Aixtron Stock

Welcome to AIXTRON's analyst conference call for the third quarter of 2026. Please note that today's call is being recorded.

Let me now hand you over to Mr. Christian Ludwig, Vice President, Investor Relations and Corporate Communications at AIXTRON, for opening remarks and introductions.

Thank you very much. A warm welcome also from my side to AIXTRON's Q3 2026 results call. My name is Christian Ludwig. I am the Vice President of Investor Relations, Corporate Communications at AIXTRON. With me in the room today are our CEO, Dr. Felix Grawert; and our CFO, Dr. Christian Danninger, who will guide you through today's presentation and then take your questions.

This call is being recorded by AIXTRON, and is considered copyright material. As such, it cannot be recorded or rebroadcast without permission. Your participation in this call implies your consent to this recording. Please take note of the disclaimer that you find on Page 1 of the presentation document as it applies throughout the conference call. This call is not being broadcast by webcast or any other media. However, we'll make a transcript available on our website after the call.

I would now like to hand you over to our CEO for his opening remarks. Felix, the floor is yours.

Thank you, Christian. Let me also welcome you all to our Q3 '26 results presentation. I will start with an overview of the highlights of the quarter and then hand over to our CFO, Christian, for more details on our financials. Finally, I will give you an update on the development of our business and our guidance.

Let me start by giving you an update on the business development of the third quarter of '26 on Slide 2. The important messages for Q3 are, we have received strong new orders of EUR 275 million, more than double the prior year level and 40% above Q2, predominantly driven by optoelectronics and increasingly also power electronics. Revenues came out at EUR 184 million within our guidance of EUR 180 million plus/minus EUR 20 million for the quarter, reflecting the ongoing ramp-up of the photonics business.

We also continue to generate strong cash flow. Operating cash flow reached EUR 138 million in Q3, and free cash flow came in at EUR 128 million. This was primarily driven by higher customer advanced payments, reflecting the strong order momentum and supporting the upcoming production ramp. We are fully on track to achieve the full year guidance with revenues of EUR 560 million, plus/minus EUR 30 million, as we target a very strong Q4 with a further increase in output beyond the level of Q3. Finally, construction of our new site in Malaysia continues and is fully on track.

Christian will now provide a detailed look at our financials on the following pages before I take over with an update on our market. Christian?

Thanks, Felix, and hello, everyone. Let me start with the highlights of our revenue development on Slide 4. Q3 revenues marked the next step in our production ramp, increasing to EUR 184 million from EUR 59 million in Q1 and EUR 115 million in Q2. This was within our quarterly guidance of EUR 180 million plus/minus EUR 20 million and 54% above the prior year quarter.

For the first 9 months, revenues amounted to EUR 358 million, 3% below the prior year period. Optoelectronics accounted for 61% of equipment revenues, GaN and SiC power electronics for 19%, and LED and microLED for 18%. After sales contributed EUR 88 million, up 10% year-on-year, and represented 25% of group revenues compared with 22% a year ago.

Now let's take a look at the financial KPIs of the income statement on Slide 5. Gross profit in Q3 increased by 88% year-on-year to EUR 87 million. The increase reflected both the higher revenue volume and a favorable product and regional mix. The gross margin improved to 48% from 39% in the prior year quarter, primarily due to the favorable product and regional mix. We do not consider the Q3 margin level to be representative of Q4 and have, therefore, maintained our full year gross margin outlook.

For the first 9 months, gross profit amounted to EUR 145 million, and the gross margin was 40%, 3 percentage points above the prior year period. The 9-month result also includes a mid-single-digit euro million one-off expense related to the personnel reduction implemented in operations in Q1. Operating expenses in Q3 were broadly stable year-on-year at EUR 31 million. For the first 9 months, operating expenses amounted to EUR 96 million, up 2% year-on-year, mainly reflecting a 17% increase in R&D costs.

EBIT in Q3 increased to EUR 57 million from EUR 15 million in the prior year quarter, benefiting from the higher gross margin and operating leverage from the higher revenue volume. The EBIT margin improved to 31% from 13% accordingly. For the first 9 months, EBIT amounted to EUR 49 million, corresponding to an EBIT margin of 14%, 3 percentage points above the prior year period.

Let me now turn to the key balance sheet and cash flow indicators on Slide 6. Working capital decreased by EUR 254 million compared with year-end 2025, primarily due to higher customer advance payments and the collection of receivables related to last year's fourth quarter revenues. Trade receivables declined to EUR 110 million at the end of September from EUR 131 million at year-end.

Customer advance payments increased substantially, reflecting the high order intake, the composition of current orders and individually agreed payment terms. At the end of September, advanced payments stood at EUR 309 million compared with EUR 44 million at the end -- at year-end 2025. Inventories increased to EUR 323 million from EUR 284 million at year-end, mainly reflecting higher work in progress for shipments scheduled in the coming quarters. Trade payables rose to EUR 41 million from EUR 34 million as purchasing activity increased to support the ramp. Both developments are consistent with the planned increase in output.

Operating cash flow reached EUR 311 million in the first 9 months compared with EUR 128 million in the prior year period. In Q3 alone, operating cash flow amounted to EUR 138 million. The main driver was the increase in customer advance payments. Free cash flow amounted to EUR 219 million in the first 9 months compared with EUR 110 million a year ago. In Q3, free cash flow was EUR 128 million.

Capital expenditure amounted to approximately EUR 22 million in the first 9 months. For the full year, we now expect capital expenditures of around EUR 40 million. This includes baseline investments and approximately 1/3 of the announced EUR 40 million investment in the Malaysia expansion with the remainder shifting to next year due to timing.

Liquidity, comprising cash, cash equivalents and other current financial assets, increased to EUR 945 million as of September 30. After deducting the liability component of the convertible bond, net financial assets amounted to EUR 591 million compared with EUR 225 million at the end -- at the year-end 2025. The equity ratio was 58%, primarily reflecting the increase in total assets following the bond issue.

With that, let me hand you back over to Felix.

Thank you, Christian. I would like to continue with an update on the key trends across our end markets. Overall, optoelectronics remained exceptionally strong in the third quarter. In power electronics, customer utilization continues to improve. Order activity is gradually returning and customers are preparing new capacity. These developments provide increasing evidence of the beginning recovery in GaN and SiC.

Let me start with optoelectronics, which was once again the most dynamic and the most important market for our business during the quarter before we briefly discuss GaN, SiC and microLED. Optoelectronics remained our largest business and accounted for approximately 68% of equipment order intake in the third quarter. Equipment for indium phosphide-based optical applications represents the majority of our backlog, which provides revenue visibility well into '27.

The structural demand driver remains unchanged. The rapid expansion of AI and data center infrastructure continues to fuel demand for lasers used in high-speed optical interconnects. Underlying device demand forecast has increased materially year-over-year.

For example, in its May '26 report, LightCounting increased its 2030 demand forecast for 800 gig and above transceivers as well as LPO and CPO ports from approximately 198 million units to 279 million units annually, representing an increase of around 40%. This corresponds to the customer demand, which is well ahead of the initial estimates that we made earlier this year.

In particular, we see several acceleration factors supporting demands in the near term commitment. The first factor is the rapid AI infrastructure build-out. Communications laser architectures are becoming increasingly sophisticated, high band requirements require higher laser power and that again requires larger areas, which means more wafers. In many cases, this also requires more complex device structures and as a result, additional epitaxy steps. This increases the demand for advanced high-performance MOCVD deposition capability, which means more tools.

It is also important to point out that we see demand from a broadening base of customers worldwide. The first large orders initially came from the U.S. and Europe. We are now also seeing broad and strong demand from Asia. This includes Japan, Taiwan, Singapore and China.

Finally, we see the industry's transition from 100-millimeter to 150-millimeter taking much longer than initially expected. So the vast majority of orders today is on 100-millimeter wafer diameter, even on 75-millimeter wafers, driven by limited substrate supply of 150 millimeters. Based on this, we expect the migration to 150-millimeter manufacturing to take longer than previously anticipated. As a result, customers are likely to remain on 75-millimeter and 100-millimeter production platforms for longer, which increases tool demand.

Taken together, these structural growth drivers and near-term acceleration factors give us increasing confidence that datacom lasers will remain one of the most attractive and fastest-growing end markets for our industry over the coming 2 years.

With this, let me move to power electronics, starting with gallium nitride. In gallium nitride, utilization rates are back to normal and first customers start planning expansions and new pool purchases. We expect this trend to continue through '27. The main driver for the increased demand now comes from the power conversion in AI data center.

With the new 800-volt HVDC architecture, NVIDIA has created a new power standard together with its partners. In this standard, the majority of down conversion steps from 800 volt and the supply of GPUs and CPUs is made with GaN powered devices. Here, we look at medium voltage to low-voltage devices in voltages such as 200 volt, 100 volt, 40 volt and 20 volts.

In AI data centers, the overall opportunity extends beyond traditional power supplies. GaN is being considered for intermediate bus converters, battery backup units and auxiliary power supplies. The focus is on improving power efficiency in the data center infrastructure. This shifts the focus of new capacity investments in power GaN towards the ICT infrastructure at lower power classes in contrast to the early GaN investments of the 2021 to '23 cycle, which focus on 650-volt consumer devices such as fast chargers.

These broad market applications for power GaN remain an important pillar, but are not the main focus of the recent activity that we currently see. We currently see the coexistence of 200-millimeter and 300-millimeter wafer sizes confirmed that we had projected earlier. Customers with large 200-millimeter fab capacity stay on 200-millimeter, while customers with existing 300-millimeter fab focus on 300-millimeter fabs for the new investments. We are discussing volume orders for both wafer sizes with customers.

With this, I come to silicon carbide. Demand for silicon carbide has been soft throughout the year, yet customer utilization rates have been continuously improving and now the first new tool orders have been coming in. Also, selected silicon carbide substrate vendors have mentioned capacity constraints for SiC substrates, and we see the growing volumes in our after sales business.

Where does this all come from? Our customers tell us that electric vehicles remain the primary demand driver for SiC power devices for the time being. The shift to 800 volt platforms in EVs and adoption beyond traction inverters significantly increases the SiC content per vehicle. The main driver of this is the sharp reduction in substrate prices. Overall, the number of produced silicon carbide units is increasing.

At the same time, industrial applications, grid infrastructure and AI data centers expand the thick application space beyond automotive. In AI data center, a big part of the conversion from the high voltage of the power grid to 800 volt to medium voltage is done with thick devices. Here, a new class of very high-voltage devices at 2 kilovolts, 3 kilovolts and even higher voltages is in the making.

Also in the grid itself, a new class of devices is emerging. With an increasing affordability of thick power devices, solid-state transformers, SSDs, have emerged as an alternative to traditional transformers. We closely work with customers on these applications. According to multiple customers, our planetary G10-SiC system delivers outstanding results in this new class of thick epi layers. This makes us more optimistic on a gradual recovery of silicon carbide throughout '27.

Let me now come to micro LED. Here, the demand remains soft, and the market has not yet reached broad volume deployment. However, we are seeing an increasing level of activity in the industry. These developments are taking place in 2 areas in particular: augmented reality applications and optical interconnect for AI data centers. Nevertheless, we do not expect this market to become a major volume contributor in the immediate term. Nevertheless, the increasing level of activity is encouraging and could see future demand for MOCVD systems in the coming years.

With this, let me now come -- continue with our footprint expansion. Our new site in Malaysia will become an important element of our global manufacturing footprint and strengthen our manufacturing flexibility, supply chain and long-term cost competitiveness. The project remains fully on track. The groundwork and piling has been fully completed, and we have almost finished the ground floor and expect to be able to use the building from fall 2027 onwards.

With that, let me now move to our guidance. We confirm our full year '26 guidance as published in April. We expect full year revenues of EUR 560 million within a range of plus/minus EUR 30 million. We expect a gross margin of about 42% and an EBIT margin between 17% and 20%. The gross margin and EBIT margin guidance includes one-off expenses in the mid-single-digit million euro range related to the personnel reduction in operations earlier this year.

Let me also add a few words on our expectations for next year. Based on the current sales pipeline, AIXTRON expects to maintain or further grow the current strong level of order intake over the coming quarters, supported by demand for optoelectronics and increasingly by growing momentum in power electronics. This lets us take a very positive look on the prospects for '27. As always, we will give you formal guidance for 2027 when we release our full year results on March 4, '27.

With that, I pass it back to Christian before we take questions.

Thank you very much, Felix. Thank you very much, Christian. Operator, we'll now take a questions, please.

[Operator Instructions] So we have a lot of questions listing out. I'm just going to start from Adithya Metuku from HSBC.

Firstly, Christian, maybe you're talking about EUR 275 million plus in order intake over the coming quarters. If I annualize that, I'm looking at roughly EUR 1.1 billion in revenue. So when I think about 2027 revenues, Street is at EUR 800 million. That's clearly not particularly demanding. But is there any reason why you might not be able to do EUR 1.1 billion in revenue next year? And if you don't think it's possible, then what could detract you from hitting those levels in revenue, I suppose? That's my first question. And I've got a follow-up.

Well, let me take this one. I mean, first of all, EUR 275 million in order intake is now what we have achieved in Q3. So please take into account that we are currently looking at lead times on the order of 12 months plus/minus. So EUR 275 million is rather translating then not starting from Q1, but rather from Q3. Please take the time shift into account. I think that's a very, very important element.

Second one, again we would not want to go now in the discussion of a precise guidance for '27. You see we've given you some indications on the order pipeline with the remarks that we see EUR 275 million and potentially a bit higher as order intake for the coming quarters, this was not an indication towards the revenue. And I think for the revenue -- so the second part of your question, for the revenue, we have to monitor how we can ramp.

As we have discussed in earlier earnings calls in the previous 2 quarters in this place, we have given you the indication that we are ramping up our supply chain. We are ramping up our production means our own people. Also, we work very closely together with our suppliers. Many suppliers are helping us a lot. We have a great group of suppliers around us. And this whole ecosystem of AIXTRON plus partners, production partners, is hiring people in part. Some of our suppliers have rented additional facilities for production, some have bought additional machines for production.

And as we have indicated earlier, it's all about the ramping speed. We don't have any limit in terms of production capacity. This is unlike our customers who have a fab and at some point, the clean room is full, and you need to build the next clean room. We always have indicated it's about the ramping speed. And I think we've done a good job over the year, Q1, Q2, Q3, and we've given some indication for Q4 to ramp up, and this will continue also for the next quarters. And this ramping speed is really what limits the increase. But again, no formal guidance for '27. Yes, go ahead.

Got it. And just as a follow-up, you're talking about power starting to recover. As far as I know, silicon carbide is largely G10 reactors. GaN, I think, is also more skewed to G10, especially if 300 millimeters is coming in or at least your newer reactor in the GaN space. I believe they should be more profitable than your older reactors where the wafer sizes are smaller. So if you do see power picking up into 2027, could that help your gross margins and help profitability trends into 2027? Am I thinking along the right lines?

I think you think along the right lines, let me shine a little bit of light on that. And as you indicate, as you rightfully asked, the G10 generation is more innovative, has a higher productivity, has a greater value for our customers. And hence, of course, part of the value also stays with us in terms of higher margins. That's clear. That's correct.

Please nevertheless, also take into account that it depends a lot on the regional mix. We have some regions, Taiwan and China, who likes to continue using the older generations, G5+ and G4, G5+ especially also in the power and the power electronics, which being legacy products continue to run at a lower margin just because it's older products, not yet on the high productivity. And so those ones take a drag on the margin compared to the G10.

But overall, looking in '27 and looking at the drive for the market segments that you have rightfully indicated, a lot of the opto business on the G10, some Western power electronics customers expanding on the G10 generation, we can clearly expect an uptick in gross margin as well.

We take the next question, please.

We can hear you, Ruben. Please go ahead.

Okay. One question related to the 6-inch. I think you talked about 6-inch indium phosphide wafers as one reason the opto market is bigger than you expected. I think in the previous call, you talked about 70% to 80% of systems shipped in 4-inch. Yes, has that share changed in the Q3 orders? And at this stage, what is mainly holding customers back on 6-inch? Is it still the sort of substrates, the yields? And at what point would you expect maybe the larger customers to convert to 6-inch?

Very good question, yes. It's funny, we actually looked at data for this one. And interestingly, for 2027, the share of 6-inch systems is going down. Normally, you would expect a more modern wafer size as time progresses, increases. In fact, in 2027, the 6-inch goes down from around 30% down to 25%. And as you indicate with your question, this is, in fact, due to the shortage of substrates. We see even towards the orders that have been already placed or in discussion with customers for '27, we see even 3-inch, very small, increasing and doubling from 15% to 30%. And this is literally because the industry is ramping so fast, the substrates are short and the customers literally take every wafer from every substrate that they can get a hold of because there's such a demand in the industry.

Okay. And if, let's say, I mean, it seems far out still, but like if they can convert eventually and they buy the new tools for 6-inch -- sorry, would they actually buy the new tools for 6-inch? Or would they sort of reconfigure the installed G10, do you think?

Customers can reconfigure the G10, which they are now buying and use for 3- and 4-inch. They can reconfigure these tools for 6-inch. I think the real question is customers have existing production lines. It is not only that you just buy some new wafers and convert the tool, but then you have to convert your entire production line, all the other tools in the line as well. And normally, those things don't go so fast.

Look what we've seen in the recent years with silicon carbide. Everybody expected 200-millimeter wafers are coming and they take volume around '23, '24. And now looking towards '27, we gradually now see this switch towards 200-millimeter wafers 4 years later. And you may have also seen that we had in the Q3, quite some sizable after sales business, if you look at the bar chart in our presentation, because that big part of that was now in preparation of the next silicon carbide ramp customers converting from 150 to 200 millimeters.

So I would say in SiC, we've seen something that was expected for '23, '24, now finally happening for '27. And I have no reason to believe why on the optoelectronics, this would go much, much faster.

Okay. Great. And just the second one, still staying on opto. Could you give us a sense of the mix between the EML lasers and how much maybe for the continuous wave lasers? And to what degree is demand already linked to the 1.6T.

I can unfortunately not because our customers take the goods and we make sure when we install the tools at our customer factories that the tool delivers all the specifications and is capable to perform according to that. But then in the end, the customer closes the door when he leaves and the customer keeps their secrets, both on the recipes they run on the tool as well as on their product mix. So unfortunately, I have no transparency into that.

The next question is from Om Bakhda from Jefferies.

I have 2 questions. Firstly, on pricing. If we sort of take the EUR 275 million order intake and you mentioned that you could maintain or grow that, it seems as though you could get towards the EUR 1 billion revenue mark in '27, and that would take you to being fully utilized before Malaysia capacity comes online in '28. So does that then change the pricing dynamic for 2027? Could we see some areas where there could be some opportunistic pricing? And have you already been in discussions with customers for pricing in the coming quarters?

So we have a very clear strategy on the pricing. We maintain our price points. So we neither give volume discounts as customers are now ordering very big quantities nor are we pushing for excessive price increases, as your question is suggesting, because we have seen that by keeping a stable price level, customers see us as a reliable supplier. It's not necessarily a negotiation up and down. It's a long-term partnership. And this long-term partnership in the very long run is then beneficial because it also secures that when the reorders are coming, they kind of by default go to AIXTRON and secures our long-term market position.

So we don't take -- and I think this may be different from us than from customers who have a hard limit on fab capacity, right? I always say my customers, right, they have a clean room and at some point, the fab is full. And then they have no possibility to extend. While on our side, we don't have such a hard limit, stop of something, capacity is full. We can always rent an additional building or some manufacturing space. So we are not limited by a capacity number of tools or certain revenue number in a year or in a quarter. It's rather we have to manage and bring the supply chain up and running, hire enough people, make sure our suppliers are ramping up and then kind of it's just a timing question of how much we can then produce.

That's very helpful. And then just a follow-up, coming back to the lead times that you mentioned earlier. I think previously, you talked to sort of 6- to 9-month lead times and now that sort of increased to around 12 months. Is this predominantly because of supply chain related? And if it is, could you give some color as to where there could be some constraints in areas of sensitivity? And if there are 12-month lead times now, does that then mean that the orders that you'd get in Q1 and Q2 of next year, you wouldn't be able to deliver those tools in 2027?

That's a very good question. So the limit, I would say, is overall the ramping of ourselves and our supply chain. I think it goes very well hand in hand. It's not that there is one thing short and like 1 component short or 2 components short is limiting us to grow. It's rather -- we hire people, our suppliers are hiring people and then we together run as fast as we can. So that's what I would say. And that gives us the quarter-over-quarter, how much every quarter can we increase, because this is how we have to think about this.

Now in terms of what orders translate into revenue in which quarter, it's not very straightforward to answer. It's not like you take the order intake from one quarter and you shift it by 4 columns and axle to the right. Each column we need one quarter because with some customers, we have capacity reservations. I spoke about the long-term partnerships. So -- and the slot reservations.

So I think it's more the general picture. If you come as a new customer, not having made any reservations and want, let's say, 5 tools or want 10 tools, then I would say, well, the slot I can give you is 4 quarters into the future. But it does not necessarily mean that there might not be a shipment of a tool in January or like in April, '27, where the order might not come, I don't know, in December. That could be the case if customers have made reservations. Sorry, it's not such a black and white picture, but it's just how the reality works.

So then -- sorry, just following up on that. And if -- so then does that mean that essentially your prepayments become more important than your orders in terms of customers, if they're willing to put down prepayments now, then that -- even though they're not translating into orders that could sort of be enough to then be ready for '27. How should we sort of think about that order and prepayment relationship?

That's a whole mix, unfortunately. Sorry for not being clear. I like to be more clear normally, but it's literally a mix of models, right? There's models about capacity reservations, there's models about prepayment staggered over the order lifetime, because many of these orders, as you can imagine, is not on order of 3 or 5 tools, but rather whatever, a dozen of tools and stretch over 2 or 3 quarters and so on. So it's a mix of different types of contracts and various shape of grid.

The next question is from Oliver Wong from Bank of America.

First question, just wanted to ask a bit more on the capacity and revenue run rate. Right now, it's around EUR 200 million a quarter. Should we expect -- as you work on improving the supply chain, should we expect this number to kind of increase in a linear fashion throughout the year next year? And I was wondering what kind of ambition you have in terms of if things go well in terms of what you're doing with the supply chain? What kind of run rate we could be looking at, let's say, by the end of next year?

Oliver, what you asked is a billion-dollar question. I wish I could answer this already. I'm not at this stage ready to answer this. Yes. I think this is the key question. We are working with our suppliers. We are working internally right now. We work on the ramp. But it's a bit too early to give you a definite answer. We will come out with our guidance in March when we have the full year results, but then we will have an answer. At this point in time, multiple factors.

As you see, it's really operational topics. Hiring people, working together with our suppliers and very close alignment, identifying which of the suppliers are the bottlenecks, what kind of limit the capacity rates, then see -- and I think we've done a pretty good job about wherever there is a bottleneck to eliminate the bottlenecks. In many cases, this was with one supplier, and we have literally together with that supplier eliminated the bottleneck and one supplier stayed on.

In other cases, the suppliers said, "Guys, we can't just can't do more." We are not unhappy if another second or third supplier is added. We need a bit more time as this goes deep into the operations to work out and then by the time we give our guidance to come up with an answer. It has multi levels of facets below it. It's a bit more complex than the orders and the customer orders and the market dynamics.

Yes. Makes sense. That's helpful. And my other question is regarding SiC. We're pretty much at the bottom this year. You mentioned demand kind of picking up. So looking at historically, 2023 was, I believe, your last peak at around EUR 200 million. And then before this year, you're kind of at EUR 100-something million a year in SiC. I was just wondering how fast do you think SiC demand could really ramp based on what you're discussing with your customers, et cetera?

A very good question. I see SiC coming slower than GaN. You see I've also changed the order and put GaN upfront. With GaN, we clearly see now a ramp starting. We see the market very clear capacity being utilized in full. I think I've indicated in the prepared remarks. Silicon carbide, we see some customers ordering. We clearly see the utilization rates going up, also in our after sales business. I've seen selected customers planning new expansion, while I've seen other customers, I would say, undecided maybe because they got burned so hard.

I mean there was a Chapter 11 in the SiC market, if we recall just 2 years back, very painful. I haven't seen the broad sentiment in the SiC yet where I would say that they go all in. I think my guess is SiC will be a bit of a softer recovery.

The next question is from Veysel Taze from Metzler.

Maybe the first one on Oliver's question. I understand you cannot really give a full visibility how your capacities will ramp up throughout the year. But nevertheless, asked differently, what keeps you away that first quarter will be not a similar revenue size in terms of capacities like Q4? I mean Q4 will be probably around EUR 210 million, EUR 220 million run rate?

I think your reasoning makes sense to say if we reach a certain capacity level, we should be able to maintain that capacity level also in subsequent quarters. I think that logic is right because in the end, as I've indicated, it's about hiring the people, getting things to run, working together with our suppliers, help them ramp the people, and sometimes we work with suppliers also helping them on operational processes. It goes very deep into operations, what we are doing. And one things are running and up and running. Normally, they don't fall back or collapse, right? But it rather stays. I think your logic is reasonable.

Great. And then the second one, Felix, a little bit on opto technological understanding. There has been now a lot of rumors about VCSEL lasers making into the AI scale-up networking. I think a few vendors have pushed forward their time lines, particularly for up to 15 meters as a sweet spot for VCSEL laser adoption. I mean VCSEL is a gallium arsenide product, and you are quite strong in gallium arsenide. But nevertheless, VCSEL has a much more mature supply chain, bigger wafers. And I think the epi complexity is not as high as with EMP lasers. Can you give a little bit color if VCSEL takes off a little bit stronger in scale up in 2027, but particularly 2028? What would that mean for your TAM or for your SAM expectations?

So we are involved in a number of the VCSEL projects for the data communications. We see the activities that we have seen so far focusing mostly on very, very short-range activities, sometimes literally on a card. We -- and nevertheless, the impression I have is those topics are still under the development on a completely different maturity stage than what we've seen on the indium phosphide, more R&D and exploratory and focusing on a complementary subsegment of the spec. This is how I have seen it for my technical team.

So in other words, I would rather translate that it may turn at some point in an upside for us rather than in a threat or a replacement of the indium phosphide-based market. This is how I see it.

Great. If I can squeeze in one more for Christian on his comment for the 4Q gross margin. I think you said that you don't see the 48% gross margin for Q3, not respective, not an indication order for Q4. I was wondering why that should not be the case. I mean, at the end of the day, your order backlog, the mix is still towards opto. And if you now have, in Q4, a little bit GaN shipments on top, why should not gross margin keep -- stay at the same level in Q4?

Felix here. We wanted to make that clear that we don't draw wrong conclusions. We have a good view on what we will ship. That's why we also highlighted it was in Q3, we had a very good product and regional mix. So the combination was just very, very good. In Q4, we will have a different regional mix and also different product mix, more older product generation, G4, G5, then also more China mix in. We don't know exactly how it will come out, but it will come out lower than the Q3.

The next question is from Gustav Froberg from Berenberg.

Just 2, please. First, a question around the momentum comments that you made that you expect more momentum going forward, either at the same level or even a little bit higher. What gives you this comfort? Is it -- or have you received letters of intent? Are there shadow orders? Is there anything more concrete to point to? Or is this more just based on customer commentary and your general feeling from discussions with customers?

And then second question is on opto -- I can do the first one first, that's fine. I'll pause.

Yes. Very short. It's discussions with customers, pipeline preparations, quotation activity. So no paperwork in the sense.

Okay. Great. And then on opto, I was just wondering if you there rather give a new assessment for addressable demand of optoelectronics tools at this stage? And maybe some words on what new assumptions you would be or you would make around that addressable demand. I know it's exceeded expectations so far this year. So just curious what your sort of new estimates or assumptions would be around sustainable opto demand.

Yes. That's a very good question. I keep asking my marketing team, and I keep not trusting the numbers anymore. Honest answer. Our first estimate was incorrect. As you have rightfully spotted, as we have indicated, right, the demand is currently much larger than we had initially indicated. And also, we see that market analysts and predictions are putting up the demand forecast, right? The numbers I quoted earlier in my prepared remarks, plus 40% from what they had estimated. I had many discussions just in the last 3, 4 weeks was around SEMICON West meeting many of my customers. The honest answer is nobody read. My customers also don't really know. So I refrain from giving out a new number. It would not be serious.

What I can tell you, nevertheless, is that customers give a very clear indication and a very strong confidence. They see the demand of tools that they order from us backed by solid orders from their customers. So it's not just to build out some capacity and then see what happens. What we've seen earlier years in silicon carbide, I think everybody was just following a dream, but in the end, the dream didn't materialize. Here, I hear from my customers that what they order to at us is backed by solid orders from their end customers.

So I hear a very strong confidence of my customers that the pipeline they are placing out and forecast they are giving us is solid -- solid and increasing. And this is what we have reflected in our remarks that we see the strong order momentum that we have now in Q3 continuing also in the next, I don't know, 2, 3 quarters.

The next question is from Craig Mcdowell from JPMorgan.

My first one was on gross margin. I wanted to ask how we should be thinking about 2027. From the order book that's been building through this year, it seems like heavily weighted to G10, opto, all for delivery in 2027. So the revenue mix seems like for delivery in '27 should be similar to what you reported in Q3 and indeed at larger volumes. So the question is, should we use Q3 '26 gross margin as a good proxy for 2027 gross margin? And I've got a follow-up.

That's a difficult one. It's really too early to give a precise guidance here. We will give a precise guidance when we publish our full year results. But overall, of course -- I mean, I would be a little bit careful in just taking the Q3 gross margin as a run rate throughout the next year. As I explained, that was really a constellation of stars, a very good product mix, very good regional mix, but you cannot expect that to continue into the next year.

But overall, of course, the higher volume will help with some operating leverage in the COGS area also in the gross margin. And overall, the mix with the opto and some GaN coming in should naturally give a few percentage points over this year's gross margin into the next year. But we cannot be more precise at this point in time.

I think, Christian, this makes sense.

Let me build on that one. I think this time of the year, everybody looks into the next year. I want to give also one more indication for next year just to make sure that as much -- I think you all see that we don't know yet. We give you as much transparency as we can as we have it. So to add to this one, we see -- I think all of us clearly see '27 will be clearly ahead of '26, a no-brainer, right? And also, we will use this good financial situation in '27 for another push on the product generation with also a significant uptick on the R&D.

If you guys start modeling, a very significant uptick on the R&D because we see we have a good experience now with the G10 generation. It's in the field for quite some time. Customers have experience. We spent some time to really make it mature and get all the topics done that customers wanted over the last years and quite successfully, I would say. And now next year will be the one to take on the next generation AXIA. So we will use the financial situation also then for a good uptick on the R&D. So we are not just run it flat.

It would mean all the additional gross profit would fall down to the EBIT, but also we are planning to up the R&D rate again to prepare for further growth, to secure our market position because we have good feedback from many customers. We've collected their wishes, what they want for the next-generation product, and then, of course, we will work hard on getting the things done to then make sure what -- I don't know, 29, 30 or whenever it is, the next generation comes and it stays as healthy and as strong for us in the market as we are today.

Very helpful. And then my follow-up was on GaN. We've seen Infineon this week announced a large dedicated 300-mill GaN facility. Can you comment on your position here? When do you expect a decision on whether you or Veeco will tool this facility? And maybe just your degree of confidence on winning that position there.

Yes. So on 300-millimeter GaN, first of all, we see our initial hypothesis that we have shared also with the market. I think we have somewhere in the backup of our investor presentation, even a slide on that in several quarters. We see that there will be a coexistence of 200-millimeter GaN and 300-millimeter GaN. And we said from the beginning that we expect the customers who have large 200-millimeter fab capacity today, silicon for 200-millimeter fab capacity will stay on 200-millimeter. This is what we see.

And we see that the customers who have large 300-millimeter fab capacity, silicon power devices, MOSFET IGBT and so on and so forth, that those guys who have existing 300-millimeter, they will gradually convert the existing fabs towards 300-millimeter fab. And I think here, we have the first evidence now that this is just happening. So our hypothesis of the coexistence of the 2 wafer sizes is, first of all, confirmed.

And secondly, as you know, we don't never comment on individual customers, also not on this one. What I can tell you is that we're preparing for a sizable double-digit number of tools and next year on 300-millimeter. So our 300-millimeter will go into very large volume next year, not only from one customer, but from more than one custom.

The next question is from Martin Jungfleisch from BNP Paribas.

I just have 2 quick follow-ups. The first one is really on the order comment you made in the beginning. Would you expect this potential step-up that you were talking about in the orders in the next couple of quarters be more driven by power semis? Or is it more by opto? Or should we assume kind of this 65% share opto, 25% share of power to be kind of the norm for the next, I would say, 1 or 2 quarters?

I would expect on the first shot that this mix that we have seen will continue. So a good level of optoelectronics, clearly the dominant one. Nevertheless, the power also continuing to be a healthy portion here.

Okay. That's helpful. And then the other question is really just on the use of cash. I mean, obviously, you have sitting on a quite substantial amount of cash with the cash generation to date, also the convertibles adding to that. Can you just give an update really what your preference is? Would the buyback make sense here? Maybe if you can share some plans on potential smaller bolt-on M&A, that would be helpful.

Yes. Somehow expected that question. Let me put that a little bit in perspective. The overall cash position, of course, is getting quite high. But please consider also the high number of contract liabilities for advanced payments where we have received that money from customers. We've not earned it, yes. I mean we need to first -- now we need to build the equipment to deliver. I cannot do a share buyback based on that.

And also we increased liability, of course, from the convertible bond. If you take all of that out, you can take the noncurrent liabilities out and the contract liabilities for advanced payments, well then you end up at around EUR 275 million, which is still very good, but we are conservative in our financial policy and not taking a share buyback now. It's too early. But we promise you we will make good use of that cash to generate value for the future.

The next question is from Martin Marandon-Carlhian from ODDO BHF.

My first one is on photonics. And how should we think about the respective contribution of pluggable transceivers and co-packaged optics to the current order pipeline? Because it seems like the real inflection of CPO volumes is for NVIDIA Feynman GPUs in '28. So I was wondering if these orders are more expected next year for co-packaged optics high volume? Or do you see already the current order intake addressing that trend for larger CPO volumes?

I expect that the capacity, as mentioned, right, we don't know in the very, very, very details what our customers ultimately do because the machines are very versatile and can be used for multiple use cases. But from what I infer from customer discussions, this is all the preparations for what we see for the Rubin Ultra and later on for the Feynman. This is really the preparation for these optical architectures, which then have the benefit of connecting many, many, many, many, many more GPUs with each other and further building up the power of the AI. So it's really in preparation of that ramp what NVIDIA is driving with their architectures and together with their partners.

Okay. Very clear. And the second one is on power. So you said momentum in GaN and SiC has improved. So do you consider Q3 order intake in power, so I guess it was around EUR 60 million or so to be already the sign of a recovery? Or you actually see more meaningful orders in Q4 and the following quarters for power? And how much of these orders do you think are directly related to GaN adoption in data centers?

That's a very good question. I expect the GaN and the SiC, and I don't have the detailed breakdown here, to be in about the same proportion as we have seen it in the Q3. We expect this to stay on a good and healthy level or to further increase. And I can clearly comment on the gallium nitride for the GaN, I attribute the biggest momentum or the biggest part of these orders clearly to the low and medium voltage part and the AI power build-out. We clearly hear that from the customers. 650 volt, which was the initial wave around consumer devices and also is adding a little bit, but the majority of additional capacity goes into the lower voltage devices, and that all goes into the AI build-out.

And the last question goes to Adithya Metuku from HSBC.

Two additional questions. Firstly, just one of your peers recently talked about -- well, they talked about a Group 4 material being used in photodetectors and then winning epitaxy layers for that step. I believe they were talking about germanium. So I just wanted to see what you're seeing in terms of use of germanium in photodetectors. And could you also talk a bit about your positioning there?

And secondly, I also wondered if you could talk a bit about how you see microLED slow and wide adoption happening in data centers. Any clarity you can give around that would be great.

Yes, very happy to. So germanium as a photodetector material is widespread and very well established. So that's kind of a well-known material, we are not addressing that. So this is not our focus. And so with that -- but I think it makes a lot of sense. germanium is the right band gap. It's well proven. It's very nice compatible with the silicon amplification circuits that you -- low noise amplifiers that you build right behind this, in terms of silicon photonics and integrated optics. So this is the way to go. This is all what I can say from my general physics knowledge. As I said, we are not in this market.

And regarding the microLEDs for the AI, we first it goes a bit in the same direction as we had earlier in this call, a question regarding the VCSELs. There is quite some heavy work on this research work. We are working together with multiple customers on that one. And again, it will be complementary and additional market segment to the indium phosphide, what we see. And what customers of ours are playing with or what the vision is, put it like this, is to have gallium nitride microLEDs. And with gallium nitride, we are involved. And gallium nitride microLEDs bonded to either the GPU or to the CPU and then to use, for example, a silicon carbide or a sapphire interposer on which you place both the GPU die, the dies for high bandwidth memory. And the vision is to make also the interconnect between the GPU and the high-bandwidth memory and optical interconnect.

Today, in all the architectures we've been discussing in the last 45 minutes, we spoke about copper interconnects between the GPU and the high-bandwidth memory. But again, the same limitation tool, which is why they get for the longer haul or within rack or within the data center replaced by indium phosphide, the same bandwidth limitation tool. And the dream would be to make the interconnect between the memory and the GPU also an optical interconnect. And you could do that with gallium nitride microLEDs, making little microLEDs and then kind of contact print them on the GPU and the backside of the high-bandwidth memory and put them on a substrate into which you write some waveguides and then have an optical connect between the GPU and the memory. That would be the vision, and there is some work ongoing.

Operator, if no more questions, then we would close the call, please.

We actually have Ruben Devos on the line again with a follow-up question from Kepler Cheuvreux.

Final question. I just had one on Penang. I think you're adding 25% to 30% capacity by '28. Is that mainly for opto? And how could the cost structure change once that site is up and running? I think you're also referring to that you could potentially expand beyond that initial build-out. Of course, curious when that decision could happen. I mean, it looks like certainly, customers are committing based upon the sort of customer advances, the lead time stretching beyond 12 months. So your visibility gets much better. So yes, curious on your thoughts here on Penang.

Let me try. A very good question. In fact, we are currently -- as we build the plant, we built for the full building, we built the shell, like means the outer part of the building, the walls and the roof kind of, and we prepare everything for the infrastructure, while from the possible cleanroom space in the Stage 1, we are only equipping about 1/3 with clean room space because as always, when you build clean room, the expensive part is not the building, the shell, but the expensive part is all the ventilation, the clean room systems and the technical apparatus that you put inside of the building. So this is the reason why in Stage 1, we only sell 1/3 of the building. Well, of course, it's economical, the heavy construction ongoing to make the shell and the roof and the kind of stuff to do that in one shot.

So that means, in other words, as we only equip 1/3, it's clear you could multiply the additional capacity that we built out there by a factor of 3, Stage 2 and Stage 3 also. We estimate that each phase will take us 6, 7 months to build, so relatively fast because it's just inside of the building, the technical infrastructure, so we could respond and react quite fast to do that. We see that for the starting point, we are well with those additional capacity that we are creating here.

And in terms of the final -- the answer regarding the cost structure, which was also part of your question, it's a bit too early to tell, because we are just building the plant now. We are ramping our team. We are learning about the supply chains in Southeast Asia. We see quite -- we get quite some good indications. We are quite happy with our decision that we made on the strategic direction, but it's too early to put that in quantitative numbers because we are just starting to engage with the local supplier ecosystem. And before I can give you a quantitative number, I think I need to have 60%, 70%, 80% of my parts and get prices and closed on them. So then at that stage, I would feel well to extrapolate and share the numbers with you.

Okay. Great. Great. And very briefly, just on the -- I think in the presentation historically, you talked about sort of the silicon carbide could be doubling the midterm tool demand and then for GaN, 3x. I mean those figures aren't in today's slide deck, I think, but should we read anything into that? Or do those figures still stand?

No, I think for the midterm, that all still holds. We've tried in this version of the slide deck, you've seen it, you caught it right. We've updated the detailed slide on the GaN and the SiC. And we thought it's more important now to shine some light as we've tried also in this call about like the near term means the trend that we see for '27 rather than the complete time frame extending towards also '28, '29, '30. So we really have focused here with the update of the slides just on the next year to start giving an indication what we see just now happen.

So that was it for today. I would like to thank everyone for your participation. And with that, I would like to hand over to your host, Christian Ludwig, for the closing remarks.

Thank you very much. Thank you all for the excellent questions. I hope to see you before end of the year at the one other conference or roadshow. And if not, go back and talk to you at the latest on March 4 when we'll publish our full year results for 2026 and then also have a more formal guidance for next year. Have a great day. Goodbye to you all.

Good afternoon, ladies and gentlemen. This is Christian Ludwig. A warm welcome from my side to AIXTRON's Q2 2026 Conference Call. I am very sorry for the technical delays that we had experienced and that had to keep you waiting. We had some major issues with our external provider. We're working on this. So next time, hopefully, we'll have a smooth communication. With me in the room today is our CEO, Dr. Felix Grawert; and our CFO, Dr. Christian Danninger, who will guide you through today's presentation and then take your questions. This call is being recorded by AIXTRON and is considered copyright material. As such, it cannot be recorded or rebroadcast without permission. Your participation in this call implies your content to the recording. Please take note of the disclaimer that you find on Page 1 of the presentation document as it applies throughout the conference call. This call is not being broadcast via webcast or any other media. However, we will make a transcript available on our website after the call. I would now like to hand you over to our CEO for his opening remarks. Felix, the floor is yours.

Thank you, Christian. Let me also welcome you to our Q2 '26 results presentation. And again, also from my side, from the Board, apologies for the delay with our external provider. Thanks a lot for everybody who kept waiting. I will start now with an overview of the highlights of the quarter and then hand over to our CFO, Christian, for more details on our financial figures. Finally, I will give you an update on the development of our business and guidance. Let me start by giving you an update on the key business developments of the second quarter on Slide 2.

The key messages are: we've received strong new orders of EUR 215 million, predominantly driven by Optoelectronics. Revenues came out at EUR 115 million, in line with our guidance for the quarter, reflecting the ongoing ramp of the Photonics business and coming out of the trough in Power Electronics in Q1. We also generated a very strong cash flow. Operating cash flow reached EUR 119 million and free cash flow came in at EUR 114 million. This was predominantly driven by higher customer advanced payments, reflecting the strong order momentum and supporting the upcoming production ramp.

Despite the soft start into the year, we are fully on track to achieve the full year guidance with revenues of EUR 560 million plus/minus EUR 30 million. We had a soft Q1 with revenues of EUR 59 million. We achieved $115 million in Q2. We plan to further increase output in Q3 to 180 million plus/minus 20 million, and we target a very strong Q4 with a further increase in output beyond the Q3 level. As you can see, we are ramping up our production with a steep ramp rate and significant increase of output every quarter of this year to serve the growing demand in optoelectronics. Finally, construction of our new site in Malaysia has started and is fully on track. Christian will now provide you a detailed look at our financials on the following pages before I take over again. Christian?

Thanks, Felix, and hello too. Let me start with the highlights of our revenue development on Slide 4. Q2 revenues marked the next step in our planned production ramp increasing to EUR 150 million from EUR 59 million in Q1. This was fully in line with our quarterly guidance of plus/minus EUR 10 million. Compared with the prior year quarter, revenues were 16% lower. For the first half, revenues amounted to EUR 174 million. The revenue mix already shows the growing importance of Optoelectronics. 54% of the equipment revenues came from Opto, 23% from LED and micro LED, 22% from GaN and SiC power and 2% from R&D tools. Our aftersales business contributed EUR 52 million and remained stable in absolute terms year-over-year. As a result, its share of group revenues increased to 30% from 21% a year ago.

Now let's take a look at the financial KPIs of the income statement on Slide 5. Gross profit in Q2 was EUR 47 million, corresponding to a gross margin of 41%, unchanged from Q2 last year. This demonstrates a solid margin performance as volumes began to recover. For the first half, gross profit was EUR 58 million and gross margin was 33%, 3 percentage points below the prior year period. The H1 margin reflects the lower production volume, particularly in Q1 as well as a mid-single-digit euro million one-off expense related to the personnel reduction in operations.

Operating expenses in Q2 were stable year-over-year at EUR 32 million. Higher R&D spending was offset mainly by increased R&D grants and significantly lower FX losses. The rise in R&D expenses primarily reflect higher depreciation and material costs as we continue to invest in our technology road map. For the first 6 months, operating expenses were EUR 65 million, up 3% year-over-year. EBIT in Q2 was EUR 15 million, equivalent to an EBIT margin of 13%. This represents a clear turnaround from Q1 and reflects a higher revenue level and improved operating leverage. For the first half, EBIT was negative EUR 8 million, corresponding to an EBIT margin of negative 4%, mainly due to low Q1 volume and the one-off expense already mentioned.

Let me now turn to the key balance sheet and cash flow indicators on Slide 6. Working capital decreased by EUR 184 million compared with the end of 2025. This was primarily driven by strong customer advance payments and the conversion of receivables from last year's fourth quarter revenues into cash. Trade receivables declined to EUR 81 million at the end of June from EUR 131 million at year-end. Strong order momentum resulted in a substantial increase in customer advance payments. The composition of current orders and individually agreed payment terms also provided a modest additional benefit.

At the end of June, advanced payments stood at EUR 197 million, more than EUR 150 million above year-end 2025. They represented around 43% of the equipment order backlog and provide a meaningful funding for the upcoming production ramp. At the same time, inventories increased to EUR 318 million from EUR 284 million at the year-end, mainly reflecting higher work in progress for shipments scheduled in the coming quarters. Trade payables rose to EUR 49 million from EUR 34 million as purchasing activity increased to support the ramp. Both developments are consistent with the planned increase in output in the second half.

Overall, operating cash flow reached EUR 173 million in the first million in the first half, an increase of almost EUR 90 million from EUR 85 million in the prior year period. In Q2 alone, operating cash flow was EUR 119 million. The key driver was the strong increase in customer advance payments, which supports the financing of the production. Free cash flow was correspondingly strong at EUR 162 million in the first half compared with EUR 74 million a year ago. In Q2 alone, free cash flow amounted to EUR 140 million. Capex in the first half was close to EUR 11 million. For the full year, we expect Capex of around EUR 55 million comprising our baseline investments and approximately 2/3 of the announced EUR 40 million investment in the Malaysia expansion. The process to sell our site in Italy is ongoing.

Our liquidity, comprising cash, cash equivalents and other current financial assets increased to EUR 816 million as of June 30, 2026. After deducting the liability component of the convertible bonds, net financial assets amounted to EUR 460 million compared with EUR 222 million at year-end 2025. Our equity ratio remained strong at 61% despite the increase in total assets following the bond issue. With that, let me hand you back over to Felix.

Thank you, Christian. I would like to continue with an update on key trends in our different markets. Overall, the picture in Q2 was very much in line with what we described at the end of Q1. Optoelectronics remains very dynamic and is currently the clear growth driver for AIXTRON, while the power electronics and LED micro LED markets remain in different phases of recovery. I will start with Optoelectronics. Momentum remained exceptionally strong in the second quarter. After the clear inflection point we saw in Q1, the market continued to accelerate in Q2, driven by strong demand from AI data center applications. In Q2 alone, optoelectronics accounted for 75% of equipment orders, making it by far the most important contributor to our order intake.

Based on our current visibility, we expect order momentum in optoelectronics to remain at very high levels in the second half of the year. The key driver remains the ongoing architectural shift in AI data centers from copper-based connections at lower speed optical links towards high-speed optical connectivity at 800 gig and eventually 1.6T and beyond. This transition requires a significant increase in the number and performance of optical links and therefore, in the number of advanced lasers needed. In the near term, demand continues to be driven by indium phosphide-based EMLs and CW lasers for data center applications. At the same time, customers are already working on photonic integrated circuits as the next step in technology road map.

Importantly, the current investment wave is no longer limited to a few large players. We now see leading laser suppliers worldwide committing to capacity expansions and the trend is also extending to smaller laser manufacturers. Based on the current pipeline and customer discussions, we expect this investment wave to remain at very high levels over the coming quarters. This gives us strong visibility and supports our confidence in sustained momentum in optoelectronics well into '27.

As discussed previously, our G10 ASP platform is benefiting strongly from this trend. Customers are increasingly standardizing on this AIXTRON platform for advanced photonic devices. The reason is that yield, uniformity and cost of ownership are critical in this application. Many systems are initially configured for 4-inch wafers, but are already prepared for a transition to 6-inch, giving customers the flexibility as the technology and supply chain evolves. Let me keep it therefore short on SiC and GaN power electronics as there's not much new compared to the last quarter. In SiC, the installed base of equipment at customers remains underutilized and customers remain cautious about new capacity investments.

At the same time, the underlying demand for silicon carbide chips continues to grow, and we are seeing utilization rates gradually increase at our key customers. When this leads to new tool orders is difficult to predict at the current point in time. Also in gallium nitride, demand for power equipment remains moderate. Also here, customer utilization is gradually increasing. However, it remains too early to determine when this will translate into additional equipment demand. Strategically, our conviction around GaN remains unchanged.

Recent announcements and discussions around Computex and PCIM have reinforced our view that gallium nitride has a strong opportunity to win designs in AI data centers, particularly for 800-volt DC to DC conversion. Over time, we also see the potential for GaN to move beyond the 48-volt level and into lower voltage point-of-load applications. In such architectures, GaN could replace multiple stages of silicon MOSFET-based power conversion, enabling more compact and efficient power delivery for AI infrastructure. AXON remains very well positioned in GaN, yet we have not seen signals for the inflection point for significant new order momentum.

Finally, a brief comment on LED and micro LED. Both applications remained soft in the second quarter. We shipped systems for red orange mini LED in previous quarters, and these systems are currently being installed and ramped up. However, the broader investment environment for mini LEDs remains limited as of today. For micro LED, most of the demand we see today is driven by AR glasses. We continue to believe in the long-term potential of this technology. We see more and more focus of end customers on innovative products such as the AI glasses of Meta Ray-Ban, which may, at some point, generate stronger customer pull. The exact timing for such order uptick remains unclear as of today, but it may materialize as early as '27.

Let me now move to our footprint expansion. Our new site in Malaysia is set to become an important element of our global footprint and represents a key step in strengthening our manufacturing base. With the new facility in Penang, we are extending our presence in one of the world's leading semiconductor ecosystems. The project is running fully as planned and groundworks have already started. With that, let me now move on to our guidance. We confirm our increased guidance for 2026 as published in mid-April. We expect revenues to come in at EUR 560 million in a range of plus/minus -- we expect a gross margin of about 42% and an EBIT margin between 17% and 20%.

The guidance for the gross margin and EBIT margin includes one-off expenses in the mid-single-digit euro million range related to the personnel reduction in operations. The measures will lead to annualized savings of a similar magnitude in the future. For Q3 '26, we expect revenues of $180 million, plus/minus 20 million. For Q4 '26, we then target a further increase in output corresponding to the steep quarter-over-quarter ramp that characterizes our fiscal year 2026. We continue to monitor geopolitical developments closely, particularly in the Middle East. This includes potential impact on energy prices, supply chains, financial markets and investment and demand behavior. At present, we do not see a significant impact on our business, but we will respond appropriately if the situation changes. With that, I'll pass it back to Christian before we take questions. Thank you.

Thank you. Operator, we are now ready to take questions.

[Operator Instructions] The first question. Question is from Adithya Metuku from HSBC.

Two questions, please. Firstly, just on optoelectronics demand, you talked about demand still accelerating. We've also seen similar commentary from your customers. I just wondered, based on your discussions, when do you see the supply-demand coming to balance indium phosphide laser landscape? That's the first question. And I've got a follow-up.

What do you mean -- let me question back. What do you mean with the demand and supply coming into balance? I'm not sure I get the question.

Yes. So if you look at your customers, if you take Lumentum's comments, they made some commentary around being 30% below demand, and they don't think they will be able to meet the demand that's out there in '27. So I just wondered -- and that's clearly driving the investments, which you're benefiting from. So when do you see their supply, which is tied to these machines you're producing coming in balance with the demand that they're seeing from their customers, if that makes sense? Is it 28? Is it 29? What are your customers telling you?

I cannot comment on individual customers. And also for us, it is difficult to assess the situation, the demand-supply situation that our customers are in because I'm unable to see the pipeline, the demand pipeline that's behind my customers. So please understand I cannot comment on that particular aspect. What I can comment on is on the AIXTRON side of things. And what we are seeing here is very clear that we see that the demand for laser, for tools for laser for the optoelectronics is a very broad market momentum. It's not only 2 or 3 players, but rather we see a diversified set of, I would say, almost 15 laser companies, many, of course, the big players from Europe and the U.S., very big players also from Taiwan and from Japan.

But also, we see a very strong momentum from China. I think let's recall, we all know China has a very strong optoelectronics ecosystem. And this whole industry, I think, almost at a global breadth, very well diversified. As I said, about 15 players who make up, we just analyzed that about 80% of our order intake is taking up quite a strong momentum. The customers are communicating to us also their forecast when they would like to have the tools for their individual ramps. -- that typically depends. They have some space left in existing factories and customers start building new fab construction projects.

And then whatever the customer fab is scheduled to get online, let's say, in the second quarter of next year, then we make sure that we reserve for the slots so that in the second or third quarter of next year, we can serve them. So we see here a broad and a continued long momentum ramping up and really building the capacity base that is needed. This is what we see to describe the demand behavior on the customer side. Maybe that's a bit helpful.

Yes, of course. And just as a follow-up, I had a bit of a technical question. So -- my understanding from recent developments is that when you go to CPOs or any other architectures that use an external laser source, you have a significant amount of losses before the light is actually fed into the fiber. And so that means there is a need for a lot of amplification lasers, et cetera. So the question is, for a given optical link bandwidth, if you were to move from a pluggable to a CPU with an external laser source, do you see a significant increase in the indium phosphide die or die size that is needed. So for example, if I take a 1 terabit per second transceiver and substitute that with a 1 terabit per second CPO, do you need more indium phosphide die and more extra machines for that transition?

Without going too deep into the technical details and also every of the customers has a different architecture behind it. What we see is generally that with increasing speed, we see increasing die size and increasing wafer capacity needed. That comes simply as the speed goes up, modulation gets more lossy and more die sizes. So the overall trend, higher speed rates is driving also further wafer demand.

It's from Mr. Martin Marandon-Carlhian from ODDO.

My first one is on the '26 guidance. Just a clarification there. I mean, there was no upgrade to the guidance, but the order intake is very strong. So do you think -- I mean, I assume that a lot of orders are going into '27. So would you say that the bottleneck for this year, it's more about execution than anything else? That's my first question.

That's a fair remark. So we feel comfortable in hitting the guidance. As you see, as we have outlined, we are steepening and ramping up our output quarter-over-quarter. Yes, you saw the around 60 in the first quarter, 115 in the second, 180 or referring to the midpoint, right, EUR 180 million for the third. And then if you just do the math, around 200 -- a little over 200 for the third quarter. We are not limited by orders. You see that also in the pie chart we had in the deck, but we are rather now really doing everything to execute to make sure that with this laser boom that has just started early in March, just not even 3-4 months out to work to ramp up our own capacity to work very closely with our suppliers. We have very good partners over the years.

And of course, the laser machines need some different parts than the power machines to ramp up the supply chain. Hence, you see this steep quarter-over-quarter increase, and that's really for the year the limiting factor. It's not. And I think the good news is we are quite able quite well to satisfy our customers' needs to get the machines out as they need it. But you can imagine such a steep ramp from 60 to 200 3x within a year in terms of output, that's quite some work behind it.

Helpful. And the second question is on the Photonics orders. I mean you mentioned that the momentum is very strong and will continue to be very strong. So looking at the run rate today, it implies that sales could be around EUR 900 million next year. So do you think we are kind of at a high plateau in terms of orders? Do you think we'll go higher? Or should we expect a little bit of volatility quarter-to-quarter for these orders?

Good question. Honestly, I think around EUR 200 million, what we have seen now is an order intake is for the next quarters as much as we have visibility what we are looking at. So we think that's a fair assumption to see. And then, of course, with a bit of a delay time, the lead times for equipment shipments, we will then see that also translate into revenues. So I think your guess is approximately right.

And the last one, if I may, on GaN adoption in data centers. I understand that it's not in the order intake yet. So -- what the discussion with customers are looking like at the moment? I mean, is it normal we don't see it yet? Or do you think the time line for mass adoption have somewhat moved?

So I think I wouldn't say the time line has moved. I think the market -- the capital markets, in particular, has misunderstood the time line and it's now just realizing how long it takes. So let's recap, right? We saw the new 800-volt architecture. And now the implementation of that architecture is going to come in steps. the first stage where we see a silicon to switch to wide band gap is going to be on the very high voltage side for the silicon carbide front end. When you do the down conversion from the overland line from whatever 10 kV, 13 kV all the way to 800 volt. This is going to be silicon carbide and that part is coming first.

And then second, the down conversion from 800-volt DC all the way to 1 volt where the switching of the CPU and GPU is happening, the gallium nitride part that will come later. And I think the industry, especially the capital market has overestimated how fast that a transition is going. And I think we gradually see now the transition starting on the silicon carbide side and then with a bit of delay coming on also on the gallium nitride side. It's a massive change in architecture and this takes time.

Okay. But would you say that gallium nitride is still coming early '27 or '27 at least?

As we said in the prepared remarks, it's too early to predict and to forecast. We see a very clear technical discussions, and we see around the trade shows like Compitex and discussions that there is a hot momentum ongoing. We have not seen that yet translate into orders. And on our side, we have -- we don't have visibility and indications yet. But we have no doubt that it is coming. I hope you could see that also from my remarks.

Yes. And just the last question, just to clarify. If the orders, let's say, would come in early '27, would you know already by now? Or is it normal that sometimes you don't have this kind of visibility and you know by it a bit later?

I think it would be normal that we get notification relatively short term. Please recall that our systems are the tools of record across the whole market for the gallium nitride power. So all the development work, all the pilot work, all the sampling and preproduction work of all the customers has been done on our systems, either the older Series G5+ or the newer Series G10 GaN. So all the customers work with our system. which is they don't need us now to support in the preparation. And at some point, I expect the customers to come along with multi-dozen tool orders and say how fast can you ship please?

Next question is from Gustav Froberg from Berenberg.

Two and half, if I may. Firstly, just on OXO and the orders in Q2, very strong, obviously. But -- is it correct to assume that the rest of the business is kind of stagnated or declined a little bit in Q2 versus Q1? Or another way of asking it, do you have a sort of base level view of orders, excluding the extra order intake you're getting from the up cycle in Opto at the moment that we can think about as a sort of sales level run rate, if you like? That's my first question.

And then second, last quarter, we talked about 80 to 120 tools needed for the Optomarket per year. It seems you're pretty much already there for H1. So I want to know in a little bit more detail if this preliminary view that you gave in Q1 has changed or if you're simply hitting that number a little bit sooner than you thought you would?

Very two very good questions, Gustav. So I think first question on the remainder of the business. I think as we have indicated, the power business continues to be stable at a low level. Silicon carbide pretty slow because of the underutilization and overcapacity still in the market. The gallium nitride, here and there are some orders, but still on a lower level before the next wave of investment kicks in, as we just discussed with a longer question of margin. So I think these -- the 2 power segments are stable on a low level. And at the same time, as I've indicated, we expect for the next quarter the Opto momentum to continue to be strong and upbeat. We see that from the pipelines and the inquiry pipelines and sometimes in many cases, with the big guys, big customers also very clear forecast. So we have a very healthy confidence on the Opto momentum for the coming quarters.

Now to your question about the tools, EUR 80 million to EUR 120 million, that was our initial forecast. We currently see a stronger momentum than that. It's too early to predict where exactly that's coming from. It could be that the overall wave is bigger than we have expected. Let's see how it continues, but we are currently seeing a bit stronger momentum than this EUR 80 million to EUR 120 million. Very well.

The next question, Bank of America, Oliver Wong.

First question is on capacity. So I understand that, I guess, if you take the midpoint of full year guide and the Q3 guide, let's say, Q4 at $220 million. And if we were to -- if this was to be stable going forward into next year, that would imply, let's say, $880 million in full year capacity. Is that the right way to think about it?

So you should consider capacity not as a limiting factor. Please recall the equipment business is relatively CapEx light. So for us, the situation is very different from that from our customers, right? Our customers need a stab in the clean room. The customers need our equipment, luckily. And for them, it's always they need, all the CapEx, all the preparation with multiple months, if not several quarters of preparation time to create capacity. For us, on our hand, we are relatively CapEx light, as you know. And we have a decent flexibility also to shift some of the assembly tasks. If you recall, what we do at AIXTRON is mostly assembly and test in our own premises, in our own clean room environment. And we have a big flexibility in times of upside swings as we are now realizing to put some more of the work to our suppliers.

Let's say, the suppliers in demand peak times then take some more additional assembly steps and we get, for example, a tool partially half assembled already when it comes to our shop floor means the tool is then not standing, whatever I put arbitrary numbers, not taking whatever 10 weeks on our shop floor, but rather taking 6 weeks on our shop floor because 6 weeks of the work has been already in an outsourced assembly space, just to shine light on why this is such. So therefore, capacities are not a limiting factor to us in.

Got it. But my understanding is lack of the full year raise in spite of the strong orders, that was I guess, somewhat related to some sort of supplier capacity constraint. Is that correct or not?

It's a ramp rate. It's not a capacity. It's a good one. Good question to clarify that. Thanks for following up on that one. It's a ramping rate constraint. It's not a capacity, right? It's the question within the given machines, premises, buildings, clean room space, what I consider capacity is how fast can we hire people on board, how fast can our suppliers hire people to operate the machines, which are then being used to manufacture the parts. So it's a purely limitation of the speed, how fast we can accelerate. So think about your car going on a highway, you have no VAC, but the question, how long does it take you to reach your travel speed.

Yes. Yes, makes sense. So I guess like we can kind of run rate, let's say, 220 million Q4.

Yes, no problem. Once you are at travel speed, you can choose easily and comfortably...

Next question is from Martin Jungfleisch, Paribas.

First one on gross margin. I think the implied gross margin for the second half is around 46%. What is your view on phasing here? Should we expect more like 44%, 45% gross margin in the third quarter, maybe up to 47% in the fourth quarter? And with that gross margin level of around 46%, let's say, in the second half, is that sort of level you're kind of comfortable with for next year? Or would the Malaysian facility ramp have any impact on that in the short term?

So honestly, we haven't done the math on the details. But what we clearly see is that the laser systems carry, of course, a bit of a better margin profile, simply coming from the fact that is highly, highly, highly complex systems with a lot of -- and all our customers have their own differentiating features that their orders. They very differentiated recipes. So the laser systems are -- I think laser epi is some of the most complicated stuff that customers make on our tools. And given from these technical requirements, of course, that translates then into a bit healthier margin profile. And from that onwards, I would also expect -- or we would also expect in '27 to clearly see an uptick in terms of the gross margins, as you rightfully said. So it will be a good year '27 that we can already see ahead of us.

Okay. But is there any ramp costs for the Malaysian facilities that you're seeing in 2027? And if that is it meaningful?

There's some ramp costs, of course. There's also some fixed cost digression effects. The first order considered is the one.

Okay. And then the second one is really on Opto. Can you disclose what kind of configurations you're mainly shipping today? Is that still mainly 4-inch? And is it increasing 6-inch? And then also, have you received any data on how the epi yields at your customers are progressing in both 4- and 6-inch.

So majority of system is shipping on 4-inch. I think probably 70%, 80% is on 4-inch, smaller one on 6-inch. All customers, of course, want to be ready for 6-inch for a conversion at a later point in time. And I think the yields that our customers achieve are probably some of the best kept secrets, I could imagine, and we don't have a transparency into those. I wish I had both.

Our next question is from [indiscernible].

One more on orders. There was still, I think, around EUR 95 million to be secured as of H1, and I think you secured those in the meantime. Still, let's say, if, let's say, ramp speed is constraint and not, let's say, the customer willingness to buy, let's say, why, let's say, the hesitance of customers to, let's say, fill the remaining slots for this year?

Well, a quarter is always an arbitrary cutoff line, right? I think as you rightfully stated, quite a decent amount of orders has already been received. We are already now 1 month into the quarter. And also, we are, at this point in time from the big guys also receiving multiple multi-tool orders. So it is filling quite fast. So no concern on that one. Order coverage is good, and also we are well on track.

Understood. And maybe on order backlog, I think there should be something like EUR 220 million now in already for next year. You mentioned Opto is quite diversified, talking about around 15 customers. Would you be able to give like a rough indication what the biggest customers make up in that order backlog?

Honestly, I don't have the data in front of me. That's a very, very specific question. But maybe let me shine a little bit light on the order backlog. Yes, there is some going into '27. There's already some going into '28 also to give a light on there. Also some part of the order backlog still relates to some power electronics where customers shifted something out. It's not all motor electronics. There's also some small portion of gallium nitride left, small portion of silicon carbide left, where some customers had placed orders. And then at some point, when the power wave collapsed, they didn't cancel or we didn't let them potential. Both cases happened. And they just say, look, AIXTRON, at some point, the market is coming back and then we will call the tools. So it's a small portion also of power.

Understood. And then one more on gallium nitride, I think there's something in the presentation saying that the technological outperformance was confirmed. Maybe just to understand how that technological outperformance is, let's say, defined and where you stand with that tool in the qualification processes?

Honestly, I don't get your question. What do you mean by technical outperformance? Which part do you refer to? I'm not fully understanding your question. Are you still there?

Yes, yes, I'm still there. So there is one bullet point actually in your presentation under your Hyperion tool, which says technological outperformance versus 200-millimeter platform confirmed by a leading customer. So I was interested in how you define that technological outperformance. This is not my phrasing. This is actually from your presentation.

No, I get it. That was important to point to the right point. So you're not referring to the update we have, but to some of the slides later on. Martin, help me find it. So what we mean is we are able at the 300-millimeter tool is a single wafer we are able compared to the 200-millimeter tool, which is a batch tool, as you can expect from the single wafer tool where all metrology measurements, temperature optimization target towards a single wafer, all the effort kind of to 1 wafer instead of spread over 5 wafers, we are able to achieve an even better uniformity particle performance in all the performance topics that you expect from an epi. And we see that customers are very happy with that because it allows them to realize the next generation of gallium nitride devices. And yes, we get very, very positive feedback from customers. And can you share with us, let's say, how many qualification processes or with how many customers you're currently running with the tool? More than a handful.

The next question is from Craig Mcdowell from JPMorgan.

My first question is on Opto orders. And I think the Opto order for Q2 is around EUR 160 million. And I understand that you do not recognize Chinese orders with a corresponding export license. I was wondering whether you could give a color on what is the sort of additional Chinese ordering in Opto on top of that EUR 160 million? Is it an additional 20%, 30%, 40% on top that's not recognized in that Q2 optical order number?

A minor topic. A minor topic. So export licenses are no longer an issue. We get them relatively quick these days. So in terms of when we receive the customer order and the export license timing delay is marginal. So not a significant number.

Okay. Understood. And my second question, just on the gallium nitride business. I look back and I think it was Q3 '24, so nearly 2 years ago that you first indicated a GaN market maybe 3x in the medium term, indicating around EUR 450 million of total revenue. I mean that was 2 years ago. I'm just wondering whether that view is still valid. I understand that orders aren't yet on the books and but whether your confidence has grown on that or do you think that forecast is still valid?

That view fully holds. We can confirm. So on the midterm strategic value of the gallium nitride, we are fully convinced. So that holds, that's fully intact. It's just a question now when is this next wave for the gallium nitride really starting after we see utilization rates increasing. And at some point then installed capacity is fully utilized and then the next wave and the next momentum is starting.

And just a follow-up on the utilization rate. I think previously, you're talking about utilization in your GaN installed base around 60% to 70%. Is that still an accurate number? Or is it edged up towards 70%...

I think it's gradually increasing, not a major shift as of now. So I wouldn't expect this is now 90 or so.

The next question is from Om Bakhda, Jefferies.

Just had a question on capacity and pricing. So if we were to look ahead, and I know that at the moment, you're not getting any signals on a GaN inflection. But if GaN was to inflect along with Opto within the next 12 months and then you were able to deliver a quarterly order run rate of above EUR 250 million. And then in that case, within 2027, you'd be fully utilized on that EUR 1 billion for the year. Do you see there could potentially be scope in pricing if you become fully utilized in 2027?

Well, as I mentioned again, right, you shouldn't think about capacity limits. I would be very happy to get additional orders, and I would then simply just ship more. So I'm not concerned of hitting any ceiling and then selling off the remaining slots at a premium, as I understand your question a bit. I'd rather say I take all the orders and ship all the orders and make sure that I fully benefit from the upswing. That's part of the strategy.

Got it. Perfect. And then just on gross margin evolution sort of over the medium term. If we assume that at some point within the next 1 to 2 years, both Opto and GaNI [indiscernible] were to be very strong together and again, above EUR 250 million in revenues per quarter. And then again, now as customers are moving to higher down wafer sizes on their production line and so then the G10 tool becomes increasingly part of the mix, could we see your gross margins trending towards 50% over the sort of the mid- to long term?

You're getting very bullish here. First of all, you get very bullish on the revenue side about projecting 250 million per quarter. So clear message from my side, we have no capacity limit. I really want to say that, yes. So from the capacity standpoint, I'm not concerned, also not concerned with your number. I think I've highlighted this several times in the call. However, I also wanted to clearly say that we don't see any signs as of today to go to 250 per quarter, right? I think to really reiterate, we have a good confidence to see around 200 per quarter for the next few quarters. And then let's see what that happens to really clarify out the assumptions because the numbers are gradually getting bigger and the longer the call continues.

So now about -- regarding your question with 50% on the gross margin, please recall that we are today at a 42% -- please also recall that getting a percentage point of gross margin is typically quite a decent amount of work. It just doesn't go so fast. There's many, many elements behind it. So what you're asking for is a very big step. And, I don't know. This is quite a long way to go. So I would say from the 42% where we are, it's -- we will clearly see a couple of percentage points in next year overall in gross margin, that's going to help us. So I think the direction of gross margin increasing is clearly there. 50%, I think it's clearly too far, maybe somewhere in between. Let's see where exactly we end up. And also we are not there yet, so we don't have detailed numbers, but just to give you an indication.

And then, of course, please also take into account, I see that you're trying to build the numbers as we continue to grow as a company, we will also then, of course, pull up the OpEx a bit. We are a very innovative company. We obviously have many ideas on the R&D side. And by investing in R&D, it was always good for us because the R&D investments we made, I think, in the past have been quite pointed. And I think all the R&D investments we've made have realized or led very nicely a few years later to a nice revenue increase and overall company growth, value growth for our shareholders as an ultimate goal.

So if we have a more financial freedom, we, of course, will also continue to increase our R&D rate, and it will allow us to pull in some projects and to realize some ideas because as you can see compound semiconductors are really at a very, very sweet spot now the properties, optical properties are great for the lasers for communications, and we need ever more communications with more data on the AI. MicroLED is coming back. There are some other optical applications alerting on the horizon, power electronics and so on and so forth. So the stronghold and the sweet spot we are in, we will continue to grow that. And I think we have demonstrated in multiple times that with our innovation ability, we can shape and we can create new markets, and we will continue to do that.

The next question is from [indiscernible].

A question regarding your Opto business. There were recently speculations that co-packaged optics, and TSMC Scoop platform, there might be delays for the NVIDIA platform and co-packaged optics potentially rather for 2029 and that NVIDIA might take a similar approach like Google using more optical circuit switching as a transition step, so to say. What would that mean for your business in Opto.

Honestly, I think you on a detailed level for the different architectures, we do not see that those optical architecture nuances affect our demand. So what we rather see is that our customers are asking us to ship the tools to provide the capability. And in the end, the tools and the wafers made on our tools can then be utilized in all sorts of different architectures, whether this is a pluggable optics, whether this is a co-packaged optics or at a stage even after that photonic integrated circuits, the PIs, whether this is a GPU to GPU rack-to-rack communications and some customers are even experimenting that's probably coming more towards the end of the decade to go to optical interconnect between the GPU and the high-bandwidth memory.

So I think the overall strategic direction of the strategic trend is very clear, which is more optical communication, more data and going from long distance to medium distance to short distances to very short distances. So on our side, we do not look into the details of the architecture that our customers are using, but I'm not concerned that any of these architecture details that clearly our customers have to take into account have major impact on us because our tools are very versatile and the customers can use them for each of these architectures. So we are very robust in respect to that.

So got it. So an OTF type of architecture would not be negative for you, at least not at this stage. And then a second question regarding the GaN opportunity. I mean you mentioned that you are quite confident that this will ramp in volumes. My understanding is the larger GaN opportunity in terms of number of guys is rather on the board level, right, when voltage regulators shifting to GaN or really the server board and then the GPUs and the compute. Would you assume that with the 800-volt DC architecture that the board level voltage regulators would be fully based on GaN? Or are you assuming a kind of mix that it will be also silicon and still not fully GaN? What is your take there?

A very good question. So first of all, the wafer area scales inverse with voltage means the lower the voltage, the more area is needed, wafer area, die area that simply, right? -- means for a given power, the lower the voltage, the higher the current and the higher the current, the more parallel devices I need to split the current over. So that's just basic physics. That's a good thing. So it will not be changed and also cannot be changed by any architecture relating to your first question on the opto. It's just what it is. So yes, the lower the voltage, the more diodes are needed.

Now to the second part of your question, what we get back as signals from our customers is that they say, look, we will convert from silicon, the old architecture is fully silicon from the big overland line and cable all the way down to the CPU and GPU by silicon. And the customer signal us we will start from the highest voltages first at the entry point of the data center and the silicon carbide. I mentioned that in an earlier question here in this call, then in the next step going from the 800-volt DC to 200-volt, 100-volt intermediate voltages in the gallium nitride and then the 48, 12 volt, 6 volt, 1 volt step by step.

So -- and that is very nice because it means that we look once the conversion to gallium nitride has started to a multiyear growth opportunity, likely starting somewhere on the 650 volt, then extending to the 100 volt and later on then going down to the 20-volt, 12-volt, 6-volt devices. And the penetration speed of gallium nitride displacing and pushing out silicon will likely depend, first of all, on the cost down curves that our customers are able to realize. The customers are working heavily on die-side shrinks. They work on devices, which are 30% smaller. Our customers are working on devices, which is 50% smaller so that out of a given wafer, you can get more dies and the die get cheaper. And then always, this is how the semiconductor industry works. As costs go down, your market grows bigger because you can displace another technology, in this case, silicon. This is roughly the perspective that we have on this market.

Got it. And then really a very quick one on the down payments. I'm not sure if I have the right number for Q1, but my impression is down payments have shifted quite in second quarter. And my understanding was you require from Chinese customers a much higher down payment, right, for the orders they are placing. So would that imply that in Q2, you had a decent or a high number of orders in Opto placed from Chinese customers? Or would that be...

No, no, this is not -- there's a misconception. So we don't request a higher down payment from Chinese customers. So what -- the down payments and the payment terms vary strongly application segment by application segment and region by region. And now given that from -- especially now with the Opto boom from many customers, we have received multi-tool orders stretching over an extended period of time. we've been able to ask quite a number of our customers for a higher percentage of down payment, but this extends all across the globe.

The next question is from Mr. Nigel Van Putten from Morgan Stanley.

I have a question, a bit of a 2-parter. From your commentary, I get a sense that historically, first quarter has been 10% to 20% of revenue. There's no real expectation that will continue. So would it be fair to say that directionally, we should look at the first quarter and '27 as sort of a sequential increase relative to the fourth quarter or at least at a similar level, not necessarily a big drop as we've seen before?

And the second part of that question would be, I can imagine that gross margins benefit from more stable or more predictable loading. So is that a major driver when you say '27 gross margins look good? I mean, on the other hand, there's more G1 in the mix. So just trying to get a sense of what are the drivers of the gross margin into next year.

I think both good assumptions. So historically, for everyone here on the call, the Q1 from the last, I think, 3, 4 years has been always a very, very slow start of the year, and then we've always finished at a super high record towards the end of the year. Looking towards '27, I mean, we can't see the full year of '27 yet, that's clear. But we have a visibility also now in the beginning of '27, of course, and we expect a very strong Q1, probably somewhere on the level of what we can see right now for the Q4, I would say.

Now to your second question on the gross margin, as mentioned and discussed a bit earlier, right, we expect clearly an uptick from the level where we stand, clearly not the 50% that we had earlier. So please, that number is, I think that will take a bit of work. I think it will be a mix of things coming together. As discussed, product mix will help, right? The laser systems being more complex, more sophisticated, hence, carrying a bit of a better margin, a bit of fixed cost digression. On the other hand, again, some ramp costs. And with all these effects together, we expect a decent uptick on the gross margin. But it's very difficult at this point in time now to tie that and split it out in the exact detailed cost.

Understand. Maybe a quick question, switching gears a little bit. You've been talking, obviously, CW, and the other high-power lasers are clearly the big opportunity. But what about VCSEL? Do you see any demand for your -- from customers for those tools? And my understanding is there's quite a bit of capacity built already. So if we see, let's say, a VCSEL growth as well for some applications, would that potentially benefit you? Or -- and what would be sort of -- I can imagine that it could be a benefit, but relative to the more high-power side of lasers, it's probably much smaller. Is that a fair assumption to make?

Very good question. So we see, in fact, there's some -- most people are working now on the traditional devices that we discussed on most of the call. VCSEL, some customers work on VCSEL for the communication as an alternative because you can just take an optical fiber and pump it onto the VCSEL, right, optic fiber and then collect the light of it. Some customers are working on that on the gallium arsenide systems. So VCSELs like we use in the same type of systems that we use for the 3D sensing a few years back. Interestingly, some customers are also experimenting with VCSELs on the gallium nitride systems. We talk about blue LEDs, if you want to say so, and see whether they can use that for doing die-to-die communications in heterogeneous integration, for example, to couple a GPU for high-bandwidth memory on a silicon carbide interposer and see whether they can replace copper interconnect by optical interconnect.

So I would say, at this stage, this is still research, but it gives us a good level of confidence that the trend of optical communication not only is a way for now 226,7,8, but rather also has some technical ideas, new innovation ideas, but probably are kicking in towards the end of it.

With that, we have the last question from Adithya Metuku, HSBC.

Actually, my last question was on VCSELs about substituting indium phosphide in certain use cases, but that's been answered. So thank you.

Perfect. Then it sounds we are done. Apologies from our side for this little hiccup. We all had to wait for 15 minutes. I hope we could make that up with a good discussion with pleasure.

Yes. Thank you all from my side as well. Most of you, I guess, will be going into the summer break. We will be starting hitting the road end of August, a lot of conferences with different participants. So hope we meet a lot of you before we meet again for our Q3 results, which will be out end of October. Until then, have a great summer break, and hope to see you soon. Thank you, and goodbye. Bye.

Ladies and gentlemen, welcome to the AIXTRON's Analyst Conference Call Q1 2026. Please note that today's call is being recorded.

Let me now hand over to Mr. Christian Ludwig, Vice President, Investor Relations and Corporate Communications at AIXTRON for opening remarks and introductions.

Thank you, Anna. A warm welcome to AIXTRON's Q1 2026 Results Call. My name is Christian Ludwig. I am the VP of Investor Relations and Corporate Communications at AIXTRON.

With me in the room today are our CEO, Dr. Felix Grawert; and our CFO, Dr. Christian Danninger, who will guide you through today's presentation and then take your questions. This call is being recorded by AIXTRON and is considered copyright material. As such, it cannot be recorded or rebroadcast without permission. Your participation in this call implies your consent to this recording.

Please take note of the disclaimer that you find on Page 2 of the presentation document as it applies throughout the conference call. This call is not being broadcast via webcast or any other medium. However, we will make a transcript available on our website after the call.

I would now like to hand you over to our CEO for his opening remarks. Felix, the floor is yours.

Thank you, Christian. Let me also welcome you all to our Q1 '26 results presentation. I will start with an overview of the highlights of the quarter and then hand over to our CFO, Christian, for more details on our financial figures. Finally, I will give you an update on the development of our business and guidance.

Let me start by giving you an update on the key business development from the first quarter on Slide 2. The important messages for Q1 are strong new orders of EUR 171 million, driven by optoelectronics. Revenues of only EUR 59 million due to seasonality, but also limited demand for both power electronics segments. Nevertheless, this was expected as revenues came within our guided range of EUR 65 million plus/minus EUR 10 million.

The gross margin came out at 18%, impacted by negative operating leverage due to low volume as well as a mid-single-digit euro million one-off cost item related to our announced personnel reduction. Despite the weaker start into the year, we confirm our new fiscal year guidance released mid-April with revenues of EUR 560 million plus/minus EUR 30 million before EUR 520 million plus/minus EUR 30 million.

We announced a new greenfield site in Malaysia. With investments of approximately EUR 40 million in the next years, we will be ready for production by middle of '27. Also, after the reporting date, we successfully placed our first convertible bond in April at very attractive terms, enhancing our financial flexibility.

Christian will now provide a detailed look into our financials on the following pages before I take over with an update on our markets. Christian?

Thanks, Felix, and hello to everyone. Let me start with the highlights of our revenue development on Slide 4. We had a weak quarter due to seasonal effects with revenues at EUR 59 million, down 47% compared to the EUR 113 million last year. But this was not unexpected as we are fully in line with the quarterly guidance of EUR 65 million, plus/minus EUR 10 million.

The breakdown per application shows that 52% of equipment revenues come from optoelectronics, 31% from LED/microLED and 17% from GaN and SiC power, and a 1% contribution from R&D tools. The aftersales business contributed to a total revenues with EUR 24 million. As this business is much more stable, it only declined by 4% year-over-year, thus its share of revenues increased to 40%, up from 22% a year ago.

Now, let's take a closer look at the financial KPIs of the income statement on Slide 5. I already talked about the revenue line. Gross profit in Q1 '26 was EUR 11 million. That implies that our gross margin in Q1 decreased by 12 percentage points versus Q1 2025 to 18%. But please recall, this is mainly due to 2 factors: the negative operating leverage of the revenue decline year-over-year and the one-off expense of a mid-single-digit million euro amount in connection with the announced personnel reduction in the operations area, both burden gross profit.

OpEx in the quarter went up 7% to EUR 33 million, primarily driven by higher R&D spending compared to the previous year. The increase in R&D cost is mainly due to higher depreciation and higher material costs. For the full year, we expect R&D costs to be about EUR 10 million higher than in 2025. EBIT for the quarter was at negative EUR 22 million. The decline in the operating result compared to the previous year is mainly due to the lower gross profit, as already discussed.

Now to our key balance sheet indicators on Slide 6. Working capital was down by more than EUR 76 million since end of 2025. Several balance sheet items contributed here. After the strong revenue number in Q4 last year, we now collected the outstanding payments and turned trade receivables into cash. At the end of March, trade receivables were down to EUR 84 million compared to EUR 131 million at the end of 2025.

Advanced payments received from customers at quarter end were EUR 79 million, up about EUR 35 million from end of 2025, primarily driven by the increase in order intake. Advanced payments now represent about 22% of order backlog. Inventory levels were at EUR 295 million, slightly up from the EUR 284 million at the end of fiscal 2025. Trade payables have stabilized at EUR 36 million, slightly up from EUR 34 million at the end of '25.

Adding it all up, our operating cash flow improved in Q1 to EUR 54 million, up by more than EUR 18 million versus last year's EUR 35 million. On the back of the improvement in operating cash flow, free cash flow also improved, came in at EUR 49 million compared to EUR 30 million last year. CapEx in Q1 was fairly stable year-over-year at around EUR 5 million.

For the full year, we expect about EUR 55 million of CapEx. This consists of our baseline investments, plus about 2/3 of the announced EUR 40 million for the Malaysia expansion. At the same time, we have started the process of selling our site in Italy. Our cash balance, including other current financial assets as of March 31, 2026, increased accordingly to EUR 270 million compared to EUR 225 million at the end of December 2025. Our equity ratio continues to be at a very healthy level of 85%.

Now, an update to the convertible bonds. Subsequent to the reporting date, we successfully completed the issuance of our inaugural convertible bonds in mid-April. We placed EUR 450 million with a 5-year maturity and a 0% coupon. This transaction is accretive to earnings from day 1 as the proceeds can currently be invested in low-risk money market instruments at yields of around 2%. We intend to use the proceeds for general corporate purposes, which may include investments into organic growth, potential M&A and also, when appropriate, share buybacks.

With that, let me hand you back over to Felix.

Thank you, Christian. I would like to continue on Slide 8, and give you an update on key trends in our different markets. I will start with optoelectronics, which clearly was the most dynamic and decisive market for AIXTRON in the first quarter before we briefly touch GaN, SiC and LED/micro LED.

In optoelectronics, Q1 marked a clear inflection point for the company. We saw exceptionally strong order momentum, with optoelectronics accounting for well over 2/3 of total order intake during the quarter. This performance was primarily driven by laser-related datacom applications, where customer demand exceeded our expectations. The catalyst for this is a fundamental shift in the data communication architecture of AI data centers.

Until now, data in these systems have largely been transmitted via copper cables and only sporadically via optical connections and if so with comparatively low data rates, such as 100 gig. Now the industry is making a technological leap. In the future, optical connections with highest data rates, namely 800 gig and eventually 1,600 gig will increasingly be used there. This is not a gradual advancement, but a genuine architectural shift, and it is precisely this shift that is leading our customers in optoelectronics to massively expand their capacities.

A huge number of additional wafer starts are required for this new type of connectivity, and this, in turn, triggers significant investments in new and expanded manufacturing capacities. Recent multibillion-dollar investments made by NVIDIA and other ecosystem players are accelerating capacity investments across the datacom value chain. This, in turn, is enabling laser manufacturers to expand their epitaxy capabilities directly benefiting AIXTRON. Therefore, we expect this wave of investment to continue over several years, extending well into '27 and beyond.

In the first step, copper connections in the AI data centers will increasingly be replaced by fiber optics and slower optical connections get replaced by faster ones, the so-called scale up. In the next step, the industry plans to significantly increase the number of connections within these data centers once again to enable even more powerful and intelligent AI systems, the so-called scale out.

And finally, data traffic between data centers will also increase sharply the so-called scale across. This too will require additional optical connections and more lasers. At product level, our G10 ASP platform continues to see very strong market acceptance and represented the majority of volume orders received in Q1. Customers are ramping aggressively to support the growing demand for EML and SiC-based lasers used in AI-driven data center infrastructure.

From a technology perspective, most systems ordered today are initially configured for 4-inch wafers, but are already equipped with 6-inch conversion kits. This enables customers to transition rapidly as soon as larger wafers become available. Datacom lasers are a performance-driven application where yield is absolutely critical. These devices require multiple architectural growth steps, meaning that even marginal improvement at wafer level can have significant multiplied impact on final yields. This is where we clearly differentiate.

Our system controls uniformity at wafer level rather than at batch level, enabling consistently superior results layer by layer, a key reason why customers increasingly standardize on our tools for advanced photonic devices. Overall, with many multi-tool orders extending beyond '26, we are confident that we are seeing the beginning of a new structural growth phase in optoelectronics.

Now, let me turn to gallium nitride. In GaN, demand remained stable but soft in the first quarter, contributing just under 10% of total order intake. While near-term growth remains limited, we see customer utilization gradually improving, and we expect demand to pick up again at some point. This may be later in the year or in the first half of the next year as fab utilization increases.

Strategically, AIXTRON remains well positioned across 150 and 200 millimeters and increasingly also 300-millimeter gallium nitride, with systems installed at multiple customers. In silicon carbide, market conditions remained soft in Q1, reflecting ongoing underutilization across the industry. Customer utilization rates in silicon carbide, however, are gradually improving. When exactly the increasing utilization converts into additional volume demand is too early to predict as of now.

Importantly, industry analyst data confirms that AIXTRON tools remained the #1 choice in silicon carbide epitaxy during 2025, even in a difficult market environment. This once again underlines that during periods of underutilization, customers prioritize tools with the lowest cost of ownership.

Looking ahead, new device architectures such as super junction silicon carbide are expected to significantly increase epitaxial complexity and layer count, a trend that clearly plays to the strength of our batch reactor technology. Finally, LED and microLED applications had no material impact on order intake in Q1 '26. Investment activity remains cyclical and at low levels, with spending largely driven by Chinese end users primarily focused on AR and smart glass applications.

Several customers are also working on devices for datacom short-link applications, and we may see incremental orders later in the year or in '27, depending on the success of this technology. During the first quarter of '26, we also took an important strategic step to further strengthen AIXTRON's global footprint by deciding to establish a new manufacturing facility in Malaysia.

The site in the Penang region will focus on assembly and testing of selected 100, 150 and 200-millimeter systems for our Asian customers and will allow us to leverage a highly developed local semiconductor ecosystem. This investment of around EUR 40 million, with start of operation planned for '27, enhances our competitiveness, improve supply chain resilience and brings us closer to our customers in Asia, while fully maintaining our strong R&D and production base in Europe. Importantly, this move has no impact on our financial guidance for '26.

With that, let me now move to our guidance. We confirm our increased guidance for '26 as published mid-April. We expect revenues to come in at EUR 560 million in a range of plus/minus EUR 30 million. At the midpoint, this is an 8% increase to our previous guidance. We expect gross margin of about 42%, and an EBIT margin between 17% and 20%. The guidance for the gross margin and EBIT margin includes one-off expenses of a mid-single-digit million euro amount in relation to the announced personnel reductions in the operations area. The measure will lead to an annualized savings of a similar magnitude in the future.

For Q1 '26, we expect revenues of EUR 110 million in the range of plus/minus [ EUR 10 million ]. Again, as stated before, the implications of the geostrategic turmoil are unclear. We continue to closely monitor the developments worldwide in the Middle East and particularly in connection with the conflict in Iran. Potential impact on energy prices, supply chain, financial markets as well as investment and demand behavior are continuously analyzed. If necessary, we will respond appropriately to risks that could negatively affect our business performance.

And with that, I will pass it back to Christian before we take questions. Christian?

Thank you very much, Felix. Thank you very much, Christian. Anna, we will now take questions, please. And for fairness reasons, I would like to ask everybody to limit himself or herself to 2 questions in the first round so that we get a chance to get everybody who wants to ask a question also to ask his question. Thank you.

[Operator Instructions] There are a few questions already incoming. The first question is from Martin Marandon-Carlhian, ODDO BHF.

My first one is on opto. I understand that you have a very good visibility, and you already talked about visibility beyond '26. But I think at the last earnings call, you did not fully commit to growth on that division in '27, considering the very high comps in '26. So, I guess maybe it has changed. But also more broadly, when customers like Lumentum or Coherent talk about more or less doubling capacity again next year, is there any reason why that will not translate proportionally in terms of orders due to improving yields or the transition towards 6-inch wafers, for instance? And I have a follow-up.

Yes. Thank you very much for the question, Martin. In fact -- so the transparency from customers now through '26 and also '27 significantly has improved. I think the key point was the announcement of NVIDIA to make those investments into Coherent and Lumentum, which we saw early March. And since then, literally, many of our opto customers, not only the 2 names, but also the other big names have been in very close contact with us with orders, multi-tool orders, multi-tool orders for '26, also multi-tool orders extending well into 2027 and others sharing forecasts in '26 or across '26 and '27.

So the transparency, as you're assuming has, in fact, significantly increased over the last, I think, literally 8 weeks. And it really started only early in March. And now as we have this transparency, we have, as you hear from us, a high confidence for the '26. We've been able, thanks to the booming market on the opto side to increase our guidance. The power market remains soft, as I just highlighted, but the opto market is really increasing. And that also gives us at least a very positive feeling for '27 due to customer forecast and from customers already, literally orders going into '27.

Okay. And I know you don't give precise forecast or anything, but is it fair at this stage to assume growth in '27? Or is it maybe a bit too early to say?

You're asking very early about the question. I think we're seeing a good momentum of optoelectronics extending well into 2027. We can clearly say that. So, that leads me to believe that we have already a solid base for '27 from the optoelectronics. And as mentioned earlier, at some point, but we -- it's too early now to predict the exact timing.

At some point, we expect the market for GaN power and silicon carbide power to come back. We don't know exactly when this is happening, whether this is at the end of '26, with revenues and orders in '27 or early '27. But I think at least there's a good probability that the power market is also coming back in '27. And if the power market -- if the power market comes on top in '27, if that really does materialize, then we would see a growth in '27 on top of it.

Okay. And the second question is on the convertible bonds. I mean, can you give maybe a bit more color on what do you think you would do with most of the EUR 450 million? You talked about M&A. So, what do you think could be the ideal target? What can it look like for AIXTRON? Is it to complement your existing MOCVD market or potentially go into different materials and so on? And also when you talk about share buyback, do you aim for this to be quite a regular policy? Or do you anticipate to be more opportunistic about it?

Martin, very good questions. To be honest, that wasn't really planned, the convertible. We just saw a really great opportunity over the last weeks that we couldn't resist. We all lived together through heavy volatility over the last years, which we not always liked. But now we have been able to leverage that volatility for a really great deal. So, I mean, overall, that gives us additional financial flexibility for basically no cost and direct earnings is directly accretive to earnings. And that's basically what it is, yes.

So it was just an opportunity, a ball laying there that we had to kick into the goal, to be honest. So, there's not too much more behind it. So, there's no direct other opportunities that we are pursuing here. On the share buyback, maybe to complete that, I mean, that is something that might become an issue -- a topic in the future, probably not at the share price level right now.

The next question is from Gustav Froberg from Berenberg.

I'm just trying to understand the magnitude of the opportunity on the equipment side in opto a bit better. And I wanted to ask, do you have a sense of the installed base of AIXTRON machines used in data center and optical networking applications today, if you completely exclude anything that goes into consumer VCSEL applications? That's my first question.

Honestly, that's a very difficult question. And also, I don't know whether that question is helpful if you're trying to triangulate something and correlate that to demand. Let me explain to you why. The optoelectronics market is a market, which is characterized by a very high technical complexity, many layers being done. And there is many, many, many old systems of AIXTRON out there. Some systems are running at customer sites for something like 30 years because the customers, if you think about optical communications for like an undersea cable across the ocean, very long qualification and a huge hesitance by customers to change.

So, there's many tools out there, which may be just running, let's say, once a week for like a day or something like that, just to produce a few devices to keep a certain design for a certain customer running. So the installed base of tools, which is out there is not really an indicator about what could be filled up. I think a certain part of the installed capacity, I would say, of the more recently installed capacity, let's say, capacity from the, I don't know, last 5 years, like last 8 years may now be filled and may be utilized for the current volume and for the current boom. But I think systems, there may be many, many, many, many, many systems from the last 20, may not even 30 years out there, but they will not be used now for the newer generations of tools.

Also, on the laser devices, the complexity of designs has significantly increased. Sophistication has increased. So, I would say, yes, there is a certain set of tools, which now still be used and filled in terms of capacity but also what I hear from customers in terms of demand, I would say, a very fast order approximation, a lot of the capacity which needs to be brought online needs to be also newly installed. Sorry for the long answer. I try to give you a bit the background of what's going on.

No, that's super helpful. And then just the last follow-up for me. Based on what you know today, how much do you think the tool and equipment market needs to grow in order to meet the upcoming demand for the next couple of years? Are we talking about a doubling of the number of equipment needed, a tripling of the number of equipment needed? Just anything ballpark-wise would be very helpful to understand.

That's a very good question. I think in the end, it depends. It's unfortunately, in our area, always a bit difficult to say, right? Because there's questions of yield of wafer size and many other variables, right, which go well beyond what we can influence or what we do influence directly with our equipment, but many elements which lie more in the hands of our customers and their processes and process flows after our tools and things which are not in our hands, we can only take guesses.

I would say, if you assume a market on the optoelectronics side somewhere in the range, my first guess, 80 to 100 tools for the laser communications, I think that might be a good approximation of order. And maybe it's 1/3 less or 1/3 more, but just to give you a house number.

But 80 to 100 needs to be installed or it's already installed?

It needs to be installed every year.

Next question is from Adithya Metuku, HSBC.

Firstly, just a clarification on the previous answer you gave, 80 to 100 tools that need to be installed. Are you talking about G10 tools? Or are you -- are the G4 equivalents? If you could just clarify that, that would be helpful. And then I've got a question.

I was talking about G10.

G10. Okay. So essentially, give or take, EUR 4 million. So that alone, you're looking at EUR 300 million to EUR 400 million opportunity per year. Is that the right way to think about it?

Not too far off.

Okay. Okay. Then maybe just following on from another question that was asked previously. You said earlier that your order intake from optoelectronics was EUR 118 million in the quarter. Visibility is rising. Clearly, if you listen to Lumentum, Coherent, Applied Opto and the others out there, they'll all be deploying more next year than what they're going to take this year. So if I take the EUR 118 million, call it EUR 120 million, nice round numbers, I can work my brain around easier mathematics. EUR 120 million, if you assume it stays flat, is there any reason why your optoelectronics revenue next year can't be at EUR 480 million next year? What would be your pushback to that argument?

Honestly, you're thinking further into the future and a more precise detail than I have been. Thanks for asking the question. Honestly, my thoughts haven't evolved so far that I could give you a credible answer.

Okay. And then just a quick clarification on the second quarter. When I think about the Q-on-Q increase, you talked about opto coming in and contributing. But would opto be the biggest driver of the sequential increase in revenue? Or could that -- or is there a significant chunk also coming from GaN?

So, I think for the year 2026, what we see at this point in time, opto is clearly the driver. And please also take into account, opto has many subsegments, right? We are currently talking about the lasers for the data communications. We also still have some customers doing some work on VCSELs. As I mentioned in my prepared notes, we have some customers playing at this point in time on a lower level on LED/microLED, which now has turned more to AR glasses, right, very small displays also with the vision to say, let's use the AI, just voice input rather than keyboard input, right, and then use the AR glasses.

There still is a market on plain-vanilla red LEDs, which also typically takes quite a number of tools, quite some wafer area. This is set on gallium arsenide. So the opto overall is the biggest part. Lasers is a subsegment of that, currently the booming and fast subsegment, but the other markets are also there. Let's not forget that. So very clear, come back to your question. In '26, opto is by far the strongest segment, very, very clear. And as I said before, also the question on Martin earlier on gallium nitride and silicon carbide will come back on top of that at some point.

We don't know exactly at which point in time, and also we don't know exactly in which order of magnitude. And this was also the reason for us to add the strategic flexibility with our plant in Malaysia. At some point, big volumes may be coming. And then we want to be ready with a flexible supply chain to really catch the wave when the next wave is coming. And our strategy is we always want to prepare for a wave. We want to be ready to catch the wave in full. We want to catch a good market share. We don't want to disappoint any customers. We really want every customer who wants to order. We want to give the customer a tool when exactly he wants it. This is also a big and important capability. And therefore, we cannot predict the future. We can just prepare to be ready when the future comes.

Got it. Very clear. Since you mentioned microLED, can I ask a follow-up or I can rejoin the queue if you want me to.

Please rejoin the queue. Be fair to the others. But we will be happy to take your question at the end again.

We are moving on to the next question. It's from Om Bakhda, Jefferies.

My first question was on gallium nitride. And specifically, you mentioned that utilization rates have been improving in the quarter. And so I was wondering, have you been getting any sort of feedback from gallium nitride customers as to if the timings have improved because you mentioned sort of second half this year, first half next year on orders. So, are you getting any increased visibility on that end from your customers?

Unfortunately, not yet. As said, we see some tools and utilization improving, but we don't know exactly. The one thing is we see how tools are filling. But the question is then, is it just the capacity installed at the customer good enough? Or is the customer preparing for a ramp because the customer of our customer is coming with big orders where the customer needs an expansion. Normally, things don't just linear improve. But normally, we know in our industry, right, at some point, somebody has a big design win. So, somebody then has a big contract to fulfill.

And then if such a big design win and contract has to fulfill, that's a trigger point for next stage of a capacity expansion. So, this is kind of how it works. And our transparency and visibility into this dynamic, which happens again besides our customers on their customer -- customer of customer side, so to say, is quite limited. So, we don't know.

That's super helpful. And then my follow-up is on sort of your inventory levels. So in the previous upcycle in silicon carbide, we saw that your inventories in 2023 increased by roughly EUR 150 million. And as you mentioned, you're sort of now preparing for the growth that you're expecting to come in the medium term from both opto and GaN. And so as we look through 2026, do you think that you'll be then taking the steps to increase your inventory levels through the years that you're ready for the growth in 2027 and 2028?

No, we are changing our model and we have done a lot of work on operations in the soft years, in the last 2 soft years. And the last upturn, we were not prepared for. We've done now in the last 2 years, done a lot of work on our operations side. We've improved our internal processes. We've closer worked with our suppliers. Yes, at some point, when our shop floor is full of machines because we are shipping more in a quarter due to the simple WIP, there may be some increase. But we have changed our operations from a build to inventory build to stock to a build-to-order model by reducing throughput times, reducing build times and so on and so forth. So, the next ramp, I don't want to go to such high inventory levels as we have seen in the past. This was not healthy. And we rather want to work with a stronger, more streamlined operations on this side. It's too early to quantify the effects in the details. But from a direction, I do not want to repeat the mistakes of the past.

Next question is from Andrew Gardiner, Citi.

I have one related to that last point. It's just around your production capacity as things sit today, what you're preparing for over the second half of this year and into the first half of next year? And in particular, as the Malaysia site starts to ramp, what will that mean in terms of the incremental capacity? If you're at mid EUR 500 million to EUR 600 million of revenue for this year, what is the -- given the changing business model that you've just described, Felix and the new site being added, what's the theoretical uplift in terms of your annual revenue capacity?

It's very, very far up. Think of us the first order from where we are today that capacity is not a limit.

Okay. And well, that's good to hear. I suppose maybe asking it another way, what the catalyst for the Malaysian site, was this something that has been in the works for some time and it's just a coincidence that the announcement is happening just as we're also seeing this huge ramp in terms of Opto demand? Or have you been able to react quite quickly to the change in market dynamics and find the site and get moving in terms of the capacity expansion?

Thanks a lot. Very good question. Well, the trigger point for the Malaysia site is not -- the Opto demand or the spike in Opto demand is not the trigger point for the Malaysia site. This is uncorrelated, if you want to say so. So, what we have seen is rather, we have seen and got signals from customers that they see over the next 2, 3, 4 years, more on a strategic level that they see potential upsides for the power electronics in particular, but also for some segments on the optoelectronics and customers came and said, look, one, if you are able to serve markets at a lower price point, well, there could be much, much, much and very significant upside on this one. We've been asking ourselves, well, how could we -- we clearly want to take those upsides in terms of additional volumes, how can we come to price points that the customers have been asking us that they said, well, this is what they need such that they can address these and open up these new market segments.

And we have then looked around the globe where -- what we could do, what we need to do in order to meet such price points. And we found the answer in the very strong Southeast Asian ecosystem, which allows, on the one hand, lower cost for the assembly of the products, but also has a very strong supply chain, a very cost-efficient supply chain, especially for more mature and older components. And that was for us from a strategic point, not from a short-term technical point, the reason to say let's open this plant in Malaysia in order to fully participate in the upcoming probably 2, 3, 4 year waves of power electronics volumes that is also set to come, as I mentioned, on top of what we see in optoelectronics now. It's a strategic move.

Next question is from Michael Kuhn, Deutsche Bank.

Starting with a shorter-term one. You still need EUR 185 million of orders for the midpoint of your guidance and that essentially to be written in the second quarter. And you're also talking about longer-term contracts, obviously. So, that sounds like order intake in Q2 might be EUR 200 million plus or even well above EUR 200 million. Is that a realistic way to think about it?

Honestly, I didn't do the math on the numbers yet. When exactly the orders are coming in.

Okay. But let's say, the logic, EUR 185 million that is still needed in terms of orders to get to this year's guidance and that this should happen in the second quarter, is that logic is right, or let's say, until early Q3?

Exactly. I think round about, you could say so, yes.

All right. Understood. And then, once back to capacity because you just said capacity is not a limit. I mean my understanding always was that there is still some limit. Still, let's say, you're, let's say, getting flooded with orders now north of EUR 200 million each and every quarter. So, would you be able in today's setup without having Malaysia yet, be able to, let's say, generate sales of EUR 800 million or well above EUR 800 million? Or let's say, where is the physical limits right now in terms of how many tools you can ship in a certain quarter, for example?

So, we always need a little bit of time to prepare our supply chain and to prepare everything and get ready for it. So, we cannot go from EUR 59 million in the first quarter to EUR 200 million something in the second quarter, right? That's something like that would not work overnight because it would be just inefficient to have people sitting around and waiting for the orders to come. And, however, as I mentioned, it's not the premises or the space or whatever that is limiting and out of our footprint, manufacturing footprint with enough ramping time, we are currently, for example, working out of a 1-shift model. We can go to a 2-shift model. We can go a 3-shift model. We can go to a 3-shift model plus Saturday work, 3 shift plus Sunday work. So, we could significantly increase the working hours, the effective working hours out of our existing premises if then the orders are there.

So, coming back to your concrete question, EUR 800 million in a year means EUR 200 million per quarter. No problem. We have demonstrated EUR 200 million, I think, EUR 220 million, EUR 230 million or something like that recently in the fourth quarter, I think, of '24 was the last time we did it. So, if you just say, well, Aixtron do 250 in a quarter, yes, multiplate by 4, then you add 1,000, yes, but that's still all in a 1-shift model. So, and if you then go to a 2-shift model or would include weekends and so on, then you understand very easily why I say it's not a limit. And so with a bit of ramping time, we can also out of today's production footprint and facilities generate significant upswings. Maybe that allows -- gives a little bit of light on that question.

Then the next question is from Oliver Wong, Bank of America. Mr. Wong, please check if you are unmuted, we cannot hear you yet.

Operator, Anna, we'll take the next question then.

I wanted to circle back on the 80 to 100 G10 optical tools Felix gave. How should we frame that exactly? Is this something -- is this sort of the demand that you're seeing in the coming years? Should we think about it as in perhaps in this next year or closer to the lower end of that and perhaps toward '28 or in the higher end of that? And is there upside to that number?

So, the line quality was very bad. Let me try to repeat the question as I've gotten it and then you please correct me if I mistaken the question. What I understood is that you were asking about the 80 to 100 tools, how we should see it? What needs to happen more on the upper versus what needs to happen more on the lower. Is that the question -- is that question correct?

Yes. Yes. Exactly, yes.

Okay. So, honestly, I think it's -- you ask for a level of precision that we do not have. I think 80 to 100 somewhere is a ballpark number and please as a market. And please also take that I said the number is -- I expect the number to be around that level, maybe a 1/3 less, maybe a 1/3 more. So, you could say also 90 plus/minus 30%. So, you could say 60% to 120, if you want to say. So, right, to really be clear what I said earlier. And so, just to give a rough indication on the size of the opportunity.

Now to your question, what needs to happen, whether this comes more to the lower end or what needs to happen for this to come out more on the higher end. Honestly, we don't know and also our customers don't know. The interesting point is, we have many discussions now in the last 8 weeks with all the major laser producers. And many say that they -- that even they don't know how much volume and output is needed because there's still many architectural questions of the network evolution, which are not so clear. We see now that agentic is shifting the model again a lot with much more CPUs needed in front of the GPUs of the network. Also agentic creating much more data traffic. On the other hand, the predictions of the industry apparently have been wrong until now. One customer explained to me that he said, look, all the data points have been derived essentially from the phase where all the LLMs were just being trained. And now we are seeing now that the AI is so good that people really start using the models at scale. And now as the models are being used, we see that's a completely different loading pattern of the network compared to the training phase. So, I think there is a very big error bar, as I understand also from our customers.

And I have customers that said, look, on our long-term plans, I need 30, 40 additional tools. But honestly, Felix, I don't know whether I need them in the first half of '27 and the second half of '27 or maybe only in the second half of '28. That's the answer I get from the customers. I'd like to pass that on to you. There is a big market. We give you a rough quantification on it. But I think anything for detailing it out, at least when I talk to my customers and ask them, "Hey, give me a precise forecast, my customers tell me, Felix, I would like to, but I don't know it myself.

Got it. That's very helpful. And my second question is, I was wondering if you could give a bit more color on the mix of your customer base for Optos. Obviously, you have the 2 major indium phosphide players in the U.S. They're ramping up capacity significantly. Do you see sort of a broadening out of the customer base in Optos? Perhaps we've been hearing about perhaps Chinese laser manufacturers that they are also ramping up capacity. Just wondering kind of how you see that?

Good question. So, as always, we never give names, right? We always preserve customer confidentiality. I think that is clear. But what I can share is that today already from the orders that you have seen in the market opportunity that I was describing and our customer base is truly widespread and truly global. Let me give you some examples. There is multiple big names from the United States. I think throughout this call, many of those names have been put out and we have very big names of the Opto demand from Europe, of some U.S. players producing in European sites. We have a very strong demand with many big multi-tool orders, multiyear multitool from Japan. We all know that there's a very strong, very, very innovative optoelectronic industry grown over many, many years in Japan. The same holds true from customers, very well-established customers and producers from Taiwan. We know Taiwan has a very strong ecosystem. And in fact, and also, as you have indicated in your question and earlier in this call, China has a very strong electronic production base, and we have also strong multitool orders from China.

So, we can say at this point in time that the opto demand that we are seeing is truly global. I think it happens also truly global in all parts of the world at the same time. That's not like a wave staging or one wave, one part has started, the others are behind and I think the whole industry at the same point is now waking up and starting the ramp. This is the view of the picture that we have.

Next question is from Craig Mcdowell, JPMorgan.

Just the first one on Optoelectronics. Just wondering from your discussions with customers, what are the main bottlenecks or constraints that they're talking about to receiving your tools? Are they talking about clean room capacity and so on? And how do you see sort of facility or clean room capacity through the next couple of years? And then second question I had was just on the middle lines of the P&L. Obviously, talking about pretty huge revenue numbers coming through from Opto. I'm just wondering what needs to go into the middle lines of the P&L to sort of capture that opportunity and to service that opportunity? What kind of degree of operating leverage should we expect from this large revenue number you put up?

Okay. Let me take the first question. I'll pass the second then to Christian. So, in terms of that bottleneck, I think it really varies customer by customer. So, some customers have existing fab capacity and clean room space. So, these guys can go directly order tools, install the tools and start to ramp. We see from other customers, and as I've highlighted with the previous question from Oliver, right, some customers -- it's global, it's truly global. Some customers don't have enough clean room space. So, some customers are now chasing clean room space or building new greenfield clean rooms or buying brownfield sites and retrofitting them. It's a truly diverse on that one.

And the other topic is what we hear across all customer base is wafer supply is a shortage. I think the topic is very well known across the industry. Indium phosphide wafers are scarce because nobody was expecting this. However, of course, as always, the semiconductor industry, once there's a bottleneck, takes a lot of resources and resolves the bottleneck. So, I'm not concerned that this is going to cut off the boom or anything. I think investments are being made, money is available there, and we know that a lot of activities are ongoing. And I think that's the 2 main topics, literally clean room space and wafer supply.

Apart from the normal challenges of growth, right, ramping significantly is short. I mean, I think everybody is quite busy and also the operations team of our customers are very, very, very, very busy these days I think, but that's clear. That's normal in a ramp in our industry.

Understood. Thank you.

And then taking over the question to the margins, the opto tools in general, running at the upper end of our margin profile. All of that development is fully reflected in our guidance and in the bandwidth of the guidance that we have put out. So, fully reflected no surprises coming from that. And please don't forget, in the guidance is reflected the one-off expense. So that's about 1 percentage point that is dragging down the margin. We have repeated that several times, so just remember what is that.

And so just on the middle lines of the P&L, maybe you're not expecting huge amounts of investment needed to sort of capture this opportunity?

And what do you mean with the middle line of the...

So, like SG&A, R&D, et cetera.

No, no, no, no, no. There's no impact on the OpEx. Of course, I mean, we are running the R&D -- in the SG&A, anyway not in the R&D, we are running our program. But giving you the number of around EUR 90 million, that's the number that we expect this year. No additional R&D expenses now expected from the Opto tool, that's fully reflected in the present.

The next question is from Martin Jungfleisch, BNP Paribas.

I just have 2 quick ones. The first one is also on Opto. I mean, based on your new guidance, obviously the business should probably more than double this year. And can you just disclose what is coming from tool sales and how much is coming from ASP or price? And should this ASP effect also last into 2027 as you grow the share of G10 tools in the mix? That's the first question.

I didn't get it in full because the complete upswing comes from tool sales. I didn't get it with ASPs. Can you repeat?

Yes. I guess like my question is basically, what is the mix between G10 and G4? And basically, what is driven by sales price this year and next year, so by ASP mix as you sell more G10, G4 tools?

I would say probably around 70% G10, 30% G4, maybe something like that. And now the G4, some of the G4s go into the laser segment because laser customers are still producing on the older series. Some of the G4s go into the -- what I mentioned earlier, red LED segment, which is a very cost-sensitive segment. So, the G4s had a -- can have very different margin profiles, and you can have the G4, which is kind of, let's call it, minimum and base configuration. If you open it up, yes, it's almost empty. And then you can have a laser G4 and it's just full. So, think of whatever put for the 150-horsepower versus a 400-horsepower into a car, different looks different under the -- in the car. So, bit of a variety on that one.

In terms of mix effect, I think going forward, it will have a positive mix effect. But again, right, the G4s going into the LED segment, they are quite a drag because that's a margin weak segment. So, honestly, I don't have the full transparency to answer the question, real granularity, I just recognize.

Okay. No worries. And maybe just a follow-up on competition, right? I mean, given there's quite a strong growth in the Opto area, what's your view on competitors there? I think Veeco has recently announced a few orders for the MOCVD tools for indium phosphide. And I guess of that kind of market number you mentioned that you have 100 tools, is this -- what is your expected market share there also over the next couple of years? Do you see any players getting stronger maybe from China?

Yes. So, we take competition always very serious, and we make sure that we got all the steps of competition. In the orders being placed in the starting of the boom, we have not seen the competition yet. We are aware of tools that got ordered -- 2 competition tools that got ordered. We are not aware of more orders. So, we take them serious, but we really take care of our customers to make sure that we have a high market share.

The next is the follow-up from Adithya Metuku, HSBC.

Just one question really. Felix, you always give very technical answers, which I like. You talked about datacom applications for microLEDs. Can you talk a bit about what the use cases will be for microLEDs as opposed to lasers? Where will lasers be used? Where do you see microLEDs being used? And how does this affect tool demand from your perspective, if you were to sell, let's say, for 1.6 terabits per second link, I don't even know if microLEDs can be used, but assuming they can be used, like does it mean one machine, you sell an equal number of microLED machines and if the laser was replaced by microLEDs when compared to laser machines? Any color around what it means for you from a tool perspective would also be very helpful.

Yes. I'm very happy to take the question. So, I mean, as always, a technical evolution and innovation continues and things change over time. I think that's part of our industry. But I think to -- in the first order, I would see microLED and the lasers, what we are seeing now, the EMLs and the PICs and the co-packaged optics, I would see it as complementary. Let me explain why.

Right now, what we are seeing for the scale up and for the scale across and is that we see the lasers being used to connect multiple GPUs, multiple racks in a data center and multiple servers in a data center. So, we talk about slightly longer length of kind of optical cables, if you want to say so. You can all imagine that, yes. We know the picture of how it looks inside and the hundreds of thousands of cables going to the backside of the rack.

Now to our understanding and in fact and also these EMLs and PICs are being used, it's indium phosphide, what we are talking, indium phosphide wafers, it's alternate phosphide tools, the G10 ASP that's being used to that. And the microLED efforts for data communications that we are involved with is mostly gallium nitride based. So completely different material system, completely different tool. So, we talk about the G5+, for example, tool, not the G10 ASP. And what we see is that some customers, multiple start-ups are working on this microLED communication, however, for very short length. So, we are not talking about connecting one rack to another, let's say, a couple of meters away or a couple of hundred meters in a data center away, but we rather talk about connecting, let's say, high-bandwidth memory with the GPU.

So, literally things that are co-packaged on an interposer with some, for example, a glass substrate. Some people also are trying to put 300-millimeter silicon carbide substrates to etch some wave guides into this, literally like semiconductor manufacturing techniques that's being deployed here, co-packaged things for the very short range in the end to have a higher bandwidth connect, for example, from high-bandwidth memory to the GPU and higher bandwidth means more speed or also benefit less energy consumed because the connection by optical can have less energy than if you do it by copper. But to a source order, I would say the aspiration for microLED in this is the very short haul versus what we talk now, the indium phosphide is rack to rack, server to server, data center to data center. It's longer haul meters, hundreds of meters, kilometers through optical cables.

Again, I think that's the starting point and we all know that then over time, innovation continues and boundaries blur and boundaries get shifted and things change. But I think at least that's the activities that we are aware of now in the first quarter '26.

Understood. And do you have any timeline around when this might happen, microLED connections with each supply?

No, it's exploratory. I think, unfortunately, probably it's going to happen like it happened now. Sometimes it happens overnight, and then we have a ramp. I'm afraid it's going to happen like that. But I think it's not going to happen within this year.

Last question in the queue so far from Nigel Van Putten, Morgan Stanley.

I hate to say, but I also want to follow up on the Opto outlook and the 80 to 100 tools or 60 to 120. I can imagine that range being quite wide on an annual basis. But perhaps it's then good to talk about some of the underlying assumptions like I guess, in terms of form factor, is there any difference between EML CW? I think we've discussed this before, but it will be good to just get your latest views. And maybe more important, I can only imagine that these numbers include like a full scale-up opportunity rather than sort of the situation where we're in today. So, I guess maybe to add to that, I can imagine that 80 to 100 number to be maybe not completely applicable yet to the very near term, but more towards the end of the decade. I'll leave it there because there's a couple of questions in there already.

Yes. I do get your question. But unfortunately, I can only repeat and shine some more light on what we've indicated earlier on that, it's too early to predict because it depends on so many factors. And the one factor we've outlined is what happens on the scale up versus the scale across, right? The other factor is on the tool side, literally what speeds are the linkages happening, right? If you go to higher speeds, for example, you need stronger lasers because then the modulation consumes more energy. And that consumes more energy, you need to pump more light in, pump more light in, you need a bigger laser chip. For bigger laser chip, you need more wafers.

Look, there is so many variables. So, we are -- I think we are probably just aware of a number of these variables. But in order to -- the error bars on each of these variables is adding up. I think we can only guess, make a rough assumption to give an order of magnitude. That's my attempt to give you an order of magnitude. But to put it in precise numbers, I think it's too early. It depends on too many factors.

No, no, I totally get that. That's why I was a bit surprised because 80 to 100 is actually a pretty narrow range. But this makes sense. So, I guess, yes, scale-up environment, scale across and the speed necessary, I guess those will be the most significant drivers from an end market perspective. And then on the supply side, it's more about yield, et cetera. So, I think you've said before that the industry is really good at solving bottlenecks. Today, it's more, as you said, clean room and especially the wafer substrates. So, have you been doing a lot of work over the last, I guess, 8 weeks and then maybe beyond that, I can imagine, of sort of working with customers or customers coming to you to see how they can improve yields in the process. Is that something you can provide a little bit more visibility on in terms of -- in the past, I think there were a view that yields are quite low in that indium phosphide space. There is a view that those can improve. But has there been sort of a push relatively recently where you guys are involved to make that happen?

Not so much. I think in the end, the processes that our customers are run have many, many, many, many, many, many steps. Sometimes not only one epi process, but sometimes multiple epi processes. And I think at this point in time, most of our customers are really focused on getting the ramp done to really achieve that. I think at some point, additional work will go and do a yield. But I think the yields are less a topic of the epi tools. I mean our tools are helping our customers to produce the best yield that's available in the industry. That's the reason why they choose the Aixtron tools. But I think when you talk about the yield or the line yield, which is all the process steps one after another, that's not only about the epi tool, but that's literally the connection of all the hundreds of process steps that they are running within a customer factory. And normally, especially in the Opto side, customers are running very different recipes. They all have their process secrets where they differentiate. I would say, especially our Opto customers are very, very secretive and are very keen on preserving their process recipes for themselves.

Thank you very much, ladies and gentlemen also from my side. As we have no more questions in the queue, I'm handing the floor back over to the host.

Thank you very much, Anna. Thank you all for listening and for the very good questions. If there are any questions still open, the IR department is at your disposal, please give us a call. We will be on the road in the next couple of weeks. So hopefully, I'll see a lot of you in person. And for those we do not see, we'll talk to you at the latest on our Q2 call, which will take place end of July. Have a great weekend for those at least who have the 1st of May holiday and for the rest, a nice Friday. Goodbye for now.

Ladies and gentlemen, welcome to AIXTRON's Fourth Quarter and Full Year 2025 Results Conference Call. Please note that today's call is being recorded.

Let me now hand you over to Mr. Christian Ludwig, Vice President, Investor Relations & Corporate Communications at AIXTRON for opening remarks and introductions.

Thank you very much, Anna. A warm welcome to AIXTRON's 2025 results call. My name is Christian Ludwig. I'm the Head of Investor Relations & Corporate Communications at AIXTRON.

With me in the room today are our CEO, Dr. Felix Grawert; and our CFO, Dr. Christian Danninger, who will guide you through today's presentation and then take your questions.

This call is being recorded by AIXTRON and is considered copyright material. As such, it cannot be recorded or rebroadcast without permission. Your participation in this call implies your consent to this recording.

All documents referred to in this call can be accessed via our website in the Investor Relations section. Please take note of the disclaimer that you find on Slide 1 of the presentation document as it applies throughout the conference call.

This call is not being immediately presented via webcast or any other medium. However, we intend to place a transcript on our website at some point after the call.

I would now like to hand you over to our CEO for his opening remarks. Felix, the floor is yours.

Thank you, Christian. Let me also welcome you all to our full year '25 results presentation. I will start with an overview of the highlights of the year and then hand over to Christian for more details on our financial figures. Finally, I will give you an update on the development of our business and our new guidance.

Let me start by giving you an overview of the highlights of the year on Slide 2. The most important messages of the day from my viewpoint are: in 2025, we have performed well in a soft market environment by achieving revenues of EUR 557 million, a decline of 12% year-over-year. That translates into a CAGR of more than 13% since 2020.

We delivered on our adjusted 2025 revenue guidance, meeting the upper end of our guidance given in October '25. Mainly due to the lower utilization in operations, due to one-off restructuring costs, and due to G10 ramp-up adjustments, our gross profit was down 15% to EUR 222 million, and EBIT was slightly down with minus 24% at EUR 100 million as a result of this.

Similar to last year, we finished the year with a strong Q4 '25 performance. We achieved 31% EBIT margin, a level comparable to last year's extraordinary Q4. This marks a great achievement of our operations team as we managed to realize all shipments that customers had asked us to deliver in Q4.

The highlight of the operating performance is our cash flow generation. Operating cash flow increased by more than EUR 180 million to EUR 208 million. And our free cash flow increased by more than EUR 250 million to EUR 182 million.

With that, we concluded the year '25 with a cash level of EUR 225 million, a good step towards rebuilding our strong cash position that we always have desired. Thus, despite the weaker net profit, we have decided to propose a stable dividend of EUR 0.15 per share to our shareholders.

Our outlook for the year 2026 is based on an expected continued weaker market environment. We expect revenues to come in at EUR 520 million in a range of plus/minus EUR 30 million with a gross margin between 41% and 42% and an EBIT margin between 16% and 19%.

Breaking this down by segment, AI will be the key revenue driver in '26, fueling strong growth in optoelectronics and lasers through rising demand for optical interconnect. In contrast, SiC, silicon carbide power will face a weak year due to overcapacity and slowing EV momentum with LED and microLED and GaN power demand remaining broadly stable.

This concludes the short highlights section. I will now hand over to our CFO, Christian Danninger. He will take you through the full year '25 financials. Christian?

Thanks, Felix, and hello to everyone. Let me start with the highlights of our revenue development on Slide 4.

As Felix mentioned, the revenues in 2025 were down 12% to EUR 557 million. Our strategy of serving various uncorrelated end markets with our equipment proved again successful in 2025. We saw strong growth in the optoelectronics area. This compensated to some extent, the weaker demand for equipment for LED and microLED as well as gallium nitride power electronics.

The breakdown per application shows that 57% of equipment revenues comes from GaN and SiC power, 23% from optoelectronics, 15% from LED and a 5% contribution from R&D tools. The aftersales business contributed to total revenues with a growth of 1% to EUR 112 million. The aftersales share of revenues grew to 20%, up from 17% a year ago.

Now let's take a closer look at the financial KPIs on the income statement on Slide 5. Gross margin decreased by 1 percentage point versus 2024 to 40%, which was primarily due to lower utilization operations, G10 ramp-up adjustment expenses and the one-off restructuring cost. Accordingly, gross profit was down by 15% year-over-year to EUR 222 million.

As we had planned, our spending on R&D in the year 2025 decreased to a total of EUR 81 million due to a reduction in external contract work and lower consumables costs. This helped to drive our OpEx down 7% to EUR 122 million.

Combined with the lower gross profit, this resulted in an EBIT of EUR 100 million, which is 24% lower year-over-year. Net profit was down 20% year-on-year at EUR 85 million. This results in an effective tax rate of 15% in fiscal year 2025, a clear positive were our Q4 2025 gross and EBIT margins at 46% and 31%, respectively.

Despite the 18% lower revenues number at EUR 187 million, we were able to beat the very strong level of Q4 2024 on gross margin level and meet it on EBIT margin level. Orders in the quarter came in at EUR 170 million, an uptick of 8% versus last year's quarter. For the full year, order intake came in at EUR 544 million, slightly weaker than last year. And thus, our backlog at EUR 258 million is down by 11% year-over-year due to the above-mentioned softness in demand.

Now to our balance sheet on Slide 6. We ended the year 2025 with a total cash balance, including other financial assets of EUR 225 million, which was well above the EUR 65 million last year. There are a number of factors driving this increase.

Firstly, inventory levels at the end of 2025 came down by about EUR 85 million to EUR 284 million compared to EUR 360 million at the end of '24. This is the result of our adjusted supply chain strategy and corresponding measures after initially front-loading the supply chain in 2024 in expectation of stronger revenue growth. We target a further reduction of inventory levels through 2026.

Second, we have seen a solid decrease in outstanding receivables compared to the last year and which generated some EUR 60 million in cash. As a result of putting on the brakes in our supply chain early on, the amount of payables have been stable during the course of the year.

Advanced payments received from customers, on the other hand, were slightly down year-over-year at EUR 44 million due to the decline in order intake, combined with a shift in the regional customer base and partially impacted by some key date effects. At year-end, down payments represented about 17% of order backlog.

As a consequence of all these factors, operating cash flow improved by more than EUR 180 million to EUR 208 million in the financial year 2025. As mentioned already in previous calls, CapEx decreased significantly in 2025 due to no additional investment requirements for the innovation center.

As a result of the significantly lower CapEx, free cash flow improved by more than EUR 250 million year-over-year to EUR 182 million from negative EUR 72 million in 2024. We expect further solid free cash flow generation in 2026.

Lastly, we are proposing a stable dividend of EUR 0.15 per share. Despite our lower net earnings, we want our shareholders to participate in the improved cash flow generation.

Going forward, following an intensive investment phase in the years 2023 and 2024, CapEx alone for the innovation center was EUR 100 million. AIXTRON plans to use the cash flow in 2026 to further build a strong cash position.

Also, I want to remind you that AIXTRON expressly does not pursue a fixed dividend policy, but rather adjust the payout ratio to reflect the respective business performance and capital allocation priorities.

With that, let me hand you back over to Felix.

Thank you, Christian. I will continue by giving you a brief summary of the key market trends we saw last year and before I move on to our expectations for '26.

I will start with our currently weakest segment, the silicon carbide power business before moving on towards the strongest segment step-by-step.

SiC. Throughout the past year, the global silicon carbide market has undergone a significant transition. In Western markets, we are seeing a temporary slowdown driven by weaker electric vehicle demand and substantial idle capacity at several customers. This has even resulted in reduced or scrapped 6-inch capacity in some cases. We expect the digestion period for silicon carbide epi tools to continue throughout 2026 in Western markets.

China, by contrast, remained a strong pillar of demand in '25 for AIXTRON with solid order intake and robust shipments in the first half of the year. In the second half of '25, also in China, SiC demand has softened. And in '26, we expect the digestion to continue also in China.

Despite this short-term softness, the midterm outlook for SiC beyond '26 remains highly attractive. Substrate prices have dropped significantly, making silicon carbide devices far more competitive versus silicon IGBTs and enabling broad market adoption, both in EVs and across industrial applications.

Even more importantly, the technological transition is well underway. The industry is rapidly moving from 6-inch to 8-inch wafers, starting with Western customers, now also in China, with a full shift expected towards '27 and '28.

At the same time, the introduction of superjunction silicon carbide MOSFETs, which require multiple thin epitaxial layers instead of a single thick layer will significantly increase epi tool demand. Our batch-based G10 SiC platform is ideally positioned for this new operating model and has already achieved major milestones with the shipment of our 100 system during 2025.

In 2026, we expect very strong demand [Technical Difficulty] at the beginning of '25 and have been steadily recovering. AIXTRON maintains a clear market leadership position with more than 85% market share across GaN device classes, and we remain deeply engaged with customers expanding their GaN road map into [Technical Difficulty] coming years.

Importantly, GaN is emerging as a central technology for AI-driven power architectures, particularly as hyperscale data centers plan the transition to high-efficiency 800-volt platforms. We expect additional volume from GaN from AI applications at some time in the 2027 and '28 time frame. The exact timing for when this happens is unknown, and we will keep you posted when signs of this are getting clearer.

In parallel, we are working with a small set of customers on 300-millimeter GaN. These customers have existing 300-millimeter silicon fab, which they desire to repurpose for GaN. Our 300-millimeter GaN tool is fully operational with our own innovation center, as we call our 300-millimeter team room and collaborations with imec and leading power semiconductor manufacturers are ongoing.

Now, let me come to the LED and microLED market. After a period of muted investment, now the market for red, orange and yellow LEDs, we call them ROY LEDs, is showing clear signs of recovery, driven primarily by development in China. This momentum from display makers who are pushing the boundaries of image quality. In fact, several major TV manufacturers are now transitioning to full RGB backlighting architectures, which further boosts demand for ROY LED as tool.

This trend underscores a broader shift. Even traditional LED backlighting is being reinvented, establishing miniLEDs as preliminary storage stage towards microLED. Enhanced local dimming, full color backplanes and ultra-high brightness panels are now becoming standard in premium consumer displays. These innovations are breathing new life into an application space that many considered mature.

At the same time, exploratory and qualification work of customers towards microLEDs continues with customers in Europe, U.S. and Asia. The focus of this work has shifted away from watch and television now strongly towards AR/VR glass applications.

We expect this market is still some time out into the future until a larger revenue contribution. And given the fact that one wafer can serve hundreds of AR glasses, the expected demand will be much, much smaller than what we would have anticipated for television applications.

Overall, we can say that for AIXTRON, ROY LEDs and microLEDs together translate into a solid revenue contribution of around 15% of group revenue for both '25 and '26.

Now, let's finally come to our strongest segment in '26, the lasers for datacom. The global indium phosphide laser market has entered a new phase of growth. And from Q4 '25 onwards, we have seen an even stronger momentum in this segment. We have served this market for many years with our proven G3 and G4 platforms historically for telecom and datacom applications, supporting the further adoption of high-speed broadband communications.

As far as cloud services with a market share we estimate well north of 90%. The demand we see today is linked to a structural up cycle linked to AI data center build-out and the development of data-hungry new generation of GPUs. And this structural shift creates the demand for indium phosphide-based lasers grown by MOCVD with a massive adoption of optical interconnect now also within the data center architecture.

As bandwidth requirements move to 800 gig and data 1.6T, the laser content per data center is increasing multifold to enable the required bandwidth. Our customers are subsequently not only ramping their manufacturing, but also rolling out new product generations with higher bandwidth that are also more integrated like photonic integrated circuit, PIC, now in order to be always faster, more compact and more energy efficient.

For the majority of our users, their road map now includes a shift away from 3 and 4-inch to 6-inch wafer size. That is an enormous step for a market that has been historically very conservative. It enables them to access the advanced manufacturing technologies for these new types of products.

Our G10 ASP product has rapidly established itself as the tool of record, as we say, for this new generation of photonic devices, replacing customer legacy system, producing higher yield and cheaper 150-millimeter indium phosphide epi wafers.

We are serving all of the top 10 suppliers to this market. And demand is coming from all regions of the world, from leading suppliers in the U.S., from the ones in Europe, but also from optoelectronic leaders in Japan, in Taiwan and in China.

Looking at demand dynamics, we expect the optoelectronics business to more than double year-over-year from 2025 into 2026. With this, it makes up for a large part of the revenue that declined in silicon carbide that I illustrated earlier.

Finally, let me now present our full year guidance for 2026 to you on Slide 19. This guidance takes into account all the factors that I just described previously. We expect revenues to come in at EUR 520 million in a range of plus/minus EUR 30 million. We expect a 2026 gross margin of 41% to 42% and an EBIT margin between 16% and 19%. The effects of a personnel reduction we have initiated in the beginning of '26 are already included in this forecast.

Now, let me comment on the first quarter of '26. As usual, sales in the first quarter of the financial year will be lower than the annual average first quarters. In Q1 '26, we expect revenues of EUR 65 million in the range of plus/minus EUR 10 million. This is comparatively low figure, fully in line with expectations and with a seasonal pattern of the business. For completeness, we have adjusted our USD to euro budget exchange rate at which we record U.S. dollar-denominated orders and backlog to USD 1.20 per euro.

With this outlook, I'll pass it back to Christian.

Thank you very much, Felix. Thank you, Christian. Anna, we will now be happy to take the questions.

[Operator Instructions] So the first question is from Ruben Devos of Kepler Cheuvreux.

I just have one on the guidance basically, pointing to EUR 520 million a year. Obviously, you started the year effects of the seasonality at EUR 65 million, which is about 12% of the total. So just curious about how you see the quarterly cadence at this stage. And what might give you maybe the confidence that orders of, I think you talked about EUR 280 million, whether that will materialize at the pace needed for a strong H2?

Yes. Thank you very much. So we expect again in '26, the pattern that we have seen in previous years, where we have the year a pretty much back-end loaded towards Q3 and Q4. I think that's a seasonal pattern, which we have already seen in 2024 and 2025.

If you recall in '24, in the fourth quarter, we even shipped over EUR 200 million. Now in the fourth quarter, it was around EUR 180 million. So it's not uncommon that we are backend or backwards loaded. I think it will not be as heavy in '26. But the Q1 is very weak. I think in Q2, Q3 onwards, we should be maybe around EUR 110 million, EUR 120 million, EUR 130 million, I don't know, something like this. So north of EUR 100 million, I would say. And then clearly, in the Q4, I think we will be peaking.

So nothing to be -- don't expect the Q2 is again another EUR 65 million and I think we would be a little dry. But that's not going to happen. Does that answer your question?

Yes, certainly. The second one is just around the G10, which is the tool of record at the leading laser customers. When a customer qualifies your tool and locks in, how long does that qualification typically last before it needs to be, let's say, recompeted? I'm just trying to understand a bit the stickiness of your opto business and whether your position today, which is very strong, obviously, whether that's a meaningful barrier already or whether there for each new product generation, that sort of, yes, reopens the door for competition?

I think in the laser business, you have probably the most sticky and the most difficult to requalify from all the segments. So with many customers, qualification efforts have been going on since 1 or 2 years already. And the complexity comes in the qualification for a laser tool from the fact that it's not just one simple laser, or one simple layer like you have in silicon carbide. In SiC, this is our most simple tool, I would say. You have one single layer, a thick single layer and every customer is doing kind of almost the same.

Now in contrast, in the laser domain, typically, each wafer gets not only put into the tool once for one layer, but the laser customers have very advanced structures. And in these modern architectures and high-speed devices that is currently now making up the market, many wafers of our customers see the tool 3, 4, 5 or even 6x from the inside, meaning the customer makes a layer, doing some other steps, the wafer is put into the tool again, makes another layer and so on and so forth.

And you can imagine if something changes in the deposition and that is repeated 5 or 6 times, an error or a change is then repeated or taken to the power of 5 or taken to the power of 6 that depends on a very, very precise repeatability. And so with many of these customers, we've been working since multiple tools, multiple years. That's also the reason why the G10 ASP, which we launched already in '21, '22 is only now getting the strong momentum from the laser market because the qualification has taken such a long time.

Okay. And just a final question is that you're launching the 300-millimeter Hyperion tool commercially in '26. Just curious how many customer qualifications are currently underway? And when would you expect sort of the first repeat orders to come in?

We work with multiple customers. I think it's important to differentiate. Some customers are, I would say, in an exploratory and research stage. And there's many of these, I don't know, I think probably double-digit or so.

However, I think from commercial relevance, 300-millimeter will, as I mentioned in my prepared remarks, only initially only for a relatively small number of customers. And that is those guys who have a very big 300-millimeter silicon fab, which they want to repurpose and convert an existing 300-millimeter silicon line to a 300-millimeter gallium nitride line.

I think all the other stuff like microLED and so on is more like playing around, researching, exploring ways. But I think those market segments probably take another, I don't know, 2, 3, maybe 4 years until they really mature. We are engaged. We work on a lot. But I think in terms of revenue and really making numbers, that's still quite some time away.

The next question is from Martin Marandon-Carlhian from ODDO BHF. Martin, unfortunately we cannot hear you anymore. [Operator Instructions]

Let us continue with the next question, please. We can take him later.

The next question is from Rohan Bahl of Barclays.

I just wanted to touch on that 300-millimeter GaN tool. I mean your peers said overnight that have gotten several orders on 300-millimeter GaN already. So I just wanted to check your progress on getting Hyperion ready for production volume lines rather than sort of your R&D quality tool that you have at the [ minute ].

I think we are very well on track with respect to that. We have also multiple orders again from these few customers that I was mentioning.

Okay. And maybe just on the 800-volt AI data center opportunity for GaN, everyone is getting excited about this. So just curious on how things are progressing here? What have customers been saying to you and whether you're still sort of expecting orders to ramp up materially in the second half? I've noticed your backlog has been building for 2027. So I wonder if there's any 800-volt business in there?

So the 800-volt is splitting essentially into multiple types along the architecture. You probably have seen the slides on the 800-volt architecture by NVIDIA and by major suppliers such as Infineon, right? So we are participating in multiple stages on that chain. The one part is coming from the overland line on the silicon carbide, which translates or transfers from over 10 kV down to 1,200 volts, 2 kV, 1 kV. This is the biggest part silicon carbide. Then gallium nitride comes into play at 650 volt at 100 volt and even at lower stages like 20 volts. So this is where we are participating.

We are with multiple customers working on 650 and 100-volt devices for exactly this architecture. And to our understanding, the qualification efforts of our customers means either IDMs or foundries, again, with their customers, being the board makers and the power supply makers for these architectures is ongoing. To our understanding, there is no clear time line on when exactly the switch is taking place yet.

And that is also the reason why I commented in my prepared remarks that we know that this is coming, and we are pretty sure that this is coming sometime in the time frame, I always say '27 and '28, but we don't know exactly when it is coming. So in the order backlog that you're referring to, I don't think there is still 800-volt orders in. I think this is other topics more like EV silicon carbide related. Of course, general tool for silicon carbide can be used for any segment. That's clear.

But I think the button when exactly the 800-volt is getting pushed and the orders are coming in, the timing is still a bit uncertain. I would not be able to give you the point in time at this period of time.

So Martin Marandon from ODDO BHF is back.

My first one is on photonics and opto, et cetera. Considering that several of your customers are talking about very significant indium phosphide CapEx increase this year. And I understand that the big acceleration in terms of orders was really in Q4 last year. Do you think we are quite early in that CapEx cycle? Or do you think that '26 could be the peak? So how -- basically, how do you think about '27 at the moment?

Thanks a lot. I think that's a very good question. Yes. Let me try to shine a little more light on it, how we see it. And again, we only have, again, a piece of the puzzle, but let me try to explain what we are aware of and what we believe. And so we see that the cycle really has kicked off towards the end of '25. So you have seen that in the fourth Q4 '25, our photonics orders have significantly increased. Q1 to Q3, they were still on a relatively low level. In Q4 '25, our photonics orders have increased.

We still, already now in Q1, we see continued order momentum from our customers. Some orders have already received. Others are in discussion with customers. And we expect -- and this is also, by the way, the reason you may have seen that our coverage of revenues with orders, our backlog is lower than we have seen in many past years because we are at the very beginning of the cycle. I think that explains this topic, so on.

However, we have indications from a number of customers like kind of their road map, their forecast, what they need throughout the year. This is baked in our guidance. So our guidance reflects that already. And we expect that the orders are coming in essentially throughout Q1 and continue to come in Q2 and covering then the revenues that we have forecasted for the year.

And as you see, it's a quite significant increase. It's more than double year-over-year for the photonics side. And I mean, this is very helpful for us because I think we all are aware, silicon carbide is really dropping almost dead this year, meaning pulling a bit hole in our revenues. This hole is now just nicely getting filled up by the photonics. It's quite helpful.

Now I think you've indicated how long this extends into '27. Of course, very difficult to predict the future. My guess is it's not only 1 year, but it's extending beyond that. However, to comment how much or to which extent it's the majority in '26 and is it even the same level in '27 or less in '27 and more in '27, that is too early. I have no indications to qualify that.

Okay. And just another one for me on GaN adoption in data centers. I think in the past, you said you could expect orders in H2 this year or in '27 for '27 and '28 revenue. But how do you think about how the ramp will happen? What I mean by this is that, it looks like a big ramp. So how it usually happens with your customers? Do you have already some discussion with when and how much you need to be ready? Or do you really see that ramp once the orders start to come in basically?

That is a very good question. I think both things come together at the same time. Typically, when a ramp for a new segment is happening, we see the first orders for that particular second or that particular application coming in from one customer or typically from 2 or 3 customers at the same period of time because normally then a segment is coming and also the guys who are using the chips are not only relying on one source, but typically on 2 or 3 sources.

So typically, we then get the first orders, particularly for a given segment to come in. And along with the orders that are coming in, we sit together with our customers, they share forecast with us, and we jointly sit on the table making a ramping plan, because normally, it's not that the customer needs only 2 tools or only 10 tools, but rather the customer has a plan and say, look here, in the first year, I need 10 and the next year, I need 15 and the year thereafter, I need 15. How do we best do it? How do we best distribute it over time and so on and so forth. This is normally what's happening.

And I expect when this 800-volt GaN ramp is really starting, we are not there yet, but then I expect to have these discussions with customers.

The next question is from Oliver Wong of Bank of America.

My first question is, again, back to 300-millimeter GaN. I understand that the -- we're not expecting huge revenues upfront. But I was wondering -- so my understanding is that whether it's with the 200-millimeter or the 300, usually customers kind of go with one major supplier, one tool of record, so to speak. I was wondering what kind of timing can we expect for the leading 300-mill GaN suppliers to kind of make a decision on that?

I think Q4, Q3 or Q4 of '26.

Got it. And my other question is regarding the lead times. I was wondering if we can get an update on currently where the lead times are for the major end markets and kind of where you expect that to trend?

Excuse me, I didn't understand your question.

The lead times between orders and revenues for kind of your major end market categories.

I think we are probably around -- I think it depends by the market, somewhere between 7 and 10 months, I would say, or 6 and 10 months, something like this. But honestly, I don't have it broken down by end market.

We are back to normal, right? If you recall, yes, in the post-COVID, our lead times were very long. We are now back to a normal lead time.

Next question is from Madeleine Jenkins.

I just had one -- another one on GaN. You mentioned utilization rates are improving. Do you have a kind of a broad sense of where they are now?

And then also on this data center opportunity, obviously, I know timing is uncertain. But sort of volume or demand-wise versus kind of the consumer business that made up GaN in the past. Do you think it's a similar size? Or do you see it being bigger? Any color on that would be great.

Utilization rates, that's always very difficult to predict, because we get more like signs from our customers, qualitative signs like: we need new tools, we don't need new tools. I would guess across the market, probably utilization rates are maybe 60% to 80%, I would say. So on a decent level now, I mean, earlier, we were probably around 30% to 50% after the big GaN investment wave where the demand wasn't there yet. So I think it's still taking a little bit of time until the next investment wave is getting triggered. But as we said, somewhere around the '27 time frame, early in '27, end of '27 or maybe even end of this year, we will see some investment trigger.

Now as for the size, and I think with GaN, it's important to note that GaN has been penetrating across all market segments. It started off, as you rightfully note, 4, 5 years ago, purely in the consumer market, chargers for smartphones, chargers for notebooks and those kind of applications where the form factor was the driving topic.

By now, we have seen GaN penetrate kind of across all the market segments, which is addressed by silicon means motor drives for battery-driven applications. We've seen it in motor drives for things like air conditioners, more like high-power, high-voltage topics. We've seen it in 100-volt and 20-volt point of loads and servers to reduce the energy consumption of servers so kind of all market segments.

So I think you cannot split GaN any longer into a consumer or non-consumer segment. I think GaN is really on a trajectory of getting a very widespread application.

Makes sense. And I know you -- in your release, you flagged that there's a decent chunk of orders for 2027 delivery. Could you just kind of provide some more color on that? Why is it? Kind of is it just lead times or -- is there kind of specific customer capacity additions going on?

No, no. What we have said is, we expect as we see the utilization rates of the installed base now gradually increasing. And as we see further adoption of GaN, particularly in the 800-volt architecture for AI, we expect that at some point, whether it's the end of '26 or sometime in '27 or at the end of '27, we don't know the exact timing. We expect at some point, utilization rates to be at a level where it triggers new investments, new tool purchases by our customers, and where especially the 800-volt architecture is then switched to GaN. Today, a big part is still on silicon.

And once that switch has happened away from silicon to the much more energy-efficient GaN, then this will trigger in our expectations, new tool orders by customers because they need to expand their capacities in order to serve this additional market segment. But when exactly it happening, whether this is end of '26 or early '27 or end of '27, we explicitly say we don't know the timing.

Sorry, I get it. So I was talking more kind of broad comment on your current backlog. I think over EUR 100 million is for delivery in '27. I just wondered why that was the case?

Well, this is a mix of applications. It's a mix of applications. A big part is silicon carbide, where customers have ordered and as the market fell down and became slower, customers said, can we have it a little later? Yes, I think the biggest part -- I would guess the #1 application amongst those is silicon carbide.

The next question is from Martin Jungfleisch of BNP.

First one is a bit of a follow-up on the guidance and the lead times. It looks like that you need around EUR 300 million in new orders in the first half to make the '26 guidance. Is that kind of the right way to think about it with lead times of 7 to 10 months? And then maybe if you can comment if you're on track to meet this kind of EUR 150 million order run rate in Q1 already? That's the first question.

Yes. We see ourselves fully on track. We sleep very well. We feel very well in covering and securing that.

Okay. Then maybe another follow-on on the moving parts. I think if I understood you correctly, you mentioned that you expect photonics revenues to double this year. So then what are the moving parts? I think you said also GaN should be up moderately. So is it like the 3D sensing part or the LED part that should be down massively this year then?

Sorry, I didn't get the last one. I didn't get the last part of your question.

Yes, I was just asking with photonics doubling, I think that's what you said this year. And what are the moving parts within that revenue guidance? I think you said GaN should also be up moderately, so is 3D sensing, LED, silicon carbide then down quite massively? Is that the right way to think about it?

Yes, exactly. That's the right way to think about it. I would say LED/microLED roughly is flat. Silicon carbide massively down. This is a big hole that's in there. And this hole, to the largest part, is getting filled up by the doubling of the optoelectronics. And that's why overall, and if you sum it up, we come at those slightly down numbers from the whatever EUR 557 million we had in the past year in '25 and now to the EUR 520 million plus multiple.

Okay. And maybe if I can, just a small follow-up on the gross margins. Can you just break down the moving parts a bit on the gross margin guidance for this year? So what is kind of the headwind from lower revenues that you're seeing, what is the better product mix and so on?

And maybe if you think about -- if we go back to EUR 600 million revenue next year, what would be the gross margins on a like-for-like basis when you assume all the benefits from the restructuring program, et cetera, should this be like 45% then?

So great question, but I don't have all the numbers prepared. It sounds like almost I would need an Excel sheet next to me to answer your question. So on a joking note. No, let me try and best to help you explain as much as I can without having a computer next to me, yes?

So you see we managed to keep the gross margins around stable compared to last year or improve even a little bit. And what you see here is already we did first a slight amount of headcount reductions early in '25, so last year already. So a part of that benefit already becomes effective in '26. We then, as you have seen, have been able to gain further efficiencies, and we do another slight headcount reduction now or have done in January already. It's completed. We did it very early in the year.

And the cost for that is, of course, included in the guidance. And we've been working a bit on our efficiency in operations, streamlining processes and operation shop floor work and all that kind of stuff, right? And all that allows us to keep the gross margin stable.

Now the question is, how should you think about it? Well, if you go into next year, into '27 -- again, I just do it on a like-for-like basis. I didn't do it the Excel spreadsheet for your hypothetical EUR 600 million. But you can then take out from the cost this what we said, mid-single-digit million restructuring cost. That's, of course, a onetime cost, and that's onetime in '26 and not again in '27, kind of. So that will help on the gross margins. And honestly, I haven't looked at the details of the product mix, which, of course, also plays a role. I haven't done that. But it will certainly help on the margin.

And just to make sure -- maybe one more comment, just to make sure that you get that, as you now probably looking to get some numbers into your model. And if you look at the R&D cost, we had in '24 an R&D cost on the order of EUR 90 million, and we had in '25 an R&D cost on the order of EUR 80 million. In the current year '26, if you do your model, we'd rather put in EUR 90 million of R&D cost.

You will come to that if you do the math anyways with gross margin and the EBIT margin, just to make sure that you get the right number so everybody gets the right numbers here because we have quite some new ideas for new products, and that always translates then for us into R&D because at some point, '27, '28, we expect the markets to pick up, and of course, our investors and you guys expect that we have then a fresh portfolio winning and securing our market position again. Now it's down. But when new markets are there, then it's a lot of fun. We want to be prepared and we want to be ready for that.

Next question is from Jarad Abed of mwb research.

It doesn't seem to be there. Let's take the next question please, Anna.

Maybe it should work now, Mr. Abed, can you hear us?

Yes. Can you hear me?

Yes, we can hear you now.

Okay. Sorry. Yes, I just have a quick question regarding Q4 backlog movement. I mean there is notably an order cancellation of approximately EUR 11 million. Can you provide some color on this?

Yes. I think that was 2 process modules. I think it was a customer from laser and gallium nitride, if I recall.

Okay. And my second question, I'm trying to understand the overcapacity in silicon carbide. Is it like structural or cyclical?

Cyclical. So we get from our customers literally the feedback that they say, look, gradually capacity is now starting to fill. I mean we looked 1 year ago probably at 30% utilization, but the adoption of silicon carbide continues in the market. A big element that helps is that the prices for substrates have dropped significantly. And due to that, the overall -- and substrates make in silicon carbide a major part of the overall cost, probably the #1 cost position is substrate. Those are getting cheaper.

With that, and the silicon carbide power devices are getting more affordable. The cost is going down. And as cost is going down, silicon carbide MOSFETs gain relative in attractiveness compared to silicon power devices, silicon IGBTs. And with the gaining attractiveness that design-in is increasing, they're getting more widespread and the demand in terms of units is increasing.

And as the units are increasing, the existing capacity gradually gets filled. And at some point -- again, we don't know the timing, but at some point, the overcapacity will be digested and then new orders will be triggered. And again, we expect this sometime in the '27 and '28 time frame. When exactly, we don't know.

Okay. But you know that like -- I mean, it's -- you mentioned previously that you expect some orders once annual EV production with silicon carbide inverters surpassed 3 million units. Is it still the case?

I didn't get your question with the numbers that you were just saying. Sorry, I couldn't understand.

Yes, sure. You mentioned previously that you are expecting like silicon carbide acceleration once annual EV production with silicon carbide inverters surpassed 3 million units. Is it still the case?

I think we've never given out a number of 3 million units for inverters. I think that's a very specific number, which is probably not from us.

I think it is referring to a broad assessment of how many cars we would need on the street to see a pickup. That was -- that's where it came from.

As a proxy.

As a proxy, exactly.

Honestly, we cannot comment on that.

Next question is from Craig McDowell of JPMorgan.

My first one is on pricing. And certainly, on the device side of opto, we're sort of seeing, obviously, a tight market, and it seems like device makers -- laser device makers are able to take price and pretty significant price. I'm wondering whether that changes the value that you offer to your customers on the indium phosphide tool and whether you're able to see price increases and specifically whether that's included in your more than doubling comments for 2026?

The main driver for the doubling is literally on the number of tools. So it's not a doubling by price, yes, that would be nice. It's literally doubling by the number of tools, by the number of shipments. But historically, optoelectronic tools are on the higher side of the pricing in our portfolio simply due to the fact that those laser tools are of a very high level of complexity.

If you compare an LED tool going into China and you take a laser tool and you open them and look at them next to each other, you feel that one tool is filled with twice the number of technology inside than the other tool. And somehow that's, of course, reflected in the price.

But given the tightness in the end market, you're not yet raising the prices of your own tools, to be clear?

No. We don't. That's never a good idea towards customers. They don't like that.

Understood. Okay. And then just on -- you mentioned that you're still in discussion with opto customers through Q1. Some of those orders might have been written, certainly discussions ongoing. Just wondering whether there's a change in tone with your opto customers, are you talking on a multiyear period now in terms of delivery? Or is it still very much sort of within the next sort of 6, 12 months that conversations are happening?

It depends customer by customer. We have both types. We have some customers discussing kind of literally the next tool. I need something very, very fast. When can I have 5 tools? Please as fast as possible. I have others more engaged in a structural discussing throughout the year '26 and then others more looking around the multiyear road map. It really depends by customer purchase team or strategic planning team. We have all of it.

The next question comes from Om Bakhda from Jefferies.

I just had a question on your silicon carbide business. I guess when we look through the course of the year, is there -- I mean is there anything that you see today that could happen, that could mean that the guidance that you've given on SiC could prove to be conservative in the second half of this year?

Well, that's a very good question with lots of buts and if. Let me think. Honestly, I think for the second half of '26, my gut feeling tells me it would be a bit too early, seriously for silicon carbide and talking about revenue, because I think there still is some capacity in the market, which still needs to be digested as we had discussed earlier.

I think if we look into '27, purely the EV demand can be a nice driver, as discussed. We see now that silicon carbide devices more and more get designed into higher voltages. So not only 1 kV, 1,000 volt, 1 kilovolt, but also 2,000 volt, 3,000 volt, 10,000 volt, so 2, 3 10 kV, and notably in the space of grid applications for solid-state transformers and applications like that. But I think this is -- would be too early to expect a tool demand, equipment demand for that in '26. I think we are clearly looking towards '27 and '28 for these new applications and new trends.

That's my gut feeling. Maybe I'm wrong. If we can ship more, we are happy to serve the market. We have capacity. We can serve the market, no problem. But I think realistically, and giving you the most realistic estimate, I would not expect an uptick in terms of revenues, maybe orders towards the end of the year, but I don't think there's a big uptick in shipments in '26.

Got it. And then just a follow-up in terms of your sort of the order momentum you're expecting in the first half of this year. When you sort of look at the discussions you've had year-to-date, how should we think about the mix in your order book? Is it sort of largely opto based in H1? Or could we see some GaN tool orders coming and inflecting in H1 potentially for shipment in the second half of this year? How should we think about that mix in the order book?

I would expect, if I look at ongoing customer discussions at this point in time, again, there can always be surprises, but I'm just extrapolating what kind of discussions are ongoing. And we know then the discussions take between, I don't know, 1 and 3 or 4 months to materialize, which kind of covers the H1 quite well. I would expect in H1 a significant optoelectronics/LED loaded order intake, whereas then in the second half, I would expect the power electronics gradually to come back.

Moving on to the next question from Michael Kuhn of Deutsche Bank.

I'll stick with, let's say, order composition. Of the roughly EUR 260 million order book you currently have, I think you gave some indications already. But could you maybe give some deeper insight into how the composition is by category, power versus non-power and maybe even going into a little more detail?

Yes. I think if we look at the order backlog of '25, I think opto is around 40%, SiC 30%, GaN 20%, LED 10%. Do you think so, Christian?

Yes. That makes sense. Approximately.

Approximately, right?

Yes.

Understood. And then on GaN and let's say, the next upward cycle, you mentioned at some point in the presentation that you expect AI data center power to drive the tool demand by factor 3. What would be the comparison base for that factor 3, just to get a better idea on how big the market could grow?

I think we look here at the comparison, the total market size is more like around '24, '25. And the factor of 3, which we've illustrated more like an upside scenario comparing '25 versus 2030 kind of a 5-year comparison, one point in time, '25 versus 2030. I think this is what we have looked at right now.

Okay. So this is -- '30, this is nothing like 4 in 2 years' time, at least from today's point of view?

No, no, no, no. I think this is a gradual increase. As we have discussed in this call already, we believe at some point in '27, there could be the first momentum starting and then it's a design in. And as always, in our applications, our markets, it's a ramp. It's a new trend, which is then happening. It's getting designed in. So our customers, our IDMs have now made devices, which is in the qualification with their customers, board makers, GPU makers, rack makers and so on and so forth. The architecture has been set.

Now the complete industry is working on it. Hopefully, it's going to be fast. We know the AI industry is a very fast-moving industry. So maybe it's faster than some of the other industries. But then at some point, it's being designed in and then the volume is starting and then gradually over time, it gets penetrating and the adoption rate goes from today 0% then whatever, 10%, 20% in the initial stage and at some point, 2030, 100% adoption rate after the adoption is completed, and then we look at that point in those numbers. So a gradual adoption.

Again, still our assumption. You never know how the adoption goes. Sometimes things go very, very fast, would be nice, but that's the assumption which is under.

All right. Understood. And then one more question. Obviously, we are not yet there. But let's say, the -- say, cycled as well and you're ramping capacity big time. When would you reach, let's say, your current capacity towards 100%? And when would you consider, let's say, reactivating your Italian capacity that is currently mothballed and what would be the potential cost associated with that? Or is that not even a planning scenario as of now?

Honestly, it's not relevant for the overall business or profitability. I would say capacity can always be scaled up in one way or another, which way we choose to take, we will decide when we are there. But I think it's nothing that affects the P&L in one way or another. It's not a constraint. It's not a limit to us. It's not a profitability limit or inhibitor or whatever it is, it's just operations.

The next question is from Nigel van Putten from Morgan Stanley.

I just wanted to follow up on some of the customer behavior in the optoelectronics end market. I mean, some of them have said that they're currently ramping supply. They see demand ahead of supply, maybe even towards next year. But do you feel that comment is directed at you?

When you speak to customers, do you have to disappoint them? Are you shipping to, let's say, 80%, 70% of demand? You've mentioned, as an example, a customer that comes in with a shipment for 5 tools as quickly as possible. Are you still able to serve those type of requests? Or do you have to sort of disappoint them and saying, well, that's going to be quite a bit longer than maybe the 6 to 10 lead time month lead time you've indicated before?

Well, in this case, good for our optoelectronics customers. The silicon carbide customers are so nice to step to the side for them in this year, leaving a lot of unused capacity, both in our shop floor and within our suppliers. And as you know, we work on a -- how do you say, modular system with our Planetary systems. So all our products are closely related to each other as a family, you can say. That is now a capacity that is not being emptied or not used by silicon carbide customers because that market is currently sleeping.

We can use the same supply chain for parts and of course, also the same kind of assembly tools on our own shop floor and the skill set of our people now to do the labor part. In other words, we have free capacity to literally serve all the demand, which is currently coming in. It might be a different game if the silicon carbide would be at the same time in the party now. But silicon carbide, as we have illustrated, is really leaving the gap, and this gap is currently just now being taken by the laser guys. It's good for them.

Got it. So when they say we can't ship, it kind of reflects your lead times, you think? Or especially when your customers...

It should not be us who's the bottleneck. Yes, it should not be us who's the bottleneck. And my team, my operational team, my sales team is handling it. I expect if there would have been a bottleneck, I would know it. I'm not aware of any bottleneck across the entire industry.

Perfect. That was my question. But then maybe a broader question. You said larger wafer size and better yield. I think one customer said it's 6 inches is 4x the product of the 3-inch, which -- or yes, the current capacity. So maybe ballpark to give us an idea in terms of the capacity you're shipping this year relative to the installed base, what do you think the increase is you can serve with sort of your view on the revenue you're shipping into '26?

Well, that's a very, very difficult question. I can only illustrate to you the various factors to that because the installed base is -- first of all, many, many tools, but many of them still on 3-inch and 4-inch wafer size, as you said, which is a much, much smaller capacity in terms of square centimeters or number of chips that you can get out of it in other ways.

The other point is, that while the installed base counts many, many tools in the installed base, many times those old tools, they would be dedicated to one product and they would only be qualified for certain products, so with huge inefficiencies. So I think we are currently like the shift in new architecture towards photonic integrated circuits to the PIC on indium phosphide, also much bigger chips, much more functionality is really -- it's a world which is not comparable to the old world I would say. Because it's different chips, different products, a much larger wafer size, much higher productivity. So I think the industry is really seeing a massive momentum.

But on the other hand, as illustrated, inside of the data centers, even inside of the racks, we go completely away from electric cables and go completely to optical data connects, which inside of the racks is really new to the industry. So the demand is massively increasing.

The next question is from Adithya Metuku from HSBC.

Just firstly, just thinking about the capacity that's coming on board for indium phosphide lasers. From what I understand, the yields are something like 50% and that the continuous wave lasers used in CPOs are about 1/10 of the die size of EML lasers. So I just wanted to hear your thoughts on how you think about the yield improvement, especially if the die sizes go down. That combined with the die sizes going down with the existing capacity that's in place or you will have put in place by the end of 2026.

I suppose the question is, it's been asked, but how much does the capacity go up? And will there be enough demand to drive further growth in your optoelectronics business in 2027 if yields go up, die sizes go down 10x because of continuous wave laser adoption. So any thoughts around that would be great. And I've got a follow-up.

I think you asked the billion-dollar question, but I don't have the answer for you, unfortunately. I think the effect you're alluding to is a typical pattern across the whole semiconductor industry that in a new market segment, you start with a relatively low yield simply because the application is there, the application needs the capacity, the application needs a ramp. But then over time, new generations of products step-by-step come in, which come with a die size shrink and higher yields, means you get more capacity out of your installed base.

Typically, such a process, so I cannot -- upfront, I cannot quantify this for you. This is -- I don't know. I think also our customers at this point in time don't know. Typically, this process that you are describing is happening over a 2.5, 3, 3.5, 3, 4-year time horizon because it takes one generation of chips and after the next generation and the next generation, typically, at least you need 1.5 to 2 years for one generation after the next, because your customers are simply not able to digest a faster succession of generations and also to increase the yield takes some time.

So what it means is my personal guess, and again, it's only a speculation, but I can share the opinion I have with you is that this is not only a 2026 trend, but at least this trend in this market will extend into 2027, that I think is very, very clear. This does not happen within 1 year. Now to which extent and how large this will extend in '27 and '28? I think that's the billion-dollar question I cannot quantify for you. But I'm very convinced that we are not talking about 1 year, but at least about 2, and I would guess rather a 3- to 4-year time horizon.

Got it. So essentially, you are expecting growth in '27, but you don't know the magnitude of the growth at this stage. Would that be a fair way to characterize it?

That's a fair way. Yes, exactly. That's a fair way.

Got it. And then just following up on an earlier question, you talked about the epitaxy machines not being the bottleneck. To my understanding, it's the indium phosphide substrates. Is that right? Or is there some other bottleneck in the system that's preventing your laser customers from ramping capacity and meeting the demand that they're seeing?

I hear also that indium phosphide substrates is a bottleneck that's currently being addressed by the entire value chain. I know this both on the side of our customers who need the substrates in their factories, and I know it also from substrate manufacturers. And I'm aware that there is a large, very well coordinated and well-orchestrated initiatives by our customers and by the substrate makers together in place to address these bottlenecks. But yes, that's, I think, a topic which is currently being worked on in this value chain and in this industry.

Understood. And then maybe just one last clarification. Are you able to give any color on the divisional growth revenue expectations for the first quarter?

Honestly, I don't have the numbers.

And the last question for today from Malte Schaumann from Warburg Research.

My first one is on silicon carbide superjunction technology. So can you maybe share your view on how the time line until adoption might look like? And then associated to that, would your tools in the existing base require an upgrade to incorporate that?

A very good question. So we are aware that all the leading device makers are currently working on superjunction technologies. To my understanding, the first devices will be launched at the end of '26 by suppliers, means in the second half of '26 or the first half of '27, volume ramps of devices happening in the market.

And we think that superjunction technologies in silicon carbide will be strongly embraced by Western players because it's a major way for them to get more dies per wafer and hence, to reduce the cost per chip. So it's a massive trend, which is currently being strongly pushed across the entire industry.

As for our tools, there's no further upgrade needed for our tools. They are able to run as is. And one point I would like to illustrate, nevertheless, is that the superjunction technology where essentially you don't take one thick layer, let's say, 10, 12, 14 microns of thickness, but you rather split this into 3 or 4 thinner layers and the wafer gets put into the tools multiple times.

Most customers embrace a technology, which is called multi-EP, multi-implant, so you do an epi step to do an implant, you do another epi step, another implant. So the wafer gets several times into our tools, a little bit like what we saw in the indium phosphide just earlier in the discussion. And that means that for one wafer of superjunction devices, you need more epi time.

You need more tool time in the epi, and we expect that this will be also one driver at some point, as illustrated in the '27, '28, '29 cycle, which will trigger additional demand from our customers for more tools because they need to expand their epi capacity in order to accommodate all the superjunction MOSFETs. So it's a market trend that we like a lot because it helps our business.

Okay. Understood. Secondly, on working capital. With the shift in the product mix away from power to opto this year, can you keep your inventory target? I think it was around EUR 200 million by the end of '26.

And then secondly, with respect to the down payments, we have seen quite a significant decline over the past few years relative -- down payments relative to order intake. So what are your thoughts where these levels should normalize going forward?

Yes, good question. So inventories, yes, we expect inventories to go further down. The shift in product mix, in fact, is an effect which is not helping. So we are still -- but we are still targeting EUR 200 million to EUR 220 million in terms of inventories. So maybe there's 20 more than we initially expected due to the shift of product mix, let's see. But still, we target a significant further reduction of inventories. It's gradually burning down, maybe a little bit slower as you're indicating, but still significant.

Christian, maybe you can take the second part.

On the down payment, it's a little bit more difficult because we don't have complete control on it. It really depends on end market mix, regional mix, customer mix and also cutoff date effect. I mean, the number at the end of the year was really low. We expect it to recover to some degree, but to predict this in detail is quite difficult. And it's also not the major negotiation point with customers, right? It's part of the deal, but not the major part. So it's a little bit difficult to predict. It should increase trend once again.

Okay. Okay. Lastly, a quick one on R&D. You indicated an increase in R&D spending this year. Would you expect another increase with the rising business volume generally over the next years and '27? Or would that volume be more or less sufficient to support your programs you have in mind?

I think we discussed already earlier. So in '24, we had around EUR 90 million. In '25, we had around EUR 80 million. For '26, we expect again around EUR 90 million.

And then beyond '26, so the EUR 90 million is sufficient for the next few years...

It look -- that always depends a little bit on individual cycles of products. At some point in the cycle, the products take a little more money. At some point in the cycle, they take a little less, it depends throughout where the portfolio stands. Honestly, I wouldn't want to predict beyond that.

Thank you very much from my side. With that, there are no more questions in the queue. So I'm closing the Q&A session and handing the floor back over to Ludwig.

Well, thank very much. Thank you very much all for your questions. The IR team and part of the management team will be on the road in the next couple of weeks, so we'll see a lot of you, hopefully, in-person.

And for those we do not see, we will have our next quarterly call scheduled for April 30, when we will report our Q1 figures. So if we don't see you until then, then have a happy Easter and talk to you end of April. Goodbye, and thank you.

Bye-bye.

Ladies and gentlemen, welcome to AIXTRON's analyst conference call Q3 2025. Please note that today's call is being recorded.

[Operator Instructions] Let me now hand you over to Mr. Christian Ludwig, Vice President, Investor Relations and Corporate Communications at AIXTRON for opening remarks and introductions.

Thank you very much, Gunner. A warm welcome also from my side to AIXTRON's Q3 2025 Results Call. My name is Christian Ludwig. I'm the Head of Investor Relation and Corporate Communications AIXTRON.

With me in the room today are our CEO, Dr. Felix Grawert; and our CFO, Dr. Christian Danninger, who will guide you through today's presentation and then take your questions.

This call is being recorded by AIXTRON and is considered copyright material. As such, it cannot be recorded or rebroadcast without permission. Your participation in this call implies your consent to this recording.

Please take note of the disclaimer that you find on Page 1 of the presentation document as it applies throughout the conference call. This call is not being immediately presented via webcast or any other media. However, we will place a transcript on our website at some point after the call.

I would now like to hand you over to our CEO for his opening remarks. Felix, the floor is yours.

Thank you, Christian. Let me also welcome you to our Q3 '25 results call. I will start with an overview of the highlights of the quarter and then hand over to our CFO, Christian, for more details on our financial figures. Finally, I will give you an update on the development of our business and our guidance.

Let me start by giving you an update on the key business developments of the second quarter on Slide 2. The important messages for Q3 '25 are our free cash flow in the quarter was EUR 39 million, totaling EUR 110 million in the first 9 months '25, while inventories are down to EUR 316 million, coming from EUR 369 million at the year-end '24. This shows we are well on track with our strategy to rebuild our cash position after we had depleted that with the construction of our 300-millimeter cleanroom, the innovation center in the years '23 and '24.

In Q3, we recognized new orders of EUR 124 million, which lead to an equipment order backlog of EUR 287 million, where we have achieved a book-to-bill of 1.04. We concluded the quarter with revenues of EUR 120 million. With that, we were in our guided range of EUR 110 million to EUR 140 million.

The gross margin reached 39% in Q3 and averaged 37% in the first 9 months. This figure includes a one-off expense related to our implemented personnel reduction earlier in the year. Adjusted for this effect, the gross margin after 9 months came out at 38%, slightly below previous year's 39%, mainly due to volume shifts and FX headwinds. As the market remains soft, we had to adjust our fiscal '25 guidance 2 weeks ago. We are now expecting revenues in the range between EUR 530 million and EUR 565 million, which corresponds to the lower half of the initial guidance of EUR 530 million to EUR 600 million, and a gross margin of now 40% to 41%, down from previously 41% to 42%, and an EBIT margin of now around 17% to 19% from previously 18% to 22%.

AI continues to be the main end market driver, especially for our Optoelectronics segment. Automotive-driven power electronics demand, on the other hand, remains soft.

Christian will now provide a detailed look into our financials on the following pages before I take over with an update. Christian?

Thanks, Felix, and hello to everyone. Let me start with the key points of our revenue development on Slide 3. In a soft market environment, we achieved revenues of EUR 120 million, down versus the EUR 156 million last year, but well in the guided range of EUR 110 million to EUR 140 million. For the first 9 months, revenues came in at EUR 370 million, down about 9% year-over-year. A breakdown per application shows that 66% of equipment revenues after 9 months come from GaN and SiC power, 14% from LED, 16% from Optoelectronics and a 5% contribution from R&D tools. The aftersales business contributed to total revenues with EUR 80 million. The aftersales share of revenues after 9 months was up by 2 percentage points year-over-year to about 22%.

Now let's take a closer look at the financial KPIs of the income statement on Slide 4. I already talked about the revenue line. Gross profit decreased year-over-year in Q3 '25 to EUR 246 million. Gross profit in the quarter was negatively affected by approximately EUR 8 million due to volume shifts from Q3 into Q4 and around EUR 2 million due to FX effects. Subsequently, the gross margin in the quarter came in at 39%, down 4 percentage points versus the prior year. After 9 months, gross profit was at EUR 136 million, 15% below last year's figure. At 37%, our gross margin after 9 months was 2 percentage points lower than after the same period last year. But please recall, as stated in our Q1 release, this includes a one-off expense of a mid-single-digit million euro amount in connection with the implemented personnel reduction in the operations area. Adjusted for these effects, the gross margin after 9 months would be around -- at around 38%.

For the remainder of the year, we calculate with an average U.S. dollar-euro exchange rate of 1.15 and the continued weakness of the Japanese euro rate. Due to high expected revenues in foreign currency in Q4, we expect an additional around EUR 3 million negative impact in revenues and gross margin with the larger part resulting from the U.S. dollar and the smaller part from the Japanese yen. Together with the above-mentioned EUR 2 million effect realized in Q3, this totals to approximately EUR 5 million negative FX impact, which corresponds with the 1 percentage point gross margin adjustment of our guidance.

OpEx in the quarter were slightly up by 4% year-over-year to EUR 31 million, primarily driven by higher R&D spending compared to the previous year. For the first 9 months, OpEx came in at EUR 94 million, a reduction of minus 6%, driven primarily by around 13% lower R&D expenses. R&D expenses were down mainly due to reduced external contract work and consumables costs. As stated before and visible in Q3 numbers, R&D costs in H2 will be higher than the H1 number. So for the full year, we expect R&D costs to be slightly lower than in 2024.

EBIT for the quarter is EUR 15 million, a significant drop versus Q3 2024. The main drivers besides the already mentioned negative factors impacting gross profit is a negative operating leverage effect resulting from lower revenues. The weaker performance in Q3 led to an EBIT of EUR 42 million for the first 9 months, a decrease of 30% year-over-year. This translates into an EBIT margin of 11%. Again, please record the one-off expense in connection with the personnel reduction I've mentioned before. Adjusted for this effect, the 9-month EBIT margin would be around -- at around 12%.

Now to our key balance sheet indicators on Slide 5. On a more positive note, working capital has continued to come down -- has come down by around EUR 100 million since end of fiscal year '24. Several balance sheet items contributed here. We continued to decrease inventories to EUR 316 million compared to EUR 369 million at the end of 2024.

Year-over-year, inventories have been reduced by EUR 111 million as we continue to work through the surplus accumulated last year. And as stated before, we expect further inventory reductions to materialize throughout 2025 and into 2026.

Trade receivables at the end of September were at EUR 129 million compared to EUR 193 million at the end of 2024. The reduction versus year-end is mainly the result of the collection of the payments related to the large shipments end of 2024. Advanced payments received from customers at quarter end were at EUR 73 million, a nice recovery of about EUR 20 million versus end of last quarter, but still down about EUR 9 million from end of 2024. This is primarily driven by some cutoff date effects and some regional shifts in the order book. Advanced payments now represent about 25% of order backlog. The fourth key element of working capital, trade payables, has now come down to EUR 24 million from EUR 34 million at the end of 2024. This reflects a now fully adjusted supply chain situation with significantly reduced purchasing levels.

Adding it all up, our operating cash flow after 9 months improved to EUR 128 million, a strong improvement of EUR 100 million versus last year's EUR 28 million. On the back of the improvement in operating cash flow, free cash flow improved even more. It came in at EUR 110 million after 3 quarters compared to negative EUR 58 million last year. This was supported by a strong reduction in our CapEx. With EUR 18 million after 9 months, our CapEx was significantly lower than last year's number of EUR 86 million. This is primarily due to the now completed investment in the innovation center.

As of September 2025, our cash balance, including other current financial assets improved to EUR 153 million. This equals an increase of EUR 88 million compared to EUR 65 million at the end of fiscal year 2024, despite the dividend payment of about EUR 17 million in Q2.

As stated before, a key priority remains the rebuilding of a strong cash position. Our financial decisions continue to be guided by this objective to ensure a robust liquidity foundation for the future. This has served us well in the past, and we see ourselves well on track towards this target.

With that, let me hand you back over to Felix.

Thank you, Christian. Let me continue with an update on key trends in our different markets, starting with optoelectronics and lasers. In optoelectronics, AIXTRON has seen a continued recovery in demand for datacom applications, which began earlier this year and has been reaffirmed in Q3. This trend is expected to continue into '26 and beyond.

Our customers are increasingly transitioning to 150-millimeter indium phosphide substrates and photonic integrated devices, PIC devices requiring advanced epitaxial performance. This segment is technology-wise very demanding. It requires excellence in the uniformity, doping control and defect management, areas where our G10-AsP platform excels.

Historically, AIXTRON has held a market share of over 90% in this domain served by our G3 and G4 planetary reactors. The G10-AsP is now establishing itself as the tool of record to the laser market, replacing legacy systems at leading customers. Q3 shipments and scheduled Q4 deliveries underscore our strong market position with repeat orders from key customers such as Nokia. Additionally, VCSEL demand is recovering, driven by LiDAR modules and automotive applications. We, therefore, expect that tools for the various laser applications will contribute significantly to our full year order intake and also into next year '26.

Now let me move on to our LED business. We are seeing first encouraging signs of reinvestment in red, orange, yellow -- ROY LED applications. Utilization rates for red, orange, yellow LEDs have been high throughout the year with double-digit system shipments for mini LED applications driven by demand for RGB fine pitch displays. Notably, some TV manufacturers such as Samsung are shifting to full RGB backlighting, boosting micro LED demand. While overall micro LED demand remains moderate, medium-term drivers are positive. We've received multiple orders for our G10-AsP platform, primarily for red pixel production in next-generation AR devices. The recent announcement of Meta's AR glasses based on micro LED technology signals a broader trend with more OEM products expected in '27 and '28. Our G5+ and G10-AsP platforms are ideally suited for these applications, which require ultra small pixels and defect-free epitaxial die. The launch of Garmin's first micro LED watch is likely to further stimulate demand across blue, green and red micro LED segments.

In solar, after years of moderate investment, we are now seeing renewed interest, including multiple orders for low earth orbit -- LEO satellite applications in constellation projects. LEO satellites are those that orbit the earth at altitudes of about 2,000 kilometers. They enable both fast communication as well as high-resolution earth observation by operating in a zone just above the earth's atmosphere, where they can maintain strong signal connections with ground stations. These satellites work in interconnected constellations of hundreds of thousands of satellites of hundreds or thousands of satellites to provide global coverage, examples are Starlink or OneWeb. We anticipate this trend to continue in the years '26, '27 and '28.

Let me now come to gallium nitride power. AIXTRON continues to lead GaN power segment with over 85% market share across all wafer sizes and power ranges. Although demand is softer compared to last year, we are seeing solid volume orders for both 150- and 200-millimeter solutions, particularly from Asian customers with ramp-up plans extending into '26 and '27.

We've also strengthened our partnership with imec. Together, we are accelerating innovation at both the architecture and device level. imec has been using both our G5+ as well as the G10-GaN platform for its 150- and 200-millimeter partner programs for quite a while. And we have now shipped a 300-millimeter gallium nitride platform to enable broader access to imec's recipes. We see first power semiconductor manufacturers adopting 300-millimeter GaN technology such as Infineon Technologies.

Regarding the overall GaN market, we are still dealing with a moderately oversaturated installed base, requiring some more time to absorb existing capacities. This digestion phase is expected to continue for some quarters before a broader recovery sets in.

With that, let me come to silicon carbide. While end-user demand remained soft, we observed moderately increased utilization rates at some of our customers. On the one hand, this is due to new EV models being launched, which drive demand. On the other hand, SiC is starting to enter the AI data center value chain, especially in voltage classes of 1,200 volts and above.

You have seen the new NVIDIA power architecture, which relies exclusively on wide band gap power devices. At the International Conference for Silicon Carbide and Related Materials -- in short, ICSCRM in Busan, Korea early in Q3, various industry players confirmed midterm adoption of super junction silicon carbide technology. This technology basically means that instead of one thick silicon carbide epi layer deposited today, we will see in the future multiple thinner silicon carbide epi deposition steps. These thinner epitaxial layers require enhanced uniformity and shortened process time.

Our G10 silicon carbide platform is well positioned to meet these needs, offering superior productivity due to the benefit of the batch concept, especially for thinner layers. We are proud to have shipped our 100 G10-SiC CVD system, marking a major milestone and reinforcing our leadership in the silicon carbide power segment in this quarter.

The silicon carbide market is still undergoing a longer digestion period, particularly in western-oriented regions. As a result, there are no major decisions for new fab investments on the agenda these days.

In summary, we can say that the soft market period still continues in almost all markets, apart from the laser market, driven by the hunger for data from AI applications. A demand pickup will not materialize in '25, and visibility in '26 is still limited.

With that, let me now move to our guidance. Due to the market situation just described, we had to adjust our guidance for 2025, 2 weeks ago. Based on the current soft market environment and assuming an exchange rate of USD 1.15 per euro for the remainder of the year, we now expect the following outlook for '25. We expect to generate revenues in the range between EUR 530 million and EUR 565 million, which corresponds to the lower half of the initial guidance, which was initially EUR 530 million to EUR 600 million.

FX effects led to an approximately 1 percentage point reduction of gross margin and EBIT margin. As a result, we expect now a gross margin of around 40% to 41% and an EBIT margin of around 17% to 19%. The guidance for the gross margin and EBIT margin includes a one-off expense of a mid-single-digit million euro amount in the relation to the implemented personnel reduction in the operations area earlier this year.

The measure will lead to annualized savings in the mid-single-digit million euro range in the future, which corresponds to an improvement in the gross margin and EBIT margin of around 1 percentage point. As previously stated, we expect our tools to remain exempt from U.S. tariffs. However, we continue to closely monitor the impact of U.S. trade policies on the global economy and stand ready to implement any necessary measures to ensure the best possible outcomes for our customers and stakeholders.

Let me, at this place, also give you a first outlook for the next year 2026. We clearly see that the medium and long-term drivers for AIXTRON's growth such as demand for GaN and SiC power devices, LED and micro LED applications, lasers and LEO solar applications remain intact. However, visibility for the fiscal year '26 remains low. And as of today, we do not see signs of a demand recovery yet. Therefore, our view today is that 2026 revenues are likely to be slightly below those of '25, maybe flat. Furthermore, assuming an exchange rate of USD 1.15 per euro, we expect the EBIT margin not to come out below the range of the current year, maybe better. As always, we will give you a firm guidance with the release of our financial year results end of February 26.

With that, I'll pass it back to Christian before we take questions.

Thank you very much, Felix. Thank you very much, Christian. Operator, we will now take the questions.

[Operator Instructions] The first question comes from Janardan Menon from Jefferies.

I just wanted to touch upon your final comments on 2026 to start off with. You said that 2026 is likely to be flat or down, but it sounded like you expect Opto to be up, and your trend -- when I look at your Q3, GaN seems to be doing quite well, while SiC is down quite sharply. So would it be fair to say that at current visibility, you would expect Opto to be up, SiC to be down and GaN to be somewhat flattish. Is that a view that -- which would be sort of a preliminary view for next year?

It's a good -- I think you got a perfect read on this one. Let me try even to quantify it for you. I think roughly in terms of percentage of revenues, we expect as a percentage of total revenues next year, we're expecting to gain about 10 percentage points for Opto, 10 percentage points gain for GaN and minus 20 percentage points in silicon carbide. So a pretty weak year for SiC, but very strong year for the Opto segment. It used to be a smaller segment. So adding 10 percentage points of the total is quite a significant one. This also helps on the margin. You have seen my comment related to margin quality. And GaN also as a percentage gaining a bit.

Just a follow-up. On the SiC side, yes, I understand that demand is quite weak right now. There's quite a bit of supply out there and automotive is still sluggish. But listening to companies like STMicro and all who are under quite severe margin pressure on the silicon carbide side, they seem to be accelerating their 6-inch to 8-inch transition because they see that as a way to improve their profitability. And ST specifically said that they'll do it within -- through the course of '26 and by early '27. I would assume that that would be true for other parts of the installed base as well given the price pressure on silicon carbide. Do you not see this as a driver at all for your silicon carbide revenue? And do you really need the end demand to recover before any improvement happens?

I think you catch it very well. Yes, the 6- to 8-inch transition is going very fast, especially at outside of China players. I think worldwide outside of China, we see the 6- to 8-inch transition progressing at rapid speed, as you have indicated with one company name, and we see the same in other players. In fact, we do hear from some of our customers that while end customer revenue is flat or down, the unit numbers are going up and unit numbers is, of course, what we as an equipment maker like, because in the end, it's about wafers and increasing numbers of wafers. So in fact, we expect that by the end of '26, the transition in the Western world, as I may call it now, including Japan, is probably concluded '27, '28, I would expect the volume to be completely going on 8-inch.

We do see on 8-inch also much better quality wafers, which helps the customers in terms of yield. That's one of the cost reduction drivers. Also 8-inch substrates are getting good pricing now. Initially, they used to be very expensive. Now the pricing for 8-inch substrates is going well. And that, at some point, means the excessive overcapacity that I was speaking about at some point will be digested. I would not dare at this point to give an exact prediction because there's multiple variables that we are just discussing. But I think we can clearly see at some point, the overcapacity will be digested and then there will be new demand.

But that transition doesn't mean buying new 8-inch machines from you, is it to generate revenue for you?

At some point, it will mean buying new demand and new tools when the existing overcapacity is consumed. Right now, we talk about existing overcapacity, which is just being converted.

Next up is Martin Marandon-Carlhian from ODDO BHF.

The first one is on something that you put on the press release on gallium nitride. You talked about utilization rate rising in data center. And I was wondering what does it mean exactly? I mean, does it mean that you already anticipate orders in the near term linked to the new 800-volt architecture from NVIDIA? Does that mean something else?

Let me explain what we mean by that. Thanks for the question. What we have seen is we have seen in the years, especially '23 and '24, we have seen quite a number of gallium nitride orders, which were happening a bit ahead of the wave, such that, I would say, early '25 at the existing volume customers, we have seen quite a significant overcapacity of installed base also in gallium nitride. That was the reason why in '25, compared to '24, our gallium nitride shipments have been slowed down quite a bit, because our existing and established volume customers literally had also in GaN, not only in SiC, but also in GaN, some overcapacity to be digested.

So as we started into '25 at some of our customers, also in gallium nitride, we have seen installed base utilization to be quite low. Now towards the end of '25 and looking into '26, we see that a much larger fraction of the installed capacity is being utilized at the existing GaN customers, while those who newly entered the GaN market in '24 and '25 in previous earnings calls, you may have recalled that we said -- well, there's still new players entering the market to gallium nitride. And those new entrants at this point in time are still in the qualification or in the device and the sampling phase of their technologies to their end customers.

You have seen the numbers that I was just commenting towards the question that Janardan was asking. We expect the GaN segment for us to be slightly up next year. Again, it's an indication, qualitative indication. as we see that utilization is increasing, and we expect due to the increasing utilization, some expanding orders from some customers kicking in. The broad market recovery, as I've indicated, with the real volume pull, we don't expect in '26. We rather expect that in '27, '28, but some increasing orders in '26. Does that answer the question?

Yes, that's very clear. But just a follow-up on this. I mean, why would you anticipate more of that volume in '27 and '28? Because we read that this new architecture from NVIDIA is supposed to be for Rubin Ultra, which is launched in H2 '27. So I was expecting capacity maybe to come a bit earlier than this. So does this mean that maybe it will not be 100% GaN for some steps at the beginning, the 50 and 12-volt steps and it will go gradually. I mean just can you explain a bit why it should come more gradually, let's say?

So this is based on our current view, what we have and the signals we get from our customers. I share the view that the new 800-volt architecture will lead to significant volumes around '27, '28. This is also our view, I share that. Now for us, it's always very difficult to predict the exact timing when customers will place the orders for new equipment because we do see certain trends, but we cannot look into the exact budgets and plans of our customers. Therefore, at this point in time, we can only comment on what we are currently seeing. If later on in the year, volume kicks in and orders accelerate, we are very happy to it. We don't see signs to that yet.

Great. And maybe a last question on GaN. I mean, you all is saying that the GaN market will be close to $500 million this year with that data centers really being really a contributor. What would you guess would be the size of the data center market for GaN compared to the overall size of the market this year, like $500 million?

So I do not have the exact timing for my message in mind. We have looked at a midterm perspective, I think somewhere triangulating '28, '29, '30, something a little further out. And in this triangulation that we've done, the data center opportunity with an upside of about 50% on top of the market without the data center opportunity.

You may recall that we have a slide out there in the investor deck, which on the X-axis has 3 time horizons. I think '20 to '23, I think '24 to '26 and whatever '28 to '30, something like this. And on the Y-axis, the different voltage levels, low voltage, medium voltage and then very high voltage. And there, we have put the AI data center opportunity, and this is the market that I'm referring to.

Maybe last question for me on the gross margin. I mean the current guidance implies record gross margin in Q4. Just can you help us maybe see the main drivers of this?

Yes. Martin, Christian here. I'll take that one. I mean, like in the last years, the Q4 will be the strongest quarter just by volume, purely shipments. Beyond that, we expect an improved product mix, especially a higher share of final acceptance revenues coming with high margins and also some fixed cost degression effects.

A little bit of color on the product mix. We expect a big share of G10 family products, around 50% of Q4 revenue so that you get an idea. So also looking at the -- comparing this with the last year, these margin ranges appear achievable for us.

Next up is Didier Scemama from the Bank of America.

I've got a couple of questions maybe clarification on the comments you made earlier on '26. And perhaps my math is not right, so please don't shout at me if I'm wrong. I think you said the SiC part of the business would be down 20 percentage points in terms of group sales. I mean, by my calculation, that would imply a pretty minor revenue contribution in '26. So is that correct? And then equally, Optos up, I think you said 10 percentage points within the group, that's going to put it at something like EUR 150 million next year. Is that the right ballpark?

I would say right ballpark, right indications, Yes. As far as we can say. I mean, it's very early, but we really want to give you some…

Yes, of course.

Yes, exactly, yes.

No, that's incredibly helpful to me perfectly honest. So I guess the question, when I look at the comments you put on the 9-month report, you said about 50% of the bookings came from power electronics. So I have to assume that the rest mostly come from Optos because LEDs, et cetera, is fairly de minimis, which if you compare to what you said last year, means that the bookings in Optos are probably up meaningfully, which is again consistent with what you said.

So perhaps when you look at history, Optos, like all the other segments have tended to be incredibly cyclical. So would you think that there is duration in that growth in optoelectronics beyond '26? Or do you think that the big CapEx cycle we see currently for silicon photonics and lasers is going to be as we've seen in the past, a big year and then it falls off a cliff.

I think you asked the trillion, the multitrillion dollar question, how long the AI bubble will last. I do not have the crystal ball for you, right? If I would, I might not be sitting in this place right now.

Okay. Well, yes, I mean, honestly, I wish you good luck.

I think it fully relates given the serious note, yes. Some joking aside, a big part of the laser part is, in fact, coming from the datacom, right? And the datacom, again, is driven by the AI and the AI data center build-out. So it's really hinges on that one, to a very big part, probably 50%, 60%. So it really depends on how exactly that's progressing. But we can only see what we have now in our visibility. But a longer-term view 2, 3 years out, I think it's as difficult as for everybody predicting the AI trend.

No, for sure. And if I may, as a follow-up, I mean, you mentioned Nokia/Infinera as a customer for your G10 platform for their peak products. Can you give us a few more examples of key customers for that division so that we understand the underlying dynamics, please?

Unfortunately, I cannot, because we keep customer names always strictly -- very strictly confidential as under NDA. We stick to that. We are extremely sensitive to that. I can give you a qualitative indication. Imagine you think who may be the top 10 providers for data communications devices for AI, you can assume that at least 80%, 90%, maybe 100% of those guys are our customers currently placing order with us and 90% of those are placing orders for the G10-AsP. Maybe I can give you that indication. And I really mean it as I say it.

Now we're coming to the next question. It comes from Madeleine Jenkins from UBS.

I just had one on utilization rates. You mentioned that the GaN power were increasing. Could you quantify that at all? And also, I guess, get a sense of what your silicon carbide utilization rates are at kind of Chinese and then Western customers?

So I understand your question about detailed utilization rates. We don't have those. And we could also not share them if we would have them. But what we can say is that based on spare part orders, based on service orders, we see a trend here, which is a good utilization increase for the GaN power, which leads us to expect some volume expansion orders in '26 at a moderate level as we have indicated.

At the same time, in silicon carbide for the overall market, I think towards the beginning of the year, we have seen very low utilizations with very low -- I mean, clearly far below 50% means far more than 50% of the capacity installed in the market was standing idle early in the market. And maybe we are now approaching a 50%, 60%, 70% utilization in silicon carbide. So we do see it increasing, but we are still far from a level on a market level where customers are really going into reorders and expansion orders. I think that's not yet on the agenda.

Then I guess all your kind of new orders in silicon carbide specifically, are those kind of new customers in China? Is that the right way to look at it?

Yes. We did have significant orders and shipments in '25 in silicon carbide into China, quite a diverse set of customers, highlighting the success of our G10 silicon carbide platform. So I think we've managed to establish that platform very well in the China market. That was all relating to the earlier question by Janardan. That was all for 8-inch or having 8-inch in mind. However, we are all aware of the large overcapacity in silicon carbide in China. Also the China silicon carbide business at this point in time has slowed down. I think the market overall is digesting the existing overcapacity. However, I think we all see the very nice success of Chinese electric vehicles. At some point, the overcapacity will be digested and there will also be new orders.

Then just a quick final question. Do you have a sense of kind of how much of your current gallium nitride revenues this year, let's say, are for data center applications?

That's honestly very difficult to predict. Sorry for having only a vague answer, because our gallium nitride customers, I think we all have a couple of very big names, leading power electronics makers in mind, right? They use our platform essentially our tools, essentially for all the applications across the board.

On our tool in the same configuration, you can produce a 20-volt, 100 volt, a 650 volt and even if you want a 1,200-volt device without any change in configuration. And therefore, we, as a maker, just send the tool as it is and the customer can do whatever the customer wants with it without a modification in those power ranges. Therefore, it's for us very difficult to predict. If there would be a different configuration by voltage range, then at least we would have an indication. But therefore, it's difficult for us to say. Sorry for that one.

Silicon carbide is different, right? 6- to 8-inch, right? It's always the customer needs a configuration and we see spare parts orders or parts orders, and we can at least give you here in the call a qualitative indication for the GaN, it's really one size fits all. And yes, customer takes it and then we don't know.

Next up is Ruben Devos from Kepler Cheuvreux.

I just had a follow-up on silicon carbide. I think you touched upon it already, but it was around your comments on benefiting over proportionally when the cycle would return. I think you talked about a more diverse set of customers. So that might be an explanation, right? But just curious around what degree of confidence you have, right, to make that statement of outgrowing the market. And even outside like automotive, how does the pipeline shape up thinking about industrial as well in silicon carbide?

Thanks a lot. I think your question hints very well towards the future direction of silicon carbide. Let me go a little deeper to expand on it, maybe some of the backgrounds, the technical backgrounds are interesting. So the first generation of silicon carbide devices, which we have seen, I would say, in the last 5 years with a very simple MOSFET consisting essentially of just one thick layer, one thick epi layer.

Now what I mentioned, the next generation of devices, which to the expectation of all market participants will be the main volume in the next wave. Everybody expects the next wave of growth, '27, '28, exact timing to be TDD to be super junction MOSFETs. So this is a device where this thick layer is split into 3 or 4 thinner layers. So each of them about 1/5 or 1/4 thick of the initial one. And it's not just one big epi, but the wafer would be put into a tool 4 times. So you make 1 thin layer, then you do some device processing and then the wafer returns to the silicon carbide epi tool comes the next thin layer and so on multiple times. And this super junction technology shifts the operating point from one thick layer, which, let's say, has in the past been deposited, let's say, in about 1 hour to 2 hour processing time, now into multiple thinner layers and depending on which type of equipment, let's say, it now takes 15, 20, 30 minutes instead of 1 or 2 hours. So the wafer gets into the equipment multiple times. And with that, the complete dynamics about the productivity of the tool, the key KPIs and so on is shifting because essentially, it's a very different operating point.

You can buy -- in an analogy, you can buy a car which is perfect as a city car, small and nice and fits into parking lots, but doesn't drive very fast, you don't care. And a perfect travel car for long-distance travel or a nice sports car for going up the mountain pathways or driving races, right? And each of the operating points has a different optimum. And this new operating point about thin layers to our calculations and also to the feedback we receive from customers is very beneficial for the batch tool which we are offering. This is the reason why we've made these positive earlier statements.

With that, let me come to the second part of your question. The other part of the market, which may provide further growth, I think it's still a little further out than '27, '28 is the market for industrial applications. That market could probably towards the end of the decade grow very big. What we are talking here is about the following.

Today, we use the silicon carbide devices mainly in switch mode power supplies or like power devices for the car in the main inverter and in voltages, 650 to 1,200 volts. We can also make silicon carbide devices, which have 3,000 volt or 6,000 volt or 10,000 volts, much, much higher voltage classes. And the industry is working on. That was, for example, one of the elements in the NVIDIA power architecture. I think everybody here in this call has the chart of the architecture. If you look at the chart of NVIDIA, on the very front end, you come from the grid and you enter the grid into the data center at voltages around 14 kilovolts, and that's 14,000 volts. And this down conversion from over 10,000 volts eventually down to 1,000, this is done by silicon carbide and then from 1,000 to 1 is done by gallium nitride.

Now you cannot only use the silicon carbide in the data center for these high voltages, but in the entire grid. And we all know as more and more renewables are being used worldwide, I think China leads the pack with driving down the cost of solar and wind, but the whole world is following. And we need much more active grid stabilization, load management, active management and so on and so forth. So the grid, the worldwide power grid will experience over the next 2 decades, massive investments into switching infrastructure.

Today, this is all being done by transformers. I think everybody knows next to the highway like these transformer stations standing. In the future, many of those will be done by active switching, and this will all be done by silicon carbide power devices.

So all the leading grid suppliers, whether this is Siemens and ABB, Schneider Electric, General Electric in the U.S. are working on such devices. And it's a nice end segment for silicon carbide to come. However, I think this is a longer-term trend. I would not put the years '27, '28 on it. I would rather put '29 onwards as a nice trend for the turning of the decades on this trend.

Just my second question related to optoelectronics, basically. I think you've called co-packaged optics as a key driver for indium phosphide adoption. How quickly would you expect the market to move there from pilot into volume co-packaged optic deployment? And you've very helpfully framed the tool market size for silicon carbide and gallium nitride in your slide deck. So may I opportunistically ask whether you've done a similar exercise for the G10 arsenide phosphide platform.

Thanks a lot. I take the suggestion. It's a good one. Let's take that on our action item list that he smiles around me here in the room, yes. It's a good one. We don't have it yet for today, so I cannot give it to you maybe in the next earnings call.

Now to your question about the sizing and what we see. For the optoelectronics market, unfortunately, it is much more difficult to predict than for the GaN and for the silicon carbide market. Let me try to illustrate to you why. In GaN and SiC, we talk at least for the low volume segment for pretty standardized segments and types of devices, right? For GaN, we talk 20 volt, 100 volt, 650 and then exotic 1,200. Silicon carbide, 650, 1,200 and now I was talking a bit about the very high voltages. So you can put it into 2 or 3 classes. Unfortunately, the optoelectronic market is extremely fragmented. We both see that in the number of players. I don't know there may be a couple of hundred optoelectronics producers and companies, while in power electronics, we talk probably about like maybe a dozen or 2 dozen, 3 dozen maybe at most, yes. So it's extremely fragmented. And such are the different technologies, which is competing with each other.

The good thing is this is physics. They all have in common. As of today, they need a wide band gap semiconductor, gallium arsenide or indium phosphide for generating the light. But then the way the light is being processed, whether this is on an indium phosphide or gallium arsenide-based photonic integrated circuit or whether the light coming on is put into a silicon photonics. You can use silicon -- silicon dioxide waveguides and switching devices. This is extremely diverse and therefore, very difficult to predict.

I wouldn't dare at this point to make a prediction where it goes. We are aware that all the guys who are working on the leading-edge CMOS nodes and also doing heterogeneous integration, all of them work on multiple technologies because even for the big guys in the industry, things at TSMC, it's difficult to really say, well, this technology is winning out against the others.

Next up is Andrew Gardiner from Citi.

I just had one on the margin outlook into next year that you provided us, Felix, saying that you thought EBIT margin next year would be in line, perhaps better year-on-year. Can you just sort of give us some of the drivers there in terms of gross margin? I mean, obviously, you've given us the mix in terms of Opto and GaN up and SiC down. How would you sort of quantify that in terms of magnitude of gross margin change next year? And also, you've done a sort of a workforce reduction earlier this year. Given the still slow market in SiC, do you see any need to continue to reduce OpEx? Or are we far enough through this down cycle now where you just sort of have to -- you weather it because you can see the long-term opportunity. So really there's not much change -- incremental change in terms of OpEx into next year?

Yes. Thanks a lot for the question. I think part of the answer you've given, let me try to give an end-to-end consistent picture. So we were referring to EBIT margins really to bottom line. I have not given indication on the gross margin, no quantitative, right? So I was really mean EBIT margin. And I think there's three drivers behind our indication towards. So we wanted to give you a very clear indication that the margins is not getting worse despite the top line suffering probably a bit. And I think there's three drivers behind it.

On the one hand, we see margin-wise, a bit stronger product mix. I indicated the gain of Opto, that helps a lot. And secondly, we will see the full year effects of the headcount reduction, which we conducted early in '25. '25, there's also cost and restructuring costs. In '26, we get the benefits of that. And the third topic is we use the slow period of the cycle right now for some operational improvements, be it working on our storage topics, be it working on logistics topics, be it currently working on our operational efficiency.

So we have quite a bunch of these things ongoing, which are just making our operations more fluent, which reduce the external spend that's going out the door all the time. And we expect some of those effects to kick in. And based on those 3 effects altogether, we expect, yes, in terms of absolute terms and a stable bottom line or percentage-wise, stable or even improved bottom line despite the probably slightly weaker top line. But I think that's important in the end for you guys also then to everybody here in this call to give an indication where does it lead on the profitability.

The next question comes from Adithya Metuku from HSBC.

Firstly, I just wondered if you could give us some clarity on what drove the push out this year, which end market drove the reduction in outlook for the year?

Sorry, I didn't -- acoustically, the line was very bad. I didn't get the question. Could you repeat it, please?

Sorry, apologies. I was just wondering if you could give us any color on what drove the reduction in guide in 2025? Where did you see this push out, which end market?

Okay. Sorry, I get it. Honestly, this was all across the board, except for the laser market. I think the laser market we've indicated is strong and continues to be strong and this is growing into next year, as we have just discussed.

We have seen a weaker-than-expected GaN in silicon carbide. Initially, as we started into the year, it's always very difficult, right, to predict the full range. And we have put the full guidance range accounting early in February '25. So looking now 7 months back. In our full guidance range, we have accounted for both a slow market scenario, which now is unfolding. So therefore, we now look at the lower half of the guidance. And early in '25 with the upper end of the guidance, we have also taken into account a more positive market environment. As we all see, the more positive market environment for power semis for electric vehicles is not yet unfolding. So the upper half, therefore, had to be corrected now down to the lower half. We are narrowing down at the lower half of the guidance.

Then just on the LED and the micro LED market, you talked about seeing signals of improvement. I just wondered if you could give us a bit more color on what exactly you're seeing, especially on the LED side? Is it driven by China? Is it anything construction related? Just any color you can give us on these two end markets in terms of the signals of improvement.

Yes. Thanks a lot. So on the LED market, this is typically almost exclusively China-only market. I think we can say, because of cost and volume effects. We all know, right, China is very, very strong these days on the display making. It used to be, as you have indicated in your question, historically, there used to be a lot of the LEDs going into construction, right? In China, they put these big, big walls on the skyscrapers. But as we all know, the China housing bubble has collapsed, right? That was also the reason why the segment was bad for us for 2 years.

Now we are seeing the classical LED market coming back with, we call it fine pitch displays means and especially display backlighting. Local dimming, local backlighting of display, you can achieve magnificent effect by either having white LEDs behind your LED display, you can create a beautiful black or you can produce quite some nice bright colors on it with that one, and that's even going now into -- turning into RGB. The good news is it is revenue already today. The bad news is it makes it much more difficult for micro LED to gain ground in the televisions because the normal displays are already getting much improved quality. So let's see what it means for the micro LEDs.

The other point, which I was indicating, we still see that on micro LED, research work is ongoing. We've seen some first devices. I was relating in my prepared notes to the Garmin watches, which is the first micro LED watch coming out at quite high prices and unfortunately, with low battery lifetime. So we are seeing that coming. And we see a lot of companies currently doing work on AR glasses and VR glasses. You may have seen the glasses launched by Meta. There's much more stuff in the preparation. I think this is a new device category, which will really come into the market quite soon. And yes, we see some moderate demand for that also next year, as I've indicated in my prepared notes. But again, it's far away, to be clear, it's far away from the micro LED massive investment wave that all of us 2, 3 years we were expecting where we would expect that micro LEDs are penetrating everything from smart watches to notebook displays and televisions, right? That one we are not seeing yet. We still see the research ongoing. So some -- many companies are still working on it, but we don't have a clear in our view when exactly that's coming.

Just one last question. With TSMC getting out of the GaN market, I just wondered, do you see a market for secondhand tools for your GaN epitaxy tools? And would that affect demand maybe next year or the year after? How do you see the implications of TSMC getting out of the GaN market?

Honestly, I see it as a bit of a reshuffle, which happens normally in all the markets where there's a bit of a slowdown in the market. I think we see the same in silicon carbide, some players are exiting, some others use the opportunity to buy some used tools to get a hold of in or to get used tool and then newly to enter the market, I think it's a normal play that happens in a softer market environment. For the overall market and for us, this has essentially no implication because whether a used tool is installed or whether a tool is installed at company A or changes the ownership and is later on installed within the factory of company B, it doesn't change the overall installed capacity in the market or doesn't change the market dynamics. So for us as an equipment maker, we are -- we support customers when they need help in either way, sometimes for moving tools, for reinstalling tools, but it doesn't change or doesn't impact the market.

The next question comes from Michael Kuhn from Deutsche Bank.

I'll start with, let's say, the usual update on 300-millimeter GaN. I think it's quite well known that Infineon is quite advanced in that context. And obviously, no big surprise there, cooperating closely with you in that regard. So when should we expect tool orders to arrive and, let's say, outside Infineon, what's your view? How many companies are currently working on the transition and preparing orders?

So I think with 300-millimeter GaN, the market unfolds pretty much as we have expected. If you recall, we stated earlier that we see the 300-millimeter GaN as a subsegment of the overall GaN market, initially targeting the lower voltage classes means 100 volt, 20 volt, maybe 200 volt. Maybe at a later time, also 650, but really starting at the lower voltage classes. And we get confirmation from many customers what we had expected early on that customers are really targeting to switch and to reuse existing silicon MOSFET or silicon IGBT capacities and to rededicate existing fabs for gallium nitride. Of course, customers need to buy new epi tool because the silicon epi tool is a completely different tool from a gallium nitride epi tool. So in any case, there's a new tool demand for gallium nitride tools.

However, the market adoption and the customer decision to the largest part depends on the installed base of factories. So customers who have today their silicon MOSFETs running in a 200-millimeter silicon fab are likely to switch to a 200-millimeter GaN tool. Customers who today are running their silicon MOSFETs in a 300-millimeter fab will want to switch and rededicate their 300-millimeter fab to a 300-millimeter GaN fab.

So that is the market dynamic. And I think based on that dynamic, we never comment on customers unless we have a joint press release with customers. So allow me to describe the trend without names as we always try to do. So we really see customers who have installed 300-millimeter silicon capacity are switching now and starting to switch and have plans. There are many, many, many other customers who have 200-millimeter silicon fabs continue to work on gallium nitride 200-millimeter. And as a result of that, our strategy going forward is that we will support both groups of customers. So GaN 300 is not displacing GaN 200. We have our GaN 300-millimeter road map. We are very happy with the results that the 300-millimeter tool is giving. But at the same time, we also maintain an active 200-millimeter GaN road map where we also work on improvements. We have multiple very close customer collaborations on 200-millimeter tool improvements or even next-generation tools for 200 millimeters.

Then on cash flow and working capital, given that you don't expect top line growth next year, how much more would you think you can further optimize working capital? Because I think you mentioned you see further potential also into 2026.

Let's focus maybe on the inventories because the rest of the working capital is always a little bit arbitrary, the receivables and the down payments. But on the inventories, our key ambition is to drive them down further. It's a little bit difficult yet to predict, not knowing the exact product mix and so on, but like at first, like high level expectation would be another 20% down.

I would be more ambitious. Let's check. So I would say by the end of this year, I would expect inventory EUR 275 million, plus/minus EUR 15 million. To give you a number, let's see how close we come. Maybe next year, EUR 200 million. Let's see, something like this.

Let's see.

Looking forward to it. Maybe you can do a little bet between the 2 of you who comes closer.

There are no further questions.

Good. Perfect. And I think we had a lively discussion. We very much appreciate as you see. And yes, stay tuned. I think this is a good exchange. And I think we all see each other latest in the February call for the full year results.

Exactly. We will be on the road at various conferences. So I guess a lot of you at one of the conferences. And for those we don't catch before end of the year already in Merry Christmas.

In October. Okay. Cheers, guys.

Thank you. Bye-bye.