AIXTRON SE (AIXA) Earnings Call Transcript & Summary

Good afternoon, ladies and gentlemen, and a warm welcome to the H1 2024 Results Call of AIXTRON SE. [Operator Instructions] Let me now turn the floor over to your host, Christian Ludwig.

Thank you, Anna. Welcome to AIXTRON's H1 2024 results call. Let me quickly introduce myself. My name is Christian Ludwig. I'm the new Head of IR at AIXTRON since May 1. So this is my first earnings call for AIXTRON. I have met or at least talked to many of you already in my first month and looking forward to meet the rest of you as well soon. Now I'd like to welcome 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 can't 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 immediately presented by webcast or any other medium. However, we will place a transcript on our website at some point after the call. And now without much further ado, I would like to hand you over to our CEO for his opening remarks. Felix, the floor is yours.

Thank you, Christian. It's great to have you on board now. Let me also welcome you all to our Q2 '24 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 the financial figures. Finally, I will give you an update on the development of our business and on our guidance. Let me start by giving you an update on the key business developments of the second quarter on Slide 2. The most important messages for today, from my viewpoint are despite an overall softer market environment, we have maintained our competitive position, and we could recognize strong orders in second quarter '24 with EUR 176 million, mostly driven by power electronics. We concluded the quarter with revenues of EUR 132 million in the upper half of the guided range. The gross margin came out as expected at 37%, which is lower compared to the previous year due to a product mix comprising a larger share of lower margin adding into it. Most interestingly, we could further expand our strong position in silicon carbide. In Q2, SiC-based systems accounted for 57% of the incoming equipment order while gallium-nitride-based systems made up 29%. Our market position in silicon carbide has been strengthened throughout the last 2 quarters with new customer wins. We were able to win an additional top 5 player in SiC. Furthermore, we have secured a substantial volume order from China, and we could win several new customers for our G10-SiC tool announcement from Japan. You may also have seen our recent press release together with Nexperia for a G10-SiC repeat order and new G10-GaN order. This is the result of the recent technical progress we have made with the G10-SiC. So while important new SiC customers drove the business in the first quarter and the second quarter of this year, we have also secured major follow-up orders from existing SiC as well as gallium-nitride customers. Mostly driven by power electronics, our equipment order backlog at the end of Q2 increased to EUR 401 million, up almost EUR 50 million from the year-end of fiscal '23. In June, we further announced the acquisition of a production site in Turin, Italy. This addition to our manufacturing footprint addresses the expected increase in demand from major customers and we'll be able to cover future order peaks at all times. Furthermore, this gives us better access to the strong manufacturing ecosystem in Northern Italy. The softer end market for power devices has impacted our delivery schedule. Customers are continuing to place orders, but delivery is pushed out in some cases. Therefore, we have adjusted our fiscal year '24 guidance. We now expect revenues in the range of EUR 620 million to EUR 660 million with a gross margin of 43% to 45% and an EBIT margin of 22% to 25%. Christian will now provide a more detailed look into the financials. And then I will take over with an update on the market. Christian?

Thanks, Felix, and hello to everyone. Let me start with the financial highlights of our income statement on Slide 4. We had a good quarter in a difficult market environment with revenues at EUR 132 million compared to EUR 174 million last year. This is due to the fact that in the last year, some shipments and revenues have been pushed from Q1 to Q2. In the first half of 2024, we recorded revenues of EUR 250 million which were exactly in line with last year's H1 revenue number. Gross profit in Q2 of 2024 was at EUR 49 million, and EBIT for the quarter was at EUR 30 million. The gross margin was at 37% compared to 42% in the year before, both in Q2 as well as H1, driven primarily by product mix. In both Q1 and Q2, we shipped a larger share of old generation systems like G4 and G5 compared to last year. These systems come with lower margins than the new G10 generation. Especially in Q2, we shipped a larger share of LED tools, which come at a lower margin. To be clear, these effects were expected and are baked into our full year expectation. OpEx in the quarter went up to EUR 36 million, predominantly driven by higher R&D spending compared to the previous year. We are currently finalizing the development work for the G10 series, while we have started the development of our next-generation systems already. As we said earlier, we will see an increase of R&D expenses in 2024 by a mid-single-digit euro million number compared to 2023. In 2025, we target R&D spending again around the level we have seen in 2023 in absolute terms. SG&A will remain flat in 2024. Now to our key balance sheet indicators on Slide 5. We've seen a further slight increase in inventories to EUR 448 million compared to EUR 436 million at the end of Q1. Referring to our last results call, the high inventory levels are still the result of our strategy to load the supply chain early enough to secure on-time delivery of our products despite tight supply chains. As we saw that supply chains are overall relaxing, we adjusted our strategy and are on the way to reduce buffer stocks. For 2024, though, we had prepared inventories to secure the initial full guidance range, including the upper point. When we saw that 2024 is coming in softer, we further decided to slow down the supply chain, but some orders were so far progressed that we decided to take delivery, leading to this further small increase in inventories. But this should now be the high watermark, and we expect the reduction of inventory levels with the effects kicking in, in the second half of 2024 when the larger shipments are planned and some effects will take until summer 2025 to materialize in full. Trade receivables at the end of June were at EUR 117 million compared to EUR 158 million at the end of 2023 and flat versus the number end of March. The improvement versus year-end 2023 was mainly a result of the collection of the payments related to the large shipments in the last quarter of 2023. The advanced payments received from customers at quarter end were EUR 133 million, representing about 33% of order backlog very much like in the last quarter. Now moving on to our key cash flow indicators on Slide 6. Our operating cash flow improved strongly both in Q2 as well as H1 as the inventory build stabilized. After 6 months, operating cash flow stood at EUR 13 million, an improvement of EUR 80 million versus last year's minus EUR 71 million. In the second quarter, we generated EUR 20 million of operating cash flow, a swing of EUR 96 million versus last year's minus EUR 76 million. Free cash flow also improved on the back of improvement in operating cash flow. In the first half, it came in at minus EUR 56 million compared to minus EUR 80 million last year. The improvement was less pronounced than for the operating cash flow as our CapEx in H1 at EUR 69 million was significantly higher than last year's number of EUR 10 million. This is primarily due to the investment into the innovation center as well as the investment in our new Italian facility. In summary, we will see returning of inventories in the second half of the year when the larger shipments are planned, which should result in an improved operating cash flow. At the same time, CapEx will come down in the second half of the year as we approach the completion of the innovation center. These 2 effects should result in a strong free cash flow in the second half of the year. Our cash balance, including our current financial assets as of June 30, decreased to EUR 79 million from EUR 182 million at the end of 2023. This reduction was mainly due to the mentioned CapEx projects and our dividend payment of EUR 45 million in May. After Q2, we have signed and closed a revolving credit facility for EUR 200 million to further strengthen our financial flexibility. These facilities are available for all purposes. With that, let me hand you back over to Felix.

Thank you, Christian. Before giving you some more details on our adjusted guidance for '24, I would like to continue on Slide 7 and give you an update on our views on the different end markets. With our G10 tool family introduced in the market since September '22, we have a state-of-the-art product offering that offers the best-in-class cost of ownership. With the G10 family, we cover all our key end markets, while we have exciting growth applications ahead of them. Overall, we are very pleased with the market traction of our G10 series. We expect that the G10 family will make about 50% of our equipment revenue in the full year '24. This is a very strong adoption, roughly 1 year after launch, given the required qualification cycles in all applications we address. This confirms that our strong focus on technology and innovation is paying off. We aspire to continue this strategy in the years to come. On Slide 8, you can see which tool addresses which application. Let me briefly touch the optoelectronic area, where we see stable demand, which is driven by datacom and telecom lasers. Datacom lasers are critical to data center to meet the higher demand of AI workload, both for intra data center and data center interconnect requirements. We maintain a clear #1 position in the laser market with a strong market share and we are very well prepared with the G10-AsP to secure this position. Now to the LED and micro LED market. As stated in the Q1 call, in '24, we see a wave of investment into traditional red LED capacities ongoing. We expect high double-digit euro million revenues in fiscal '24 from this market. What we see happening is that players which have been traditionally only serving blue LED are now investing into red capacity as well. Our G4 is here the tool of record in this segment due to its track record. On micro LED, where the G10-AsP is the tool of record, the industry continues to work on micro LED technology despite the news from Apple and ams OSRAM in Q1. Micro LED revenues for fiscal '24 are expected to be at high double-digit euro million driven by several customers building R&D and pilot production lines to commercialize the technology. The manufacturing of micro LEDs is challenging, particularly the mass transfer aspects. This makes timing predictions difficult, but multiple customers have confirmed their continued interest. Let me continue in more details with the GaN power electronics market on Slide 8. There have been some exciting developments lately that I would like to share. We see the market accelerating, fueled by more and more applications and provide you an update as we realize that market analysts capture most but not all trends, especially the future development in this application appears not to be well understood until now, and we believe that they are important for you to understand our recent strategic moves, such as the investment in our Italy location. The Slide 9 depicts the deployment of our GaN power technology application by application. We have ordered them by voltage class on the y-axis and by time of market adoption on the x-axis. Please recognize that these specifications are approximate and don't aspire for precision. You may recognize a number of past developments that have driven down volumes in the past and that we have discussed in this place. In 2019, 2020, the first wave of products were fast chargers for consumer applications such as smartphone and also notebook power supply. 650-volt with the first voltage took a large market traction. The first generation of GaN products by most players was on 650 volts. The high-voltage, high-power AC to DC power converter used, for example, in data centers, where the next application, which continues to grow in volume even today. And a derivative of that were Class D audio devices, which some of you may be enjoying at home. Most recently, we've seen the emergence of low-voltage GaN products such as 100 volt and 200 volt products. Initially pioneered by smaller players, these products now have become part of the portfolio of most of the big GaN players. This comprises DC to DC power stages used in data centers, but also in automotive. Solar inverters and specialty devices such as wireless charger and [ liner ] applications. The most important volume driver here is probably the usage of GaN in low-voltage motor drives also called [ BLDC ], be it a battery tool, battery-driven consumer applications or in low-speed electric vehicle. This is a huge market, which might be 3 to 5x larger than the high-voltage market from my estimation. Most recently, we have seen GaN devices emerge for high-voltage motor drives. Texas Instruments launched in June of this year at the PCIM power electronics trade show, a 3-phase 250 watts power inverter. With such devices, you can massively boost the energy efficiency of home appliances. And we all know that efficiency has become one of the key selling points in white goods. Furthermore, we see gradual adoption of GaN in electric vehicle onboard charger in short OBC. Here, gallium nitride competes with silicon carbide and with traditional silicon power devices, but it has benefits when it comes to bidirectional charging. Overall, we can say that GaN is set to replace silicon application by application and step by step. Many of these applications, but not all, are captured by market analysts. What is not fully captured yet is the market prediction but in the market predictions by analysts in our view, is the next and upcoming wave of GaN applications. The most exciting are GaN power delivery for AI chip and highest voltage GaN at 1,200 volt for electric vehicles [indiscernible]. Turning to Page 10. Let us have a look at GaN power delivery for AI chip. We all know that one of the biggest challenges for AI is the energy consumption. The power consumption of NVIDIA AI chip is growing by a factor of 3 from 2020 to 25. A recent study published by the German business newspaper Handelsblatt expects global power consumption driven by AI to grow from about 40 terawatt hours in '23 to about 140 terawatt hours by '28. This is in 5 years or 30% CAGR per year. At the recent imec Technology Forum this year, Lisa Su showed the power consumption needed to train the next generation of LLM, which equated to the power output of an entire nuclear power plant, quite a big number. In short, the industry must address the power consumption of AI chip to enable the continued evolution of this technology. We can and we will contribute with GaN for power delivery to address this challenge. Throughout the next generation of AI GPUs GaN switches will gradually be deployed in the conversion chain from the wall-plug at 240 volts to the individual CPU and transistor and 1 volt. The conversion and distribution across the chip occurred in multiple steps. First down to 48 volts, then 12 volts and then finally, to 1 volt where the individual transistor is sitting. In all these conversions, you have silicon power devices today and they're all set to be replaced by gallium nitride chips in the future to increase the conversion efficiency. This is particularly interesting because the die size increases with current. In an oversimplified manner, you might say that for 1,000 watt power supply or power you might need 4 to 5 switches of 1 amp to carry the load at 240 volts, but close to the individual CPU and the individual transistor, you might need 1,000 switches of 1 amp to carry the same load. And because at 1 volt, 1 volt is the voltage used within the chip. This is not precisely the math, but it serves you to illustrate the idea that the further you go down in voltage, the more switches you need and the more die size you need. We expect this growth opportunity to kick in '26 or '27 onwards with gallium nitride. The other trend that we would like to bring on your radar today is the usage of GaN in very high-voltage application on the other side of the spectrum. We have illustrated this at Page 11. Until today, these applications were exclusively served by silicon IGBTs and silicon carbide MOSFET. Now gallium nitride has also achieved quite some voltages of 1,200 volts, which you see on the left. Multiple customers are working on GaN technology for the electric vehicle main inverter also in collaboration with car OEMs. In this application, gallium nitride takes on the direct competition with silicon carbide. Those players who focus on GaN for highest voltage assume that gallium nitride will realize a deeper cost down curve compared to silicon carbide and eventually be the more cost-competitive solution. This trend would be beneficial for AIXTRON because in silicon carbide, the substrate makes up a big part of the cost, while in gallium nitride, the SC takes the largest share. We don't know today how this race will eventually turn out, but we are preparing for it in terms of potential volume upside. When the wave comes, we are ready to catch it, and we want to address the resulting customer demand. Hence, overall, we see ourselves very well positioned with our G10-GaN to address these applications and the resulting requirements of our customers, be it in terms of technology, or in terms of high-volume manufacturing capabilities. With that, let me now turn to our view on the current silicon market, silicon carbide market dynamics. We observed that at this point in time, the overall market worldwide has built more capacity than there is demand. Nevertheless, we and our customers remain very positive about the mid- to long-term outlook. The electric vehicle market will continue to grow, driven not only by regulatory demand, but much more importantly, by better product offerings. In segment, where the car offering is attractive, the EV adoption is currently increasing. We will see that we expect more and more cars with next-generation battery technology coming to the market offering lower cost and better range. And all this also requires silicon carbide inverter. Furthermore, the situation at each of our customers very strongly. Despite the overall market slowdown, the volume ramp of a number of larger players continues, while smaller players appear to be consolidating or slow down significantly compared to the initial plan. We had anticipated increasing cost pressure and commoditization in silicon carbide market right from the beginning, and our expectations are now being confirmed. The trend increases the focus on cost per wafer, which plays to the strength of our product. The G10-SiC has established itself as the most productive tool in the market. meeting the required technical specification and offering the lowest cost per wafer. We maintain a clear #1 position with a growing market share. Our very strong silicon carbide orders booked in Q2 are the best evidence for our view on the market and our competitive position. As I said the last time, money talks. Now let me move on to our infrastructure investments, which we have made in light of the expected market development that we outlined just before. In June, we announced the purchase of a new site near Turin, Italy. An existing building was acquired for a single-digit million euro investment, including the facility investments, total spend will initially be a low double-digit million euro figure. This will allow us to rapidly expand our production capacity up to a potential twofold increase of manufacturing volume in the future. Such increase would require further investments in infrastructure of the facility and hiring of additional staff. With this, we are strategically preparing ourselves for the future, which I just described in power electronics. The Turin site especially gives us the flexibility to address the expected increases in demand for major customers, and it will be able to cover future order peaks at all times. Finally, in May of last year, we announced that we will be expanding our facility at our headquarters in Herzogenrath, Germany with the new innovation center. This is adding about 1,000 square meter of cleaning space to our R&D operation. It will allow us a much deeper collaboration and co-development with our customers. The construction is progressing well. We expect to move our first system into the innovation center as planned in the second half of the year. With this, let me now give you an update on our adjusted full year guidance for the year. As stated before, the stated significant amount of orders received in Q2 is already for shipment next year. Hence, we have adjusted our guidance for '24 as follows. We expect total revenues in the range between EUR 620 million and EUR 660 million before EUR 630 million to EUR 720 million. We expect a gross margin in the range of 43% to 45%, which is unchanged. And we expect EBIT margin in the range of 22% to 25% from before 24% to 26%. Our revenue guidance for the subsequent quarter is as follows: for Q3 '24, we expect revenues between EUR 150 million and EUR 180 million, where production plans will be quite balanced between Q3 and Q4, shipments will be more Q4 heavy based on customer plans to take delivery of the system, which are mostly derived from customers at completion date. We have seen this in the last year, so nothing new here. And with that, I'll pass it back to Christian before we take questions. Christian?

Thank you so much, Felix. Thank you, Christian. Anna, we're now ready to take the questions.

[Operator Instructions] The first question comes from Martin Marandon of ODDO BHF.

My first question maybe is on the order momentum. Can you maybe comment what you said a few weeks ago mean that you expect similar kind of orders over the next few quarters? Is it reiterated? And I have a follow-up.

Yes, definitely. Thank you. Yes, we look -- when I made that statement, and I can reconfirm that statement today. We looked at our pipeline and the discussions we currently have with customers. The pipeline is quite strong. We have ideas when some of the customers want to have the tools. We know our lead times. And based on all these factors, we have given out our expectations that the momentum in the Q3 '24 and also the Q4 '24 will be somewhere on the level of what we've seen in the second quarter. And it was important to us to give that message to really show you that the second quarter is not just a one-off. And then afterwards, it goes down, but we clearly expect Q3 and Q4 to be on the same level.

Okay. And maybe now a question on GaN. Since you talked a lot about it during the presentation. What is your visibility on GaN adoption for AI power delivery, meaning that can you help us understand all the cycle between qualification, adoption, how long does it take before the chip is being qualified and you have some idea of when GaN will be really adopted in a massive adoption in AI?

Thanks that you asked a million dollar question. So we put in the presentation that the best guess we have as of now that this is going to start in '26 onwards, so to say, then a trend ramping up in volume. Again, a lot of this happened at our customers, who use our tools at G10 or some also older generations, yes. So to say, to do this -- so not everything is transparent to us, but that's the best guess we can give. And we all know that the energy consumption is a very pressing point to the industry. So we see that there's a big demand, a big need and I think everybody is now running as fast as they can. And this year, '26 is our best estimate. '26 onwards, starting in '26 and then in the following years to draw volume.

Okay. Maybe the last one, if I may, on the laser business. How do all the innovation we are seeing today on photonic and 3D integration of photonic cases directly on the packaging translates into demand for AIXTRON laser business. I mean, what I'm trying to understand is that, does that necessarily mean that laser demand will be structurally higher compared when the photonic is not on the packaging.

Thanks. That's a very, very good question. What we do see is, overall, that the data volume is continues to exponentially grow. Again, lasers and optical data communication can be one factor contributing to the energy challenge of AI, not to the compute within the chip for sure, but to the compute from one card to another card, yes, for the data transfer within data center, but also from data center to data center. So we expect overall an increase in demand. Now how exactly that plays out with multiple and various architectures being currently being discussed, the question you raised. We don't have full transparency on that one yet. As soon as we have transparency, I'm very happy to provide an update here on this play.

The next question comes from Olivia Honychurch of Jefferies.

My first is on the 2025 outlook. Now that we're at the half point in the year, it feels like a good time to be checking in on where you're seeing 2025 growth versus what you guided for us the FY '24 results in February. Back then, you said that you're expecting strong growth in revenues in 2025. And given that since then, you've downgraded your 2024 guidance because of some shipments pushing out of this year and into 2025. Can we therefore assume that the growth outlook for next year hasn't been stronger than what it was as those shipments switch from this year to next? And then I have a follow-up.

There is a lot Olivia. Honestly, it's a bit too early, given the dynamics in the market to comment now reliably on '25. I think a lot is currently in dynamic and changing. Just look what today's day has brought to the market. I would want to wait a little bit before providing indications for '25. What I can tell you is and what we have already outlined here is that we do see, especially in silicon carbide, some of our large customers continuing to invest through the cycle with a clear ambition to have factories up and running capacities, not only installed, but really fully operational in full swing with full yield and full operational efficiency and that some of those customers are clearly gearing up to say, look, today, there's many, many, many players in the silicon carbide market. There is going to be a shakeout. We've just seen some news this week. And if I'm the first one to invest through the cycle to put up the capacity to have it up and running and I achieve a superior cost position and at some point, I can then gain market share. And some of those players who have this cost focus and volume focus have shown deliberately our tool. And they do continue to clearly place orders now with clear shipments already in '25 so they don't care so much about the current environment, but they rather put those capacities in the ground. So that I do confirm. However, how that fills up to the total picture, which is needed to give an outlook for the overall guidance, that's a little bit too early given how many moving parts are around us.

Okay. That's fair. My follow-up is around the full year margin guidance. So 37% is what you achieved in H1 on the gross margin level. You ask why '24 guidance is 43% to 45%, you'll be needing to get above 46.5% in H2 to get to even the lower end of that target range, which I think would be the highest gross margin you've ever achieved. What gives you the confidence that you can get to that level, particularly in such a low growth year on the top line?

Yes. Olivia, I'll take the question. Well, I mean, we are quite clear on the shipments in the second half of the year given our order book. The key here is, of course, a much better product mix then some better fixed cost aggression due to the higher volumes. And all of that together then will result in this improved gross margin that will bring us to the target. So the key here is really the improved product mix also the end market mix, but also the new generation versus old generation mix.

Next question comes from Michael Kuhn of Deutsche Bank.

Firstly, one on order intake, which was obviously strong in the second quarter still a major part of that seems to be for 2025. And that, let's say, despite the fact that you have, let's say, substantial inventory, which I think your clients also see. So my question would be what makes your clients order now with such a substantial lead time? And were there any, let's say, concessions on your side made towards the clients to make them order now be it on prepayments or any other contract conditions?

Great question. I could see you also being a purchaser at one of my customers. No concession with me, sorry, I have to decline. So on a serious note, Michael, some customers have clearly they have build-out plans, especially those silicon carbide guys I was talking about and they put their factories on the ground. They have their clear plan when they want to ramp when they want to get volume up in place and they see no need, so to say, to push out ordering of the equipment. They rather put the orders in place when they have their plans and the plans are getting ready. And we would not be putting concessions on getting nice numbers, but in the end, get lower margins, sorry, not with me.

That is good to hear, Felix. Then staying with related topic, inventories. You mentioned only a slight increase in the second quarter and you obviously said earlier that you only expect an improvement in the second half, let's say, as we are now in the second half, and you probably have a clearer picture. Do you feel comfortable to, let's say, roughly quantify what kind of reduction you could achieve in the second half, and you mentioned some measures only taking effect into '25. So maybe also a rough outlook on '25 and let's say, what level of inventories you would feel comfortable with over the upcoming quarters?

That is a lot. Let me also take that. So by the middle of '25, we want to be back to normal. So when we talk in 12 months from now, we want to be back to a normal level. The reason is we have decided to take a soft breaking strategy, rather hard breaking strategy because we clearly want to be in a collaborative situation with our suppliers. You see despite the weak environment, we are doing well. We have a very good cash position. And we didn't want to break it too hard because at some point, the next upswing is coming. We are clearly seeing already that towards that. And then we will be on the other position, not slowing down with our suppliers, but rather asking them to accelerate. And that was the reason why instead of doing a hard break strategy, which, of course, we could have done let's say, we decided to do a soft one. And you will see a gradual degression by the Q3, not too much, don't over-expect you will see and then you will see a digression of the inventories Q4, Q1, Q2. And by the middle of next year, we should be back to normal. It really follows from our strategy.

Okay. And could you, let's say, like quantify it back to normal?

Well, like the levels of -- I would consider normal levels of 60%, 70% of order backlog. With some good optimization of our -- like our internal processes, maybe we can get down even further, but that will be my best guess right now.

That's a clear quantification. And then one more on, let's say, LED/micro LED. Could you provide a rough split for the first half, what the contribution of each of the 2 was? And let's say, how much more, let's say, margin-dilutive old tool LED business you expect to generate in the second half of the year?

So let me just check the numbers here. I think in the second quarter, the whole was 40% or over 40% of the total revenues. So that was quite a big one. I think throughout the first half, maybe 25%, something like that, 25%, 30%, something like this. So let's say, a little over 30% as a total. And honestly, the split in what it made is we don't have the numbers here in terms of margin effect. I mean you saw the overall gross margin where we came out. That's clear.

And the major hit, of course, came from the traditional red LED part, which comes at a lower margin.

All right. No, understood. And then very last question, promised. I think in some point in the release, you speak about ongoing upgrades of the G10-SiC. Is that also applying to tools that are already installed at your client [ fabs ]? And was there any cost associated to that in the P&L in the first half?

Yes, that definitely applies to installed in our customers. Typically, that involves a certain qualification time. And after that, the customer then buys the upgrades. And that -- given that it improves the tool -- it's not free of charge, but it comes at a charge because it's improvement that we provide to the customers. So no, it doesn't come at a cost, but rather it comes as a benefit to [indiscernible].

The next question comes from Madeleine Jenkins of UBS.

We hear your customers referring to the consumables for silicon carbide [ as toxic ], is that a key factor to consider when using a supplier? I was just wondering if you could talk a bit more about that and what that involves exactly? And also how it might differ between a batch tool versus a single wafer?

I didn't fully get the question accurately, Madeleine. Could you please repeat?

The consumables for silicon carbide [indiscernible]. Just kind of what that involves exactly and then how it differs from a batch to a single wafer tool.

Yes. Okay. So I do understand the question is how consumables contribute to the different batch to single wafer tool, is that the question you're asking?