Review recorded.
The end of quarter. We did, for the quarter, EUR 24 million, a significant improvement versus Q2 2024. The main drivers, in addition to the operating leverage effect resulting from higher revenues, were a more favorable product mix, which led to improved gross profit and the above-mentioned lower R&D expense. The current performance in Q2 led to an EBIT of EUR 27 million for the first half, an increase of 18% year- over- year. This translates into an EBIT margin of 11%. Again, please recall that the one-off expense in connection to the personnel reduction in the operations area is included in this figure. Adjusted for this effect, the H1 EBIT margin would be around 13%. Now, to our key balance sheet indexed on slide five. Working capital was down by EUR 65 million since end of fiscal year 2024. Several balance sheet items contributed here.
We continued to decrease inventories to EUR 328 million compared to EUR 369 million at the end of 2024. Year- over- year, inventories have now been reduced by EUR 120 million as we continue to optimize our working capital. As stated before, we expect further inventory reductions to materialize throughout 2025. Trade receivables at the end of June were at EUR 130 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 very large shipments end of 2023. Advanced payments received from customers at quarter end were EUR 52 million, down about EUR 30 million from end of 2024, primarily driven by some cut-off date effects and some regional shifts in the order book. Advanced payments were planned about 18% of our order base.
The fourth key element of working capital, trade payables, have now come down to EUR 22 million from EUR 34 million at the end of 2024. This well reflects the now fully adjusted supply chain situation, which will significantly reduce purchasing levels. Adding it all up, our operating cash flow improves in H1 to EUR 85 million, a strong improvement of more than EUR 17 million versus last year's EUR 13 million. On the back of the improvement in operating cash flow, free cash flow improves even more, aiming at EUR 71 million compared to negative EUR 56 million last year. The improvement was even more pronounced as our CapEx in H1 at EUR 14 million was significantly lower than last year's number of EUR 69 million. This is primarily due to the now completed investments with the innovation fund. Our cash balance, including other financial assets, as of June 30th, 2025, improves to EUR 115 million.
This equals an increase of EUR 16 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, our top priority for the year’s test will continue to be the implementation of our strategy. We will apply our core competencies and abilities to market high growth differentiation and market potential in order to sustainably increase the value of the company. 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, as this is currently our most dynamic market. In optoelectronics, AIXTRON continues to lead the market, maintaining a clear leadership position with a strong and sustained market share over many years.
Our technological edge and earned customer trust have enabled us to secure additional top-tier engagements for our G10-AsP platform. Recently, a major global customer has extended its G10-AsP order to also include their operations in the U.S. Besides this, as communicated in the CR in April, we have won Nokia as a customer. Beyond this, we are seeing strong traction amongst other top 10 industry players, as well as a growing number of customers across Europe, the U.S., Japan, and Taiwan. The increasing demand in optoelectronics is primarily driven by the rising need for laser technologies in data and telecom applications. This momentum is supported by several key industry shifts. The rapid growth in data demand, driven by AI, data center, and 5G, is causing bandwidth needs in transport networks to double roughly every two years.
In paper field data centers, this trend is accelerating the shift towards co-packaged optics to support AI workloads. As data links become more parallelized, the demand for datacom and lasers is expected to suffer, prompting increased investment in optical infrastructure. At the same time, photonic integrated circuits are gaining traction over traditional discrete laser setups. By integrating lasers, modulators, and detectors on a single chip, PICs offer better performance, reduced size, and lower energy consumption. This transition is closely tied to the adoption of 150 millimeter indium phosphate substrates, where our G10-AsP system delivers industry-leading yields. The PIC market is forecasted to reach $41 billion by 2031, growing at an annual rate of around 16%. As PICs incorporate over 100 components, manufacturing precision becomes now critical.
Manufacturing processes must be tight, must meet tight specifications across wafers and production cycles, while backend manufacturing must adapt to higher volumes and complexity, further reinforcing the move to 150 mm indium phosphate substrates. Overall, we expect lasers to account for approximately one-third of our full-year order intake. This growth is fueled not only by the demand for datacom lasers to support AI workloads, both intra data center and interconnect, but also by the increasing adoption of multi-junction PICs for LiDAR applications in the automotive sector, namely in China. Silicon carbide. The silicon carbide market is currently undergoing a longer digestion period, particularly in Western-oriented regions. As a result, decisions for new fab investments are not on the agenda these days. However, in the first half of the year, we saw continued momentum in Asia, namely in China, where demand remained robust and investment activity continued at a decent level.
In the first half of 2025, silicon carbide shipments totaled up to 45% of our equipment revenues. However, this strong momentum is not expected to continue in the full year. Overall, we expect to be roughly flat year- over- year in silicon carbide revenue. Despite the soft market, AIXTRON's G10 platform benefited significantly from this environment, gaining strong traction due to its outstanding performance and cost efficiency. In H1 2025, AIXTRON has received the order and completed shipments of a major volume order for its G10-SiC system for a China customer. This order supports both six-inch and eight-inch production capacity, underscoring the flexibility and scalability of the G10 platform. This major deal reflects the continued demand seen in China during H1.
For the second half of 2025, we expect slower demand also from Chinese SiC customers, while customer qualifications and strategic planning for future capacity expansions are ongoing around the world. Whenever the SiC market will recover, AIXTRON is well-positioned to benefit strongly from new investment activity, thanks to its proven technology leadership and strong customer relationships. With that, let me come to GaN power. The GaN market remains soft in the years up in the Western world, as far as we can judge. However, the regional dynamics have changed significantly. Western markets remain largely soft, with investment decisions continuing to be postponed. In contrast, China demonstrated strong momentum in the first half of the year, driven by sustained demand and active investment behavior. This regional strength has helped to balance out weaker activity elsewhere, resulting in an overall year-on-year GaN tool shipment performance that is roughly flat.
Despite these headwinds, AIXTRON has secured a significant volume order in the GaN power segment from its tier one customer from Asia. The order underscores the continued trust in AIXTRON's technology leadership. The market is currently facing a moderately oversaturated installed base, requiring time to absorb existing capacity. This digestion phase is expected to continue for some quarters before a broader recovery is expected to settle. Nevertheless, AIXTRON maintains a clear number one position in the GaN power market, supported by its proven performance, strong customer relationships, and readiness to scale once market momentum returns. One of the drivers for the expected market recovery, which will come at some point, is the GaN and AI power supply opportunity. We had already highlighted this as a potential major market driver in our past calls. At the time, this opportunity was not included in market researchers' models.
Recently, we have seen quite some news flow on this topic. The expected rapid growth of AI data centers using NVIDIA's new 800 volts high voltage DC HVDC architecture is creating a significant opportunity for GaN power technologies. Companies like Infineon, Maritas, and others are leading the charge by integrating GaN and silicon carbide semiconductors to enable high efficiency and high density power supply, reduce copper usage and infrastructure complexity, and finally improve energy efficiency and lower cooling and maintenance costs. The shifts to 800 volt HVDC and centralized power architecture demand advanced GaN devices for both primary and secondary power conversion stages. We believe that our role in enabling high-performance GaN device manufacturing makes us a key enabler in the AI power revolution in the future.
The timing of that wave, nevertheless, is very difficult to predict as of now, as the architecture was just released and now the design-in cycle and the device manufacturing has to start. With that, let me finally come to the markets of micro LEDs and LEDs. As most of you are certainly aware, AIXTRON continues to hold a clear number one position in the market for red, orange, yellow, and ROY LEDs, underpinned by consistently strong market share. However, also in this market segment, demand is currently weak. Last but not least, micro LEDs. The market has not yet materialized at scale. In 2025, order activity in this segment is limited, with no signs of meaningful volume adoption. The few orders AIXTRON has secured this year are all focused on R&D and pilot lines rather than commercial production.
While interest in applications such as TVs and augmented reality (AR) devices persists, the transition to high volume manufacturing continues to be held back by unresolved cost and process challenges. In summary, taking the sum of all markets, we can say that the soft market period continues in almost all markets apart from the laser market, driven by a hunger for data from AI applications. In all the other market segments, AIXTRON remains very well positioned and we use the soft market period for very close engagement with customers on preparing the next level of technical innovation and next generation products. Once the market demands will pick up again in the future, AIXTRON will benefit from these activities. Timing for the next update, nevertheless, cannot be predicted yet. Let me finalize the update with a look at the U.S. tariff situation.
The U.S.-EU framework has been signed on July 27, thus shortly, in which semiconductor equipment is being declared as a strategic product with zero percentage tariffs imposed. Although the preliminary text has not been published yet and details need to be worked out, it appears highly likely that the semi-equipment exemption as of today will stay. With that, let me now move on to our guidance. Based on the current market development, the current tariff situation and the budget rate of $1.10 per one euro, we confirm our guidance for the full year 2025 as published in February. We expect revenues to come in at a range of EUR 530 million-EUR 600 million. We expect a gross margin of 41%- 42% and an EBIT margin between 18% and 22%.
The guidance for the gross margin and EBIT margin includes a one-off expense of a mid-single-digit euro amount in relation to the implemented personnel reduction in the operations area. The measure will lead to annualized savings in the mid-single-digit euro range in the future, which corresponds to an improvement in the gross margin and EBIT margin of around one percentage point. As Christian explained before, there could be some potential impact from FX risks. An average USD per euro exchange rate of 1.20 in the second half of the fiscal year 2025 could reduce the full year growth and EBIT margin by around one percentage point. As previously stated, we expect our tools to remain exempt from U.S. tariffs. However, we continue to closely monitor the impacts 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. For the third quarter in the year, we expect revenues in the range of EUR 110 million-EUR 140 million. This range is slightly wider this quarter because of an unusually high number of shipments scheduled right at the top of days between Q3 and Q4. They will for sure all ship in 2025, but the exact timing, whether Q3 or whether Q4, of some of these limits is not so easy to predict as of today and the slightly wider range compared to the normal range that we are given. With that, I'll pass it back to Christian before we take questions.
Thank you very much, Felix. Thank you very much, Christian. All correct, we are now ready to take questions.
Yes, thank you very much. Ladies and gentlemen, if you would like to ask a question, please press nine and star on your telephone keypad. In case you wish to withdraw your question, please press three and star. Please press nine and star to register for a question. One moment for the first question, please. First up is Martin calling from ODDO BHS . Over to you.
Hi, thanks for taking my question. My first one is on optoelectronics or photonics. What do you think this year is the level of incremental tool demand that is coming from AI and photonic integrated circuits? I mean, when I think the last call you talked about 20% of sales this year to be linked to optoelectronics, 20% of equipment sales, how much do you think of this is linked to these photonics trends that we see today?
That's a very good question. I think the number of about 20% for the full year revenues is a decent number, yeah, for opto, telecom, datacom altogether. I think that number still holds. If you then add the LEDs for the datacom, I think we may have come up all the way up to 30% for the full year. I would say rough indications maybe one-third to one-half being linked to the datacom opportunities.
Okay, thank you. I see the last call you said you expect this to grow again next year. Is it the expectations there?
I think that's a very, that's a billion-dollar question. It's very difficult to predict for us. We have seen a very, very strong onset right now of this market. We are really seeing, as we have written also in the press statements, that the G10-AsP is almost captively taking the market. It's a beautiful situation. The tool has really been designed for this market and is delivering fully against the expectation we had on that one. Now the question, how big is our customers' market going to grow? I think this will to a very large extent depend on how is the further rollout of AI data centers continuing. It could be that there is a further growth momentum. It could be also that it comes to a new steady level and it continues to be on a much higher level, but on a steady for us year- over- year.
Let's not forget, right? AIXTRON, the equipment is always the first derivative of the overall market growth. If the market is just linearly growing, that means the flat revenue line for us, we always stay flat, but flat is something. Only when a market is undergrowing and end market is growing through an exponential growth, that means for us a further growth yea- over- year. To predict exactly how is the rollout of data centers in the year 2026, 2027, 2028, and 2029 going, for AI, that's a bit difficult.
Okay, and the last one on photonics, how should we think about the end products which are driving the demand? Is it the vast majority pluggable transceivers, or is there even a startup between capacity for co-package optics, switches, etc.? Yeah, how should we think about it?
It's all of it. You can think of it as a mixed bag. I think that's the best description. Sorry for the highly emphasized thing, but in the end, it's literally all of that. Some customers still for long-haul communications make etch and make as lasers, and some guys make VCSELs, some make even super advanced VCSELs. We see now eight multiple resource steps, some are even doing 10 multiple resource steps. We see the VCSELs not only like we know them all from the iPhone, right? We see the VCSELs now going heavily into extending devices for the car. In China, I mentioned that also in my speech. We're seeing massive, massive efforts in self-driving cars there, and cars being equipped with multiple long-range VCSELs, longer wavelengths, higher power levels. We see the datacom opportunity, which a large part is becoming photonic integrated circuits. We see the co-package optics.
It's all of it. That's an extremely diverse market. It's a very broad group of customers, and typically we like a broad market, because a broad market means that it's very diversified for us. We see the market both unfolding. I think we see the market unfolding in all regions of the world at the same time. We have demand from multiple customers in Europe, multiple customers in the U.S., multiple customers in Japan, plus many customers in China. It's a broad market momentum. It stays a mixed bag on the level of each individual application, I have to say.
Okay. Maybe the last one for me, and I will get back in the queue, is just a clarification on how you build your guidance because I was, let's say, slightly surprised of the mix on the guidance. If we think about you getting to the midpoint, the order that you communicated at the midpoint was around EUR 110 million, EUR 113 million in Q3. If we add, you know, services, that means EUR 116 million of orders. That seems like an acceleration versus Q2. At the same time, on the backlog included in the guidance, it seemed quite low. I mean, it's only not even 50% of the backlog that is included in the guidance for the year. Yes, the clarification of ECP, maybe the backlog more coming to next year for silicon carbide, I don't know, or gallium nitride, and acceleration of orders in the short term for other applications.
Just to understand the mix there.
Let me try to clarify. Maybe we were not clear enough. Let me clarify to sort that out now based on your question. It's just good that you shine some light on it. In relation to the Q3, we have given a revenue guidance, a full quarter, a full revenue guidance for the Q3 of EUR 110 million-EUR 140 million. That's revenue. That's not order details. Also, for revenues for the full year, we have confirmed our full year guidance of EUR 530 million-EUR 600 million. I think the best guess is we are shooting at the midpoint of that range. If some positive effects come along, we may be more closer to the high end. The high end very much is still well possible. That's why we are not narrowing down.
If we face some headwinds or some customer discussions are getting a bit more delayed, or the customer order comes in, but you know customer only wants the system in Q1 of next year, then also it could be towards the end of the year. The full revenue range of EUR 530 million-EUR 600 million is fully possible. Our best guess, and probably also your best guess, is literally at the midpoint. I have to clear that out. We have, as of now, not given any guidance for the order intake for the year. Nevertheless, what I tried to bring out in my speech when I was speaking about the high first half of the year silicon carbide revenue shipments, revenue in the first half, silicon carbide was 45%.
I wanted to accelerate or to bring out the point that we are not expecting in the second half, again, to have 45% of revenue to come from silicon carbide. The second half will be much more dominated, for example, by optoelectronics. I think the big part of the revenue is silicon carbide we have seen in the first half. Maybe that was a bit confusing. Apologies for that.
Okay, thank you.
The next question comes from Gustav Froberg from Berenberg. Over to you.
Hi, Ron. Thank you for taking my questions as well. Two things on demand. First on China. You talked a little bit about silicon carbide for us in China, perhaps softening a little bit towards the second half. How does the build in China or Asia on the silicon carbide side compare in your eyes relative to real demand? Could you maybe draw some parallels between the build you're seeing there and now versus what we saw that resulted in overbuild on the Western side a couple of years ago? It's really a question on the run rate of growth in China and Asian markets. The second question on demand more broadly. You're talking about demand and market demand for most of your applications, bar opto, for example.
I'm wondering, from your point of view, do we need a new killer app or something to get the market to reaccelerate again? Do you think that the reacceleration to come that you talked about as well can be achieved based on current markets, current technology, no killer app? Thank you.
Perfect. I think very good questions. Thanks for posing the two of them. Let me come first to silicon carbide in China. I think what your question, what's underlining your question already is, I think, very, very well too. In China, we have seen over the last two, three years a massive build-out in silicon carbide wafer production capacities. I think China today has six-inch wafer production capacities, which is probably good enough to serve 2x or 3x the world's market demand. In six-inch, we have in China a massive, I'd say truly massive overcapacity. There's literally entire factories filled with tools, and they're all standing idle and being turned off, and companies starting to scrap tools. You wonder why in this situation there is any demand at all.
This additional demand is coming solely and purely from the transition to eight-inch wafer size, and particularly from the technological advancements that have been made with eight-inch wafers. You can buy today eight-inch wafers with a much better surface quality than you could ever have it on six-inch wafers. This better surface quality allows you, and at the same time, we are seeing that the silicon carbide chips are getting much, much larger than it was the case in the past. Just to give you some numbers, we started off three, four years ago with silicon carbide chips typically having a size of two by two millimeters, so four square millimeters. Most recently, the chips were shipping mostly on five by five millimeters, already 25 square millimeters. We are now seeing most customers and customer demos running on 10 by 10 mm, so 100 square millimeters.
You can see a chip size going from four to 100 square millimeters, that's a massive movement. This massive movement allows you, when you put a power module for a car together, that inside of your module, you no longer need to have six, seven, eight chips that you all have to connect with bonding and wires and so individually, but you rather have a single chip or maybe you have two single chips, and all the power can run through it. You have a lot of packaging cost reduction by going to a single chip. The single chip is enabled by the fact that on eight-inch, you get a much better surface quality as it was never, never been the case in six-inch. That is the technological improvements and advancements that have been made.
At the same point, as the chips are getting much larger, I'm on eight-inch, yeah, a chip is square, a wafer is round. On an eight-inch wafer, you just can put more chips and have a better utilization of the surface area. Sorry for the long technical explanation, but I wanted to bring out the point that there are clear drivers to what eight-inch wafer size. This is now in China also driving the market towards an eight-inch silicon carbide wafer. This is the reason why there is an additional demand on new demand despite the fact that there's a lot of dense six-inch capacity in China. I hope that explains it a little bit. Now, to your other question about the demand, and Jess, you picked it up very well. We wanted to give a very clear message about, I think, the view we have on the market.
Most of the market segment except from the opto market, it is, I think, I call it a zoom. It went down from the demand size. Now it's kind of flat. The question is, when does it pick up again? What's the signal, when does the pickup happen again? A clear answer, we don't know it yet. Especially, we don't know the timing yet. It's very clear, with a very clear view that the market will come back. I think one of the topics at some point, the silicon carbide capacity, which is out in the market, will have been digested. The switch to eight-inch wafer size is ongoing, also in the Western side. All the customers in the Western markets are now qualifying super junction devices, which need also the better wafer quality.
By the time the existing capacity is being digested, I think we are seeing now, I call it a classical Gartner hype cycle, where the initial phase is a total overthrowing and total overinvestment in silicon carbide. The market collapsed, and at some point now, the market will come back in a much more long-term, much more healthy, much more steady way in silicon carbide. I think over the next one, two years, we will see electric vehicles gradually gaining traction around the world. Electric battery technology allowing now, I think, driving ranges, which meet the expectations of consumers somewhere in a realistic 6,700 km range. Costs for battery technology coming down, silicon carbide, and the whole value chain getting cost-effective and mature. We expect some decent uptake from that market. There are also new markets on the horizon.
We've been shining some light on our transition value chains for AI. If you read the chart that we have on the presentation, from the NVIDIA, Infineon, Maritas, the architects are quite in the details, you'll see that there is no word of silicon anymore at all on the entire chart. You literally come in from the high voltage from the power plants to your overland lines. You are at the multiple kilovolts, 6,000 volts, 3,000 volts. You convert it down to the 800 volt DC. You do the conversion inside of the data center. All of this will be driving both silicon carbide demand on the very high voltages and gallium nitride demand on the high voltages and on the low voltages. That's going to be a major driver.
Furthermore, in the long term, but that's further out, there's a bunch of new applications for silicon carbide in the making, additional applications that no analyst or analyst report has on the radar. It's more like innovation topics where the whole transmission architecture, which today is made out of transformers and the stuff which you see next to the highway when driving, where the overland lines are crossing. All these things are being put now into power devices, being made out of silicon carbide. At some point, this entire switching architecture goes to silicon carbide. There are new applications in the making. The enabler for that is that the wafer sizes are getting bigger, the wafer surfaces are getting better, and the cost points are coming massively down. I think everybody is aware of the massive price drop on silicon carbide wafers.
As always, in our semiconductor industry, when the prices are dropping 2X to 3X, that allows that completely new applications are opening up. Sorry for the slightly long answer, but maybe that gives a bit of a perspective of what's coming in the midterm.
Great. Good product. Thank you.
The next question comes from Didier Scemama from Bank of America.
Thank you, Kevin. Thanks for taking my question and good afternoon. I just wanted to ask you your early thoughts about 2026. I really feel like the order intake in the second half is not going to be particularly exciting outside of datacom based on your commentary about demand and overcapacity in photonic integrated circuits and GaN. I just wondered if you can run us through sort of the puts and takes for 2026. I think you've highlighted some, let's say, long-term positive trends in datacom, but I was a bit hesitant to extrapolate the trends. I think you also said that photonic integrated circuits and GaN wasn't clear as to when you would turn. Can you reference through the puts and takes for next year? I've got a follow up. Thank you.
Thank you. Thanks a lot. Very good question. I think you have seen the essence of our message already out from what we said earlier. Unfortunately, the message is we don't know yet. I think that's the best we can give you because there is no clear indicator on a tipping point yet. As your question is indicating, at some point, it will pick up clean. The midterm horizon is very clear, and there will be demands coming, but we are still in the point where our customers have a decent amount of idle capacity, unused capacities, both on the silicon carbide and in the gallium nitride. Now it really depends on when is the end market that our customers are serving taking up again.
Gradually, the capacity utilization of our customers is rising, and our customers then gain the confidence by themselves to say, "Okay, now whatever we had an 80%, 85% utilization level, let's push the trigger button for new purchase orders for equipment." Of course, it will come to us, but we simply don't know as of now when that's coming. As we don't know, we also cannot, unfortunately, shine any precise light on the year 2026.
No, that's completely fair. I mean, it's difficult obviously to know what's going to happen in the next three to six months. It's hard to know what's going to happen in the next 12 months.
I hate to say the message, but that's the truth. I want to say it, but it's just as it is.
I hear you. No one really knows what's going to happen tomorrow. My follow-up is on my favorite question, which is on working caps. Just going back to the question I asked last quarter, your payables are down to 25 days, which is roughly half the level they were before, while your DSOs are sort of stable to slightly higher than they used to be. Perhaps we interpret that, as it seems like rent and freeze are progressively coming down.
You were.
If I understood correctly, Christian here, you were talking about the payables, right? Yeah. I mean, payables are coming down.
Payables are coming down a lot versus history. Inventories are coming down. I understand that. Yeah. Your receivables, let's say, are not increasing, but they are a bit higher than normal. I just wondered why the payables are so low.
The payables are just so low because we are buying much less material because we are still consuming the two high inventories that we have. We've slowed down our current fluctuating basing of orders to the absolute minimum, basically components in need, because we do not have any inventory. That is just, with a little bit of a delay, then going into the payables level. The payables level should remain at a lower level for quite a while until we reach the normalized levels of our inventory, and then we need to see where they develop. Maybe at some point we need to increase the sourcing level, and then payables could also go up again. For sure.
I guess that's a good instance with the uncertain outlook for 2026.
Yeah. I mean, so far, we are really limiting the different sourcing to what we do not have in stock, like customer-specific items, and focus completely on using up existing inventories.
Okay, this is what you're saying.
Yep.
All right. I would say thank you very much, right here, Christian.
Great. Thanks, Didier.
Next in line is Madeleine Jenkins from UBS.
Hi, everyone. Thanks for the microphone again. Apologies for the last word. I dropped off for a bit, but I just have a question on your four-year guide, 2025. The new orders you publish that you need in H2 to reach a guide that's significantly higher than they have been in the last few years. I just kind of wondering what to change and also what gives you the confidence that you'll receive these given, obviously, the market conventions you're talking about. Thank you.
Thanks a lot. Yeah. I think, as I mentioned, we are shooting towards the midpoint of the guidance. Yeah, that's what we are clearly aiming with some potential deviation to the lower side or some potential deviation to the high side. Let's see where it comes out. Yeah. If you take that, we are expecting once again a very strong Q4, as we have seen it also in the previous years. I think you know the seasonality that we have in AIXTRON and have had in many, many years in AIXTRON. As you see in the chart that is published also on page 11 of the presentation, you see that we are expecting still to ship between EUR 80 million and EUR 150 million of new equipment orders in the year. Yeah.
We are expecting the order intake on a level as we are, potentially a bit higher, potentially the level where we are. Yeah. We are not expecting the order intake to collapse or anything. Yeah, I think we are on a stable level right now where we are. Based on that, you see EUR 80 million-EUR 150 million of orders to come in in Q3 and/or early Q4, and then still shipping until the 31st of December is very well in range with what is possible. Yeah. We feel well with that number. There is, of course, a very clear pipeline, a named pipeline with individual customers, individual orders, and projects behind it, not just the guesswork, of course. Yeah. That's the number of what we are shooting for.
Okay. Makes sense. Thanks. On the commentary about kind of overcapacity, silicon carbide in six-inch, but we're also seeing incremental demand for eight-inch. Just to help me understand, why are customers not reusing your tools from six-inch to eight-inch if they just sat there idly, or is there a slightly different configuration? Yeah, just be interesting to get your color on that. Thank you.
Customers who have from AIXTRON the G10 series can nicely switch the tool from a six-inch configuration to an eight-inch configuration and use it and get all the benefits of that, and produce, in what I've illustrated a bit like, the increased technical demand. That's very well possible. However, all generations of equipment in the market had only been suitable for six-inch, both our own very first generation. It was called G5 Warmboy, but also there's still generations of, you may remember, the onset of the six-inch RAM. The initial tools were all from Tokyo Electron, from TEL. This was a six-inch tool only. There's a bunch of other tools in the market which were made for a six-inch configuration and are usable only for a six-inch in DC performance.
That means there is a bunch of equipment out in the market from the previous waves which cannot really be used on the eight-inch production. Means for the coming years, it's not really usable and it needs to be replaced.
Okay. That's helpful. Thank you.
The next question comes from Michael Kuhn from Deutsche Bank.
Good afternoon. Thank you for taking my questions. Essentially, follow-ups. The first would be on currency. Thanks for providing this H2 scenario with $1.20 and a day one 1% potential margin headwind. Maybe looking further out into the future and into next year, let's assume we have a permanently weaker U.S. dollar. What would that mean for your profitability and what, let's say, mitigating measures could you take?
Great question. Thanks, Michael. Some are expected that question from you. It makes a lot of sense. Honestly, I would highlight that this exposure for this year explicitly because this year is a little bit higher. Why is that? Because normally, we are selling revenues in U.S. dollars between like 20%- 25%. That range always changes a little bit. Normally, we had a fairly good net revenue from sourcing in U.S. dollars. However, this year, that net revenue is not working that well. Why is that? Because we are still sitting on quite some inventory levels from the past where we have bought material in U.S. dollar at a higher U.S. dollar rate, which we are working down now. The sourcing level is a little bit lower, and that's because of the slightly higher exposure to the U.S. dollar this year.
Also, we have scheduled slightly higher shipments in the second half of the year, and we started to highlight this. Going forward, once that inventory is worked down, our sourcing level in U.S. dollars will also increase again, and the net revenue should work better again. Beyond this year, I would expect a lower exposure. Is that clear?
Yes, that's all very clear. One more to talk about, let's say, tipping points and what could trigger better orders again. If you, let's say, want to track the market, for example, in EV, but also in the area of data center, is there, let's say, good indicators to track? In the car space, for example, a number of cars with big components shipped, or in data center, let's say, any qualification milestones that we should be aware of that could trigger a new, let's say, wave of orders for your equipment as well?
I think that's a very good question. I think in silicon carbide, due to the idle capacity and the wafer size change from six to eight inch, it's very difficult to foresee right now because there's these multiple effects ongoing, which I think we've been discussing here in the call a little bit. I think that's quite difficult. I think on the gallium nitride side, the key topic is going to be when this AI value chain, which we have spoken about in this call, which is now kind of just being agreed as an architecture. You can think of now the companies have agreed on, let's say, on a blueprint, I would say. Now everybody inside of their companies is working on various projects to do their part and make their part of this blueprint a reality. It's going to take some time until this blueprint materializes.
I think by the time the first one or the first multiple players with these groups, who are part of this alliance, give signals that they are really seeing volumes and shipments, I would expect now that this alliance also jointly and publicly has made announcements that this is the architecture. Of course, they want the whole industry aligned according to what they've agreed to. I think by the time we see first announcements from players of this alliance that real projects are coming on that one, that could be a good point to say it's really starting, not only like as a handshake between the architecture, but in terms of dollars, business, and P&L value.
All right. Understood. You mentioned in the context of micro LED once more the technical challenges that the customers are still facing that are running the pilot lines. Any update there? Maybe also on what AMS-OSRAM ultimately did with their tools that they received?
Honestly, no, no very clear picture. Unfortunately, again, I hate to say it. I don't like those things, but it's the reality. I think we see continued activity, as I've indicated a bit in my speech. We do see that customers continue to work on the topic in many very diverse directions, I have to say. It's not just one mainstream direction and everybody follows that, but multiple directions and multiple end applications, that's in the focus. We hear from customers and multiple customer projects that they make some nice progress. My typical question, of course, also to my customers, is when does it translate into orders? So far, I haven't been getting a satisfactory answer, which I could be passing on to you here in this call. I would say it's too early to tell, unfortunately. The work in the segment goes on.
I think now we have to wait and see which of the many segments, one segment being AR glasses, AR/VR glasses, quite an interesting one. Some work is continuing to go in the direction of televisions, other projects to go in the directions of wearable devices. Automotive, as it's smaller, gets a very high dollar for our customers, high dollar end market being another application. All this is being worked on. When it translates into orders and of how big of an amount of orders does it translate to? I think there's two scenarios possible. It stays a premium, sophisticated, high-end, expensive niche market where our customers potentially can charge a premium value on some high-value displays because it has a super unique feature that goes into a high-end car. Great, nice, good for our customers, but in the end, a very small number of tools for us.
There is a breakdown and it comes really into one of the volume segments, let's say, televisions or let's say smartphones that, of course, would translate to a high volume of wafers, needs a high volume of tools. It's too early to say that it can take any of these directions.
Thank you. Thanks for the detailed answers.
Next up is Ruben Devos from Kepler Cheuvreux .
Yes, good afternoon. Thank you for letting me on. I just had a follow-up on this year's guide. I just wish to tell I have that fully understood. I think the backlog is about $280 million. Half of it will be out the next six months, I assume the other half will go out next year. With regards to the $80 million- $150 million for new orders, my understanding was always that because of a high invention and stock, you could be able to assemble much quicker than your usual lead time. What is sort of the latest, let's say, point at which you could accept an order and you could still ship it this year? I understood you were saying that there's some likelihood you could still reach the high end.
If that were to happen, just curious whether, based upon your utilization rate or just your customer engagement in general, you would think that's likely coming from your Eastern customers or your Western customers. Thank you.
Thanks a lot. I think very good. First of all, the understanding of the situation that you have depicted is very much correct. I can confirm. Nevertheless, one minor adjustment. Yes, we expect to ship, as we have indicated, around $140 million out of the equipment backlog still in this year. Nevertheless, I want to clearly point out that some of the backlog is both for 2026, meaning next year, but also some backlog already for 2027. I think it's a minor part of it, but just want to be correct because you were only mentioning one of the years. That's that upfront. In terms of what is the latest point in time when we are able to ship, that really depends on which of the products.
We have some products where essentially we are on an inventory level where I would say an order may come in pretty, pretty late, I would say even beginning of Q4, and we would still be able to ship it out. If it's a standard configuration, if it's a tool where most of the parts are available, you know, plain vanilla, and nevertheless, there's other parts of our portfolio where essentially if an order comes in, we need to secure almost the entire tool through our supply chain. It really varies. That's also what we give to our customers. Of course, we ship into our sales teams around the world. Now, in this particular situation, very clear indications which of the product series in our portfolio is shippable with which lead time. I think it varies.
Very clear, we do expect that some orders still coming in beginning Q4 are still possible to ship in the December timeframe for some of the series. Mostly the older series, which is more like a plain vanilla kind of stuff, while a general, newer series is more sophisticated and less of an investment. I think there was another part of your question. Did I get it all, or was there something else?
It sounded a bit like you did not rule out, let's say, the high end of the range. What is, I guess, a bit your feel of the indication from which this demand could come from if you hit the high end? Is it rather Eastern-based or Western-based?
I think, honestly, I don't have the information in front of me. It's literally a pipeline of customer opportunities, named each individual. Honestly, I don't have it in front of me which is more left or right. I think mostly it's going to be less a question of, is the demand coming or is the demand not coming? I think it's mostly a question of, is it coming at a point of time that the customer also wants to have the shipment done in December, or does the customer say, "Come on, I take it whatever in February because my facility isn't ready," or, "What is my RAM plan?" You know, you typically can plan on a timeline of one or two quarters, but the exact date and the cutoff line is always a bit difficult to predict.
Oh, okay. Great. Thank you. Just a final question. I was thinking about your growing installed base of G10 systems. I presume that your recurring revenue from service and spares should become more important. Yeah, could you provide some metrics maybe on this business? You know, what is the attached rate for service contracts on these new tools? Or maybe what percentage of your gross profit do you expect to come from service in the next, you know, let's say, three years given the installed base on G10?
Yeah. I think, in fact, you name it. The after-sales business has been growing quite nicely. Now, in this year 2025, that's, of course, a bit muted because of the low utilization of tools. You know the tools which are installed but currently not running. They're not consuming anything. We expect next year when the markets are hopefully coming back and at least the utilization, first of all, is growing up. When the utilization is high again, at some point, of course, new orders for new equipment are coming back. The after-sales market is going to grow again. I think, Christian, where are we now? I think $120 million, $150 million per year, right? Something like that. $120 million. Yeah, something like that. I think next year probably is more than $150 million, $160 million if I would guess, if the utilization comes up. Let's see. Let's see. All right.
Thank you very much.
Before we continue, just because of the time, please keep your questions to two per analyst, because otherwise, we'll not make it. We have a long list of questions. Thank you.
Next in the line is Martin Jungfleisch from BNP Paribas.
Yeah. Hi. Good afternoon. Thanks for letting me on. I just have a few questions. First one is really on the change in carbon market share in China. What do you estimate is your market share for heated source there? How do you think about local competition in China, given that AMEC has recently brought a tool to the market? Have you seen local competitors pop up more often, or are those still more on the fixed inside? That is a different question.
I think market share is in silicon carbide in China. It's always a bit difficult to predict because the transparency is not so high. I would really guess our market share as of now north of 50%. Maybe it's typically 60% in China as of now on the eight-inch market because our tool has gotten very good traction on that one. Now, as it comes to local China competition, I think it's no surprise that China is driving heavily in efforts to localize the equipment. I think it's a strategic initiative led by the government. We are watching that very carefully, and we don't have a picture yet how that's going to shape the market shares, and how good the local competition is going to be. We just have to watch that before we can seriously comment on that.
Okay. Makes sense. Thank you. The second question is on gross margin. It looks like to reach the low end of the gross margin guidance, do you want to keep it around 45% gross margin in the second half or maybe even closer to 50% in Q4? First of all, is that the right way to think about it? Also, what would drive this, given that the share of silicon carbide is most likely lower in the second half? Would opto also have a similar margin than power to make up for it?
Yeah, clearly. If you just run the math and compare to what we achieved in the second half of last year, it's a similar performance. Of course, a little bit lower revenue, of course. We expect to clearly get a better sort of mix. I mean, the optoelectronics business, G10-AsP-driven, is running at high margin. There's also some operating leverage effect with a larger second half, similar effect than we've seen in the last year. If you just take a look at Q3, Q4 last year, nothing out of the extraordinary that we expect now for this year. The executives between Q3 and Q4 will receive depending on the business. Overall, that is what it is.
Okay. Great. Thank you very much.
Next up is [audio distortion] .
Hi. Thanks for taking my question. My question, the first question on GaN, and specifically on the DSMC decision to back out of the GaN foundry market. Over the years, I think this has been a significant customer for you on the GaN side. How do you see the effect of that on your GaN customer base? Is that good for you because other people will have to step in and therefore order equipment, or will TSMC be selling that equipment out and therefore, you know, that is more of a headwind than a tailwind for you? I have a short follow-up.
I think, of course, we all have monitored this topic, right? We are very much informed of TSMC pulling out of the GaN market. I think it's simply an individual company decision, right? I think TSMC is extremely busy by building out $1 billion after the other, right, with super high revenue opportunities. I think it's just the portfolio pooling and taking off that. I think they not only pulled out of the GaN market, but of multiple smaller market segments, as we are aware. Yeah. I think it was just the portfolio pooling exercise and the market is the market. Yeah, the volumes will go somewhere else. We are supporting the market. No big deal.
On another market, which is a red, orange, yellow market, this has been a lumpy market and it sort of comes once every couple of years and then goes off. Is there any signs, any increase in discussions with Chinese customers, etc., which suggest that this could be a market which could come back in 2026?
We have some volume also running in that market in 2025, so it's not dead. There's some volume in it. As you say, it's lumpy, and suddenly a customer comes around the corner and wants 20 tools, 20 tools and converting to a high double-digit billion amount, right? It's difficult to predict. We do see the market is not dead. The market is going. Here and there, a customer makes a capacity expansion or like a sub-renewal decision. Sometimes it's also renewals now of old capacity being taken out, being replaced by newer, more productive tools. We take it as it comes.
Understood. Thank you.
The next question comes from Andrew Gardiner from UBS.
Good afternoon. Thanks for taking the question. I'll keep them brief because they're follow-ups. I'm just trying to see on the 2025 guidance, Felix, you mentioned that, you know, the top end of the range is still within sight. I'm just wondering which end markets need to come through for you to achieve that. As you said, opto has been the most dynamic. Maybe that's some of it. Given its relative size within the mix, it would seem to me like you would need some of the other end markets to come back. At the same time, you're saying you don't have great visibility. Can you help me understand what the driver would be to see the top end of the range? Just following up on Gennardin's question on TSMC, is there any risk of that GaN equipment coming back into the market second hand?
I don't know, I see this 550 tonnage equipment. Yeah, it's because the offensives were early in GaN. Is that equipment fairly old and not really upgradable? Thank you.
Okay, let me answer your first question first. I think what do we expect for the second half? Let me phrase it rather like that. As I've indicated, we had in the first half a very, very strong revenue share from silicon carbide power. We do not expect that to repeat as I've indicated before. We expect silicon carbide to be weaker in the second half. Nevertheless, in the first half, on the other side, the gallium nitride was a bit weaker, and we expect the gallium nitride to come out in terms of revenues as stronger than in the first half. Also, in the second half, we expect that many of the laser orders, which we've been speaking about broadly in this call, which were ordered in the first half, to shift and convert into revenues in the second half.
That's what we're clearly expecting for the second half of the year to come as revenues. Also, some of the other, we call it other optical segments, we also expect them to come, some of those. I think it's just the cyclicality, all our markets, and we are happy to have multiple end markets. Some are going down, some are going up. In the end, the question about do we reach the upper half of the guidance or even the upper quartile of the guidance, or is it rather the lower quartile? That depends less on markets or dynamic of end markets. It's more individual customer discussions, and it's more the cut-off line effect at the end of the year for individual customer orders. Maybe that gives you a little bit of light on what we expect for the second half.
Thank you. Is this the potential for used tools from TSM?
Yeah, from TSMC. First of all, the tools are in use, yeah. I think the question is, will they continue to be in use, or will they be optimized and be replaced by brand new tools, yeah? In the base case, it stays as it is. In an upside case, it's a new tool revenue opportunity for us.
Okay, thank you.
Next up is Adithya M etuku from HSBC.
Hi. I didn't get asked in those two questions. Firstly, just maybe a slightly technical question, looking at the transition from plugables to CPOs. I just wondered, if you go on a, if you take a single fiber and you go from using the plugable to plugging the fiber in straight into a CPO, does your content grow materially? Does the indium phosphide dioxide actually increase in that scenario, or is the opportunity for PIC really coming from CPOs driving a significant increase in fiber usage? Any color you can give around that would be helpful. I've got a follow-up.
Today, what we have is you have a communication of an electronic connection between a CPU and a transceiver module, yeah. The transceiver today is bundling a lot of the electrical signal. If you would go and directly have an optical connection to a CPU, you would be connecting to many more points, yeah, in the CPU. You would be connecting to many more points, yeah. As you cut out the electric intermediate layer, I would expect the overall content for indium phosphate compound to go up. That would be my expectation.
Understood. For a single fiber, when you go from plugable to a CPO, you get more content?
I would expect that, yes.
I understand. Just as a follow-up, I noticed your revenues from the Americas were up year on year. I just wondered if you could give any clarity on what drove this.
Honestly, I think that must be just an individual customer order. It's not like a dedicated market segment. I mean, we clearly see that in America, we have many laser customers and the datacom. I think it's a good market segment. We have quite a decent customer base there, yeah. Honestly, I hadn't looked. It's going to be a mixed topic of mostly customer dependent, I would say.
Understood. Thank you.
Okay. Due to the time, we will close the call today. I know there's still some follow-up questions in the queue, but we have to move on. Please contact the IR department afterwards. We're happy to help you out. Thank you very much for listening in. I wish you all a great summer break. Talk to you in the conferences and the roadshows in the September-October timeframe, and otherwise, we'll hear you in our October call for Q3 results. Thank you very much, and goodbye.
The recording has been stopped.