Hello, and welcome to the Soitec H1 2025 results analyst call. My name is Saskia, and I will be your coordinator for today's event. Please note this call is being recorded. For the duration of the call, your lines will be on listen- only; however, you will have the opportunity to ask questions at the end. This can be done by pressing star one on your telephone keypad. I will now hand you over to Pierre Barnabé, CEO, to begin today's conference. Please go ahead.
Hello, everyone, and welcome to Soitec H1 2025 results conference. I'm Pierre Barnabé, Soitec CEO, and I'm very pleased to be with you today, along with Léa Alzingre, our CFO, and Steve Babureck, our EVP strategy. Then let's look at the disclaimer, you know, by heart. Then we have a lot to cover today. I will share the main highlights and drivers behind our H125 performance, and Léa will address the financials. Let's start with the highlights of our first semester. Our H1 2025 financial performance was in line with our expectations. Revenue reached €401 million, down 15% year- on- year, as we had announced at the beginning of the fiscal year. This essentially reflects significant RF-SOI undershipment to some of our foundry customers in the context of inventory absorption across the smartphone value chain.
Our 33% EBITDA margin demonstrates the positive contribution of new products and the resilience of our operating model in the context of lower revenue. While we strictly control our cost base, we continue to significantly invest in R&D for new product development and in our capacity expansion. Finally, the significant improvement in our cash generation with € 129 million operating cash flow supports our positive free cash flow of € 35 million in the H1 2025. This strong operating cash flow allows us to continue to invest in R&D and agile capacity deployment. Looking at our revenue from a quarterly perspective, our Q2 performance confirms what we announced in May. We have started to rebound from our Q1 2025 bottom, with more than 85% sequential increase in Q2 2025. Our revenue performance reflects different dynamics across our three strategic end markets.
Let's start with our Mobile Communications division, which delivered a significant rebound on Q2 2025, up 164% over Q1 2025. The H1 2025 revenue performance in our Mobile Communications division, down 32% year- on- year, reflects the inventory absorption among some of our foundry customers, which continue to impact RF-SOI as anticipated. Customer inventories are stabilizing around one year on average, which seems to be the new normal, and we are confident that the improvement in RF-SOI inventories will support our rebound in H2. Beyond RF-SOI, we continue to make significant progress in the adoption of new products in 5G smartphones, Wi-Fi 6, 6E, and 7, as well as applications beyond mobile, such as 5G infrastructures and satellite communications. POI continues to ramp very fast, with 10 customers in productions and more than 10 in qualification.
Our product has been endorsed by all key U.S. players, and volumes are driven by a strong appetite worldwide. FD-SOI adoption for both sub-6 GHz and millimeter wave continues to expand, as all major players have selected our products. On the end market side, we anticipate the smartphone market to return to growth in 2024 and 2025 in the low single-digit territory, while 5G continues to grow. Let's move to Automotive & Industrial, where the weakness across the market has impacted our H1 2025 revenue, down 20% year- on- year. In a flattish, light vehicle environment, the automotive semiconductor market is set to grow at a modest mid-single digit this year, before recovering above 10% in 2025. We continue to see increasing adoption of our products and growing content per car, driven by the increase in infotainment, autonomous driving, functional safety, and electrification.
On Power-SOI, inventory adjustment at customer level has impacted our revenue in H1 2025. FD-SOI adoption continues to progress rapidly, with leading foundries and IDMs developing products, such as GlobalFoundries design wins on 22FDX and STMicroelectronics' recent 18 nanometers adoption, with production expected by H2 2025. On the SmartSiC side, we continue to see a growing appetite. We are pleased to announce a fourth customer in qualification with the addition of a significant player in Asia. On top of these four customers, the number of prospects evaluating SmartSiC has grown from around 25 before the summer to 35 today. Electric vehicle penetration continues to grow at a lower pace than last year. Qualification cycles on the SiC value chain are lasting around two years longer than anticipated, as we already shared in July. Our fab is ready to start ramping up in line with customer qualification.
Our edge and cloud AI divisions continue to grow, up 57%, driven by a strong appetite for low-power computing devices and edge AI applications. The increase in revenue was driven by higher sales of Photonics- SOI, which benefit from the AI-driven growth in data centers' capabilities and by a positive phasing effect on imager SOI. On Photonics- SOI, we continue to leverage the strong cloud AI momentum in the industry, evidenced by high CapEx commitments across the AI value chain.
We are accelerating our partnerships with leading-edge fabless to support AI/ML developments. Co-packaged optics enable new data center architectures with better performance, energy consumption, and lower costs, a critical challenge for the artificial intelligence value chain. On FD-SOI, our product portfolio continues to expand, supporting new generations and extending beyond general purpose into AI computing. Strong customers' appetite with new product adoption and committed capacity investment support our product roadmap.
Finally, on Imager-SOI, we are sampling second-generation wafer dedicated to higher-performance sensing applications, as we will progressively be phasing out Imager-SOI first generation. Let's now have a look at the sustainable deployment of our expansion. We have set very ambitious long-term targets on the key pillars of our ESG roadmaps, dividing by two our water consumption by fiscal year 2030, reducing our Scope 1 and 2 emissions in absolute terms by 37% by 2030, improving the diversity among our employees to about 40% by fiscal year 2030, and delivering on our five ethics key pillars. We continue to progress on our targets. On the water front, we have already passed 40% of recycled water in our processes, on our way to our 50% targets by fiscal year 2030.
We also shared important news on the governance side yesterday, with the appointment of Frédéric Lissalde as Chairman of our board of directors, effective on the 1st of March 2025. I am delighted with this decision, and I look forward to benefiting from Frédéric's experience and vision. I would also like to express my deep gratitude to Christophe Jégoux for his trust and guidance during this transition phase. As of March 1st, 2025, Christophe will resume his former roles as referent director and Chairman of the Audit and Risk Committee, and he will continue to serve on the Strategic and ESG committees. Let me now hand the microphones to Léa in order to comment on our financials.
Thank you, Pierre, and good morning to all of you. We have met our revenue guidance for H1 of minus 15% year- on- year, as we still navigate a complex environment. We are monitoring very closely market conditions and inventories at our customers. We continued our financial discipline to maintain profitability while preserving investments, and our four strategic priorities remain unchanged: one, extend our technology leadership; two, accelerate our diversification beyond RF-SOI products; three, ensure we have a relevant positioning in terms of products and the appropriate cost structure; four, optimize value creation for all our stakeholders. Let's start with the main takeaways of our H1 2025 earnings. We delivered €338 million in revenue, down 15% year- on- year, at constant FX rates. We maintain strong profitability with a 33.4% EBITDA margin.
Our EPS is down year- on- year , reflecting the lower loading of our fabs and the level of revenue during the period. At the same time, we managed to generate positive free cash flow at € 35 million, with strong monitoring of our working capital while we continue to invest in our future. And we maintain a healthy balance sheet with a moderate net debt at € 51 million. Pierre already commented on the revenue performance. Let's move directly to our gross margin. We delivered a 30% gross margin, down 6 points year- on- year , in line with our expectations. As anticipated, we faced three main headwinds: one, higher depreciation expenses, as expected, two, underutilization of our SOI fabs, especially in France in 200 mm and in Singapore in 300 mm, three, a lower revenue, as explained earlier.
These unfavorable effects have been partially offset by a strong industrial performance, with improved yields and efficient agility in resources allocation between our fabs, a highly effective cost control program, and favorable effect of grants. Gross margin is and will remain a key priority for us. As explained in May, we anticipate our FY 2025 gross margin to be almost stable year- on- year, between 33% and 34%. The main drivers of the gross margin improvement between H1 and H2 will be a higher operating leverage coming from a better fab loading and a much higher revenue, and a more favorable product mix. Let's move now to the current operating income, with which EUR 28 million, an 8.2% margin, as compared to 21.3% in H1 last year.
We continued to significantly invest in innovation, with gross R&D expenses before capitalization up EUR 12 million, a 90% increase, reflecting our priorities to: first, reinforce our technological leadership in each of our three end markets and across all product lines, second, increase efforts in upstream R&D, hire talent, and continue to build collaboration with innovation platforms. At net level, R&D expenses were up EUR 9 million. SG&A expenses amounted to EUR 31 million, up EUR 6 million year- on- year .
This increase is mainly due to an unfavorable comparable basis with non-recurring effects in H1 2024. Without this, the increase would have been less than EUR 2 million. At the same time, we maintain our financial discipline. Finally, our 33.4% EBITDA margin is almost flat year- on- year . The decrease in EBIT margin has been offset by the increase in non-cash items, mainly depreciation expenses and non-cash effect on inventories.
Taking into account a EUR 4.3 million depreciation on Dolphin Design Goodwill, our operating income reached EUR 23 million. Moving now to the bottom of the pyramid. Regarding financial results, it was a loss of EUR 8 million compared to a gain of EUR 2 million last year due to the FX results, a loss of EUR 6 million this year versus a gain of EUR 3 million last year. Our income tax continued to benefit from tax loss carry forward, but was also impacted by the effect of the new Pillar 2 regulation. As a result, net income reached EUR 14 million, representing around 4% of revenue. Let's move to our cash flow. We generated a positive free cash flow of EUR 35 million in H1 2025, a significant improvement of EUR 120 million compared to the EUR 85 million negative free cash flow last year.
Despite a lower EBITDA, this strong increase essentially comes as a result of a much improved working capital over the period while sustaining capital expenditure. We generated a €129 million operating cash flow compared to €45 million in H1 2024. This strong improvement was due to the working capital decrease by €27 million, explained by a €130 million decrease in trade receivables due to the seasonality of the activity and the good performance in terms of cash collection. This favorable effect being partially offset by: one, an increase of inventory in anticipation of the H2 deliveries for €65 million, same as last year. Two, a decrease of €48 million of trade payables versus a decrease of €105 million last year. As you may remember, in H1 last year, we had non-recurring down payments to suppliers to secure long-term supply agreements.
Last, we had a favorable effect on tax paid, with a decrease of EUR 9 million as compared to the same period last year due to tax reimbursements over the period. Let's continue with cash out from investing activities at EUR 94 million, which reflects EUR 120 million of cash out from CapEx, partly offset by EUR 17 million of leasing contracts and EUR 9 million of cash in from cash investments. We invested our CapEx for capacity increase for POI, SmartSiC tools to complete our two production lines for 150 and 200 respectively, in Singapore for the continuation of the first step of our fab extension, and we also had ongoing investments in innovation, sustainability, and automation. Moving to the balance sheet, our financial structure remains very healthy, with solid equity, strong cash position, and loaned debt.
Let me highlight that we reclassified EUR 65 million assets and EUR 33 million liabilities from our subsidiary Dolphin Design following the decision to sell this business. The assets related to the IP business have been sold to Jolt Capital on the October 31st , and we signed an agreement with NanoXplore on the November 11th regarding the sale of the assets related to the ASIC business. Regarding net debt, we ended the period with a moderate net debt at EUR 51 million, up EUR 12 million. The generation of positive free cash flow for EUR 35 million has been offset mainly by the increase in gross debt due to new leasing contracts for EUR 49 million, including the second tranche of our Bernin 4 Fab. This concludes my comments on H1 2025.
You can see that despite the decrease in revenue, we maintain a good level of profitability, and we generated positive free cash flow while continuing to invest for the future. We are taking actions on all items we can control, such as operating expenses and CapEx, and we are well prepared for the revenue rebound in H2. Let's focus now on the outlook. I will now hand over to Pierre to comment on our revenue and EBITDA guidance confirmation. Thank you for your attention.
Thank you, Léa. And now let's have a word on our guidance. We confirm our fiscal year 2025 guidance for a stable revenue and a fiscal year 2025 EBITDA margin at around 35%. We are confident in our ability to extend our rebound in the second part of our fiscal year, notably for the fourth quarter, as the RF-SOI inventory level started to improve. Our confidence is supported by around 90% backlog and contract coverage today, and strengthening customers' intimacy as evidenced by new products launched with key foundries on 300 SOI. Our strong H2 recovery will reflect different dynamics across our division. Year-on-year growth in Mobile Communications will be driven by continued strong momentum in POI filters and RF-SOI volumes recovering from their Q1 2025 bottom. Customer inventories are stabilizing at around one year on average, which seems to be the new normal.
In automotive, and industrial, the growing penetration of our products will continue to be offset by the ongoing weakness in the automotive market. Edge AI and cloud AI will continue to benefit from a strong momentum driven by artificial intelligence, both at the edge and in the cloud. We are also confirming our EBITDA margin around 35%. As Léa mentioned, we anticipate a strong rebound in our H2 gross margin. Fiscal year 2025 should be in the same range as fiscal year 2024, between 33% and 34%. We'll continue to control our costs while benefiting from a higher operating leverage driven by a better loading of our fabs. Continue to invest significantly in R&D to prepare for the future and continue to manage our SG&A expenses.
Finally, the third element of our guidance, we are adjusting our fiscal year 2025 CapEx guidance to around €230 million from €250 million in order to reflect a moderate rebound in end demand in calendar 2025, both on SOI and SmartSiC. To conclude, we are very confident that our fundamentals will allow us to grow next year in the context of challenging market conditions. For calendar year 2025, we anticipate different dynamics across our three end markets. The Mobile Communications market is expected to continue to slightly improve. The automotive market weakness should persist through the first half of the year, and investments across the Cloud AI infrastructures are expected to remain at elevated levels.
In the context of these temporary challenging market conditions, we continue to enhance our technology leadership to strengthen our SOI positionings with both existing and new customers and to deploy our expansion into compound semiconductors with the acceleration of POI volumes and a fourth qualification initiated on SmartSiC. As you see, we continue to deliver on our diversification strategy. At the end of fiscal year 2025, four product families will contribute to more than $100 million revenue each, compared to only one a couple of years ago. Two other products are strong candidates to pass this milestone in the midterm. This concludes our remarks. Thank you very much for your attention. Now it's time to move to the Q&A.
Thank you. Ladies and gentlemen, if you would like to ask a question, please signal by pressing star one on your telephone keypad. Again, that is star one for your questions today. And our first question comes from Aleksander Peterc from Bernstein. Please go ahead.
Yes, good morning, and thank you for taking my question. So I'd have a couple. The first one will be on mobile. I just like to understand where you stand now with the sell-in versus sell-out of the channel overall. So can you confirm that the level of inventory now at foundries is such that there will no longer be a massive under-shipment into the channel in the second half? We see that equilibrium now in place in H2. And just to give us an idea, in H1, to what extent have you under-shipped? I.e., how much more of SOI was sold in the broad smartphone market versus how much you have shipped? Then secondly, on SmartSiC, could you give us an idea at what point in time you think you'll enter into volume production for this product line?
I'm under the impression we have a two-year delay here in qualifications, at least the major ones. And I'm just wondering if this gives a longer runway to mono-SiC products that can progress lower on the cost curve of the substrate, and that would surely put pressure on your SmartSiC product economics. So if you could just give us a little timeline and idea of the economics here. And just lastly, very quickly, what are your plans for the 25 convertibles that mature in October? Will you roll it to interim financing at all, or will you just pay it down as it's now out of the money? What are the plans here? Thanks.
Thank you, Aleks. Then for the first part, the Mobile Communications, what we observe is by the end of this year is a stabilization of inventories around one-year average, meaning that, of course, we see a positive trend for the second half in the RF-SOI because consumption rates continue to be there, of course, to sustain a 3% smartphone market increase this year and 3% next year. This is the visibility we have so far, and we state, let's say, a progressive recovery in this area. Taking into account, again, that it's RF-SOI in Mobile Communications, as you know, we have two other components that are POI and FD-SOI that are doing very well, and with a strong growth in POI adoptions with more and more customers.
Regarding SmartSiC, as we already stated, there is a two-year delay compared to the original plans we had a few years ago due to the time of qualification and adoption process of silicon carbide, generally speaking, and SmartSiC. We are, as you know, ready for, let's say, ramping up our productions. The different milestones regarding technology, supply chain, operations, and customers are past, and the next one is going to be past. What is super important is to get a fourth qualification now with the customers in Asia, certified under evaluation. We don't expect the certified to shift into qualification. Why not? But some of them are going to go there, and we are on a good traction, but the market is taking a bit more time. It's not an issue to get these delays regarding the pressure on the price.
As we said already, first of all, we are benefiting. We are observing the pressure on the prices in Silicon Carbide on the 6-inch and 8-inch. First of all, we are benefiting from this drop because we are buying a part of Mono-SiC product. Second, we are working on the two other elements of the SmartSiC, meaning process and Poly-SiC, to really be in a position to follow the curve. Then we are working on plans to remain competitive, including a premium, because, of course, we are not making, I would like to say, a material solution that is Mono-SiC. We are making a product that is bringing added value. And this added value is more and more recognized.
The fact that we have a better concentration of dies, the fact we are reducing, we are making CapEx avoidance, we are reducing CO2 consumption is something that is totally stated and understood by our prospects and customers, and we are really working very hard. We are not at all immune or blind. We observe the market, and I can tell you we are extremely active to follow the trend and to continue to be super competitive while investing in R&D in this domain. For the last question regarding the convertible, I will leave the floor to Léa.
So yes, currently, we are working on several scenarios and based on both our needs of financing for this convertible bond and also mainly on the market conditions. So we will decide, I would say, until the end of March what will be our refinancing scenario, and we will execute the first half of next year.
Thank you. Thank you very much.
Thank you.
Thank you, Alex.
We now move on to a question from Emmanuel Matot from Oddo BHF. Please go ahead.
Hello, Pierre, Léa, and Steve. Emmanuel from Oddo. I have, let's say, three questions. First, the level of sales growth is very impressive in Edge AI. Is it sustainable, and when do you think Photonics-SOI will reach EUR 100 million of annual sales? Second, it's too early to guide for sure for next year, but do you think that the tech can be back to high growth, I mean, double digit? Is that a possible scenario or too early to be so optimistic? And third, you mentioned that customer inventory is stabilizing around one year on average for RF-SOI, which seems to be the new normal. But what was the normal historically, let's say, before COVID? Was it very different? And if yes, why such a change in the value chain of RF-SOI? Thank you.
Hello, Emmanuel. Thank you for your questions, and let's start with Edge AI. Edge AI, there are several components. We have photonics, and silicon photonics is growing very fast. And we see this trend. Should it be in the range of +60%, perhaps it's a very exceptional growth, but we will get a strong growth and intense growth in silicon photonics because there are more and more needs to sustain the data centers deployment, next-generation data centers deployment. Today, we can state that there is no new next-generation data centers AI deployment without silicon photonics. Then each time you have a next-generation data centers bringing AI, you have silicon photonics. All the transceivers today are going to this type of technology. Then it is sustaining the growth. But of course, there are other components in Edge AI. You have FD-SOI.
FD is doing quite well with the general purpose adoptions, but there are some new design wins to come. It will take a bit of time, and there is also the modulator for which, as we said, we are now in the first generation that will progressively come down, and we are working on the second generation, then it's going to be, of course, a mixed analysis of the trend in edge AI. But what is clear is that silicon photonics is going to continue to grow very largely for data centers, but also, we need to think on other applications around edge, around co-packaged optics for interconnection within GPU in the future.
Then in the coming years, we see many applications around AI that are going to adopt silicon photonics because it is quicker, it is more secure in bandwidth, and also it's supporting very high bandwidth, and it is consuming less energy. That is today at stake, as you know, to sustain the AI explosions. And that's something that is a key element. Regarding fiscal year 2026, as you can guess, and you got the answer in your question, it's too early to give you really more colors. What is very important is what we said about the market because we are, of course, observing the market for 2025. Then for 2025, we see a slight recovery on mobile. We see still a first difficult semester on auto, and we see good traction on edge and cloud AI thanks to silicon photonics. That's it for the moment.
But as you know, we are in a process of, after having analyzed the market, we are looking at division by division, what are the trends for the coming years, and then we're going to deliver to the board our view for the budget fiscal year 2026 by end of March. And then by springtime, you're going to get the color of the growth. Regarding the last point for the inventories, as you know, it's a new normal we are observing. It is higher than what was the kind of normal before COVID. The normal before COVID was more in the range of eight months, eight to nine months. Why we see this as around a year new normal as a kind of stabilization of the inventory?
It is coming from the fact that you need the six to nine months from the time you put the engineer wafers in the factory of the foundries or the IDM to have the devices packaged and shipped. Then you need six to nine months. And then, as you know, more and more fabless or integrators are asking for getting a design wins, the foundries or the IDM, to justify that they have in their inventories the product, the wafers. Then if you take a buffer of security, you reach this one year as a new normal. This is the explanation of this observation we have made.
Thank you very much.
Thank you. And from Kepler, we have Sébastien Sztabowicz with our next question. Please go ahead.
Yeah, thank you for taking my question. The first one is on the guidance for 2025. So you are confident, but it requires a very strong acceleration in the back half of the year. And you are mentioning a 90% backlog coverage, but this is more or less in line with the traditional backlog coverage that you have at this point of the year. So what is making you, I would say, confident to reach the outlook as you missed it a year ago with the same backlog coverage? That will be the first question. The second one is a positive one on silicon photonics. We have seen a very big buzz in the market coming from Asia, Taiwan, Semicon, TSMC yesterday, quite bullish on silicon foundry services and so on.
Just to understand, do you see some upside on the previous message that you forecasted or past two years, 30%-40% growth on silicon photonics for the coming years? And second one, do you have any indication on the potential addressable market by the end of the decade? What could be the number of wafers that you can ship by 2030? Thank you.
Okay, Seb, on the first question regarding the coverage based on backlog and contracts at 90%, it is an improvement compared to last year at the same period because we are more in the range of 80%. Then it is increasing the level of confidence. That's the first key element. The second element is the sentiment we are getting thanks to a new type of conversations with customers that are coming back to us to really engage discussions for long-term agreements, multi-year agreements. That's something we are observing for a few, let's say, weeks and months now and quite encouraging. And regarding the remaining 10%, of course, it is based on, let's say, identified and clear business in terms of product volumes, customers, and value. Then it's not something, some deals we need to raise and to create. There are already identified elements.
Then that is giving us a certain level of confidence to reach the guidance fiscal year 2025. Regarding the silicon photonics increase, it's too early to give you a figure today, year after year, and it's going to be part of course of the disclosure of our BP. But what we see clearly is a need that is increasing. First of all, for what we are addressing today, it's mainly transceivers activities for the data centers where we see, you know, the value chain very well. Then there are more and more asked by players in data centers, AI constructions. And we see this growth to be sustainable on the long run. But they are not the only applications. There are other applications that are going to use more and more silicon photonics.
We are talking about cloud, particularly today, but at the edge, the use of silicon photonics and particularly the co-packaged optics are going to be very important drivers. And it's going to go to the GPU level because the interconnections between GPU today on the top of the components, on the long run, will be difficult to be sustainable because it consumes a lot of energy. You have a very small but still latency. While having silicon photonics interconnecting GPUs at the substrate level is, of course, a very big gain in terms of latency, bandwidth, and power consumption. That is, as you know, the risk of bottleneck of AI explosions. Then that's something in which we are confident.
And we're going to give you more, let's say, color when our business plan exercise is going to be complete on the different products. To the question of would it be $100 million product drivers, it's again too early to say, but we're going to give you more details in the coming months. Very encouraging and promising products for sure. As I said, in the long term.
Okay. Thank you.
Thank you. And up next, we have Olivia Honychurch from Jefferies. Please go ahead.
Hi. Thanks for taking the question. Just a couple from me. Another one on the RF-SOI wafer side. You talked about the new normalized level at your foundry customers. Looking ahead then, to what extent do you expect those foundries to start building inventories back up through H2, particularly if there is still some more clearing to go, as I think you mentioned in the press release yesterday? Perhaps that's only at a couple of customers. But generally speaking, how much replenishment do you need at those foundries to get to your FY 2025 guidance? And then I have a follow-up. Thanks.
It's important to, I believe, underline the word of stability. Okay? Then we are at a new normal that is higher than years ago. Okay? But this new normal seems to be stable. Is it also an expectation by the foundries and the IDM of higher consumption rates for the future? Maybe. What is clear is that progressively we see the improvements, and we remain, of course, cautious but ambitious. It's also very important to keep in mind that we are coming also to a kind of new normal in terms of smartphone growth. Because we experienced double-digit growth in the past years, in the year 2010 and the decade. Now, if we take this year and next year, we're going to be more in the range of +3%.
Then there is also a kind of overall stabilization of the supply chain in the range of this low-digit % of growth. Okay? Of course, still the elements under, let's say, discussions is how AI is going to impact the refresh of smartphone, I would like to say, inventory in the world. Will we get AI applications that will make the individuals obliged to change their hardware to sustain new types of revolutionary applications? This is another discussion. But short and very midterm, it is clear that we are coming to a progressive recovery based on a 3%, few % increase in smartphone business, while 5G continues to progress, but at also a progressive path. Then this is really something we need to keep in mind.
Great.
I'm talking about RF again. It's really around RF because POI is in another dynamic. POI is in a penetration mode. Then POI is getting shares in the existing cake, okay, while RF is expanded and is, I would like to say, mature in terms of positioning. POI is penetrated quite rapidly and eating really segment shares more and more.
Great. Thank you. I guess maybe a couple more, actually, in that case. The first is just based on your comments just then. How do you see content growth going forward? Is it sort of in line with the years' levels previously, or do you see that going up because of the various dynamics that you talked about in the next year or two, perhaps with the AI smartphone trend? And then my second question is a little bit broader.
So you've obviously maintained your FY 2025 guidance. At your Q1 results, you gave sort of slightly more color at the segment level. So you were talking about 0% revenue growth in the auto side and I think double-digit in Edge AI and cloud AI. Are those segment-level outlook statements still valid, or has there been a change in the mix given what we've been seeing in some of your end markets? Thanks.
Okay, then back to your first question on the content, then we see still an increase, of course, in the RF content because there are more and more applications to be sustained, and 5G is making its job, of course. This increase in content in RF is also helped with additional millimeter wave, let's say, increase thanks to, again, POI, and also a bit of FD-SOI for the millimeter wave shares that is today stabilized around 10%-15%, and of course, also applications like satellite communication, envelope trackers, and so on, then we see the content to continue to increase by adding these different products within the smartphone.
If we look at the structures of, let's say, and the weight of the different divisions within the fiscal year 2025, let's say, outlook, of course, RF-SOI at the end of the year is going to weigh less than 50%, more or less, because the dynamics of the auto product is quite impressive. Because when you imagine that this year, four product lines are going to hit more than $100 million each, that was not the case even two years ago. Of course, there is a change in the weight of the different divisions reflecting the markets and reflecting also the penetrations of our new products that were very low, that was very low two to three years ago. Then it's really a change in paradigm because this diversification now is going. It's a pure and simple reality.
Then it will make, of course, a change in the weight of the divisions. We need to think also midterm. Midterms, whatever automotive market this year and the first semester of next year are going to be soft, the need for electric vehicles is still very active. I mean, if you look globally at the growth of electric vehicles needs and sales, it's still quite high. Then the needs for components to sustain this type of developments are eligible to our engineered substrates, whatever it is, SmartSiC, Power-SOI, FD-SOI. But also we could imagine to have RF-SOI and Photonics-SOI in the cars tomorrow. Then it is clear that we need to look on midterm on the balance between the different divisions.
We still believe that we're going to be in the range of 50%-60% for mobile and 20%-25% for the two others divisions sustained by the growth of the different products within these divisions. Then it's very important to keep that in mind. But we're going to give you, again, more, let's say, accuracy and a date on these dynamics after the BP exercise by springtime.
Okay. Great. Thanks. And just to clarify, you are still expecting flattish revenue growth in the auto division for FY 2025?
We are expecting, yes. Our guidance is to have a stable revenue fiscal year 2025.
Okay. Great. Thanks.
You're welcome, Olivia.
Thank you. Thank you. And as a brief reminder, that is star one for your questions today. And up next, we have Adithya Metuku from HSBC. Please go ahead.
Yeah. Good morning, guys. So a couple of questions from me as well. So firstly, we've seen a lot of mixed data points from the RF guys in the last quarter. We've seen negative headlines around iPhone demand. We've seen negative data points around automotive, industrial, and markets. We've seen your ST Micro, your customer cut outlook, warn about seasonality in the March quarter, which is your last quarter where you're expecting a significant recovery. And yet you're keeping your guidance unchanged. So I just wondered if you could help me reconcile the data points I'm seeing and what you're seeing that is leading to you keeping your guidance unchanged. And then I have a couple of follow-ups. Thank you.
It's a novel comment on the different product lines, but what we see is for the fiscal year 2025, we see RF, let's say, development and shipment as planned to the initial, that's what we were expecting initially. Okay? Then RF is in line with the way we built the guidance early this year. Automotive, as we say, is a bit weaker than what we are expecting. But it is, and we said it already in July this year, but it's going to be compensated by good tractions on other products, particularly edge and cloud AI, and particularly silicon photonics and FD-SOI. Then you have RF really in line with expectation. POI in line with the growth expectations we have designed at the beginning of the guidance.
Auto, as we said already, is nothing new, a bit weaker than the initial plan, compensated by good traction on edge and cloud AI, and particularly silicon photonics. It is showing also that the company is able to be flexible. This is also one of the strengths of the company. We are resilient in terms of, as it has been underlined several times by Léa, we are resilient in terms of profitability and cash because despite some ups and downs within the quarters across the year, we are able to be flexible in terms of typology of productions and in our capacity to react to customer demands. That's, I believe, a very important signal that around our diversification, our flexibility is bringing a lot of resilience and adaptive, let's say, capacities, capabilities to the company and to the group, whatever it is in Bernin or in Singapore.
Understood. And just as a couple of follow-ups, are there any non-cancelable orders or is there anything like that that you're shipping into your customers, which your customers are taking because they have to and not because they want to? Is there anything like that that we need to be aware of that's going on?
No. No. There is absolutely no cancellation of whatever POs or whatever change in contracts. There are clear needs for the products we are preparing for them, for our customers. Then this is not at all something we are observing. We are observing, I would like to say, the other way around. We are observing customers calling us, calling me directly to discuss long-term, let's say, commitments and even some common ideas to develop next features or next applications in future platforms. And this is something that is quite new and giving a very interesting and promising signal.
I got it. My question was specifically around non-cancelable orders, so are there agreements that your customers have made previously that require them to take the wafers even if they don't need them? Is there anything like that that is going on?
No. No, no. Nothing like that.
Just lastly, so would it be fair to assume that your expectations by division have not changed specifically in the last three months for this year?
Yep. Yeah, yeah. Yeah, exactly. Then the latest view and pictures we gave you on July with soft automotive division compensated by edge and cloud AI. It's the right pictures that are going to drive us to the guidance delivery.
Have they changed by product?
By product, it's a bit the same because we have not, for the moment, 15 products by division, but then it's three products by divisions with some key ones. RF is dominating mobile while POI is getting more and more, of course, territories. Automotive is largely driven by power, whatever. FD-SOI is getting more and more traction while SmartSiC is in adoption mode, as you know. Edge and cloud AI is driven by, of course, a good momentum on FD-SOI, but as I said, a very good growth and traction coming from silicon photonics. Those are the big dynamics within these divisions. It is a good demonstration of the large diversity of our product, but also for customers because we didn't talk that much about customers. We were expecting around 20 customers generating 90% of our revenue. We are on that trend.
That's also a big change. It was not the case three years ago where seven customers were making 90% of our revenue. Then it's also a way to be diversified and to solidify also foundations and also reducing the risk and increasing visibility to some extent on the top of, of course, China that is becoming more and more important.
Understood. Thank you.
You're welcome.
Thank you. And we move on to a question from Robert Sanders of Deutsche Bank. Please go ahead.
Yeah. Good morning. Yeah. Three questions, if I may. The first question would just be, could you discuss your position in imaging? Looks like ST, one of your partners, has got a new socket at a large customer. Just wondering if that is part of your plan to serve that socket win out of Crolles. And then second question would be on filters. Can you just talk a bit about your market share if we include Broadcom as a BAW filter player within the potential served addressable market?
How much share penetration have you had so far, and where do you think that can go, including potentially winning Broadcom? And then the last question would just be on the smaller diameters. I assume that the loading in those lines is pretty low, probably below 50%. Is there any scenario where you could reduce capacity, do a bit of restructuring in order to deal with that underloading? Thank you.
Okay, Robert. Then on the first question regarding the imager, I will not comment, of course, on any design wins and any customers, let's say, discussions in the overall supply chain. What I told you is that we are in the first generations of imager that has been very successful. Of course, there are new types of substrates to sustain these, let's say, imaging capacities and capabilities. We are working on these second generations. The first generation is going to progressively, let's say, ramp down starting now for the coming, let's say, quarters. We are working on the second generations. They're going to come later on. We'll have a transition period. Of course, we want to continue to be active in the imaging imagers world till end of the decade.
Regarding the filters market, we always said we want to reach 30% market shares by 2030 in terms of overall filters market with our POI. We see clearly this target in the line of sight because we are getting more and more traction. As we said, there is a big shift particularly appearing this year where we are now equipping Western world providers. You have heard some announcement by some key integrators like Qorvo mentioning a next generation POI.
I don't say it's us, but you can guess that these types of announcements are extremely appealing and extremely promising, and there are other players, particularly in the US, who want to really make the filtering elements and components more accurate and able to support a large scale of frequencies. You can imagine that with our POI range of products, we are answering exactly the needs of the markets as it is today. I continue to believe that this product is going to become a standard in our industry. For the loading of our factories, I'm going to pass the mic to Léa. I could give you some points.
So you are fully right. During the first half, the loading for 200 mm fab in Bernin was quite low. But for the second half, we are anticipating a much higher loading, mainly thanks to Power- SOI products for our Q4 deliveries. And keep in mind that for 200 mm SOI product, we are also working with our partner in China, Simgui. So this is a way for us to create flexibility and agility in terms of production.
Thanks a lot.
Thank you. And that concludes today's Q&A session. So with that, I'd like to hand the call back over to you, Mr. Barnabé, for any additional or closing remarks.
Thank you all for following our H1 2025 analyst call and for the quality of your questions this morning. The next date in our agenda will be the release of our Q3 2025 revenue on February the 5th after market close. This ends our call for today.