solutions. We're pleased to have with us today Steve Oliver, Vice President of Corporate Marketing and Investor Relations, and with that, I think, Steve, you have a few slides for us.
Sure. Thank you, Tristan. So hello, and welcome to Navitas. It is a small group, so if anyone wants to jump in with a question, just raise your hand. So a little background, the silicon chip does two things. It processes data, so things like the Grace Hopper chip from NVIDIA, that's where a silicon chip is used to process ones and zeros. But silicon has, in the past, been also used to process power, converting from AC from the wall down to 12 volts for your laptop. And in that function, silicon has now been superseded by two new materials. One of them is gallium nitride, and the other is silicon carbide. They both have different capabilities. GaN is incredibly high speed. When we talk about high speed, that means the switching frequency of the power converter.
High speed means you put energy into a transformer or into a capacitor faster but at a smaller quanta, quantum. So the transformer doesn't have to be as big, as heavy, as expensive. So the faster you can go, the smaller, lighter, cheaper you can make the whole system. Silicon carbide has amazing high voltage capability, so you can use it for railway locomotives, for wind turbines, big solar applications. Navitas has both of these technologies, and we don't have any of the legacy silicon parts, so we can grow much faster than the rest of the market.
So while you may have heard of people like Infineon, that's a German semiconductor company, they're the number one in silicon today, and they have growing business in silicon carbide and GaN, but the silicon business is either flat or decreasing, so it's very difficult for them to grow, and we've committed to grow faster than the market. That market is a $22 billion per year opportunity. As you can see here, silicon carbide goes to really high power, high voltage things. Gallium nitride, more for the things that you plug into the wall.
As we have both of these technologies, it means that we are agnostic as to which one the customer chooses, and in many cases, for example, if we're working on an onboard charger for an 800-volt battery in a car, you would need silicon carbide to address the 800 volts, but because it still needs to plug into the wall, gallium nitride is the best solution. So you can have a no-compromise hybrid using these technologies. And going back to that processing ones and zeros, this is information from the NVIDIA GTC from a couple of months ago. I think it was Sunday night our time, Jensen stood up and talked about the new AI chips. The Rubin is the new family coming after Blackwell. He made some statements about power as well. Power is really critical.
The new chips are very hungry for power, and he said that, "Hey, a new data center using AI chips would be 32,000 of his GPUs." For us, that means if you include, energy efficiency, cooling, the switching, everything that goes around a system, it's about 75 MW for a data center, which is a pretty good sized city. So we're really keen to improve the performance of the efficiency and also the size, and within one of these racks, these big filing cabinets, it used to be that there was 30 kW of demand. With the AI chips, that's now 100 or 120 kW, so 4 times as much demand with no more space allocated for power conversion. So it's our job to squeeze more power in a small box, and that's what we've done on the right-hand side with our roadmap.
I'll jump quickly to the end because Tristan will ask me a few questions coming in. But we are definitely growing faster than the market. Last year, we doubled revenue, getting up to $80 million, and you can see here the Q1 and Q2 estimated figures from our Q1 release. We have cash in the bank. We have $130 million, and we have 0 debt, and we are on a roadmap to EBITDA breakeven when we get to about $55 million revenue per quarter, and we believe we have enough cash in the bank to get to that point with margin. The line at the bottom there, the customer pipeline, this is our CRM system with some severe haircuts. So we have everything in... We use Dynamics 365, but it could be Salesforce. We take everything in there.
It has to be a real project. It has to be the purchasing person who says, "Yes, it will start on this date, it will have this many pieces, and you're in the ballpark on pricing." It can't just be an R&D, kick-the-tires hypothetical. We also take a serious haircut because it might get delayed by a quarter or two. We also take a haircut because some of the big guys, like the Cupertino company, they insist on two sources. So we take these slices, and that $1.6 billion represents the lifetime revenue of our visible and reasonably confident projects that we have going forward, and that plus 30% is since a December date. So we've grown another 30% in about six months. Over to Tristan.
Thank you. I wanted to start with something that's probably close to 50% of your revenue right now, which is silicon carbide. And if you could give us an update on how you see the trends. I know at your Analyst Day in Torrance in December last year, during the silicon carbide presentation, I think it may have been the CTO mentioning that perhaps we could get to a supply-demand balance later this year. So he was earlier than consensus in terms of when he see an equilibrium of that, and is that something that you're seeing? How do you see demand evolving? I mean, we know silicon carbide orders have come down already a couple of times, and the first time was really with Tesla industry-wide in October last year. How do you see that? How do you see the second half, and, and how do you think you're positioned there?
Yeah. So the ups and downs in the market are sector by sector. So, for example, we believe we've already seen the trough of the mobile charger market, and we're on the way back up there. EV, solar, look to be bouncing along the bottom. We're not quite predicting when exactly that will be. But we do see, because of our design-in pipeline, our customer pipeline, that we do expect to see a stronger second half of 2024 and a very strong 2025, and that's based on these design-ins have a certain gestation period. For a mobile phone charger, it's about nine months, but then it blips for a year and then goes back to zero, so you have to keep winning that kind of project.
For a home appliance, it's about a two- to two-and-a-half-year design window, but then it lasts for about five years. We've been doing a lot of the kind of hard work in this gestation period, and we see a lot coming back in 2025. Specifically in solar, we're already in that market with silicon carbide. That industry is chewing through some inventory of finished goods. For the gallium nitride, companies like Enphase have given us testimonials in the past, and they've committed to use gallium nitride instead of the old silicon in their next generation of what's called a microinverter. That's a per-panel converter block. They are working through some silicon inventory. They will be quickly moving on to the GaN product from Navitas. EVs, we have relationships with Geely, with Shinry.
These are some large Chinese companies. Originally with the onboard charger, but also now going into some of the traction drive applications. The roadside charger as well is a key one for us. These are the big kind of filing cabinets with Tesla and a hoop written at the side. One of the big companies is called SK Signet. That's a Korean company. We've got a very healthy market share within that company. We are actually potentially benefiting from the Inflation Reduction Act. The Biden administration has said that we'll have a whole rollout of EV chargers across the nation, with an asterisk, and the asterisk says, "If we're paying for it, 55% has to be built in the U.S." And we make our silicon carbide in Lubbock, Texas. So even...
So that gives us a heads-up, compared to some of the European or Asian manufacturers. The home appliance is a big one as well. Different markets. When we first went through the IPO, we weren't sure if gallium nitride would be good for turning wheels, turning motors. Turns out, yes, it is. We hired some guys from our customers and also from our friends in competition, and we've got a really nice roadmap now. We've got a lot of parts in production with, whether it's haircare, vacuum cleaner, washing machine, dryer, a lot of different applications where we're using gallium nitride's high-speed approach and also integrated functionality because we're not just a switch. We're a real IC. We can do lossless current sensing. We can have 2,000-volt protection diodes. It's a really tough little system.
Great. Thank you. Maybe you could go into more details between the traction and the use cases for silicon carbide versus gallium nitride. You mentioned solar, for example. I thought that was mostly gallium nitride and also maybe in automotive. You know, kind of tell us how you benefit from being a supplier of both, which not every company does that, obviously, and what application specific to each.
Sure, yeah. So in the automotive world... This will go backwards. The automotive world is split by voltage. So with passenger electric vehicles, you have a 400-volt battery. Tesla runs on 400 volts. Other companies have 800 volts. If you have a Porsche Taycan in your driveway, that's an 800-volt battery. And then we're actually looking also at long-haul EV trucking. The U.S. has signed up to the COP 27 agreement that says by 2035, 50% of all long-haul trucks, these are the Kenworths, the Peterbilts, the Mack Trucks, have to be fossil fuel-free at point of use. And by 2040, no more diesel trucks. So it's a great inflection point there. Also the same for eVTOLs, that's electric, vertical take-off and landing. This is the Joby, the Lilium, the Archer Aviation.
For those applications, that's a 1200-volt rail. So the market is split by voltage. 1200 volts, silicon carbide, whether it's charging or spinning wheels or propellers. 800 volts, still again, silicon carbide, because that's a sweet spot of voltage. 400 volts, that's a mixed area. Again, if you're plugging into a wall, gallium nitride is the best solution. If you've got a 400-volt vehicle like the Tesla Model X, which is pretty big and pretty heavy, the traction is very much silicon carbide. It's a current density thing. But if it's a 2-passenger or even 4-passenger city car that goes into, you know, India, Indonesia, China, that's a great place for gallium nitride in the traction drive. So it's a bit of a split, but mostly by voltage.
Okay. You've dedicated, uh, capacity supportive of, I think, longer term, $200 million in silicon carbide revenue. How is that capacity ramping? Are you scaling down a bit, given, the incremental weakness in the market relative to plans announced last year? And generally, how should we look at the growth for silicon carbide for your company over the next several quarters?
Right. Yep. So today, Navitas is a fabless company. We use TSMC for our gallium nitride chips, and we use a company called X-FAB, which is Lubbock, Texas, for our silicon carbide. We did announce a plan to do what's called Epi processing. That's where you grow a very thin layer of silicon carbide on top of a base silicon carbide wafer in our Torrance, California, facility. That would give us more control over that bottleneck in the process. It's also still a little bit of an art than a science, and we've got some great guys that we recruited for that. But because of the slowdown in the market, there's now extra capacity at X-FAB, and so we've been able to push off the project in-house, and all of the payments actually went with that push out. We've been able to have a very good control of cash, and still grow our business.
Okay. In terms of the demand, I mean, obviously in silicon carbide, you're facing fairly large companies like ST and onsemi. We've heard at industry conferences that you tend to rank extremely well in terms of silicon carbide performance. So does that really matter? How far is it from being a commodity market? What makes you win? Is that the performance? Is that just the fact that people don't have enough capacity, so they are going to order from whoever offers them silicon carbide? Or is it also because you provide gallium nitride, and you're able to do cross-selling opportunities? So in other words, you know, what is kind of the edge, you know, that you offer that leads you to get design in from customers?
Well, that's a good series of questions, and I think the answer is it's a mix. When we started our joint development center in Shanghai with Geely. Geely is a company that owns Polestar, Volvo, a few other companies as well. We started with gallium nitride, and that's on track to be in mass production in 2025, next year. Then, when we took on the silicon carbide technology, Geely had an issue with another vendor's product and said, "Hey, you've got that silicon carbide now, can you help?" And we said, "Yes, we can." So that was a great example of we're already an approved vendor as a company, bring in this new technology and to refer to performance. If you do the same voltage and RDS(on), which is a measure of conduction capability, our chip is smaller than anybody else's.
It's a no-compromise hybrid between what's called a planar approach and a trench approach, and it means we can have very good, die per wafer in manufacturing. And for the customer, when you push power through a device, it gets hot. It's not perfect, it gets hot, but ours stays remarkably cool compared to ons emi, ST, Infineon, and that reduction in temperature equates to reliability. And so we've done a head-to-head with ons emi, for example, and, 25 degrees Celsius reduction in temperature equates to about three times longer lifetime. So there's a temperature benefit, there's a lifetime benefit, and we also have extra protection features around it that other people don't do.
It's how's the IP around this? Is that the way you design the module? 'Cause you're getting your silicon carbide from the same sources as everybody else. Is some of the technology that makes you lead in gallium nitride apply to silicon carbide as well, or is it completely different? If you could explain, you know, what makes you run your chip 25 degrees lower than the competition.
I'll step back a little bit and say that the model is the same for Navitas GaN as it is for Navitas SiC, silicon carbide, in that the IP is in the design of the chip. We then will give a mask set, which is the series of stencils that the factory uses. We'll give a stencil set, a mask set to TSMC, and they will deliver a complete finished wafer. We'll also give a mask set to X-FAB, and they will deliver a finished wafer. So the processing is not anything remarkable, but we're able to design in features that improve the performance and the reliability. You had a second part of that question, I've forgotten it.
Yeah, just the IP that you have y ou know, that allows you to have a performance and temperature.
Right. Yeah, so for silicon carbide, yes.
Yeah.
Thank you. Yeah, so silicon carbide, part of it is for robustness, which is, there's a function called avalanche capability, which is how much energy the device can stand under duress. It is a fault condition, so it shouldn't happen, but if it does, our part survives, other people's don't. And then you have the temperature reduction under load. That is part of our IP, how we've been able to create a functionality within the device to stay cool.
Okay. You've talked about some inventories, notably in solar... But at the same time, you know, mobile phone rebounding. Do you see excess inventories elsewhere outside of solar panel, or is everything pretty much... Or, and also what do you see in automotive?
Yeah. In automotive for silicon carbide, things have been slower. So, for example, there was a unit with GE that we said could be up to, I think, $20-$25 million per year. We've reset that to maybe $10-$15 million. We didn't lose the business, but not that many people were buying cars, so it's gone through a little dip. In the mobile market, we actually don't see a lot of inventory in the market. The technology that we have allows an incredibly fast charging system in a very small size, and we're in 10 out of the top 10 OEMs. If it's not listed here, it's because we're asked to be conservative and not discuss them in detail. There are some companies who are more aggressive with their technology. That's the battery technology, actually.
So people like OPPO and Xiaomi have a new type of lithium-ion battery that can handle a lot of power in a short space of time, and it stays cool when you charge it, and it doesn't wear out. Some other companies are more conservative. Now, what does that mean? That means that some people are still on 20-watt, 30-watt charger. If you have an iPhone 15, it maxes at 28 watts. If you have a Xiaomi 14 or 15, that's a 140-watt charger. And there's a thing called the GT3, which is a 240-watt charger for a phone, which is nuts. You can literally see the little needle move as you plug it in. It charges from 0 to, I think, 100 in 9 minutes, so it's incredible.
Basically, what it means is that the market is moving into higher power chargers, going away from the classic sugar cube into very small, very portable, very light technologies. Now 100- watts is standard, 140 watts is standard, moving higher in technology, and we dominate that market area. So we are the number one player in mobile fast chargers. There is another thing to mention about adoption as well. If you think about any technology, if you think about retractable seat belts in a car, that was first introduced into the very high-end product, I think it was a Mercedes C-Class, and everyone said, "I don't need that. That's too expensive." And now every car has retractable seat belts.
So it started off where people used GaN in the high-end phones, and it was maybe 5% of, for example, OPPO or Xiaomi used GaN. Everything else was still the old silicon. This year, they'll go up to about 30% of all of their chargers shipped will be GaN, and we have the majority share in that. So the adoption of this technology is coming through the phone charger market. It's going into solar market with GaN, it's going into EV market, it's going into home appliance.
Could you talk about the opportunities in data center? You touched on this during your introduction with NVIDIA. I know you've mentioned a few million initial ramp in data center, second half of this year and then ramping. You've talked in the past about Europe at some point requiring the most efficient power supply for new data center deployments. Where are we there? Is that something that we see in the U.S. as well, and, and generally, how, you know, could you describe the opportunity and the ramp that you see in data center on the basis of your design wins?
Great. Yeah, so, efficiency is key. For the European Union, they've said that you have to be 96% efficient or better, which is a pretty tough thing to do. Otherwise, you just can't ship into Europe. They won't, won't allow it. That is primarily from an environmental aspect, because otherwise you have to, you know, have more power stations to create the electricity. But there's also a huge financial benefit. You know, it doesn't matter how many cents per kilowatt hour, if you can reduce your energy bill, that saves a lot of money. And, a typical data center, not even an AI one, spends as much on electricity per year as all of the hardware that goes into that facility. So efficiency is key.
Now, you can do it, well, you could, in the past, do it with the old silicon technology, but you literally had to throw money at the problem. You had to put more and more silicon chips in to get the efficiency. With gallium nitride, you can actually have a 3x smaller box. It's called a silver box because it's silver, and it's a box, and it converts from 220-volt AC down to 48 volts, and that goes on to the NVIDIA motherboard. And we can do that at a lower dollars- per- watt than the old silicon technology. So a higher efficiency, smaller size, and cheaper. That's a big, big uptick. You can see here where we sell to what's kind of a middle guy, which is mostly Taiwanese companies in this case, like Delta, Chicony, Lite-On, Flex, and they will then sell into Dell, Azure, Amazon Web Services, these kind of customers. So it's a really big uptick for us.
Could you expand a little bit on the content opportunity? So if we take the example of Blackwell, on your slide here, you have 2000 GPUs. How many power supply? What's the ASPs for which? How should we look at the $ content per, either per rack or per installation?
So it works out at about $5 per kW, so $5 per 1,000 W. And that actually roughly translates whether it's a 100-W cell phone charger or it's a 1-MW data center. So it's about $5 per kW.
Okay. For a general purpose data center on average versus accelerated data center versus AI, if you can remind us the numbers of what the potential total content is for you?
I haven't done that calculation yet. I'll have to get back to you.
Okay. Okay, but clearly, all incremental. Do you see an expansion beyond the three major wins at tier one power supply customers, and are we able to quantify that, or we don't know yet how many of those are going to ship and.
Yeah. So we've, we've got 30 customer projects right now. For us, projects could be the 3,000-watt block, 5,000-watt, 10,000-watt block. So it's definitely multiple customers. For us, people like, Dell, a very attractive customer because, our initial market into the mobile fast charger, one of the first companies was Dell for the laptop power supplies going through Chicony, Lite-On, and Delta in Taiwan. When we turned to the Dell server division, it's also specified in Austin, Texas, then transfers to Taipei for Dell, and it's made by the same people. We're already qualified with the same base GaN material. And now with our GaNSafe range, which is primarily focused on people, designers being able to sleep, being able to design in this GaN technology with, with confidence. It's robust, it's efficient, and it really makes sense on a dollar-per-watt basis.
Okay. Anything we haven't talked about that you would like to highlight?
Sure. I think I mentioned it at the start, but investors are starting to see the daylight between Navitas and other power semiconductor companies. In terms of, because we're not anchored, we're not dragged back by the old legacy silicon, we don't have internal arguments, you know, one person's profit center against another person's. So we can purely target the customer, find out what they need, whether it's silicon carbide, whether it's gallium nitride, it's always the new tech, so we're not dragged back. So our growth trajectory is very strong. We have said that we'll grow 5x more than the market will grow. So while other companies previously said, "Oh, we'll do, you know, +5% for this year," and now they've revised it to -5% or -10% for the year, we're still forecasting growth.
Yeah.
In the market.
Great. Thank you very much for presenting with us.