X-FAB Silicon Foundries Aktie

Ladies and gentlemen, welcome to the X-FAB Third Quarter 2026 Results Conference Call. The speakers today will be Damien Macq, CEO; and Alba Morganti, CFO. [Operator Instructions]

I will hand the conference over to Damien Macq, CEO. The floor is yours. Please go ahead.

Thank you, operator. Ladies and gentlemen, good morning, good afternoon. The third quarter provides further evidence of the recovery we started to see earlier by the end of the last quarter is gradually taking shape. Revenue amounted to $194.4 million. Excluding the impact of revenue recognized over time under IFRS-15 Revenue reached USD 200.7 million, up 2% sequentially and within our guidance of $195 million to $205 million. More importantly, bookings increased 30% quarter-over-quarter and 38% year-on-year to $224.7 million. The highest level -- this resulted in a book-to-bill ratio of 1.12 and brought our quarter end backlog to USD 313.6 million.

Our core automotive, industrial and medical markets generated $187.8 million in revenue, representing 94% of our total revenue, excluding IFRS-15. Automotive revenue return returned to sequential growth, increasing 3% to $119.1 million. Even more encouragingly, automotive booking increased significantly as the inventory correction that affected auto no longer weighted on demand. This is an important indication that conditions in our largest end markets are progressively normalizing.

Industrial revenue increased 10% sequentially to USD 49.6 million, supported in particular by the strong development of silicon carbide. Key customer reported as well gains on the robotic markets. Medical revenue amounted to USD 19.1 million in the third quarter. While this was 9% below the previous quarter, underlying demand remains strong in ultrasound and pacemaker applications, partly offset by lower demand from a major customer lab-on-a-chip device and for temperature sensors.

Looking beyond the quarter, the medical pipeline remains promising. During Q3, we secured important design wins for a silicon photomultiplier used in medical scanning as well as for handled ultrasound application further strengthening our position in this attractive market. Together, the improving order situation in automotive, the sequential growth in industrial and the opportunity developing in medical reinforce our confidence in a gradual recovery.

Capacity utilization across our manufacturing network is also increasing [indiscernible] prerequisite for converting higher demand into improved operational performance. At the same time, we are seeing encouraging evidence that the growth opportunity we have been developing beyond our traditional markets are starting to materialize.

In our Smart CMOS & SOI business, strong demand for battery management system and AI data center test application supported production ramps during the quarter. We also secured an important design win for an activator used in optical circuit switches for data center. Our CMOS business -- sensor business is gaining momentum as well with customers looking for alternative supply sources. All these developments illustrate how its core competence in analog and mid-single semiconductor manufacturing are opening additional opportunities in medical and data center application.

[indiscernible]. Revenue reached USD 16.2 million, increasing 68% year-on-year. Silicon carbide accounted for most of this business with its fab recording an all-time high number of silicon carbide wafers sold during the quarter, up 45% year-on-year. At the same time, our gallium nitride activity continued to gain momentum with a new number of gallium customers doubling from 6 to 12 over the past year. This development demonstrates the growing contribution of Wide-bandgap technologies to its diversification strategy.

Microsystems & Photonics represents another important growth pillar. Microsystems & Photonics revenue amounted to USD 25.9 million in the quarter. Demand for MEMS sensor was particularly strong, especially from China. We also installed a pilot line for an industrial inkjet printing application based on [indiscernible] material further broadening the range of applications that can be industrialized in our manufacturing platform.

In Photonics, we are seeing an important transition from technology development towards commercial engagement. Our heterogenous integration technology for integrating light source is generating strong customer interest and photonic prototyping revenue is growing. Paid engineering activity is increasing as well with photonic NRE expected to grow by approximately 40% year-over-year.

Customer engagement are developing in areas such as light source integration and optical interconnect for data communication and AI infrastructure. This is an important development for X-FAB, the technology and capability we have built over the recent years are increasingly translating into concrete customer projects. We are supporting this opportunity with the right industrial platform. Construction of [indiscernible] has started, further expanding our capability in micro system and related technologies. At the same time, the transformation of our existing manufacturing footprint in [indiscernible] continues according to the plan.

The end of line activity for Side B in [indiscernible] are executed in a controlled manner with completion currently targeted around the end of the first quarter 2027. This development are 2 sides of the transformation -- these developments are 2 sides of the same transformation. While we are progressively phasing out an older manufacturing footprint, we are simultaneous creating the foundation for the future of [indiscernible]. This transformation will further strengthen the sites focus on differentiated technologies and applications and reinforms its long-term role within IFA global manufacturing network.

I will now pass the word to Alba, our CFO, to give you more update on the financials.

Thank you, Damien. Good evening, ladies and gentlemen. Let me now take you through our financial performance for the third quarter.

In the third quarter, our revenue and EBITDA margin landed within our guidance and our profitability improved sequentially. The stronger bookings also support our expectation of a gradual recovery. However, we remain focused on our financial discipline. Our EBITDA reached $37.5 million, which shows an increase of 12% compared to the second quarter. Our EBITDA margin was 19.3% compared to 16.8% in the second quarter. If we exclude the IFRS-15 impact, the margin was still within our guidance of 17% to 20%. Operating profit was $4.3 million compared to the $2.1 million realized in Q2. Our financial focus remains on translating higher activity into stronger profitability, while maintaining cost controls over cost.

Our CapEx totalized almost $32 million in the third quarter. We also received the first funding payment under the European Chip Act for our new microsystem fab [indiscernible]. If we deduct this funding, the net CapEx is a bit less than $30 million. Our priority remains to balance investment in future capabilities with disciplined capital allocation and liquidity management.

At third -- at quarter end, cash and cash equivalent stood at $196 million, representing an increase of $32.5 million compared to the previous quarter. Cash generation, working capital management and investment discipline remains central financial priorities for us.

And now let's have a look at our fourth quarter outlook. Looking ahead, we expect our fourth quarter revenue to come in between $205 million and $215 million, with an EBITDA margin of 19% to 22%. As in the previous quarter, this guidance excludes the impact of IFRS-15 revenue recognition over time and assumes an average exchange rate of USD 1.14 per euro. The guidance point to further sequential improvement supported by the stronger order intake and increasing capacity utilization.

And to conclude, I would like to highlight that our third quarter performance confirms the business recovery. But nevertheless, our financial priorities remain to convert order momentum into profitable revenue, maintain investment discipline and strengthen our cash position.

And now I will pass the word back to Damien.

Thank you, Alba. Overall, we are entering in the final quarter of 2026 with increasing confidence while remaining focused on execution. Our core markets are recovering, bookings are at the highest level in 2 years -- in the past 2 years, factory utilization is improving and new opportunities in medical, data center, Wide-bandgap technologies, Microsystems & Photonics are progressively moving from development towards business. Over the last several years, we deliberately invested in differentiated technologies, capabilities and capacity also through the long term. At the same time, we continue to adapt our manufacturing footprint to support its fabs further grow.

Our priority is now clear, execute this transformation successfully, convert the recovery and our growing opportunity pipeline into production, increased utilization of our manufacturing assets and translate this into sustainable and profitable growth. The third quarter represents another important step in that direction.

And with this, I wanted to thank you for listening and open the floor for questions. Thank you all.

[Operator Instructions] The first question is coming from Aleksander Peterc from Bernstein.

First of all, can you tell us a little bit more about the backlog that continues to increase nicely. You have a good book-to-bill. Is this feeding predominantly stronger growth into next year? Or is it going to be largely consumed already in the third -- in the final quarter of the year, where is this going? And I also know that you have a nice inflection in prototyping in most businesses, is this also an indication of activity into next year? And then I have a small follow-up.

All right. So regarding the backlog, yes, we got a solid backlog ahead. Typically, the backlog that is created is now lasting for 3 to 5 months typically, our cycle time remain under control, which means that most of our customers are ordering as late as they can. And so our visibility in the backlog for the full year 2027 is only going for us towards Q4 and partly in Q1. But all in all, if you look at the backlog combined with the forecast, we see a nice trend coming in our way at this point. So that is confirmed by the number that Alba just guided and how this evolves into Q1 is something that we will expose later on, but it's not a one shot. We see a trend.

Regarding the prototyping, I'm not sure I got completely the question. So we see on the prototyping some evolution that is a bit flattish in the sense that if you compare our prototyping revenue over the past few quarters, you can observe a kind of stabilization around a lower level versus what we used to have in '23 and '22. Now remember, the prototyping numbers that we provided back during these years included also some of the some of the revenue related to ramp-up of products in preproduction. So the NRE revenue curve that you see that is reported in our data includes this overshoot for the year '23 -- '22 to '24, let's say.

What we see, and that's not visible in the data that we provide, we are evaluating also the underneath evolution of our CMOS & SOI and we see some trends towards an increased number of tape-outs and also an increased number of value for tape-outs. So far, we do not publish that. We want to consolidate it and build a solid history. But as soon as we have a bit more information about that, we will also share it with you. I hope that answered your question, Alex.

Yes, yes. Just a quick follow-up on the wide-bandgap, could you tell us what is the breakdown between your data center customers and automotive customers in silicon carbide?

Compared to Q2, it's not changed drastically. So we still see 50 -- close to 50% in what we call industry and within industrial data center some volatility in automotive. So we have a customer in China generating a bit of volatility in our numbers. but we see very solid growth in the data center area. So that's -- moving forward, we anticipate to see continuous growth in data center. It's really quite impressive what some customers are achieving there. And from interaction with them, we see also a solid forecast coming in our way for 2027. Automotive, we will have to see how this business continue to evolve over the next quarter.

The next question is coming from Emmanuel Matot from ODDO BHF.

First, in automotive, do you now have enough visibility to call a sustainable recovery than just inventory normalization? Second, still in automotive, there any areas where fab has lost content or market shares during the previous negative cycle [indiscernible] that lasted 2 years. And do you see more competition from China for some specific key products [indiscernible]. And my last question, could you please let us know the total amount of LTA advanced payments that you still need to reform to your customers at the end of September?

Thank you, Emmanuel. So in automotive, the visibility right now is good. If we talk to our customers, they come with a positive view on how automotive is evolving. Remember, we mentioned this inventory correction. We believe it's behind us now. We believe we will now see the basic growth that we anticipate in automotive that is linked to electrification, that is linked to customization and also the upgrading of electric functions through the segment in automotive. So even with a constant number of cars being sold in the world. The number of chips sold [indiscernible] are expected to continue to increase and we believe this increase is what we will continue to see. The visibility in terms of numbers is low for the reason I mentioned earlier we have now customers ordering late. We have optimized cycle time. And so I think that's a good and healthy for the business overall. But unlike the times like 2, 3 years ago where the fab were very, very loaded and where our cycle time was extremely long. The visibility has reduced from that effect. But from indications that we have from customers, we get a positive view on the evolution for the technologies that we are deploying in automotive.

If you look at the second question regarding the business that we might have lost, it's difficult to answer to that question. We see some customers telling us that they see renewed interest from the fact that we have capacity available. And so this year renewed interest from Tier 1s. This renewed interest [indiscernible] for the technologies of fab.

I must also say that over the past 3 years, a tremendous effort has been made in improving our overall quality and operational efficiency in a general sense. And that is also helping our customers and their business.

Regarding competition from China, we see some customers, some Chinese customers ordering at [indiscernible]. So I mentioned the BMS opportunity that is coming directly from China, for example, but also some of the MEMS and there, we see a strong demand. We know that competition will arrive and is preparing also some products on their side. But at this point in time, we have not lost opportunity or we cannot -- I cannot report any big opportunity that we might have lost from competitors in China.

On the contrary, we see a fabless in China coming to its fab because they face a challenging situation. All the fab in China are reporting a fully loaded situation. They are very actively deploying chips for AI and around AI and so I think also fabless in China are securing their supply chain by continuing to interact with us. So right now, this competition from China is not touching us. That's about your second question.

And for the third question on LTE, I will let Alba answer.

So at the end of Q3, our total balance, the residual amount due to customers under LTA prepayments was $173 million. So you can consider that by end up with another $20 million to be repaid within Q4.

And the [indiscernible] will be repaid next year.

No. It's roughly $20 million per quarter. So it will take another -- it will take us to -- up to 2028, but the vast majority will be already repaid in the next year.

We should be halfway end of the year. That's how you should look at [indiscernible].

Yes, that's the message. End of '26, we should be around 50% repayment done in total.

The next question is coming from Robert Sanders from Deutsche Bank.

I have a few questions. Yes, first question would just be around your bookings. Could you sort of estimate what bookings -- what percentage of your bookings are data center related? And I have a follow-up.

That's a good question. I would not be able to provide you the number. I think if you look at the booking right now, the majority of the data center booking is still on silicon carbide. But as I mentioned, we have also some some peripheral activities in the CMOS for pre-drivers and also, yes, for the the chips that are involved with the comparative monitoring and temperature management of data center. So based on motor drivers, either find drivers or cooling drivers for liquid cooling of data center. So we know about the application, but basically the same product or similar products are sold either in automotive or into data centers. So it's a bit difficult to get the exact value that would go directly into the data center.

Got it. And can you talk a bit about the situation in the industry in terms of capacity allocation challenges, some nodes being full, IDM shutting down 200-millimeter, TSMC and Samsung cutting wafers to too, western companies. How is that affecting you? And where are you seeing the demand by node?

It is somehow helping us. I was mentioning to some extent, we invested probably at a good time. So right now, there are not so many factories who are claiming to have capacity. And I think it's very, very challenging for foundries today to create capacity because we see the price and the lead time for equipment increasing also the effort that are made by mega factory in Texas that is very well known, seem to absorb a lot of the bandwidth of -- and a lot of the orders of tooling for equipment suppliers. So there is some challenge for factories right now to increase their capacity.

So from that perspective, we have a good position and the consequence of that is that, yes, we see new customers coming in our way to look at our technology to support their supply -- to expand, in fact, their supply chain. So some customers who used to work exclusively with Taiwanese factories start to talk to us, and we are well engaged on some design wins with some of them. I cannot be specific about that. But all in all, it's really helping us.

We see also some customers starting to request a new type of LTAs with us. So that's also somehow an indication that there are some challenge on the supply side. And I think from our knowledge of what's happening in Asia, I mentioned to our knowledge, most of the factories in China involved with -- also involved with mature nodes are fully loaded. They created for some capacity in 300 millimeter, but they are loaded. And the problem is aggravated, as you mentioned, by TSMC, they closed 2 factories already announced last year in [indiscernible] they transformed them into factories capable to support the packaging and assembly for data center. And likewise, Samsung also closed one factory. That's -- these are the things that we know about.

And this is helping because some customers come to us for their new developments. But this is also helping our customers because they start to see that they are regaining some designs because some of their competitors are unable to supply or have challenged ahead to supply these customers. So all in all, I think we are in a good position from that perspective.

Sorry, just to be clear, the mega factory in [indiscernible], you're talking about [indiscernible]? I mean, surely, those would be buying [indiscernible] not mature tools because I have not...

They don't buy, of course, mature tool. They buy advanced tools. But from time to time, we need also -- we don't need advanced litho in our fabs. But when it comes about ordering to the same suppliers, talk about [indiscernible] and others. I mean all these guys, they see big orders coming in their way for the upcoming factory and the mega factory is one of them. And this is creating tension on the equipment supply. So the impact, what we see is that the -- we anticipate that the lead time for equipment will increase.

So given you're one of the few people to have spare capacity, how quickly could you fill your fab?

That's a good question. Overall, if you look at the fab filling, it depends if it is about customers who have existing product and they gain market share with their existing products. This goes extremely fast. So there, we can see demand raising. And we believe this demand raising is a combination of our customers gaining certain market share plus the fact that the intrinsic business in automotive, for example, keeps increasing. So there is this effect that plays also in our favor that comes quick.

When we talk about customers that are developing new product with us or considering to use its fab for the next generation that takes much more time. So it's there is no simple answer to your question.

On our side, what we are doing, you remember maybe in the previous call, we were mentioning that we were doing active management of our capacity to adjust the capacity to the demand. We are now in the process of reopening our dynamic capacity to be able to support this upcoming business.

So the thing to track from here will be, you can potentially get line of sight to the 1.05 goal if you get to full loading but the thing to track here is your depreciation, right, because there's a lot of tools that you've already bought that are sort of not kind of switched on right now. So the depreciation will presume go up as you activate spare capacity? And that's it for me.

Yes. I will not comment on -- numerically, I will not give numbers on this, but yes, we have pushed out the depreciation of tools in -- mostly in [indiscernible]. But in the second half of this year, this will be fully done. So most of the depreciation of our factories that is linked to the $1 billion investment we did over the past 3 years will start to be depreciating by the end of this year. So this is the tool, in fact, what we do typically is that we validate that the tool and some basic recipes are working on it, and so we can quickly activate these tools. So this is not slowing down or refraining any issue there.

The next question is coming from Ruben Devos from Kepler Cheuvreux.

Just maybe not thinking maybe mid to long term, but right, but next year, would love to have a bit more sense where you could end there? I mean $220 million of bookings, you talked about a good trend, not a onetime thing. What you've been characterizing throughout the call, sounds like book-to-bill will remain above 1%. So $220 million annualized could be [ 880 ], right? Is that -- is that wrong to have that in our models, high 800, something like that?

And I guess, I mean, you talked about visibility 3 to 5 months, right? So typically, you know more in Q4 and Q1. But what are maybe the drivers you have most visibility on today?

Yes, I think your calculation is correct. That's on the [indiscernible] right now, we are -- we would not guide for 2028. But indication, I think all your statements are sensible. You have to take into consideration also that next year, we will execute the end of life. So the end of life of the business in Site B will be completed around March, April. So we still see orders coming in our way. So the exact date of closing this site is still around this end of Q1, early Q2. And so this last time buy will reduce the overall revenue in the range of $50 million for 2027.

So you can assume on a yearly basis that we run around $80 million in Site B that we are closing in April let's say, March, April. That's one nuance on your numbers. And for the rest, how the year will evolve, I think that's going to be disclosed in -- once we get a bit more visibility during the Q4 results.

Okay. That was 5-0 or 50?

50. Sorry, it's 50 and on a yearly basis, it's $80 million. That's what's the business that is coming out of Site B. [indiscernible]

Okay. And then maybe if I just think about your Q4 guidance, right, first like the midpoint sort of suggesting you could have $9 million more sales and a $6 million more EBITDA. So it's just about 60% drop-through. Is that sort of the number we could be looking for going forward, like you've also had a cost program, I think that's ramping up, right? Not much higher depreciation to expect. So yes, that drop-through rate, is that something we can work with? .

In terms of what -- which rate? EBITDA margin, you mean?

Yes. EBITDA.

I think it's reasonable. Of course, we have different effects. We still have. So our cost optimization -- a big portion of the cost optimization comes from managing the headcount. So we have some reduction on the headcount over the year. And this reduction of the headcount in 2026 calls also for some severance, okay, I was looking for the word, severance payments that are impacting 2026. So that plays on the cost saving. And regarding the depreciation, keep in mind the depreciation for the site in [indiscernible] will be full in Q4. So we can see, yes, if it is end of Q4 or beginning of Q4, maybe that also can [indiscernible]. But I think the overall model is reasonable.

Okay. And then just a final one on the design win for an actuator in the optical circuit switches. Can you maybe say a bit more about what you're actually making there? Is it a sole source position? Or are you qualifying as a second source next to an existing supplier?

Yes. I think it's a technology that is used in the architecture of Google. So it's not that we -- it's a Google alike type of switching on the backside of the GPUs, which is all optical based on MEMS mirrors. And so our chip is capable to drive the high voltage that is required to drive this MEMS device. So they rely on the SOI capability, and our opportunity to drive relatively high voltage there. So that's the opportunity. That's what I can say about this opportunity. Now we see multiple interests coming for this type of capability, not only from one customer, we talk to multiple customers about this one.

Okay. And the production volumes for the one you've won, when will that...

Yes. It's too early to talk about volume. We are in a design win phase. So volume will be 2, 3 years from now, but this type of architecture is quite attractive, I think, overall. So I don't want to talk about volume because if I [indiscernible] the volume that I get from the different customers there, I get quite impressive numbers. So I think as soon as we see how the things are coming across, we are not the sole source for this type of technology. But as soon as we see are the volume ramp up, we will be able to disclose this. Also what you see as an effect is that multiple customers go probably for a similar opportunity. So it's difficult to sort out what exact volume will get there. But it looks like all the numbers that we look at are big.

The next question is coming from Luke [indiscernible]

First of all, we see after a long time, indeed some encouraging improvement. So that is very pleasant scene. But my question is, could you highlight a little bit more on the cost savings, which were program actually for the fourth quarter, which was announced in the first and second quarter?

Yes. So the cost savings were, as Damien mentioned, the main driver came from the headcount reduction that we started in several months ago already and we continue to go ahead with. So we are in hiring freeze. We make quite some evaluations on the team. We do -- we encourage also some pre-departures. As mentioned by Damien, we paid also some severances to some of the people. And all this -- the cost savings per se are not always visible immediately because sometimes you have to to pay the severance, which is an additional cost and the savings will be seen a bit later.

But on and on, we are on track with all the measures. We also -- before in the first quarter, we also [indiscernible] tools. So we put them then down, if not needed. We try to optimize our ballpark of tools. We also pushed out some CapEx that were not absolutely needed. And this kind of normal measures that are quite classical, I would say, but the conclusion is we are on track.

To complement what Alba says, so as you realize the visibility on the saving will come more next year once the severance will be fully paid and the people will be of the payroll I want to highlight also on the tools. So there is the standby of tools that is helping as long as we don't need that tool. But what we have done also, and that's a bit more difficult to quantify overall. What we are doing is also to make sure that we concentrate the activity and we qualify the most modern tool.

And so overall, if you look at our [indiscernible] of tools, we are also removing older tools from our factories. By the way, we managed to sell these tools because the market for tools is now attractive again. And so we now are considering selling older tools to certain factory also in Asia. But this impact of running our production on more modern tool will also have -- will also be positive on the cost side because the same tool is using a much more -- is providing much more performance and so much, much less -- much less chemicals, electricity and more efficiency per operator for the same tool. So this is also going to help.

And then actually, we need to see and now coming through the new orders through the full [indiscernible] and also to see it at the bottom line.

Yes. Absolutely. So from our side, as I mentioned, it's -- the -- I don't want to say the challenge because that's our job, but the challenges execution of the tools that we have and making sure that we ramp up orderly to secure the demand that is coming from our customers.

At least the information today is very positive and very encouraging.

There are no more oral questions at this time, so I hand the conference back to the speakers.

Yes. We got offline also 3 questions that we might address. There was a question about booking and utilization. How much of the booking surge is real end demand versus customers rebuilding inventory or pulling in orders and how much of the Q4 guide is already covered by backlog? So that was one question from [indiscernible] from KBC.

What I can say is that how much of the bookings serves is real demand and demand versus customary building inventory. To our knowledge, we do not have -- we do not see a clear indication of customer rebuilding inventory. I think they are just inventory to the demand. But I cannot provide a quantified numbers of what is the truly end demand versus active inventory management.

How much of the Q4 guide is already covered by backlog? In fact, if you look at our forecast for Q4, we are already in the backlog. I think, I'm checking the numbers, yes. If you look at the Q4 guidance, we have almost the guidance almost fully covered by backlog. So if you look at our [indiscernible] back, we are within -- or even on the upper side of the guidance. So at the end of the day, it's all about execution for Q4.

There was a second question about automotive. So automotive is still down 90% year-over-year at 119, even though inventory correction no longer weighted on booking, when do you expect automotive revenue to turn positive year-over-year? And what is the lag between booking and revenue in this segment?

So the lag between booking and revenue is typically 3 to 5 months, let's say, 4 months average. And when do you expect automotive revenue to turn positive year-over-year? Indications are there for 2027 that we should see some good numbers coming in our way in automotive. So right now, we see strong orders also in automotive. Maybe one nuance on this is that, at this moment, we are combining in automotive products that seem also to be going directly into robotics -- into the [indiscernible] robots.

To give you an example, a motor driver chip that is typically used for air conditioning management in a car, but also for temperature thermal management of the car. Some of these embedded motor find also applications in robots. The same apply for position sensors. They might find also application in robots. So at this moment, we combine in automotive, the demand coming from our largest customer, [indiscernible], for example, and they report a good demand from robotics. So there we need to work it out also with our customers to figure out what is really automotive, and what is robotics.

I want to highlight also on robotics, we have other historical customers of X-FAB who are also reporting very strong demand in robotics. So that's an exercise that we need to do by ourselves to sort out from automotive, what could be going into robotics because the growth there seems to be quite significant.

And then there was a question about margin and CapEx. Q4 EBITDA margin guidance of 19% to 22% is only a modest step-up, while EBIT is just $4.3 million and net debt is [indiscernible]. What utilization level is needed to bring EBIT margin back to mid-single digit or better? And what is the CapEx outlook for 2027 at the pace of Chip Act Funding?

So I mean, these are good questions. Overall, if you look at the utilization that is required, we do believe that to be around the 25-ish percent EBITDA, we would need probably a loading in the range of 75%. That's what our model shows. But we need to be extremely cautious in the sense that this model are assuming a certain mix of product.

And what we observe is that now our mix of product is transforming quite a bit with the rapid growth that we see in wideband gap, for example. and also anticipated things that will come from a new market like GaN and robotics and photonic forest sorry, by 2028. So that's what we could say about EBITDA in the 25% range. I think EBIT to mid-single digit or better, we can come back on this later on. I don't know if we want to to be too specific about this because for the reason I just mentioned about variation of the mix.

And then what is the CapEx outlook for '27 and the pace of the Chip Act funding? So CapEx outlook, I think Alba mentioned that we are -- we want to stay in the 10% to 15% of revenue with our CapEx. And we are considering the net CapEx, which means we will always report the CapEx minus the Chip Act funding. The Chip Act funding is a range of 30% of our investment. So that's what you can use in your estimates.

I think these were 3 offline questions, I'm sorry for -- I hope I covered that properly. Otherwise, we can have follow-up discussions on the topics.

All right. Then with this, I want to thank everyone for listening to our call this evening. And I want to highlight our next appointment for the Q4 and full year results that will be on February 11, 2027. And I wish everyone a great day or great end of the day. And thank you all. Bye-bye.

Goodbye. Thank you.

Thanks for participating to today's call. You may now disconnect.

Ladies and gentlemen, welcome to the X-FAB Second Quarter 2026 Results Conference Call. The speakers will be Damien Macq, CEO; and Alba Morganti, CFO. [Operator Instructions] I will now hand the conference over to Damien Macq, CEO. The floor is yours. Please go ahead.

Thank you, Gaia. Thank you, everyone. Good evening, everyone, and thank you for joining us today. The second quarter of 2026 showed another quarter of disciplined execution in a market environment that remains mixed. The revenue came in at USD 199.8 million, down 7% year-over-year and up 2% sequentially. Excluding revenue recognized over time, revenue was USD 196 million, within the guidance. Bookings reached USD 173.3 million up 2% quarter-on-quarter, while backlog ended at $291.8 million. Importantly, excluding the temporary impact of the Erfurt end-of-life program, our underlying book-to-bill ratio returned above for the first time since Q2 2024, providing an engaging sign that demand conditions are gradually improving.

Our core markets of automotive, industrial and medical represented 93% of revenue totaling USD 182.1 million. Automotive revenue was USD 116 million, down 19% year-over-year and 5% sequentially, reflecting continued inventory normalization. However, booking improved significantly and automotive book-to-bill reached its highest level in almost 2 years. Combined with a growing number of design wins, this support our view that the market is gradually recovering. Industrial revenue was USD 45.2 million down 4% year-over-year and 13% quarter-on-quarter, primarily due to temporary order volatility from a major silicon carbide customer.

Medical continued to perform strongly, reaching USD 21 million, up 39% year-over-year and 9% sequentially. This is driven by pacemaker and ultrasound applications. In Smart CMOS and SOI, revenue reached USD 156.7 million. During the quarter, we saw a significant increase of a new automotive opportunity, particularly in Europe. And one of our top 10 customers renewed and extended its long-term commitment in pressure sensing. We continue to see healthy demand in battery management system in China, while opportunities like linked to robotics are becoming increasingly visible. We are also seeing growing demand from AI-enabled infrastructure where our technology address power management, sensing, infrastructure control, cooling and connectivity applications.

The broader manufacturing landscape is also evolving. AI-related demand is driving the reallocation of capacity at several 8-inch fabs in Asia while supply resilience and geographic diversification have become increasingly important for customers. With the completion of our Malaysia expansion and availability capacity across Europe, the U.S. and Asia, X-FAB is well positioned to benefit from these trends. In Microsystem and Photonics, revenue amounted to USD 28.7 million up 14% year-over-year. We secured a new high-volume microfluidics application for blood analysis, we continue to see healthy MEMS demand and expanded opportunity for our aluminum nitride platform.

Photonics remain a particularly exciting area. We are making progress in copackage optics, data communication, augmented reality, virtual reality and quantum computing-related applications. Through our collaboration with Ligantech, we have established advanced business interaction with emerging photonics and quantum computing leaders in Europe and the United States. We continue to expect photonics volume production to start in 2028. A major milestone during the quarter was the award of EUR 127.4 million under the European Chip Act to support the expansion of microsystems and photonics manufacturing in airport. Together with the launch of -- together with the launch of -- Micro, this investment reinforce our specialization strategy and positions Erfurt as a key growth engine for MEMS, microfluidics and photonics.

Turning to wideband GAAP. Revenue was USD 10.6 billion, up 34% year-over-year, while quarterly demand remained somewhat volatile, we secured 3 new silicon carbide design wins, expanded our opportunity pipeline and continue to advance customer programs towards future production. Execution of our gallium nitride road map remains on track, and customer interest continues to increase. Looking ahead, one of the most exciting opportunity for X-FAB is the rapid growth of AI-driven data center infrastructure. We see opportunities across all 3 business units including power conversion, sensing photonic connectivity, silicon carbide and gallium nitride technologies. Based on today's pipeline and customers' engagement, we estimate a long-term revenue potential of approximately USD 300 million annually for data center applications.

Finally, the Erfurt end-of-life program remains fully on track and we continue to expect substantial completion during the first half of 2027. Together with FAB for Micro, this transformation support our specialization strategy and creates the foundation for future growth. At the same time, we continue to execute our cost control and productivity initiatives with discipline across the group. And now I will hand over to Alba for financial updates.

Thank you, Damien. Good evening, ladies and gentlemen. We will now continue with the financial update. From a financial point of view, Q2 was a quarter of disciplined execution in a still challenging market environment, which weighed on revenue development and capacity utilization. Nevertheless, we delivered revenue within our guided range. Our EBITDA was of $33.6 million with an EBITDA margin of 16.8%. If we exclude the impact of revenue recognized over time, the EBITDA margin would have reached 17.6% within the guided range of 17% to 20%. Our second quarter profitability mainly reflected the softness in the automotive end market, which limited our capacity utilization. As already mentioned by Damien, our cost savings program is progressing as planned with expected positive effects towards the end of the year. .

We also continue to manage our cash carefully with cash and cash equivalents of $163.6 million at the end of the quarter. In the second quarter, we recorded a noncash tax impact of $11 million related to the derecognition of deferred tax assets in Malaysia, which affected our net result but had no impact on our cash. Overall, our financial priorities remain unchanged. Preserving liquidity, improving operational efficiency and supporting the group's long-term growth opportunity in automotive recovery, data centers, microsystems, photonics and wideband gap technologies. Looking ahead, I can only repeat what Damien said. We remain focused on cost efficiency, cash discipline and on growth opportunity that support our long-term strategy.

Our business continued to be naturally hedged and our profitability remains unaffected by exchange rate fluctuations. At a constant U.S. dollar euro exchange rate of 113, as experienced in the previous year's quarter, the EBITDA margin would have been 0.1 percentage points higher. Our capital expenditures in the second quarter amounted to $24.2 million, bringing first half '26 CapEx to $50.2 million. Cash and cash equivalents totaled $163.6 million at the end of the quarter, while our net debt amounted to $312.3 million, which is an increase of $20.9 million from the previous quarter. And to conclude this financial section, I would like to share our next guidance.

For Q3 2026, our revenue is expected to come in within the range of $195 million to $205 million, with an EBITDA margin in the range of 17% to 20%. This guidance is based on an average exchange rate of USD 114 to euro and does not take into account the impact of IFRS 15. I would like to conclude this financial section by saying that we remain confident in the medium-term outlook and our strong operating leverages, diversified technology portfolio and long-term customer relationship position us very well to benefit from a recovery in the demand. And now I would like to give the Board back to Damien.

Thank you, Alba. As we step back from quarterly fluctuations, 4 messages stand out. The first one is our demand is stabilizing. We reached the bottom with underlying book-to-bill returning above 1.0 for the first time since Q2 2024. The second is data center opportunities are accelerating across CMOS photonics, silicon carbide and gallium nitride. Third, our specialization strategy continues to gain momentum, supported by FAB for Micro, Erfurt transformation and our growing photonic position. And fourth, execution remains strong, demonstrated by the new silicon carbide design wins, our progress in gallium nitride and the successful cost control initiative. And last but not least, the EUR 127.4 million Chips Act funding award that we confirmed. Taken together, this development reinforce our confidence that its fab is increasingly well positioned for the next growth cycle and for sustainable long-term value creation. Thank you, and we are now happy to take any of your questions.

[Operator Instructions] We have no questions at the moment. [Operator Instructions] The first question is coming from Luke [indiscernible] Luke, we cannot hear you, you can maybe try again later. In the meantime, we will take the question from Narahari Narsapur.

Thank you so much for wonderful call -- wonderful updates. My question is, I think we are still not using 100% capacity of the fab. When do you expect to go, I think the full capacity is like [ 1.3 billion ]. When do you expect this to go up. And second question is, recently because of some social media posts fixed stock has seen a lot of iterating units. How does the management see this and can the company benefit from this somehow?

Okay. Thank you, Narahari, for the question. So on your first question, when do we see a full utilization at this stage, end of Q2, the utilization was at the 60% level. So we anticipate a gradual increase of this capacity utilization. The model that we are developing for the company goes for a steady capacity increase, and we want to reach levels in the region, 85% to 90%, 85% will be a good capacity level. The question is about when I think it's all about the market conditions and how fast is the market evolves. We have seen in the past some substantial evolutions. You know that we are on a relatively long supply chain, particularly, for example, in automotive, we have a set of tiers and -- between ourselves and the OEM. And this can create swings in the demand. My anticipation at this stage and from all the KPI that we have, we see that the demand is recovering.

Now how fast this recovery takes place, it's difficult to predict. If you look at our model and based on the growth that you see on our markets, we believe that we have enough capacity for the next 2 to 3 years. and we continuously monitor how the demand is evolving. And based on that, we see how we need to take into consideration possible next capacity increase. But I would say, for the next 2 to 3 years, the capacity that we have in hand is sufficient.

Regarding the question about social media, yes, we were -- I must say this, this is a sudden interest of -- that the post created was a big surprise to us. Reading the post, there was a lot of information that was in fact not totally new. The post created some combination and some yes, it was a compilation of a lot of news that were already on the market. And this has created maybe a significant interest in our company. We see that, to some extent, positively because we are working on the different areas that were covered by the post regarding photonic regarding wide-band gap also the interest for some of our high voltage and CMOS technologies that could also be relevant for a growing application like in data center. That's why here, we provided a bit of an update. And if you look at the PowerPoint material that is provided for this call, there is a bit of more data on this. the feedback that we provided to the post was from our side, no specific new news and I think it's up to the market to make their own conclusion on how to utilize a different information.

You could say that some of the information was maybe created some shortcut between certain actions that we are taking, for example, in the context of the photonics fab and certain large companies. but nothing was fundamentally wrong in the statement that we have posted altogether in one shot in one single post. So that's my comment about this specific post on -- does that answer your question?

He disconnected. But he can enter the queue again if he has some more questions.

So maybe we can try with Luke now.

Yes. The next one is coming from Luke Deserta. Yes. In the meantime, we have another question is coming from Trion Reid from Berenberg.

It's Trion here from Berenberg. I just had a couple of questions. The first was just on the data center opportunity that you highlighted. You talked about $300 million long-term opportunity. Obvious question would be, could you give us a bit more detail on the timing. Is there any particular area you said a few of any particular area which is going to be a bigger contributor and how much revenue do you have today? So what does that $300 million compared to today? And then the second question was just on the robotics opportunities that you mentioned are becoming a bit more real. Just be interesting to get a view on that? Is that essentially via Melexis or is it other customers? And if you could help us to quantify the opportunity that would be, that would be great.

Okay. So today, our estimate on the data center is that we have a relatively modest revenue, mostly coming from what we do silicon carbide also an area that we have in photonics and some business that we run on CMOS estimate is in the range of $20 million to $30 million. It's a relatively large range. But if you look at it compared to the global revenue of its fab, it stays relatively modest. The $300 million is based on the number of opportunities that we tag and that we flagged that are directly connected with data center evolution. How fast and how quick could we get to this $300 million. That's a good question, and I will not be able to answer today. It is significantly linked to also the success of some of the customers we are interacting with in this domain.

And also, we know that the data center deployment is particularly the new architecture of data center is driven by multiple dynamics that we do not control directly from an perspective. So that's the reason why we stay relatively prudent about providing more accurate view within our different business units and in terms of timing. But as soon as we get some substantial and significant news on this, we will definitely be able to provide you more insight on this. So that's the answer on the data center.

On robotics, it's not only Melexis. So we see multiple customers in automotive, but also in the industrial area reporting the interest in robotics for the device that we are producing. So we produce sensors, we produce a position sensor, we produce also our customer produce with our technologies, activators and so we have also customers on the industrial sector, we are also reporting a nice growth in robotics. And it's a global growth. So it's also happening in China, but it's also a global growth that we've seen in robotics. So it's difficult to size this at this point trend. But likewise, as soon as we get a bit of more visibility, we will share that with you. It's only -- we thought it was worth mentioning that we see some of these traditional automotive players. I think Melexis gave also some insight on what they see in robotics, but we see also the same feedback from industrial customers. Hope that answers your question.

[Operator Instructions] We have the next question coming from Arnaud.

Get on to the other question?

[Operator Instructions]

Do you hear me now?

Yes.

Okay. So I do have a question around the financing. I think the uptick in the sector is a bit later than anticipated regarding the announcement of the initial $1 billion investments and the prepayments of the long-term agreements are -- paid. And the first bank financing of $200 million is, I think, is due at the latest -- at the end of next year. What do you see as the best way to finance this gap? I think that will be there in the coming 1 to 2 years.

Okay. So I will take first long-term agreement repayment, so we are gradually repaying the long-term agreements, prepayments. We still have a portion which will be repaid this year and another 1 next year. So it's not all done. And then regarding your question on the first revolving credit facility of 2021 of EUR 200 million. Yes, it's correct. Well, actually, the maturity is scheduled for November this year, and we already asked for an extension of it by 1 year as per contract. We had that clause in the contract, which we activated and this will help us already to extend by 1 year.

Now you are also correct by saying that we -- only with the extension of that line, we want our financial indebtedness won't be sufficient. I mean, we won't have enough capacity on credit lines to -- for the future. But we already gained 1 year, and this gives us some room for maneuver to see what would be the best options to either go for a new revolving credit facility as of next year when this one will expire or another way. There are a few things that we are looking at -- well, some alternatives as well. And we, as every time, we'll take the best option from a financial point of view. So we are working on it. That's the conclusion.

Yes. There is appetite in the market from the banks to.

Yes. Yes. Even for the extension of this line, we had really a lot of appetite from the lenders. .

The next question is coming from Luke [indiscernible]

I see that you are raising the outlook modestly, let's say, from EUR 195 million and then from EUR 200 million to EUR 205 million. As you say, the order intake is now accelerating. Would you expect that for, let's say, the fourth quarter that we would and once again, would it be about the same level?

Yes. We don't provide guidance for the fourth quarter. But yes, as I mentioned earlier, we reached the bottom, we believe we reached the bottom in Q2. And the way we look at the business right now for the second half as we said already a quarter ago is that second half will be strong than first half. So anticipation is that there will be a prolongation of the recovery, how fast and how high we are still a bit cautious to see what our customers are planning to do. Talking to our main customer, they seem quite optimistic about Q2 -- Q3 and Q4, sorry. And this will be -- this will likely be reflected in our numbers as well. But so far, we don't -- we are not in a stage to provide guidance for Q4. Just a reminder or cycle time are quite short.

And we want to be cautious also, as I mentioned earlier, in the automotive industry when the cycle are long the supply chain is long, you can have really significant swings that can hit. So we want to make sure we get the full visibility on how the year will go before giving a guidance for Q4. But I'm quite optimistic about the fact that the recovery is not just a single shot for Q3.

And then I have a second question. Are you also in a position to be able to start increasing prices?

Yes, that's a good question. So I think right now, in general, we are following how the market evolves from a supply perspective. So we see that on the wafer supply, there might be some risk intention coming in our way. So we see that there is a high demand for SOI wafer. We see there is, in general, demand in a 200-millimeter wafer. So any price increase that comes there will be gated to our customers. At the same time, we want to make sure our customers are gaining market share and winning new markets. So it's always a trade-off that we need to make. And it's done on a case-by-case basis opportunity for opportunity. So at this point, we are not in a position to announce a global price increase. But clearly, new demand is coming in Norway, and that will place us in a different situation compared to what we have seen over the past 6 to 9 months.

There are no further questions at this time, so I hand the conference back to the speakers for any closing remarks.

Thank you, Gaia. Thank you, everyone, for participating to this call, and we remain available for any subsequent questions that you might have, and we need to give you an appointment for the next conference call, which is -- let me -- I'm sorry, I'm a bit -- with the notes here. So we want to give you an appointment for the next conference call, which is planned on the October 29, 2026 where we publish our Q3 2026 results. Thank you for your time. And with this, I wish you a great day and a great evening. Bye-bye.

Thank you. Goodbye. .

Thank you for joining today's call. You may now disconnect.

Ladies and gentlemen, welcome to the X-FAB First Quarter 2026 Results Conference Call. [Operator Instructions]

I will hand the conference over to Damien Macq, CEO. The floor is yours. Please go ahead.

Thank you, operator. Ladies and gentlemen, welcome to this quarterly conference call of X-FAB. Representing X-FAB in this call are Alba Morganti, CFO of the group; and myself.

In the first quarter of 2026, X-FAB generated revenues of $195.6 million, down 4% year-on-year and 12% quarter-on-quarter. Excluding revenue recognized over time, quarterly revenue was $205.8 million, above the guided range. Our core markets, automotive, industrial and medical represented 94% of the total revenue, reaching $192.9 million. Order intake amounted to $169.4 million, up 3% sequentially, while backlog stood at $308.4 million, reflecting the continued cautious ordering behavior in automotive rather than underlying end market demand.

Automotive revenue came in at $121.6 million, down 10% year-on-year. Inventory adjustment and limited short-term visibility persisted, but the structural trend towards electrification, customization and ADAS remains intact. Semiconductor content per vehicle continues to rise, driven by applications such as battery management, thermal regulation and advanced sensing.

During the quarter, we secured multiple sensor element design wins, particularly in Asia and China, including gas pressure and acceleration sensor as well as a new road noise cancellation application for electric vehicles.

Our industrial business delivered $52 million, up 32% year-on-year, supported by strong silicon carbide demand and growing momentum in data center power management.

Medical revenue reached $19.2 million, up 39% year-on-year with solid contribution from pacemaker, ultrasound, DNA sequencing and contactless sensing alongside new medical ultrasound opportunity from our 110-nanometer BCD on SOI technology.

From a technology perspective, CMOS revenue was $156.9 million, impacted by automotive softness. Nevertheless, we achieved design wins on our 180-nanometer and 110-nanometer BCD on SOI platform, spanning data center and medical ultrasound applications.

Our SPAD technology continued to gain traction across industrial, medical and mobile markets and interest in our next-generation 110-nanometer DCD platform for automotive sensors is confirmed from design wins by our top customer.

Our Microsystem and Photonics business delivered a record $33.7 million, up 42% year-on-year, driven by strong customer traction in diverse applications. For example, we successfully established aluminum scandium nitride-based piezo-MEMS production for inkjet printheads and saw growing customer engagement in MEMS resonator and ultrasonic transducers.

Finally, wide bandgap revenue reached $15.1 million, up 152% year-on-year. Silicon carbide wafer shipments increased to 14,300 units, nearly tripling year-on-year. This confirms the strong customer pull in this technology. In GaN, we delivered the first 1,200-volt prototype on a 200-millimeter substrate and initiated a vertical GaN customer project supporting increased prototyping activity in 2026. Overall, our technology pipeline, design win momentum and diversification progress clearly position X-FAB for growth as market conditions normalize.

And now I will pass the word to Alba for the financial section.

Good evening, ladies and gentlemen. And now let's walk through the financial section. In the first quarter, our EBITDA was $34.2 million with an EBITDA margin of 17.5%. Excluding the impact of revenue recognized over time, the EBITDA margin would have reached 18.4%, which was -- which would have been then within the guided range of 18% to 21%.

As you could see, we had a significant reduction of the net sales due to an impact coming from the IFRS 15 revenue recognition over time in the first quarter of minus $10.3 million. This is the natural consequence of the expiration of most of our long-term agreements with some of our customers at the end of 2025. So from 1st of January this year, we are now back to the usual business free of LTAs, with these customers. But the expiration of these contracts triggered according to the IFRS 15 rules, a significant reduction of the quarterly sales.

Our net debt went from $285 million as of the end of Q4 2025 to $291 million at the end of Q1 this year, so rather stable. Meanwhile, we reimbursed $26.4 million prepayment of the LTAs in the first quarter, and our total net debt decreased from $480.4 million as of the end of Q4 2025 to $436.1 million at the end of Q1 2026.

In terms of profitability, our first quarter was primarily affected by softness in the automotive end market, which impacted on the overall revenue. Given the substantial share of fixed costs, expenses could not reduce proportionately. The positive effect from the ongoing cost savings programs are not yet visible, but we anticipate them to become tangible by the end of 2026, including a headcount reduction in the high single-digit percentage range and a gradual decrease in operational expenditures. Additionally, increased research and development activities during the first quarter led to higher-than-usual R&D costs.

Cost structure is actively managed to protect profitability in the short term while maintaining the capabilities required to support future growth when volumes recover. As you know, our business is naturally hedged, which keeps our profitability unaffected by exchange rate fluctuations. At a constant U.S. dollar-euro exchange rate of $1.05 as experienced in the previous year's quarter, the EBITDA margin would have been 0.9 percentage points lower.

Our CapEx in Q1 were well in line with expectations and totalized $26 million. CapEx has significantly moderated as we move past the peak of our investment cycle, supporting cash generation and preserving financial flexibility.

Cash and balance sheet discipline remain top priorities. Cash and cash equivalents at the end of the first quarter totalized $144.7 million.

And to conclude this financial section, I would like to share our next quarter's guidance. As automotive [ visibility ] remains limited in the near term, we reinforce our prudent guidance approach. We expect our Q2 2026 revenue to come in within the range of $190 million to $200 million with an EBITDA margin in the range of 17% to 20%. We anticipate a steady recovery of our automotive business in the second half of 2026. This guidance is based on an average exchange rate of 1.15 USD/Euro does not take into account the impact of IFRS 15.

I would like to conclude the financial section by saying that we remain confident in the medium-term outlook. Our strong operating leverage, diversified technology portfolio and long-term customer relationship position us well to benefit from a recovery in demand.

And now I would like to give the word back to Damien.

Thank you, Alba. Let me start by reaffirming the 3 pillars that continue to guide X-FAB's strategy and execution. Specialization, diversification and disciplined execution. These are not new priorities for us. But over the past quarter, we have taken several decisive steps to reinforce them structurally, operationally and commercially. On specialization and diversification, these are the heart of the transformation we have just completed.

During the past quarter, we finalized a global reorganization of X-FAB around 3 dedicated business units, each focused on one of our core technology families. We have a Smart CMOS and SOI BU, a microsystem and [ photonics ] and a wide bandgap technology, including silicon carbide and gallium nitride.

These moves reflect a clear strategic intent. First, it allows us to accelerate the specialization of our technology portfolio by bringing development engineer, technology road maps and business development much closer to our customers. This proximity is critical to enable true co-creation, developing differentiated technologies together with customers and accelerating product ramp-up from R&D to volume manufacturing. In parallel, it strengthens accountability and focus within each technology family, ensuring that innovation is both market-driven and execution-oriented.

Second, this structure supports our ambition to further diversify our markets and customer portfolio. By aligning our organization around fast-growing end markets such as electrification, digital health care, industrial automation and AI-enabled infrastructure, we are better positioned to capture growth beyond traditional cycles and to serve customers with very different applications, needs and life cycles. This diversification is already visible in an important customer mix evolution.

During the past quarter, our top 20 customers ranking, representing around 75% of our revenues changed meaningfully. The share of our largest customer to our total revenue declined from 41% to 35% due to inventory corrections, but also due to the significant increase of sales generated by customers ranked from #2 through #20. Altogether, they increased their revenue by 21% from Q4 '25 to Q1 '26. This evolution is strategically important for X-FAB. It reflects the diversification of our business and the larger growth of our silicon carbide and microsystem businesses as well as the continued success of highly differentiated X-FAB technologies in automotive applications such as lighting and sensing.

We are also seeing continued momentum from our Chinese customers, particularly in the automotive segment, MEMS, battery monitoring systems, silicon carbide and also for new applications like technologies for noise cancellation within the cabin.

Turning to disciplined execution. Operational excellence remains a top priority. We continue to improve cycle times, yield and quality, ensuring that our fabs remain competitive, resilient and reliable partner for our customers, particularly in automotive, medical and industrial applications where quality and supply security are paramount.

Finally, let me highlight the ongoing transformation of our Erfurt site, where we are accelerating the deployment of microsystems technologies. This site plays a central role in our microsystem and photonics road map and is critical to supporting growth in medical, sensing and next-generation industrial applications.

To conclude, our commitment to innovation, strategic business development and disciplined execution remain unwavering as we continue to build the sustainable long-term success of X-FAB for all stakeholders.

This concludes our Q1 results presentation. Operator, please open the lines for questions.

The first question is coming from Aleksander Peterc from Bernstein.

I just have a few. The first one is on automotive where you seem to be in a double-dip configuration now with the second quarter of roughly minus 10% declines in your revenue year-on-year. And this seems to track quite closely your #1 client, Melexis. So I'd just like to know if you see a broad-based weakness in automotive? Or is it primarily your #1 client there?

And then the second question is, do you see any signs in your order intake patterns that would highlight any potential inflection in the coming quarters? We have book-to-bill that remains quite low and your backlog keeps reducing. It's now 40% below peak. I'm just wondering if you can give us any hints on how this would bottom out? And I have a very quick follow-up, housekeeping one afterwards.

All right. Thank you for the question. On the first question, yes, as I mentioned in the presentation, it is correct to say that the #1 automotive customer reduced their demand quite a bit in Q1, and this is linked to somehow inventory management on their side. At the same time, we saw for the other customers, stabilization of -- in general in automotive.

If you look -- if you zoom in on certain applications, we see some good strength, as I mentioned in, for example, battery monitoring systems. And we see also some good strength in front lighting applications. So I would say there is a kind of quarter correction from our #1 customer managing their inventory, and there is some good stabilization and slight increase in certain applications for automotive.

Regarding the forecast from discussions with our customers, we see a recovery expected in the second half of the year. So it's not going to be visible in Q2. That's why we remain cautious in Q2, but we see some sign of recovery in the second half of the year.

Regarding the backlog, you have to remember that our backlog is also impacted by the End of Life that we made on our site in Erfurt. So this End of Life continues to trigger orders on our side. But yes, the visibility that we get from this End of Life is directly there, but it is not generating some good evolution of the booking overall because all this business was already booked in the past. Does that answer the question?

Yes, it does, yes. That's great. Just a quick housekeeping one for Alba. Can you help us understand where your G&A is going to track for the remainder of the year and into next year, given your CapEx plans as they are presently, where do you see G&A?

G&A, you said? Sorry?

Yes.

Well, G&A is a bit of a tricky one, especially this year because we are implementing a new ERP. So yes, these years G&A is also impacted by the implementation of the new ERP SAP in our system. So of course, the G&A costs are going up. So we are facing some years of increased G&A. If you looked at the evolution, you might have seen that already. And this should go down then after the implementation of SAP. We should be running in 2028 -- early '28, we should start with our new ERP. And then, of course, the cost will significantly go down. This is...

Towards '28 and then go down onwards.

Yes, yes, of course, the cost will fluctuate a bit over the years. Now we have quite a lot of external costs that are normal in this space.

The next question is coming from Ruben Devos from Kepler Cheuvreux.

I just had one on utilization across your footprint. Curious whether you could give us a bit more sense around what that might be, the loading, either by site or by technology? Yes, I think that's the first question will be very helpful already.

Okay. Utilization globally is in the low 60%. We don't provide utilization per site. So -- but yes, we have -- I think that was also linked to some messages that we provided in the past. We have established additional capacity for the upcoming future. So we are ready for the growth of our business. But the low 60% is also an explanation on some levels that you see today in terms of cautiousness on profitability.

What I want to highlight is also that linked to this reduced loading that we see today in our factories. We are taking these actions to continuously monitor the necessary capacity, and we are [ idling ] some tools. We are making sure that the cost control in the factory is aligned with this limited utilization at this moment. So low 60% is the number.

Okay. All right. And I believe you talked about the utilization for silicon carbide back in the Capital Markets Day in September. That was then at a rather lower level. But of course, you've seen quite -- you've reported strong growth now in Q1. Would you mind disclosing what -- how that now is?

Yes. So we have built quite some capacity in silicon carbide. So we can run up to 10,000 wafer a month. And at this moment, we are right at the 60%. Also in the low 60%, we load around 6,000 wafers of silicon carbide in this available corridor. And there, we want to make sure we are ready for the growth that we see quarter-over-quarter. So it's a bit...

And on, let's say, AI-related data center power management. I was wondering what is really pulling you into that opportunity? Is that like sort of the higher voltage and power density angle or is it more of a push for localized supply? Is it simply a shortage of specialty capacity out there? Or yes, a bit more color here would be great.

Yes. I think it's a lot linked to the silicon carbide technology. So we have engaged very, very early with different companies on silicon carbide. I think Navitas is a publicly known customer of X-FAB. And we see really a renewal of the interest of silicon carbide going from applications like power switch for solid-state breakers, but also utilizing energy conversions from 800 volt to system voltage down to 48, 12 or 6-volt. So it's silicon carbide drives a lot of this in terms of production.

Also the interest on GaN is for a portion linked to data center. I want to highlight also that the data center on our side is also -- the growth on data center is driving also some additional growth for other technologies like, for example, our technology is used for timing clock to synchronize GPU in racks. So the idea is to reach really atomic precision with the technology, and this is driven from a chip that was built out of [ X-FAB ] technology.

And moving forward, we also some design wins regarding high-voltage driver ICs to drive the power transistors. So there are multiple applications that are driving some growth on the data center, mostly for energy management, thermal management, but also timing management. The reason why we talk about AI is also and data center is also that we have also an activity on photonics. And in the photonics space, we see also some interest for application of our technologies. But there, we are more in the development phase than in the revenue phase.

All right. And then just a final question. I think we've been reading about some foundries pushing further into power and specialty, right, and not only in China, but also outside China. Also have been reading about some IDMs adding some analog capacity. Is that showing to some degree in your pricing or actually on the contrary, I mean, we've also been reading about pricing actually being increased, right, in certain mature node areas. So yes, how do you think about that?

Yes. We read the same press. We have seen also these articles about price increases in this area. I think there is still a strong push on automation and silicon carbide for automation where probably the price pressure is extreme. We try to stay on applications where the high voltage and the customization that our customers are able to execute on our technology give us a certain advantage.

So there is price pressure always everywhere, but I would say we try to get -- to stay in applications where this pressure is sustainable. So I would not see any specific abnormal push towards price down or price up. We see that more from an end application perspective, probably more pressure on automotive and a bit less pressure in industrial, data center, renewable where we like to play.

The next question is coming from Emmanuel Matot from ODDO BHF.

Some of my questions have already been answered, but I have 3 questions to ask you. First, should we expect a reversal in demand for the industrial business as the positive effects of introduction of certain legacy technologies come to an end?

Second, can you confirm savings of $6 million by the fourth quarter [indiscernible] from the Q4 '25 earnings release? And could you repeat what happens to R&D in Q1 [ it was up ] significantly and I've been disconnected at the time for the call. Is that structural the new level of R&D spending or just something specific to Q1?

Thank you for the question. So first question, reversal in demand in industrial. No, we do -- in fact, we see significant increase in industrial for certain segments. So data center is the biggest one. But I think across the industrial business in general, we see some growth that we have not seen in the past. So basically, the industrial business, if you look at it from a global perspective, is nicely recovery across the board, but with significant growth on data center.

I will answer the R&D question and let Alba answer the other question on the $6 million. On R&D, so what we have done in Q1 is a specific effort to improve the quality and the robustness of certain technologies. So that has additional cost of mask and wafers. We are also in the process of creating a second phase of 110-nanometer technology, BCD-on-SOI in France and BCD in Malaysia. And this development -- this joint development that we have in France and Malaysia are creating a phase of additional wafer requirements to qualify these phases. So this is not something that should prolong. It's more linked to some specific effort that we made regarding the improvement of the robustness of our technologies. Do you want to take that?

Yes. And yes, Emmanuel, indeed, we had a softening of the cost reduction effect or results visible. That's clear because, we are just at the end of Q1, and we put some measures that will take a bit longer to be visible, but we still -- we remain confident that our cost savings plan will materialize as expected by the end of the year.

The next question is coming from [indiscernible].

First of all, I see that, of course, in different parts of the business [indiscernible] good to see. However, if we look at the prototyping in automotive, it stays always at the same level or it even weakens. Are we having some design missing? Or are we not able to get to the customers to business there because we still have to take into account after the big investments, we normally should be able to double the volume and double the sales figure. But I see that prototyping is lagging behind a one-to-one relationship between prototyping and the sales of automotive. But if it is in the long run low, then I'm very concerned.

Yes, that's a good comment, and thank you for the question. So if you look at our historical prototyping, there was a period where the levels were extremely high. The years were the year '22, '23 -- yes, '23 let's say. And that is linked to different effects. [Technical Difficulty] So there was this effect of allocation, but also that during this phase, some of the preproduction that we had in microsystem and also in silicon carbide were assigned as prototype. So we were doing preproduction. In reality, it was production of conditionally qualified material and this was [indiscernible] prototype.

So if you look at the long-term curve, there is an anomaly. I think that was already flagged in earlier calls. If you look at more the recent time, the last 6 quarters or 5 quarters, there has been some steady growth after a low level in Q1 2025. But since then, the prototyping, the NREs of our CMOS kept increasing quarter-over-quarter. And Q1 this time is a bit lower versus Q4 last year.

But if you compare quarter-over-quarter, the growth is there. So this is definitely a KPI that we continue to monitor. The fact that customers are coming back to us because they see that we have capacity, they see also operational performance improvement is a good sign. What I would like -- I would love to see that this translates into a true quarter-over-quarter growth.

The point is if you compare Q1 '26 with Q1 '25 from the NRE on the CMOS side, there is a significant growth. So we will continue to monitor these things. I think overall, the effort that we are doing in sales and the effort that we have done by the creation of the business unit will pay off. And my anticipation is that the level of prototyping will continue to increase. And this is linked also to the interest that we see in our technology.

So the BCD over SOI that we have in 180 and 110-nanometer is really gaining traction. The new technology that we are developing a BCD standard BCD on 110-nanometer is also getting some interest from our major customer. So I would say it's a very good point. It's a very good KPI to monitor. But so far, if I look at the plan that we have for 2026, the Q1 results that we have on our prototype are according to our plan. So it's an important KPI. Thank you for asking the question. I hope it answers your question, by the way.

What is also a little bit of concern, of course, is the gross margin is also going down or at least that's the impression that I have. Is there anything happening on that? Or is that just due to the fact of the elimination of the [ $10 million ]?

No, it's mostly driven by capacity and utilization, right? So right now, as I mentioned earlier, our utilization is the low 60%. We would be better to have a higher level of utilization. So we are taking the actions to look at the loading of our factories and that will naturally increase the gross margin of our products.

If I may complement, our cost base is rather a very high fixed cost base. And for the moment, we are rather 70% fixed cost versus 30% only variable cost. Therefore, yes, our leverage is really with -- as Damien mentioned, is really with a higher utilization rate.

It can only be solved if the top line is improving it.

Yes...

That's what we need to get...

It can be mitigated. Alba mentioned the cost optimization effort that we are doing. So on all the sites. We are taking actions to adjust the different breakeven points and the profitability at lower utilization. But as Alba mentioned, these actions, they take time, right? So we should see some positive results in the second half of this year from this cost optimization, but the major effort is on loading.

Okay. So of course, first on getting all the new equipment in and getting all those things done. Now the focus really has to be on getting volume at the right margin and so on so that we can benefit from the big investments.

Absolutely.

Next question is coming from Trion Reid from Berenberg.

It's Trion from Berenberg. yes, I just have 2 last questions to ask. The first is just on the microsystems and photonics division, which you renamed from the previous MEMS. Obviously, that's growing strongly. I just wondered if you could give us an idea of how much of that division is photonics rather than microsystems and how much that might be contributing to growth?

And then the second question was just on the working capital. It was a small outflow despite the fact you mentioned, I think, about $26 million of repayments to Melexis. So that sort of seems to suggest an underlying working capital improvement. I wonder if you could comment on that and whether that's sustainable going forward.

Thank you,. I will take the first question. So today, photonics is mostly development. So on the revenue, it's all about NREs. We mentioned at the end of last year, the importance of NREs that we got for photonics. So microsystem without photonics broke end of last year, the ceiling of $100 million. And yes, with photonics, I think we were in the range of in the range of $5 million to $6 million NRE end of last year. Sorry, photonics...

Photonics was $7 million.

$7 million...

Sorry, 7% of the total NRE of the group.

Okay. That's -- yes, that's 5.5% to 6%. So that was -- so photonics is mostly and we anticipate product revenues from photonics to hit more around end of '27, '28. So not yet visible outside of NRE at this moment. And the second question, I will let Alba answer on working capital.

So first of all, Trion, good to hear you. For the repayment of the LTA repayment, it was not only Melexis. Of course, largely, they brought a large portion, but it was not only Melexis. Anyway, yes, this is indeed the repayment we made in Q1. And if you look at working capital, yes, it improves indeed. The fact that we reduced the CapEx significantly because we ended up with [ $0.25 million ], so to say, only compared to the high numbers we had in the last 3 years, of course, helps a lot the working capital needs to get back to a normalized level.

There are no more questions at this time. So I hand the conference back to the speakers for any closing remarks.

All right. So I wanted to thank the audience. I want to thank you for participating in this press conference to this Q1 results. And I wanted to give you also an appointment for our second quarter results that are planned for July 30, 2026. So thank you, everyone. Have a good day, a good evening, and speak to you soon.

Thank you. Goodbye.

Bye-bye.

Thanks for participating in today's call. You may now disconnect.

Welcome to the X-FAB Full Year and Fourth Quarter 2025 Results Conference Call. [Operator Instructions]

I will hand the conference over to Rudi De Winter, CEO. The floor is yours. Please go ahead.

Thank you, and welcome, everyone. We have today in the conference call with me Alba Morganti, CFO; and Damien Macq, my successor and CEO.

In the fourth quarter in 2025, we recorded revenues of $222 million, up 18% year-on-year and down 3% quarter-on-quarter. The fourth quarter revenue in our core markets was $204 million, up 13% year-on-year and down 5% quarter-on-quarter. The core business represented 94% of our total revenue. Full year revenue amounted to $870 million, up 7% year-on-year. and within the guided range.

In the full year of 2025, our core business was $814 million which is 7% growth compared to the previous year, and our core business represented also 94% of the total revenue. Our order intake in the fourth quarter came in at $164 million, and it understates actual underlying demand by approximately $30 million to $40 million as first of all, we have shorter cycle times resulting in customers ordering later. Secondly, we have higher wafer yields resulting in order -- reduced order quantities. And third, because of the absence of new bookings in the 0.6um CMOS technologies that we are terminating in early 2027. Customers already placed these orders following the last time by announcement more than a year ago. The backlog at the end of the quarter was $318 million, down to -- from $347 million and at the end of the quarter. The backlog relative to our revenue is still high compared to history and pre-COVID situation.

In the fourth quarter, automotive revenues came in at $133 million, up 3% year-on-year and down 10% quarter-on-quarter. The sequential decline was mainly due to inventory corrections in various channels of the supply chain and also influenced by the end of major LTAs and the underlying end market. This trend is also evident in the full year automotive revenue, which recorded only a slight increase of 1%. Although the shift to full electric mobility in 2025 advanced at a slower pace, it remains an important megatrend, underpinning our automotive business. Key automotive applications driving the growth in 2025 included battery and thermal management systems as well as on-board chargers for electric vehicles.

In the industrial end market, we recorded a quarterly revenue of $50 million. This was up 40% year-on-year and 6% quarter-on-quarter. The strong growth was driven by the recovery of the SiC business, the recovery of the fragmented industrial market and the prototyping revenue for Photonics, while an elevated level of production in last-time-buy technologies, also contributed. The fourth quarter medical revenue amounted to $21 million up 28% year-on-year and flat sequentially. The demand for DNA sequencing, ultrasound applications as well as contactless temperature sensors was strong in the past quarter.

In the total of 2025, our medical business achieved a record revenue of $71 million marking 26% increase over previous year.

For a further update, I would like to pass the word now to Damien.

Thank you, Rudi. Good morning, good afternoon, everyone. Let's start with CMOS and SOI revenue. In the fourth quarter, this revenue was up 7% year-on-year and down 5% quarter-on-quarter. For the full year, revenue grew 8% compared to the previous year. Quarterly microsystem revenue was up 24% year-on-year and down 9% sequentially. In 2025, X-FAB microsystem business achieved revenue exceeding USD 100 million for the first time, representing an 11% increase compared to 2024. X-FAB silicon carbide business demonstrated a remarkable recovery, achieving robust growth in the fourth quarter, driven by solid demand for data centers, electric vehicles and renewable energy applications. Revenue increased by 77% year-on-year and 6% quarter-on-quarter. Silicon carbide wafer starts raised 60% sequentially. This was the largest ever silicon carbide wafer start in the quarter. The weaker quarter-over-quarter revenue growth reflects the much higher share of customer-supplied silicon carbide wafers, which carried lower billings due to the less pass-through of substrate costs. For the full year, silicon carbide revenue reached USD 33.8 million, which constitute a 34% decrease against 2024 when the first quarter was still exceptionally strong.

Quarterly prototyping revenue was USD 20.3 million, down 14% year-on-year and up 3% quarter-on-quarter. Over the past 3 quarters, its fab recorded a notable increase in CMOS and SOI prototyping revenue reflecting renewed customer confidence after capacity constraints were resolved with last year completion of it's fab capacity expansion program, and from significant operational improvement in terms of product yield and cycle times.

Let's now zoom in specific products and development highlights achieved in 2025. In Q4, we secured a major design win for 110-nanometer CMOS technology, particularly in sensing application, which should contribute to revenue growth from 2028 onward. For our 110-nanometer SOI platform, we see confirmation of the strong ramp-up for motor control and automotive LED drivers, and we booked an important design win for new ultrasonic applications.

In microsystem, interest in our through-silicon via technology continues to grow, especially for photon-counting CT scanners. On top of the technology capability of its fab, another key factor in this engagement is our customers' ability to establish a fully localized supply chain for this next-generation scanner platform. In MEMS, we achieved the first design win for our next-generation inertial sensors. For power application, we launched as well an innovative snubber technology, this device is integrated inside silicon carbide inverters to review the switching losses by up to 70% and therefore, improving the electrical vehicle range. This solution has already been adopted by 1 OEM and is under evaluation by several orders.

In the photonic space, we teamed up with LIGENTEC on the lowest lost silicon nitride platform for various applications, main one being quantum computing. This already contributed to 7% of our total NRE in 2025. In GaN, we secured several major NRE. A first one for the development of industrial protection devices. 2 additional protraction inverters for small and medium EVs and a fourth one for 800-volt data center applications.

In silicon carbide, we continue to make excellent progress with a major Asian Tier 1 in EV traction inverter, achieving record-setting electrical performance. Our latest XSICM03 platform has also enabled multiple customers to reach leading-edge electrical performances, driving further design wins across renewable energy, automotive, data center and circuit breaker applications. Its fab technology portfolio with the emphasis on power sensing and microsystem technologies is strategically aligned with global mega trends, including the electrification of everything with worldwide decarbonization initiative and advancement in health care for aging populations. This alignment creates substantial opportunities within X-FAB key end markets, automotive, industrial and medical, driving sustainable growth in the long term. The short-term visibility remains limited, primarily due to continued inventory adjustment by automotive customers and persistent geopolitical uncertainties.

Let's now go through the short operation update. By mid-2025, X-FAB concluded its major 3 years $1 billion capacity expansion program. In September, we celebrated the grand opening of our new facilities in Malaysia, following the launch of our production in this new cleanroom. All equipment related to this expansion have been installed and qualified. Production in X-FAB popular 180-nanometer technology is being ramped up gradually there. The site's target capacity of 40,000 wafer start per month will be fully operational by the end of 2026. The increased capacity and reduced cycle time enable X-FAB to become much more attractive for new business opportunities and to respond more quickly to market opportunity when they arise.

In the fourth quarter, significant progress was made in securing financial support under the EU Chip Act for the growth of X-FAB's microsystem business. The funding will be used to further advance the MEMS and microsystem offering and more specifically to support the ongoing transition of the site in Erfurt Germany, to the microsystem hub of X-FAB Group.

Capital expenditure in the fourth quarter reached USD 25.2 million, bringing the total CapEx for the year to USD 204.1 million. This is lower than the initially projected USD 250 million as some expenditures were postponed to the current year. New capital expenditure in 2026 are projected to come in at around USD 100 million. This CapEx will be allocated to enhance our process capabilities, to facilitate the transition of the Erfurt Germany and Lubock Texas sites to microsystem and silicon carbide, respectively, and to support necessary maintenance and further autoimmune activities across all sites.

In response to the short-term challenge and limited visibility, we are also introducing further cost efficiency measures. This initiative includes a planned headcount return in the high single-digit percentage range for 2026, as well as a gradual reduction of operational costs. By the fourth quarter of 2026, cost savings are estimated to reach USD 6 million per quarter. Concurrently, we remain well positioned to respond swiftly to increasing customers' requirements and growing demand.

I will now pass it over to Alba Morganti, CFO of X-FAB for the financial update.

Thank you, Damien. Good evening, ladies and gentlemen. We will now go to the financial update. I would like to start this financial section by highlighting that in 2025, we totalized $870.3 million sales, meeting the yearly guidance of $840 million to $870 million. This represented an increase of 7% compared to 2024. The sales in the fourth quarter totalized $222.3 million, which is an increase of 18% year-on-year, also in the guided range of $215 million to $225 million.

In the fourth quarter, our EBITDA was -- of $42.3 million with an EBITDA margin of 19%. If we exclude the impact from revenues recognized over time, our EBITDA margin would have been of 19.2%, which is still below the guided range of 22.5% to 25.5%. The main reason for that is the miss -- that we missed, our guidance is that the fourth quarter profitability was impacted by a one-off item totalizing $9.3 million, out of which $6 million resulted from the renegotiation of a long-term agreement for the SiC raw wafers while $3 million were due to the reevaluation of our silicon carbide substrates inventory in Lubock and we had renegotiated lower prices. If we exclude these exceptional items, the EBITDA margin would have been of 23.6%. I would like to add that with the productivity improvement plan Damien was referring to, we are in a trajectory to achieve a 30% of EBITDA margin with a quarterly revenue level of $240 million. Since a few years now, our business is naturally hedged and our profitability remains unaffected by exchange rate fluctuations. At a constant USD/Euro exchange rate of USD/EURO 1.07 as experienced in the previous year's quarter the EBITDA margin would have been up 19.3%.

Cash and cash equivalents at the end of the fourth quarter amounted to $194.3 million, up $20.1 million compared to the previous quarter, while our net debt decreased by $4.5 million quarter-on-quarter. Our cash position improved despite the fact that in 2025, we repaid several bank loans for more than $75 million as well as leasings for more than $24 million and we also repaid some prepayment that we received from customers, including LTA contracts for about $43 million.

And to conclude the financial section, I would like to share our next year's guidance -- next quarter, sorry. Our Q1 '26 revenue is expected to come in within a range of $190 million to $200 million with an EBITDA margin in the range of 18% to 21%. This guidance is based on an average exchange rate of 1.17 USD/Euro and does not take into account the impact of IFRS 15. For the time being, we are not providing a full year 2026 guidance due to the limited visibility and the current macroeconomic environment.

And now I would like to give the word back to Rudi.

Thank you, Alba. While we remain cautious about the near term, we are observing encouraging developments across our business. Momentum in our CMOS and SOI prototyping revenues is building as our operational improvements make us again more attractive. We see a high level of interest in our microsystem capabilities that is opening substantial new opportunities. Our silicon carbide business is on track for recovery. And we see strong traction for our photonics on the one hand and the gallium nitride on the other hand, demonstrated by several new contracts. I firmly believe X-FAB is excellently positioned for robust growth. The fundamentals of our business remain strong, and I'm confident of X-FAB's long-term sustainable growth. X-FAB is ready for what's next.

With that, I'm very pleased to hand over the leadership of X-FAB to Damien Macq, who will succeed me as CEO of the group. And with that, we close the opening, and we are opening for questions.

[Operator Instructions] The first question is coming from Robert Sanders from Deutsche Bank.

Can I just get a bit of better understanding of the trends by end market? It looks like industrial, medical is performing well, auto clearly soft. Is that what you expect to be the story of the year as it were, even if you can't give specific guidance. Is there any reason to think that that's not the story of the year.

And then the second question would just be around your lack of 300-millimeter footprint and your largest customer's business in China, they seem to be under big pressure to use local foundries. What can you do to mitigate the risk that they have to source locally as opposed from you?

Yes. I think your understanding of the end market -- this is Damien speaking -- is aligned with what we see right now. We see uncertainty on the automotive. So -- but strong industrial. So I think the industrial will stay strong, medical -- was a good year in Medical and likely this will be prolonged as well with important design wins as well in the pipe.

Regarding the 300-millimeter footprint. At this point in time, we still see customers coming from China to look for our product. So basically, we being a high voltage, being a automotive qualification. And right now, we do not see the absence of 300-millimeter in our site as a handicap. So we keep seeing strong business coming out of companies that are headquartered in China, and we will need to monitor the evolution over the coming quarters to see where other situation are evolving there. Does that answer your question?

Operator are we still...

Yes. I think 0- Robert Sanders, I was looking for a feedback, but I guess, this replies to his question. So we have the next question coming now from [ Luke DeSoto ].

First of all, I would like to thank Rudi De Winter for the huge investments which you have done with X-FAB, and this without any dilution for the shareholders. I think this is an incredible good job and how that has been managed financially. I see that there are now a lot of platforms and technologies available at X-FAB. What are the actions? That would be my first question. What are the actions to increase actually the sales because that is still a little bit lagging behind?

Yes. So that's a very good point. Maybe this was not disclosed in this presentation, but we are reorganizing ourselves also across 3 business units. So we have moved some people closer to end customers, and we want to reinforce our co-design capabilities with customers. So business development is and was along 2025, a critical topic that is going to be prolonged in 2026. And if you see some of the results for 2025, so we observed quarter-over-quarter an improvement of our design win in the CMOS technology. So this is something that we need to [ prolonged ]. You have also to realize that we are coming out of a period of allocation with a lot of stress and a lot of stretch with our customers. This period is now over, and it's really time to go hand and [ try ] to collect additional business. And I think the items that were described also in this call, also in regard with microsystem are a good demonstration of the result that this can provide. So it's a very good point, deploying now the technologies and all the good deals that we have in hands to our customer is essential.

Also I want to repeat what I said earlier regarding photonics. So photonics is also a technology where we see a lot of traction from our customers. I want to repeat that 7% of the total NRE of the company was realized on photonic -- with photonic technologies.

Okay. My second question is more towards finance. I see that the current liabilities are very high, $700 million, $702 million, while the current assets are at $648 million. Are we going to have some stress on a financial basis?

Look, no, we have credit lines available to support the working capital needs. We have our first revolving credit facility of EUR 200 million, which will expire end of this year, but we are in renegotiation with the banks of the syndication to use the clause of -- which is included in the contract to extend it by 1 year, and we will still have the other one running until -- well, it's of -- we have another 3 years for the other one. And so for the time being, we see rather a decrease of the net debt, thanks to the fact that our major CapEx expansion plan is now over. Therefore, we see a relaxation actually is the other way around of our indebtedness. We have been able to repay several debts which were due, so absolutely in line with expectations in 2025, and we will continue to do so in the future.

Okay. So overall, actually, we just need to focus now on getting more sales and then everything will leverage out, and we will get the benefit of the investments.

That's clear...

Yes. It will be marvelous company.

Thank you for this lovely compliment. But yes, you're absolutely right. Thank you.

[Operator Instructions] There are no further questions at this time. So I hand the conference back to the speakers for any closing remarks.

Thank you. Damien speaking. Before closing the call, I wanted to take a moment to acknowledge the leadership and instrumental contribution of Rudi, as it was already mentioned in this call. I'm grateful for the strong foundation that we have built under his tenure over the past 15 years and for the support from the Board and from our global team as I step into the CEO role. My focus will be on further specialization through continuous innovation, offering unique capabilities on market and customer diversification and on disciplined execution in our operations to serve our customers with the level of performance and quality required to make them successful. For the past 3 years spent within X-FAB as COO, I had the opportunity to interact with our global teams. I trust we are already and committed to building and delivering our robust growth on the momentum already in place. Uta, Alba and myself remain available for any follow-up, and we look forward to speaking with you for our next quarterly conference call on the results of the first quarter 2026. This call is scheduled on the 30th of April. Thank you very much to everyone. Bye-bye.

Thank you.

Thank you.

Thanks for participating to today's call. You may now disconnect.

Hello, and welcome to the X-FAB Third Quarter 2025 Results Conference Call. My name is George. I'll be your coordinator for today's event. Please note, this conference is being recorded. [Operator Instructions]. I'd like to hand the call over to your host, Mr. Rudi De Winter, CEO, to begin this conference. Please go ahead, sir.

Thank you. Welcome, everyone. In the conference call today, we also have Alba Morganti, CFO.

In the third quarter of 2025, we recorded revenues of $229 million, up 11% year-on-year and 6% quarter-on-quarter, which is well above the guidance of $215 million to $225 million. We also progressed well in our core markets: automotive, industrial, medical with a revenue of $216 million, up 14% year-on-year and 5% quarter-on-quarter. Our core business now represents a share of 94% of the total revenue.

Now breaking it down by end markets. In the third quarter, the automotive revenue was $147 million, up 1% year-on-year and 2% quarter-on-quarter. The third quarter industrial revenue was $48 million, up 51% year-on-year and 1% sequentially, reflecting the overall recovery of our industrial end markets. The gradual recovery of the silicon carbide business contributed to this positive evolution.

Now for the medical business, the revenue in the third quarter hit a record high of $21 million, up 74% year-on-year and 40% quarter-on-quarter. The growth was mainly driven by contactless temperature sensor, DNA sequencing and echography applications that altogether did very well in the past quarter.

Now looking at it by technology. In the third quarter, the CMOS revenue recorded a growth of 10% year-on-year and 4% quarter-on-quarter, mainly due to the extra capacity that came online for the 180-nanometer BCD-on-SOI and 35-nanometer CMOS node demand was weaker.

Microsystems revenue was up 27% year-on-year and 9% sequentially. This is based on a broad set of customer-specific microsystem technologies that we co-created with our customers. Also, demand for new developments in the micro systems remain strong, and this is an area where we will continue to see above-average growth.

Our silicon carbide business continued to recover and revenue grew strongly by 30% year-on-year and 20% -- 21% quarter-on-quarter. The number of wafers produced in the third quarter more than doubled compared to a year ago. The revenue did not follow the same way due to the product mix and also the ratio of consigned substrates that was much, much higher last quarter than a year ago.

The positive trend in the evolution of our silicon carbide business is underpinned by increasing bookings attributed to sustained demand from data center, electric vehicles and renewable energy applications. We also see good traction on our new technology platforms that we released at the end of 2024. Many customers are developing their new generation products based on this platform that will give improved performance and lower system cost.

The fact that we offer full supply chain for silicon carbide in the U.S. is well perceived by our U.S. customers that are designing in products for high value-added assets such as data centers, industrial equipment and electric energy systems. Quarterly prototyping for the past quarter was $20 million, down 16% year-on-year and 6% down quarter-on-quarter.

The order intake for the third quarter amounted to $163 million, down 25% year-on-year and down 21% compared to the previous quarter. The booking in the industrial segment was good. The weakness is primarily due to inventory corrections by automotive customers as well as broader macroeconomic uncertainties resulting from geopolitical tensions and trade disputes.

These factors have led to a more cautious ordering patterns while customers also take advantage of shorter cycle times, placing orders later than usual and with reduced lead time. As a result, feasibility is still restricted. The backlog for the third quarter came in at $347 million compared to $413 million at the end of the previous quarter.

Let's now move to the operations update. In September, we had the inauguration of the new cleanroom in Malaysia, which will increase the site's manufacturing capacity from 30,000 to 40,000 wafer starts per month. Production at the new facility is being scaled up progressively with the full increase in capacity anticipated by the fourth quarter of 2026.

The expansion will effectively double our capacity for the popular 180-nanometer BCD-on-SOI technology, which is particularly suited for applications such as smart motor drivers various drivers such as piezo actuators, LED drivers and battery management systems. The recovery of the silicon carbide business is supported by the existing capacity at our Texas facility.

The current installed capacity will enable us to do more than double the wafer starts. The capital expenditure for the third quarter was $23 million, bringing total year-to-date CapEx to $179 million, and the full year capital expenditure is projected to be less than $250 million. Let me now pass the word to Alba for the financials.

Thank you, Rudi. Good evening, ladies and gentlemen. We will now go through the financial update. I would like to start this section by highlighting that the third quarter, we succeeded in increasing our sales by 6% quarter-on-quarter, which were the highest since almost 2 years, totalizing USD 228.6 million and which is well above the guided $215 million to $225 million.

Our EBITDA grew by 4% quarter-on-quarter and 7% year-on-year, being the highest this year. Our EBIT was almost $24 million, down 5% year-on-year, but increasing by 10% if we compare it to Q3 last year.

Third quarter EBITDA was almost $354 million with an EBITDA margin of 23.6%. If we exclude the impact from revenues recognized over time, the EBITDA margin would have been 24.2%, within the guided range of 22.5% to 25.5%. Our profitability remains unaffected by exchange rate fluctuations as we continue to be naturally hedged.

At a constant U.S. dollar euro exchange rate of 1.10 as experienced in the previous year's quarter, the EBITDA margin would have been unchanged at 23.6%. In the third quarter, we reported a financial result of minus $5.6 million, mainly due to interest result of $4.3 million and realized foreign exchange losses arising from the reevaluation of the euro-denominated debt amounted to $800,000, but of course, it's a noncash item.

Cash and cash equivalents at the end of the third quarter amounted to $174.2 million, which means an increase of $16.5 million compared to the previous quarter while net debt decreased by $21.1 million quarter-on-quarter. Despite the peak of CapEx expenditures payments in the first half of '25, our financial situations remain solid.

As anticipated and now visible, our CapEx are now significantly decreasing which translates into an improvement of our net debt position, which trend is inversely for the first time since a while. Especially in the current context of uncertainties and geopolitical tensions, it's important to keep our financials strong.

And to conclude this financial section, I would like to share our next quarter's guidance. Our revenue is expected to come in within the range of $215 million to $225 million with an EBITDA margin in the range of 22.5% and 25.5%. This corresponds to a full year revenue in the range of $863 million to $873 million for the full year 2025. This guidance is based on an average exchange rate of USD 117 to euro, and does not take into account the impact of the IFRS 15. And now I would like to give the back -- the word back to Rudi.

Thank you, Alba. I'm glad about the solid increase in revenue for the third consecutive quarter amid a challenging macroeconomic environment. This is a significant interest or there is significant interest in our specialty technologies.

Our silicon carbide business has made measurable progress with growing design activity on our latest silicon carbide technology platform. Additionally, our microsystems division continues to advance with collaborative co-creation projects enhancing our growth pipeline, while visibly continues to be limited, I'm confident in X-FAB position that supports sustained long-term business expense.

Besides the third quarter results, I announced today that I will be passing on the CEO role to Damien Macq, today, COO; on the 6th of February 2026 after the full year results call. Damian is very well prepared and the Board of Directors and myself have full trust. He is the right person to lead X-FAB. I will make sure there is a smooth transition. I will be supporting him in my role as a member of the Board and of course, further future. Operator, we are now ready for taking questions.

[Operator Instructions]. Our first question this afternoon or this evening, will be coming from Mr. Michael Roeg of Degroof Petercam.

Yes, Well, first of all, congratulations on taking the next step and well, spending a bit more time in the Board of Directors, looking down on your successor and guiding him. But of course, I will leave you with a couple of tough questions.

So the first one, bookings have come down strongly, and prototyping sales has also come down based on the chart in your PowerPoint presentation. So should we expect a slow start in 2026?

Well, first of all, the prototyping is something that fluctuates. There are milestones on projects and so forth that -- so I think the prototyping is at a good level. Remember, this is discussed from 0 every quarter, and it's all about new contracts and new activities. So it's still a substantial business development activity ongoing, so on. Quite happy about that.

The production bookings, they indeed are weak, that we still have quite a good backlog. That represents roughly almost to our 2 quarters. So it's a bit less -- it is, of course, too early to say, but it's a sign of weakness in the market. So our guidance for the Q4 is still good. It's in the range of where we were now Beyond that, visibility is low. And yes, so we could see maybe a weaker start from next year.

Okay. And is there a decent amount of backlog for way deep into 2026? Or is most of it typically scheduled for Q1, Q2?

This backlog is -- so customers, they order now typically, when they need the goods. So typically, all this backlog is mostly to execute on deliveries in the quarters to come.

Okay. That's reassuring at least. Then I have a question. If I compare your sales in Q3 with those of Q2, then they've grown by $13 million, yet the sequential trend in gross profit is minus $2 million. And this is not explained by depreciation and amortization because that was the same in the 2 quarters. So something was in your cost of sales, something strange. Can you explain that?

Yes, there is an effect that -- of inventory. So the work in progress, I mentioned we have shorter cycle times because we have improved capacity. The cycle times come down in the -- in the fab that is very much appreciated by our customers, so we can deliver faster. But as a result, the work in progress is lower and that has -- in the quarters where we decreased this WIP as a negative effect on the profitability.

Is that something that only hits your P&L wants to the lower work in progress level and then if it remains at that level in Q4, and then you will not have that cushion?

Yes. yes. So this is -- this is an effect that we have when the WIP -- so this is a typical effect when the activity in the factory decreases or the cycle times as short and the valuation of the WIP decreases when the valuation is -- when we come back to a steady state, then this effect is not there.

But you can also have the opposite effect if bookings go up, loading in the fab goes up, you have the opposite effect where the revenue is not yet there, but the WIP increases and the WIP is valued and therefore, it has a positive effect -- could have a positive effect on the margins.

And if I do the calculations, it's around $4.5 million to $5 million impact in the quarter. This is above average, I guess, normal trends, correct?

Yes. So this is -- if we look at the full year, it's even bigger. So we also have inventory or WIP corrections in the previous quarter. I think it is coming to a stabilization in -- by the end of the year.

Okay. Then my next question is about Texas. You mentioned in the press release that there will be capacity expansion in 2026 towards the end of the year. How much CapEx is there involved with this particular expansion program?

There is no CapEx involved. What is mentioned there that these equipments that are already delivered and paid for that will be qualified and will be added to the operating and the production lines.

Okay. So this is part of the existing program that has just been completed. So there is no additional expansion program currently planned?

For now, there is no expansion with additional cash out planned.

Okay. And during the Capital Markets Day, you mentioned that there would be discussions about further automation of 4 of the 6 fabs. Is there anything that came out of that or a midterm plan for that?

Well, this is an ongoing process that we're working on more automation. This is mostly labor and IT and that kind of thanks to a lesser extent, related to CapEx. There might be some CapEx that is minus compared to the equipment investments.

Okay. Then my final question, a very quick one. There was $2 million of other income in the OpEx. Can you explain what that was for?

There was a sale of $2 million.

Next question coming from Emmanuel Matot of ODDO BHF.

I have 3 questions. First, already, could you comment about your decision to step down from your role of CEO, what are the motivation behind that very important decision for you?

Second, too early to guide for next year, for sure. But with the churn you have and the ramp-up of significant production capacities, are you confident at this stage for further sales growth next year? And third, I wanted to know if you compete in some spaces with GlobalFoundries because it has just announced a significant capacity expansion in Germany, and I wanted to have your view on that.

Yes. So first of all, with respect to the organizational change and me stepping down as CEO. So as we as a manager or as the founders of the company, we always, yes, have in mind that, yes, it's important to grow succession and then move on. I think while the team is very well prepared and Damien Macq is ready to take that role in my view. And therefore, I decided or started thinking about this a while ago. And I think now it's the right moment to for him to take over. And I think this is very well placed to do it.

Second question, so I didn't -- now the question was GlobalFoundries. It was a question on GlobalFoundries Dresden. So GlobalFoundries Dresden, they're in different -- I don't see this as a competition to X-FAB. They're maybe also doing automotive first, but that's more into ECU type processors and FD SOI for further applications. It's a different type of SOI. So X-FAB is also doing a lot of SOI and we are very successful in SOI, but it's high voltage SOI. That's different characteristics and FD SOI is more used for lower voltage systems and lower power applications. And the third question was?

It's about the visibility you have now, which seems to be limited, but also we can expect significant production capacity next year on products, you are fully loaded. So growth next year or Gary, you don't want to comment at all.

Well, today, the capacity, we are not in allocation anymore as compared to a year ago. So the revenue, the output is mainly driven by the demand in the market that we are following 1:1 now. And so if demand is there, we'll be growing if -- so we will follow the demand. So we're able to -- if demand is there, we're able to anticipate only take advantage and grow. If it's not there, of course. So we are now dependent on the market and, of course, all the new projects that we are in the pipeline, and we gradually rolling out.

We'll now move to Robert Sanders of Deutsche Bank.

Best of luck with your next role and welcome to Damian. I just had a question about Nexperia. So you're in the European automotive supply chain. There's been a lot of warnings around line stoppages from both European and non-European OEMs talking about significant risk weeks of inventory. Can you just give us your take on how severe and serious the situation is based on what you can pick up? And then I have a few follow-up questions.

Yes. I'll following this, of course, also closely, but I also -- most of it is what I also pick up in the press. So what I know is that Nexperia is indeed in it's more -- seems like low tech components, but they're very good at it, and they're producing that in very high quantities. And so they're having some areas of significant market share.

And I think for most of those components, there are replacements, but ask the question whether these -- if there are shortages, whether the replacements can be ramped up quickly enough. I have no to my knowledge, it is not yet line stops, but I cannot tell how far it is of.

Yes. I mean, based on your experience, I mean, what they do is they do small signal logic components like diodes and BJTs and stuff like that, but they are used in like body, comfort, lighting, BMS, interfaces, sensors, safety, just a lot of low-value components. As you say, they have very high market share. I mean based on what you've seen in your previous job, I assume you would be looking at 6 to 9 months to requalify on a competitor. Is that the sort of time line?

Well, I think that also, if you look back at the COVID situation, normal -- the market is normal, then all these things take time. However, if there is a threatening line stop things can -- things can change quickly, and there is a lot of agility and creativity.

So I think it is more -- I think it's not so much a matter of qualifying it. I think there is -- most of the cases, people are very agile and flexible to move if it's really needed. But it's more a matter of the components in sufficient quantity there from alternative sources.

Now I also typically see there have been cases in the past also when Sumitomo plant was blown up years ago. That was producing 50% of these particular chemicals that were absolutely needed everywhere in the semiconductor industry. We had 50% market share. But -- also there, the dynamics, the agility of the whole market, then that came in motion. And finally, it's sourced it out. So let's see how this will turn out.

Right. And then just a question about China. Obviously, since we last spoke China has turned downwards. There's been production cuts at BYD and Li Auto and all these other guys, too much unsold inventory. How have you seen that manifest in your business, whether it's direct with Chinese customers or indirect through Melexis?

I think that's too early, too early to say with somewhat further in the supply chain. It's difficult. We don't have a good visibility, except that we see our bookings in the third quarter that were lower than the previous quarter, in particularly in the automotive segment a bit across the board. As I mentioned, the bookings in industrial, they were good. Yes. So it seems to be more on the automotive side, as I see the weakness.

And just one last one on OpEx. How should we think about the impact on OpEx of all these various different expansions, whether it's on G&A or just on your cost base more generally?

That we do not expect an effect except to the fact that once these expansions are active and producing, then they come into the depreciation. So it will have an effect on our depreciation.

[Operator Instructions]. We will now go to Guy Sips of KBC Securities.

Most of my questions were already answered. There were also experience-related. I have one question on the data center. Do you see there the positive trend, you see that accelerating? Or is it just on a continuous pace as it was over the last quarters? Or do you see a real improvement since, let's say, the Capital Market Day?

Well, I think it's -- in the revenues, we see a gradual progress. So in the beginning of the year, we saw strong activity in data center that continues, but it is complemented now also with better demand for industrial and renewable so inverters for renewables and also a bit of automotive.

So it's the silicon carbide activity is broadening. And I think that a lot of the data center -- yes, growth still has to come. So it's not yet -- so first of all, the architectures architectural change that uses more silicon carbide in data centers still is coming. And I think also all the announcements on CapEx and so forth start to build buildings and they're not yet installing the infrastructure yet.

And can you put a kind of a time frame on this?

No, I cannot answer. It's -- I think some of the customers of us who have announced activities with the data center companies on 800-volt architectures and so are rather talking about 20 -- real ramps in '27.

Next question will be coming from Mr. [indiscernible] who is a private investor.

My question is actually the following. If AI is applied, let's say, on the development of prototypes, is it possible to substantially reduce the throughput time for the development and the prototyping.

It's a good question. I don't -- not really -- that's not really the case. So there is -- in digital -- in the digital world, there is more activity on automating design environments and so forth. So I think there, it could have an effect on the analog mixed signal design that we're doing so far, there are people looking at it at research institutes, but nothing that is practically usable to my knowledge.

Okay. And then I've got a second question. Is it the automotive business part of it? Is it coming to end of life, and that has to be replaced by new products, new components or in which stage are we?

Not particularly. So we -- there is, of course, a continuous flow of innovation, but we have existing products that are -- that exist already a couple of years that will -- that are also being designed in, in electric vehicles in certain functions. So that continues.

But -- it's not that there is an abrupt end of life of certain components. Of course, if combustion engines are used to a lesser extent, if you have like applications like lambda sonda or pressure sensors that go into an exhaust system of a combustion car, that will gradually phase out. But yes, as you hear, there is push in Germany to extend lifetime of combustion engines and so forth. So I do not see an abrupt change in the next cities rather as a gradual change over the next 10 years.

Okay. And could you elaborate a little bit more on the Chinese market for our business?

Yes. So as X-FAB, we have direct Chinese business. It's around 10% of our business. But indirectly, we have also customers are selling into China. So I think the direct and indirect business of X-FAB, it's maybe closer to 35% or so of our revenue that goes into China.

Now what our customers do have less that's a bit further away, as I would have to refer to the conference call of our customers. What we sell directly in China are predominantly technologies that are, to a lesser extent, available in China. So that's typically our BCD on SOI for high-voltage smart systems, like motor drivers, battery monitoring systems and so forth. And there, we see continued interest also for new designs with our customers because this technologies are not immediately available in foundries in China.

[Operator Instructions] We'll now move to Trion Reid of Berenberg.

Trion Reid from Berenberg. Just also wanted to add to my best wishes really for whatever you're going to do in the future. I had 2 questions on the results. The first was just around the fact that you talked about the results being strong, but the order intake weak. I'd be interested if you have any comments around the pattern of orders through Q3.

You saw a sort of any trend or you're just seeing more lumpiness or short-term ordering? Just be interested to get any more color on that. And then my second question was just on the CapEx, which was pretty low in Q3, but your Q4 guidance seems to suggest it will actually ramp up quite significantly in Q4, just to understand why that is and what that sort of implies for next year as well.

Yes. So your first question with respect to the bookings yes. So the -- yes, that is what it was in the third quarter. So far, if you ask what happened in October, yes, that's somewhat a continuation of the Q3. The peak or the low amount of CapEx in Q3 is indeed less than what we forecasted.

This has to do rather with -- we're not pushing the throttle on our expansions in Malaysia. So the -- this is just a delay. So the cap -- we have not canceled anything. It's just that there is a slower rollout and therefore, also the invoices come in slower and so the cash out is also somewhat less.

So the -- in Q4, we expect that it will increase. This is not new CapEx, but it's just a shift of things from Q3 and Q4, we'll have to see maybe there will be some shift from Q4 into Q1 next year. But the increase that you -- that we can expect in Q4 is not a sign for next year. It is just the balance from Q3 that shifted in Q4.

Okay. Great. And just to follow up on the order intake. You're suggesting that there was not a decline through the quarter that the quarter was weak for order intake in all 3 months.

No, it was somewhat flat over the quarter. It's not that there was a sudden effect at the end or so. It was somewhat flat over the 3 months.

We have a follow-up question. This one coming from Mr. Robert Sanders of Deutsche Bank.

Just on the Kuching ramp and the EUR 1 billion expansion that you've done. Is the idea to continue to push that ramp, even though the demand is not really there just to sort of get the economics going because of -- right now, it's a lot of tools but not a lot of revenue? Or is the idea to basically free that plan until the demand is there?

The plan is to continue with the rollout of the equipment and to install it, qualify it to be ready for demand that can come. So the -- the equipment is mainly for our 180-nanometer BCD-on-SOI and there, we have good design wins. We have a unique position in the market. So we have -- we feel confident that this will pick up, and we want to be ready.

Got it. And is there any plan to reduce or exit any legacy facilities? I think you've already said you're going to end-of-life effort by the end of '26. So I'm just trying to understand how much of your capacity that's quite old now is sort of going to get wound down as part of this upgrade.

Well, this is an effort we're making the transition from already for many years, we're transitioning to more micro systems business. and gradually exiting the CMOS there. The end of life that we announced there was for 0.6 micron CMOS signal. CMOS is also a in there and that will run until end '26, a little bit in '27 and then there will be one cleanroom there that we -- there are 3 cleanrooms on the site there, and there will be one cleanroom that will be closed.

Got it. And maybe a question for Alba. What is the France run rate annualized at the moment for revenue in just Corbeil?

So yes, Corbeil is currently at $54 million per quarter at the moment. So yes, it was the highest quarter that we recorded now in Q3. So it's ramping up significantly. We also invested there in additional tools, as you know, and we are now running only X-FAB technologies since a while. So the efforts start to pay off.

Got it. And what's the EBITDA of that facility?

We don't give EBITDA breakdown, as you know, because we don't produce everything on one side. Some of the products are shifted from one side to the other. And -- so it's quite difficult to -- it's misleading to give EBITDA breakdown by site. You know that we start -- some products we start in one factory, then we continue in other factories and so on and so forth. So yes, as I said, it's misleading.

Got it. And just last question. If Melexis is able to get BCD-on-SOI processes from 8-inch Grays in China. What does that mean for you? Does that mean that you will lose 30% of Melexis business? Or is that something you don't think is a likelihood?

Yes. I don't know if grays has BCD-on-SOI, I think definitely not in the quality and the features that we are offering. Anyhow, it's this -- if a customer designed a new product in another fab, it takes a lot of time to qualify and transition. In the meantime, all the existing business typically stays until the products are really end of life. So I don't see an immediate effect.

As we have no further questions at this time. Mr. De Winter, I'd like to turn the call back over to you for any additional or closing remarks. Thank you.

Yes. Thank you very much, everyone. So I'm looking forward to hearing you again on the 6th of February for the Q4 results. That will then be together with Damien and I will be then handing over to Damien. Yes. Thank you very much, and have a nice evening.

Thank you. Goodbye.

Thank you. Ladies and gentlemen, that will conclude today's conference. Thank you for your attendance. You may now disconnect. Have a good day and a good night.