Navitas Semiconductor Corporation (NVTS) Earnings Call Transcript & Summary

Great. Welcome back. I'm Joe Moore, Morgan Stanley semiconductor analyst. Very happy to have with us today the management team of Navitas, CEO, Gene Sheridan; and CFO, Todd Goodman. Thanks, guys for coming.

Thanks, Joe.

You have a range of knowledge about the company, for those listening, can you give us an overview of your business and what the story has been since the de-SPAC in 2021?

Yes. Navitas makes next-generation power semiconductors. For most of my career, and in our industry, we made power chips out of silicon. Kind of like Moore's Law, silicon is running out of steam in the power electronics space. It's being replaced by 2 compound semiconductors: one is called gallium nitride, or GaN; the other one is called silicon carbide, or you might say SiC. Those two have similar qualities compared to silicon. They can switch much faster and much more efficiently than traditional silicon. And speed and efficiency is sort of the name of the game in the world of power electronics. Silicon carbide does it extremely well at the highest voltage and highest power levels. Gallium nitride does it at more moderate voltage and power levels. So they're actually quite complementary in replacing what is a $20 billion to $30 billion heavily silicon-based power electronics world that can get displaced by these new materials to create power electronics that are faster charging, more efficient, smaller size, higher power density, lower cost and ultimately, more reliable. So a lot of benefits there. That's the basis of our company. We did the IPO with just gallium nitride. We used the proceeds from the IPO to acquire a silicon carbide company and a couple of other smaller complementary ones to now become the industry's only pure-play next-gen power semiconductor company. By that, we're actually the only company that has gallium nitride and silicon carbide without the distraction or dilution of traditional silicon power devices.

And can you talk about that a little bit? I mean, the advantage of being a pure-play would be focus. You have some agnosticism between different types of substrates. But you're also competing with bigger companies that might have more resources than you. Just how do you balance those things out?

Yes. I think you hit on the most important, which is focus. And so given our small size, we should have focus, agility, intimacy, innovation, risk taking, all of that to our advantage given our size. But that's key, to be focused. And where we're focused is actually where the combination of GaN and silicon carbide gets the best results. We started originally with GaN for mobile chargers, and we can talk about that. But with the combination, we're focused on AI data centers and electric vehicle onboard chargers. Both of those applications benefit from the combination of both, something we have unique strength in. And we go very deep at the application system level to teach our customers how to drive the best system innovation, system results, using that combination of GaN and silicon carbide and specifically in those 2 markets.

Great. So you've talked about a $450 million design win total in 2024. You've talked about a pipeline of $2.4 billion. Your revenue is a lot smaller than that. Can you talk about the definition you have for design wins and pipeline and kind of how you have so much robust growth coming?

Yes. And it's a blessing and occurs. Power is a really big market, so there's no shortage of opportunities. And frankly, $2.4 billion is too much of an opportunity for our company. That's everything we could touch, qualified opportunities of where our sales team has confirmed a real production-intent program that has a good technical fit and value proposition for our GaN or silicon carbide. So that's an early-stage qualification. It's not a design win. It's certainly not a purchase order. But $2.4 billion is big. The 3 segments I talked about, mobile, EV and data center, is about half of that. That's still a lot. And even in those worlds, we're really laser-focused on the best customer opportunities, the best value prop within mobile, EV and data center. And when we can actually take those opportunities and get them to a technical approval stage where our chip is being selected to go into production, into end hardware, that's what we call a design win. And that's what we're most proud of and most influential for our future growth, is $450 million of lifetime wins last year, which will really be kicking in later in '25, '26 and a little bit in '27.

Okay. Great. So shifting a little bit to that revenue growth profile. You've talked about some weakness in the March quarter, both seasonal and some demand weakness, and really through the first half. Can you just characterize what the headwinds are that you're seeing? And when you've given confidence that it will recover in the back half, what's driving that?

Yes. Certainly, the last 12 months, the whole industry saw a slowdown in the end markets of solar, EV and industrial. Any time, even if those are not flat or down, they're just slowing the growth rates, you're going to create some inventory congestion in the pipeline at the end customer's equipment, the component level at the customer, the distributor, the channel. So I think the whole industry has felt some inventory congestion associated with the slower growth rates of solar, EV and industrial. It feels like we're on the back end of that. We see stability. We see now some uptick in those markets. So it feels good for us that we're getting close to the end of that, but we still project Q1 is likely the bottom, Q2 a little bit more, and then we're off to a normal healthy growth rate, especially as all those design wins kick in, in the second half of the year.

Okay. And you're characterizing that. I mean, it's not just an inventory thing. There is some demand weakness that sort of triggers the inventory dynamic that you're seeing.

That's right. I mean EV, for example, is growing 50% a year. When they first took it down, it was like 30%. I think Tesla was actually flat there for a while, even slightly down. So any time you go from 50% to 30% or 50% to 10% or 30% to 0, the whole supply chain has got to hit the brakes. And you're going to get pockets of inventory that we all have to work through, in some cases, even stopping launch of new products to work through the selling of the old products before they launch the new ones. So I think most of that now is behind us, and we see a lot of growth going into the later quarters.

Okay. Great. So maybe you could talk about the GaN market a little bit, starting with your position in the market, the high end of the market, technology advantage. Just how are you feeling about your prospects in GaN?

Yes. Well, GaN started with mobile chargers. And as much as we've been at it for a few years, we really created the tipping point there, first when Anker launched their first aftermarket GaN chargers, after that followed Belkin, Amazon, all the aftermarket guys. Then Xiaomi's CEO stood up and said, "Look at this GaN charger. This is the future of chargers." And then all the mobile players followed from there: Xiaomi, Oppo; but not just in China, but all around the world, Samsung and many others. So we've now gone to where 10 of the top 10 mobile players across smartphone, tablets and laptops are all using Navitas GaN technology to supercharge their chargers. With that said, we're only at about 10% adoption rate. So we've had an incredible growth that took us to where we are today, but we still got 90% of those chargers shipping out there with silicon that over time will switch to GaN. So I think that's promising. And that's just the first vertical. Now we're recreating that kind of a tipping point, as I said, at AI data centers into electric vehicle, even GaN into solar is coming later this year.

Okay. Great. So on smartphones, at that level of penetration, it seems like there should be quite a bit of growth. Obviously, you have some big customers that aren't users of GaN. So just can you talk about that trajectory? How quickly can you see that penetration improve?

Yes. I think every year, it's going up and up. It used to be just, 3 years ago, you had that 5-watt Apple cube that was typical for all of us. We are delivering in that same form factor about 30 watts today, 6x more powerful, 6x faster charging, about the same size, which is an incredible experience. Started out with a price premium, double the price compared to the charger. You get the high end of the market. Now it's almost no premium at all. So it becomes the default way to charge. As you get these phones going to about 45 watts and higher, we think GaN is inevitable. We saw that already happen with Xiaomi and Oppo. Last year, they were at 30% GaN adoption rate. We see that same trend moving to others, Samsung, Apple and many others over time. So I think the trajectory is really strong to go from 10% adoption to capturing a big percentage of that remaining over the next few years.

Okay. Great. Where are we with GaN adoption in autos? And I guess we talked about the sort of Tesla moment when Tesla started using silicon carbide in size a few years ago. Do you see something similar on the GaN side?

Yes. In fact, that's really the tipping points you're classically looking for. Just like Navitas created that in the chargers first with Anker, Xiaomi and now 10 out of the top 10, tesla did that with silicon carbide 5 years ago in the traction motor. We're now creating that tipping point for GaN in the onboard charger for EVs. Just in the last month, one of the top EV players in China called Changan, C-H-A-N-G-A-N, they're the #3 EV player in China, put out a report announcing their intention to go across the board to GaN for next-generation onboard chargers that will launch next year. We have multiple other Tier 1 players like that, that are planning similar launches in GaN onboard chargers for next year. So we think that tipping point is right around the corner.

Okay. That's helpful. Can you talk about competition in GaN both from the more traditional players, from companies like Transphorm, which was acquired; GaN Systems, which was acquired, but also from Mainland China.

Yes. GaN Forever (sic) [ Forever GaN ] has had a very incredible potential in its speed and efficiency, like I mentioned, compared to silicon, but it's also had an Achilles heel. Driving them is extremely difficult. A silicon MOSFET, you can give it a voltage input that's very flexible, plus or minus 5 volts is enough to drive it. But a GaN device to operate at these high frequencies, you have to give it very precise and accurate, both in voltage and in timing. So this drive challenge has been there for the whole industry. Navitas uniquely solved this drive challenge by integrating the driver right into the GaN chip. So we can perfectly control the timing and the accuracy of that voltage to allow it to operate at the highest frequency, the highest efficiency without any reliability or robustness challenges. In a nutshell, we call that the world's first and only GaN power IC because it's an integrated circuit, not just a transistor, but the driver. And since then, we've opened up even more integration of sensing functions, level shifting, auxiliary power supplies, protection circuitry. So we've gone down a very exciting role of almost like a system on a chip, with a GaN chip integrating analog circuits all around it. And we've done this really in a class of one. We're the only company to do this type of innovation, and we've got 200 patents and counting kind of protecting it. So to your point, most of our competitors are really GaN discrete players. We offer a great discrete, but so much more wrapped around it with this integration, which offers a lot of system-level benefits.

So when I look at that competitive set, you really are differentiated. You're not seeing them into the types of the parts of the market where you're generating revenue.

That's right. I think if you want a low cost -- just get a small efficiency upgrade, but you're really cost-driven, you can probably take a discrete and do a low-cost driven. But if you want the highest efficiency, the fastest charging, the highest density, that system-level integration allows us to really lead in the high end and medium end of the markets.

Okay. And then within China, the Innoscience IPO raised some eyebrows, right? It had several billion of market cap, negative gross margins. That's interesting, right, because there's more than 1 Chinese company that fit that profile these days.

Not the first or last.

But that seems tough to compete with. So it's important to stay differentiated from them. I mean, can you talk about that? And to the extent that the China EV market is an important market for a lot of this onboard charging, does the definition of automotive grade tend to favor local suppliers over time?

Yes. I think the markets are getting more homogeneous in some ways. Automotive grade is a tough grade to hit, EV or traditional ICE, that's sort of universal around the world. I think in China, though, you've got additional pressures. We have a China-for-China strategy that I think has served us really well. China-for-China really just means give them the support, the capability, not just sales and marketing, tech support, we have finance, operations, R&D, all the things needed to serve the Chinese customers in China that can do business with our China subsidiary. So I think our China-for-China strategy has worked really well. Combine that with highly differentiated technology that they can't buy from an Innoscience from the local supplier, I think it explains why we've done so well in China so far.

Okay. Great. Could you talk about your strategic partnership with Infineon and how that benefits you?

Yes. So everything I've talked about so far is high-voltage power chips in GaN or silicon carbide, 650 volts and higher. This is going into high-voltage AC-to-DC power supplies typically. But in some applications, there's also power supplies that are lower voltage, DC-to-DC converters, especially true in the data center. The first stage of power conversion is high-voltage AC to DC. The output is 48 volts. Second stage is a 48-volt power supply that converts it down even lower to power the GPUs and CPUs. Today, we're not participating in that next 48-volt converter, but that's about to change. We formed a partnership with Infineon. So they're both offering dual footprint, great GaN chip technology at the lower voltage to target the 48-volt market. So it's kind of two benefits for us. One, we're opening up a whole new market, doubling our data center potential for the 48-volt power supply, similar to the EV, by the way, the onboard charger powers from the grid from your home. But there's also a 48-volt battery inside the EV. We can now open up that market for 48-volt power supplies to charge and discharge the secondary battery in EV. But in both cases, we know customers love to have 2 sources to go really big with GaN. We've seen that where we've been proprietary and win the high end, but you want to also get the mainstream markets where the volumes are very big. By creating proactively 2 sources with Infineon, who's also a well-respected quality, reliability company, not a bottom feeder on price, so it's a really nice match in bringing 2 really good credible sources to accelerate the adoption of this lower-voltage GaN in 48-volt systems.

Great. And then maybe you could segue to the data center power opportunity and tell us how you're going to participate in that market.

Yes. So data center, as I mentioned, starts with the high-voltage chips. We're finding the combination, as I said earlier, of GaN and silicon carbide gets us the best result for the AC-to-DC power supplies, or PSUs. That's off to a great start. We just started that business in Q3 of last year. Last year, we had 40 design wins. 30 of them came in Q4. So a really nice trajectory. All 40 are ramping throughout this year. It's coming from a 0 base, so the numbers aren't huge, but they're really ramping nicely this year. That's just for the high-voltage AC to DC. Now we started sampling the lower voltage for the 48-volt. That will double our opportunity as we start those customer engagements this year. We expect the mid-voltage GaN to start ramping next year. So that will double our total opportunity, but these things take time as we're working through the development cycles with our customers.

And when you talk about that number of design wins, that's with the power supply companies.

Just the high voltage, and with the power supply companies. We don't sell directly to Amazon, AWS or Facebook, but we sell to the power supply companies typically in China and Taiwan, and then they resell to the rack companies who then sell to the data center integrators.

And what's the line of sight to like timing of usage for that stuff for production?

Yes. Q3, Q4, we started shipping. So we even mentioned low single-digit millions right out of the gates in the second half of last year, and that will keep ramping throughout this year.

Yes. Okay. Great. Maybe shifting to silicon carbide. If you could talk about your technology advantage there with trench-assisted planar gate MOSFET. What do you bring to the silicon carbide market? Because it seems like a very competitive space these days.

Yes. As compared to the GaN where we talked about integrated circuits, here on silicon carbide, they're all discrete. Doing the integration because it's vertical is extremely difficult. So we don't have that same level of integration, but most silicon carbide suppliers are either planar suppliers or trench suppliers. They have their pluses and minuses. Traditionally, you would call trench the best performance, but more expensive; planar, a little bit more simple, more reliable, easier to manufacture and lower cost. GeneSiC, one of the things we loved about the technology they created before we acquired them, is this trench-assisted planar. To not get too technical, it's got the performance of trench with the simplicity, cost, robustness and manufacturability of planar. So it's sort of the best of both worlds. And that's really important, obviously, as a smaller player, an earlier player, to bring significant technical advantage. What does that translate to? It translates into better in-circuit efficiencies and cooler running temperatures coming from that trench-assisted planar, but we do it with really great robustness. In fact, the avalanche energy rating, which is a robustness indicator, is the strongest you can have in the industry. And we 100% test for it in production, which is also unique, combined with an attractive price point given the planar manufacturing. So that's a really nice combination in the silicon carbide world. The other big advantage of our silicon carbide technology is its voltage range. It's very hard to take silicon carbide above 1,200 or 1,700 volts. Our GeneSiC technology goes to 3.3 kV, even 6.5 kV. And that becomes really important as you think about upgrading the grid or major energy storage installations, big equipment, that operates at very high voltages.

It's a lot of volts.

It's a lot of volts.

And how do you think about the scalability of the Navitas silicon carbide solution? And kind of when do we start to see more revenue growth out of this business?

Yes. definitely, this last year was a tough year with the slowdown in inventory corrections. We talked about GaN was up 50%, but the silicon carbide was weaker. And so luckily, together, we still grew when others didn't even grow last year. But we see that trend where both GaN and silicon carbide are going to be growing nicely throughout this year. And part of the markets have recovered, and we got through some of that inventory correction. But more importantly, a big part of the $450 million of design wins last year was with silicon carbide in these markets that have been a little slower, solar, industrial and EV, coupled with AI data centers kicking in with both of them. So I think we'll see nice growth in both GaN and silicon carbide throughout this year.

Okay. And then you talked about a lot of the growth coming from the intersection of the 2 technologies. And one example, obviously, is the data center business. But that does seem like a unique capability that you have. Does the fact that Infineon has acquired GaN technology and has the capability, do you see over time other people trying to replicate what you're doing there?

I think they are the closest in that they have -- they're actually one of the few that have GaN and silicon carbide, but also they have a lot of the older traditional silicon as well. But I'd say, again, we are a lot more focused. They're trying to win that battle on so many different fronts. We're very laser-focused on mobile; EV onboard charger, as I said, where we leverage the GaN and the silicon carbide; but also the AI data centers, as you brought up, again, leveraging the GaN, the silicon carbide. But also our unique system design capability, I'd say that's another key differentiator that goes far beyond what Infineon or other semiconductor companies may do in terms of application support.

And does the customer -- I mean, are you able to deliver combination products that the customer can't figure out on their own from buying GaN and SiC from different vendors?

I think they can do it, and the market is maturing as people are learning how to get the most out of these 2 technologies, but we certainly accelerate that. And it's almost like a free service we're giving to our top customers to show that path with these reference designs that are sort of 80% production-ready, so they can speed up the learning, speed up the time to market. That builds a lot of trust and a lot of opportunity to really closely partner with our customers.

Yes. Okay. Great. And then just there was one thing that we got questions about. You disengaged with the silicon carbide distributor in the December quarter. It seemed a little disruptive numbers-wise. Can you just talk about that?

Yes. Our silicon carbide distributor wasn't performing as well as we'd like on the design win side. And that distributor, in particular, had a lot of concentration with silicon carbide across not just us but other customers as well. So at the end of the day, we decided to disengage with them. But we've secured a new channel partner there, and that's driving a lot of these design wins and also going to help us with that second half ramp.

Okay. Great. So maybe we could talk about financials. As your revenues come under pressure, you've been disciplined, taking OpEx down. Can you talk about the choices that you're making there? Is there any capability that you lose as you spend less? And just how do you think about the spending levels relative to revenue opportunity?

No, that's great. I mean when we looked at OpEx, it sort of was 3 sort of areas. We initially wanted to refocus the company on 3 main sectors: EV, mobile and data center. And so once we reorganized the company, restructured into those 3 sectors, we can then start looking into those sectors and say, where were acquisition synergies. We acquired a couple of companies in the past, and now is a good time to sort of drive those acquisition synergies to lower our OpEx. And then the third element there was customer acquisition costs. Initially, when we were teaching everyone about gallium nitride and silicon carbide, the acquisition costs were higher. But now that companies have started engaging in them, understand the value prop there, the acquisition costs have gone down, which has allowed us to sort of rightsize the organization to this $15.5 million OpEx on a quarterly basis going forward. And we think we can sort of maintain that amount and drive a lot of leverage in the business.

So there is a prioritization in the areas where you're spending a little bit less as part of that.

Yes, absolutely. But I mean, I think it's equally across the organization. If you look at like our R&D-SG&A split, it's going to be the same split on a go-forward basis. We've just sort of dropped that baseline down.

And I could add, Joe, when we bought GeneSiC, they were broad line selling into a lot of places, solar, industrial, energy storage, data center hadn't started. So you put these together, and we were in 7 different sectors, in almost all the major electronic sectors because power is everywhere. But of our size, back to focus, right? We had to focus. So part of what Todd described is really double-downing on where our value prop, our strength, our strategic advantages are best: mobile, not just EV in general, like traction. That's where most guys are focused on silicon carbide. We focused on that onboard charger, not the traction. And then, of course, AI data centers is just ramping up. So that allowed us to really sort of double-down and focus our dollars in an area that I think in the end won't lose anything on the top line. I think more focus will actually deliver greater sustained leadership in those 3 sectors, even if it means at the expense of a few others.

I actually wanted to circle back on some of the automotive questions around this. Given for 2 years, we obviously faced one of the more severe shortages that the automotive industry has seen, and there's become sort of a risk aversion getting built into at least the Western automotive OEMs where they're very focused on longevity of supply. You competed with 2 smaller companies that both got bought by big companies. Does that put you in any kind of disadvantage? Is it hard to get buy-in from the bigger Tier 1s and automotive OEMs?

I would have thought more difficult, not to say that it's a slam-dunker, it's easy. But I think being an independent company, financially clean balance sheet, good cash position, we are in front of and engaged with almost every Tier 1 out there. So I think we're in the game and making incredible progress. But I think we're going further than that. This is back to like the Infineon thing, proactively setting up dual sourcing, looking at dual sourcing inside our company to give our customer confidence, but dual sourcing outside our company to give our customers confidence. So I think all that works together, combined with obviously having the best technology, which is really what's going to motivate them first and foremost.

Yes. Okay. So you don't need to get to a certain scale to have a switching point where you get more revenue opportunities.

Yes. I mean I think we'll look -- I think as you scale up, you go from fabless. We're kind of in this fab-light model. We're making manufacturing investments, how you get the cost structure, the capacity, the multi-sourcing. I wouldn't rule out an IBM model at some point or partnering or being part of a bigger company that allows that scale to continue to go.

And you mentioned your SiC manufacturings in the U.S., which is probably helpful from a trading perspective.

Yes. So we use X-FAB in Texas for silicon carbide. So that's a real positive when it comes to geopolitical concerns, tariff concerns. A lot of our customers are in Europe, U.S. And for GaN, we use TSMC out of Taiwan.

Yes. if you put tariffs on TSMC, that's a whole another matter.

Right. The whole world will feel that one.

Okay. Great. And then from a balance sheet perspective, you've talked about having a strong balance sheet, a couple of years of liquidity there. Can you talk about the strains that you might see in this weaker first half? And just do you need to put any contingencies into place around that?

Right. I think our balance sheet is in the best position it's been in a long time. We have $87 million in cash, no debt. Our inventory level on our balance sheet is at a very low level right now. And so we burned maybe $12 million of cash in Q4, and that is only going to go down on a quarter-to-quarter basis. We expect to sort of drive the single-digit cash usage to eventually drive ourselves to EBITDA positive in 2026. So the balance sheet looks in a great position today.

Okay. Great. And then I'll finish, and I'll open it up to the audience. But on gross margin, you talked about gross margins improving in the back half of the year. Can you remind us what that looks like from a business mix perspective? What are the needle-movers in terms of gross margin?

Right. So I think historically, our gross margins have been hovering around that 40% margin level. Mobile traditionally is below that, where the high-power markets such as EV and data center will be above that. So I think long term, we still believe we're going to get to that 50%-plus, and that's going to be driven by us moving our mix more towards from mobile up to those other markets.

And how much lower is mobile, just to get a sense?

Sure. Well, we don't say specifically, but it's closely approaching our average of 40% as we've sort of driven some new cost reductions and haven't passed that on to the customer. So that's allowed us to drive more of it.

Yes. Great. Okay. Let me see if we have questions from the audience.

So I just want to ask a quick one, actually, if I could. We talked a lot about the OBC in the car as a play for gallium nitride. Do we dismiss the inverter as ever possible for GaN at this point?

No, don't dismiss it, but just practical about it. Power is a pretty conservative industry. You really need a field track record. I think we did that really well in the mobile space, now shipped over 150 million chips with parts per billion type quality and reliability. We're going to repeat that confidence building in the other sectors. So I think you're going to want a couple of years of OBC field track record just being very pragmatic about it. In parallel with that, doing the work. A motor is pretty different than a power supply. How do you optimize it, have the short circuit capability? There's a lot of unique things there that need work, coupled with, I think, that GaN OBC field track record. That will get there, but I think it's a few years before it's really big for traction.

And I thought it was interesting that you have this wraparound play for systems. Would that play into the inverter as well? And can you remind me, what specifically do you have that you play into the systems approach?

Yes. So if I back up for a second, in all these power systems, to really get the best benefit of GaN or silicon carbide, you need to take advantage of not only its high efficiency, but it's fast switching speed. That switching speed is the key to the system design challenges. As soon as you want the entire system to switch, say, 2x faster or 5x faster, almost everything in the system has to get redesigned. The controllers have to operate at that high frequency. The driver has to operate at that frequency. The magnetic components, inductors and transformers, all need to be redesigned, even reconsidering the core materials that are optimal for that higher-frequency operation. The capacitors need to be reconsidered. The filters to deal with the EMI, all of these things. So basically, the entire system has to sort of be reconsidered and rethought. And so designers that have been designing power supplies the same way for 30 years, all of a sudden find themselves with a whole new technical challenge. That's hard for everybody. And so when it came to the chargers, even a small little charger, we had to do all that work. At least we chose to do that work to accelerate our learning and then ultimately accelerate the learning of all the power supply companies. And ultimately, when you do that, we're building trust. We're sharing knowledge. We're creating joint labs. I'm helping you on this, you're helping me on that. Together, we're building the next-generation power supply. That worked in spades in chargers. Now we're repeating it. We had announced a joint lab with Geely, one of the fastest-growing EV guys in the world, to actually not just be our lab to design this, a joint lab that we do together. And we love the joint lab concept because, again, it takes a village to pull that together, figure out those hard problems. And then you've got really sticky solutions when you're done. You're not going to get kicked out for $0.25 or $0.50. And so all of these needed, I don't know where we end up on motor. There's a lot of technical work to be done before we finalize the go-to-market strategy. Is it the same go-to-market or a little different? We'll have to see how that plays out. But I think our system approach and application-specific approach has served us well so far. Yes. Good question. Thank you.

Can you maybe talk a little bit about the visibility you have into your end markets, right? I think you mentioned that Q3, Q4, you expect some return to growth. What makes you confident in that? Can you maybe comment on that a bit further?

Yes, 3 factors. Certainly, number one, we see stability and reduction in sort of the inventory adjustments. So I think kind of the removal of a negative is encouraging in itself. The stability in some of these markets that have been pretty bumpy, working through any growth changes and therefore, inventory correction. So that's good to see. Number two, the strength of our $450 million design wins. Last year was the first time we formally did a full add-up of a full year, to have all of this data to look at our win-loss ratio. And our win ratio compared to loss was over 50% on that $450 million. So I feel pretty good about that win ratio. I feel really good about the $450 million. That is lifetime revenue. So that will be running over a 1- to 4-year period. So the annual revenue is a subset of that $450 million, but a lot of that starts kicking in, in the second half of next year and into '26. And the third thing is we do a 4-quarter rolling customer forecast every month globally around the world. So we get customer forecast, not just the design win data. We're adding that up. And so we're looking at those trends and constantly scrubbing and challenging them. And so all 3 of these factors point to the soft guidance we've given on Q1, Q2, but more importantly, Q3, Q4 growth like you brought up.

I mean when you talk to [ Lee ] about these longer-term shifts in motor control and the traction motor, what's the lead time on those decisions being made? If we're investing now in a new technology for EV-centric cars, when does that get implemented? Is that different in China than it is in the rest of the world?

Well, different regions and customers can move at different paces. But I'd say, fundamentally, you could spend a year or 2 in what you might call advanced research. It's not tied to a specific program. In this case, I would say the real opportunity is how to optimize high-frequency motors at that kind of power level and scale. So it's actually quite a bit of work, not on the motor driver as it is on actually getting high-efficiency, high-speed motors to operate at higher frequency that really leverages gallium nitride, and then what do you wrap around that. And once you optimize that motor and all the circuitry around it, then there's going to be a lot of talk about reliability, robustness. and then it's kicking off a product and going through an entire product development cycle. Even as EVs development cycles have come down, it used to be 5 years to make a car, now the fastest guys are maybe 3 years, 2.5 years. So I think you've got an advanced research phase to get through, is that a year or 2, and then 2.5 years of product development, including some pretty robust reliability studies as part of both of those. So that's why they kind of add up. There's still quite a bit of work for the industry to do, I believe, before we'll see mainstream adoption of GaN for electric motors.

Helpful. Any more questions? If not, we'll wrap it up there. Okay, guys. Thank you very much. Appreciate it.

Thank you.