Hello. Good morning, everyone, and welcome back to the Industrial Technology Conference. It's my pleasure to introduce AIXTRON. It's my pleasure to welcome Christian Ludwig, Vice President, Investor Relations and Corporate Communications at AIXTRON. He will guide us through the presentation and share deeper insight into the company. Before we begin, a quick housekeeping note, the conference is being recorded. All participants are in a listen-only mode. After the presentation, we will move on with the Q&A session. Please submit your questions via the chat box at any time. My team at mwb research follows AIXTRON closely. I will share a link to our most recent research update as well. Christian, we are looking forward to your presentation. Please go ahead. The floor is all yours.
Thank you very much. A late good morning to everybody on the call. My name is Christian Ludwig. As said, I'm Vice President of Investor Relations Corporate Communications. I will walk you through our investment case in the next 20 minutes. Then I'm happy to answer all your questions. Before we start, just for housekeeping, a quick look at the disclaimer. Then let's jump right into it. AIXTRON. Who are we, actually? We are really a German startup founded as a spin-off of the RWTH in Aachen a little bit more than 40 years ago. Today, we're a company with a little more than 1,100 employees, active in nine countries around the world, really serving global, our customer base. We are the technology leader in deposition systems. You may ask yourself, what is a deposition system?
A deposition system, you can see one here, is a tool that applies a very thin layer of material on top of a wafer. A wafer is the basis for the semiconductor industry. Normally, it is made from silicon, but it can also be sapphire, it can be silicon carbide, it can be indium phosphide, it can be gallium nitride. A lot of different materials are possible. Our tools are capable to deposit the material groups that we supply on all different kind of surfaces. Important to know is that we're the deposition leader in what we call compound semiconductors. We're not playing in the run-of-the-mill silicon-based industry, but only in the niche of the compound semiconductor industry. Compound, meaning you combine two different materials to a new material. Then deposit that in a very thin layer on top of a wafer.
As you can also see here in this slide, the blue line kind of describes the virtual, very simplified fabrication process of our customers. You can also see that the deposition step is at the very first step of the manufacturing process in the fab. We start with the chip manufacturing, or our tool is the first step to be used by our customers. Then a lot of other steps follow. You can also see, if we do not do our job well at the beginning, we put in a lot of defects in the layer. Of course, you only realize that at the end, which makes the whole chip process making very expensive if we do not our job well. Our USP is to be as defectless as possible to begin the whole wafer-making process. Now, where do our tools come into play?
We basically serve four different end markets, the end markets are kind of defined by the material system that we supply. On the top side, you can see silicon carbide and gallium nitride. These are both what we call power electronics materials, so they go into power electronic chips, which can be used for various applications. I'll come to that later. On the bottom half, you can see our, what I would call it, the legacy business, the optoelectronics and the LED. This is what initially started the company. As you can see on the MicroLED/LED side, two different material groups are being used, gallium nitride and arsenide phosphide. You need different materials to supply all three colors for the LEDs. Gallium nitride is required for green and blue. Arsenide phosphide is required for red. From this core, I would say, we expanded over time.
We discovered that gallium nitride for the green and blue LED can also be used to apply gallium nitride for power electronics. From the red LED, we also expanded into the laser-making business, which today is one of our most exciting end applications. I'll also come to that in a minute. We serve these four end markets with three different tools. A dedicated tool for silicon carbide. You can see that on the left-hand side. This tool, our G10-SiC, sells for around EUR 4.5 million, just to give you an idea. The tool for the gallium nitride application sells for around EUR 4 million. You can see that in the middle. Then the tool which we predominantly use for the laser applications, but it can also, of course, be used, especially for MicroLED, that also sells for around EUR 4 million.
These three tools all have a very similar gross margin, a product gross margin of close to 50%. Especially in gallium nitride as well as in the laser business, we are still selling older tool generations like the G5 and the G4, which have a lower gross margin. Overall, the gross margin levels of these two segments are lower than for silicon carbide. If you have a look at the competitive situation today, we have split it basically according to the core process. Straightforward is the silicon carbide business. We actually do a CVD process here. This is based on 2024 numbers. We had a market share of 27% in the global market, with the number one being ASM, a competitor based in the Netherlands. According to the latest numbers that we have just received, in 2025, we surpassed ASM.
We now have a market share of around about 35%, with ASM still around 30%, and we've got some other players there as well. The big opportunity for us in this silicon carbide business is that as the industry moves to eight-inch, it is really only us and ASM who have competitive eight-inch tools. TEL, NuFlare were very strong in the six-inch business, but are losing out on eight-inch. We hope that with the next wave of investments for silicon carbide kicking in at some point in the future, we have a chance to increase our market share, maybe even to 50%. There will always be room for others, again, ASM will not be going away. That is the kind of outlook we can give here on silicon carbide.
On the MOCVD side, which is the process we use for both GaN, LED, and the optoelectronics, our market share is considerably higher. If we would do it at one level deeper, just looking at GaN market share, we believe is around 90%, and the same thing is applicable for the optoelectronic business or the laser business. We also are at 90%. Only in LED, we basically share the market with Veeco and AMEC. As for the LED side, we only serve the red LED today. Green and blue, we lost to Veeco and AMEC 10 years ago. They are now dominating the market there. Overall, I can say that we have a very strong market position in basically all our end markets, being market leader across the board with a very considerable market share. What does that mean for our revenue, you may wonder.
Here you can see our development over the past six years and the outlook for 2026. As you can see, we had a very strong run up to 2024, mainly driven by the, at the time, the power electronics business, both silicon carbide and GaN, our two newest end market additions, grew very strongly. As I already said, we saw a downturn in the power electronics market, and especially on the silicon carbide side. The industry had over-invested, expecting a much higher growth from mainly electric vehicles, because that was the main end market for silicon carbide applications. We have been dealing with a declining market since 2025. Also in 2026, we saw that continuing. 2025, we got a lucky break on the MicroLED side. We had some, I would say, one-off purchases for R&D systems, which did not reappear in 2025.
That's where we saw then the downturn in 2025 being a little more pronounced. As you can see, in 2026, we were becoming a little bit more optimistic. We're now looking at EUR 560 million ±EUR 30 million. The main driver here is the optoelectronics business. As you can see, the purple color already increased massively from 2024- 2025, a growth of 60% from roughly EUR 60 million -EUR 100 million. For this year, we're looking at maybe even EUR 250 million, just to give you an idea where the growth rate for the optoelectronics business could take us. The reason for this strong growth is, I'll jump to this slide, is what you can see basically here. The data center build-out for AI is really driving demand for optoelectronic connections. Here you can see the hyperscaler CapEx, which is bound to continue.
This is, of course, one main driver for us. More and more data centers built, you need more optical connections. The industry is also moving from 400G- 800G- 1.6T, which means you always double the amount of lasers you need for this kind of service. That is one big driver, but it's not the only one, because if you look at this slide, the build-out is basically what you can see as the scale-out. That is happening as we speak. It is just a growth of the data centers where there's kind of a one-to-one relation between our business and the build-out of these data centers as more and more are built. Now we're seeing a strong inflection point with the so-called scale-up, which you see at the top end of this graph.
Scale-up means that on the very short connection side, so it is one to five meters, what today is still completely copper, starting next year, we will see replacement of copper by optical fiber connections. That, of course, opens up a whole new market. The market experts are giving very different numbers here, but I would say it is at least five times as big as the scale-out for us. This is what you see in our order intake for the opto business this year.
We already showed you in Q1 that we received almost EUR 120 million orders for the optoelectronics business. I told you already, we believe that the optoelectronics business for us will grow from EUR 100 million to potentially as high as EUR 250 million this year. Although we have no clear guidance for next year yet, we would expect that this growth will continue next year as well.
Further down the road, we even have another opportunity, which could become very interesting for us, and that is the so-called scale-across. This is the very long-haul connection of up to 1,000 kilometers, which today is also mainly copper-based. There is another potential to go to optoelectronic fibers, and this will expand the market growth for us for at least another year or two. Now, moving on from the very exciting optoelectronic markets to the power semi market. We touched it before. We serve this with two materials, silicon carbide and gallium nitride. As you can see, the overall power electronic market is a niche business. On the upper right side, you can see it is only 3% of the overall semiconductor device market. If you look into it, at the moment, the still dominating material is silicon.
Our two materials, silicon carbide and GaN, are only responsible for roughly 16%. If you look at the left-hand side where we try to elaborate a little more on the specific applications, you can see that GaN actually is the smaller part, with only EUR 500 million, and silicon carbide already significantly bigger. Silicon carbide is really driven by the EVs. Tesla introduced silicon carbide for their inverters six, seven years ago, and that is 70%-80% of the end market demand for silicon carbide applications. You can see here in the circle that there are, in principle, a lot more applications possible for silicon carbide. Unfortunately for us, the number of chips you need for these applications is relatively small. As you can imagine, how many locomotives are being built every year, how many electric construction machinery is really being built every year?
Those numbers are low. They may be, from a revenue number for our customers, very high because these chips are very expensive. At the end, what we need is volume on our tools, and that is only really possible, we believe, for the next years on the EV side. The good news here is that we are seeing a lot more of what we call 800-volt battery architecture systems for EVs coming to the market. We hope that this will drive a continuous demand for silicon carbide and help the situation on our customer side, because the customers today are running at capacity utilization levels between 40%-50%.
It is starting slowly to increase, which is good news for us, until it reaches the 80%-85% threshold that normally is required before customers start ordering tools for capacity extension, we expect it will be mid to end of next year. Before that, we would not expect a strong recovery of the silicon carbide business. On the gallium nitride side, this is a little bit different. Although here it looks like less applications, in the truth, we have more applications for gallium nitride that make us hope. I'll jump a little bit here. You can see here that a lot of applications have already been put to the market. It's just taking a while until these applications start being used in volume by the end customer. Here, the capacity utilization situation is significantly better than for silicon carbide. We're looking at 70%, roughly, across the board.
Here, from a slightly down year this year, we expect moderate growth next year already. There is one big application where we had a lot of hopes for, and this is also here driven by the AI opportunity. In this case, it's specifically NVIDIA, which is driving this. With the introduction of the Rubin Ultra chip end of 2027, NVIDIA is changing the whole power delivery architecture. They're moving from a 240-ish/480-volt architecture to what they call an 800-volt HVDC architecture. With this move, they will start replacing the silicon-based power chips, which are today used to a 100%, with, at the high-end silicon carbide, and at the medium and low voltage ends, gallium nitride. We believe this is a significant business opportunity for the market and obviously also for us.
As we have 90% market share, we believe we will participate one-to-one, regardless whom NVIDIA will give the business. We have tried to assess this opportunity, and we believe by 2028- 2029, we're talking of additional 25-30 tools per year for this opportunity. As I stated before, at EUR 4 million per tool, we're talking between EUR 100 million-EUR 130 million of additional revenue just for this AI data center power delivery, kicking in end of next year. Hopefully, with the usual lead times that we have of seven to nine months, we may see first orders end of this year or beginning of next year. That's kind of a view here. For this year, it will have no impact on our revenue or numbers.
Hopefully, as I said, by end of the year on the order intake, with first impact on revenues next year, really ramping into 2028- 2029. One final word on the gallium nitride side, because here we're going to see a change in technology for gallium nitride due to the expected growing demand and the replacement of silicon by gallium nitride. The market has been asking for a 300 mm tool, just to be open here. This race is still open. I told you before, we have a 90%-plus market share in gallium nitride. This is true for 150 and 200 mm. For 300 mm, we have a competitor in Veeco, and they're also trying to get in that market. A technology change is always a window of opportunity for competition to get in.
We'll be shipping our first volume tools end of this year, but we expect Veeco will do the same. In the course of next year, we'll probably see who is going to win out. We have received very positive feedback from our customers, but probably Veeco will claim the same, so we cannot promise you anything here. What I can say is, from a risk perspective, the market for 300-mm tools will be relatively small. At best, maybe 20% of the total market. The rest will remain with 150-mm and 200-mm GaN tools, so where we'll keep our dominating market share. I would say overall, the risk that we'll lose significantly is fairly low, and we actually believe that we will be able to keep our high market share also on the 300-mm side. Here you see again the silicon carbide tool.
I already mentioned that we expect here a recovery of the EV market, hopefully kicking in sometime end of next year, when we see even more of the 800-volt models gaining traction. Again, it will not help us this year on the revenue side. We expect silicon carbide tool sales to probably be below EUR 50 million, and I would not expect any change in the revenue assumptions for next year. By 2028, we could see this starting to kick in. Hopefully, it will help us out two years from now. With that, I'm coming to almost my final slide. Just giving you a quick update on the short term. For the second quarter, we've given a guidance EUR 110 million ± EUR 10 million as a range.
That means for the full year, with the upgraded guidance, we're looking at EUR 560 million ± EUR 30 million. If you do the math, with the weak Q1, that means we roughly need EUR 200 million per quarter in Q3, Q4. That is unfortunately a little bit for us because we only got these orders so late from the optoelectronic industries, that we are underutilized both in Q1 and Q2, and they need to try to get it all in in the second half of the year. We're very comfortable that we will achieve that, and also the number you can see in the graph, we still need EUR 185 million of revenue, of order intake, to make ends meet for this guidance. We're quite confident that we're able to show you that in the second quarter.
All the additional orders we will get in the rest of the year will already be revenue for next year, due to our seven to nine months lead time. You can also see our margin assumptions. They are still a little bit subdued versus the previous margin we showed, but at EUR 560 million of revenue at midpoint, we have a certain operating leverage. Obviously, that still burdens our margin potential, but that should improve as top line continues to grow in the next years as we expect it to do. With that, I'm done with my presentation. I would now open the floor for Q&A.
Christian, many thanks for the insights you have provided and perfectly in time. Just a brief reminder to all the participants, please submit your questions via the chat box. I will start with the first question we have received. In former times, a system for order recognition was in place. 50% order received with down payment, 50% license and shipping details received. How is that currently being handled?
Yeah, that's a good question. We have not changed the recognition of the order intake. It is still the same, the case for orders where we need an export license. We only value these orders when we have the export license in place. We're still waiting for that until we show it in the order intake and also in the order book.
Okay. Thank you.
Basically, the rest, we record the order when we receive the purchase order.
Perfect. The next question is a quite technical one. Do you see MicroLED-based optical interconnects as a realistic future MOCVD demand driver, or still too speculative versus InP-based solutions?
As you may recall, because of the time, I did not touch the LED, MicroLED business because at the moment, there's not a lot of exciting things to talk about here. I know that some companies are working on MicroLED as a potential very near-term communication methodology, and indeed it could then take away business from indium phosphide-based laser business. Today, it's all copper, so either way, it's going to be additional business for us. The fair answer today is we don't know if it will see a breakthrough and will be possible or not. It is in the R&D stage. From our perspective, 2029- 2030, that's the first realistic timeframe when this technology may become industrially attractive and usable.
We don't know if it's going to win out against the traditional, in a sense, indium phosphide laser-based fiber optic.
Thank you for answering the question. The next one. Can you give some color on the potential market size of G10-AsP tools with CPO adoption and ramp-up?
Yeah. We've tried to assess the opportunity for us on the optoelectronic side. It is not that easy. There's a lot of moving parts. Outside the predictions from our customers, we also have two thematics that will go against our potential TAM, and that is the wafer size. Today, still the majority of the market is on four-inch, but we're seeing the first tools being also shipped in the six-inch configurations, and a lot of customers still buy the tool with a four-inch configuration, but already buy the upgrade kit for six-inch at the same time. Over time, the market will move from four-inch to six-inch. That is our expectation. The difference between a tool in four-inch and six-inch configuration is very simply 50% more processable capacity for the customer. Of course, that will eat against our TAM.
The other thing that we have difficulties is the line yield. Today, line yields for the wafer production are very low. We believe between 25% and 60%, 25% being six-inch, 60% a mature four-inch process. Of course, with all the resources now flowing into this sector, we would expect that yields will improve over time. Our best guess at the moment is that the market over the next two to three years, so for starting 2027, 2028, 2029, will be between 60 tools-120 tools for us per year. This broad range already tells you that we have a limited visibility because the error bar is fairly high. At least gives you for your calculation purpose, a floor. I mean, 120 tools at an average price of EUR 4 million gives you EUR 480 million at the top and EUR 240 million at the bottom.
That's kind of the range we're looking at today. It may change next quarter. I have to admit that because the visibility is fairly limited. As I said before, we don't even have all the orders for this year. For next, we have very limited orders yet. It all happens very short-term. Of course, makes the predictions for us very difficult.
Prediction and visibility is a good word for the next question. What are the most material challenges or risks you are seeing right now that could have the biggest impact, let's see, positive or negative, over the next 12-18 months?
Yeah. I think the biggest challenge we have to make sure that we ramp this as smooth as possible. I mean, we have a big facility in Germany where in principle we could do up to EUR 1 billion in revenue, if we have all the parts on time at our facility. I think getting the supply chain ready to deal with this kind of demand, that is one big challenge. Of course, the second challenge is getting our own production ready for this. We have done two rounds of layoffs over the past 15 months, we will need a lot of temps to support our ramp, because we said with the volatility of the business, trying to keep a certain layer of personnel on our own side does not make a lot of sense because we will have too much time when they're underutilized.
We said this next ramp we will do with a lot of temps, but they also have to be on board on time. Then we have one additional thing that we have to deal with, and I didn't mention that so far. We have decided to build a greenfield plant in Malaysia. The building has started, construction has started as we speak, and we want to be production-ready by end of next year. Of course, this is a big project for us that has to be managed carefully, to ramp that up in time to be ready by end of 2027. Of course, with all the demand that we're looking at, it would be quite nice to be sure that we have this additional capacity up and running because we may actually need it.
Managing this while managing the ramp, I think those are the most critical topics that we currently have, that we have to watch very closely.
Many thanks for this answer. Everybody is looking at the hyperscalers and especially at the CapEx spendings. The question is, do you sell to hyperscalers directly or via general contractors, or what is your relation to hyperscalers?
Yes. Thank you. Very good question. The answer is no. On the optical and electronic side, we sell to companies who do the laser. That would be a Lumentum, a Coherent, a Sumitomo Electric, SMART Photonics, AOI. Those are the names that are our customers. I don't even know if they sell directly to the hyperscalers or if there's another middleman in between. I have to admit. On the power electronic side, again, we sell to the Infineons, STs, TSMCs, Innosciences of this world. Classic chip manufacturers. They then sell to module makers. I think only then that part is purchased by the hyperscaler. We're not even a Tier 1, we're a Tier 2 or Tier 3 in this whole data center buildup scenario.
Perfect. Let's talk about China. What are your views on the InP export restrictions from China potentially limiting your customer production?
Well, luckily, AIXTRON, I think that is the company people are referring to, which is a big indium phosphide exporter, is not the only one. We have Sumitomo Electric, which is Japanese. We have Freiberger in Germany. There are a couple of smaller ones. Overall, it is clear that indium phosphide shortage is one of the bottlenecks for the buildup of optoelectronics. That is very clear. We are seeing that everybody in the industry is building up capacities. We would expect that bottleneck to be taken care of by mid-end of next year. Of course, it helps in the sense that it will probably smooth a little bit our ramp as well, because obviously customers are not going to buy tools when they just sit idle in their fabs because they don't have no substrate. It does not worry us too much.
It helps us a little bit. Again, we believe by mid-end of next year, this shortage will be taken care of when all the additional capacity the industry is building up is coming on stream.
Thank you, Christian. Maybe a last quick answer just before, we are running out of time. Would it be too aggressive to assume order intake around EUR 250 million in Q2 2026?
I cannot comment on upcoming numbers for Q2. I am sorry for that.
Sure. Actually, as mentioned before, we had 30 minutes. We are running out of time as the next presentation with LPKF, also related to this broader topic, will start. Thank you all for your questions, and Christian, a big thank you for the presentation and the helpful insights. All participants will receive a short email regarding the feedback, so feel free to share it with us, and thank you once again. The next presentation, we have provided the links in the chat box already, and now that is time to say goodbye and take care.
Goodbye, and thank you very much