in their models.
So I would say that the best thing to do is really look at us, evaluate our business from ATS revenue, wafer services revenue. That's really where you should be evaluating the growth and how we're doing and also the profitability of the company. One of the metrics that we've thrown out there was a couple of years ago. Now we said that our break-even point was around $45 million in quarterly cost. Anything above that amount on a revenue perspective would be having at least a 50% flow through to gross profit margin. We still stand by that. And the history has shown over that two-year period that we've been delivering at least 50% flow through every additional revenue dollar above $45 million. The other element that we, since we've grown since then, we said from a bottom line perspective, it's roughly around $70 million of quarterly revenue that it takes for us to break even on the bottom line. We're still standing beside that. Again, it somewhat depends on the mix of the revenue flowing through, but roughly around $70 million is when you can see that break-even.
$70 million of ATS and W S.
Exactly. So really it goes back to, I mean, you answered the question, you get it. Look at us from Wafer Services, ATS. That's really what we refer to. Look at the revenue, look at the margin from that, and then just put the tool revenue over top.
Perfect. All right.