Hi everyone, and welcome to Soitec Fiscal Year 2025 Results Presentation, whether you're here with us in Paris or joining remotely. I am Alexandre Petovari, Investor Relations Director at Soitec, and on behalf of the team, thank you very much for joining us today. Before we begin, let me remind you that today's presentation includes forward-looking statements. I encourage you to review our disclaimer available in the presentation materials. Today's session will begin with our CEO, Pierre Barnabé. You will then hear from the rest of the executive team on strategy, innovation, operations, and finance. These prepared statements will last about an hour and 10 minutes, and we will then move into Q&A and close the session with a wrap-up from Pierre. Thank you again for being with us today, and with that, Pierre, over to you.
Thank you, Alex. Good afternoon to everyone here in Paris, and a warm welcome to all of you joining us online, wherever you are in the world. Thank you for attending our fiscal year 2025 results presentation today. I'm joined by the management team here, and it's a pleasure for all of us to be with you again. Fiscal year 2025 has been a year of continued transition. We made meaningful progress despite persistent volatility. We have continued to execute with discipline, to accelerate our diversification, and to prepare Soitec to deliver on its long-term growth ambition. Today, I'd like to discuss how we have strengthened the company, how we are navigating this complex environment, and why we remain confident in the strong fundamentals of our operating model. Let me start with an overview of the context.
Fiscal year 2025 was impacted by the ongoing inventory correction across the smartphone supply chain, coupled with weakness in the automotive market. These headwinds have affected our top line, but not our trajectory. We stayed resilient, maintaining a strong EBITDA margin, solid cash generation, and a healthy balance sheet. We also used this period to get stronger, accelerating diversification, expanding our customer base, and sharpening both industrial and financial execution. Our foundations are stronger than ever, which helps us navigate this volatile environment. Our operating model positioned us for two-times revenue upside, significant operating leverage, improvement in cash generation potential. We will give you more details on that model through the presentation today. Our revenue declined by 9% overall, in line with our revised guidance, with different dynamics across divisions. Mobile Communication was down 12%. Strong growth in POI and continued FD-SOI adoptions were offset by the RF-SOI inventory correction.
Automotive and industrial declined 22%, reflecting the broader weakness in the automotive market, but product adoption continues to progress. In edge and cloud AI, we saw very strong traction, especially from demand related to AI. That growth was partly offset by lower Imager SOI sales, as we saw the first impact of the phase-out. Let's now dive into our first end market, mobile communications. Let's start with RF-SOI, our reference substrate for RF front-end modules embedded in 100% smartphones. We continue to strengthen our leadership. Our segment share moved above 90%. New design wins confirm our strong position, notably with GlobalFoundries 9SW, UMC 3D IC for 5G, and Broadcom and Tower for Wi-Fi 7 RF-SOI. Regarding RF-SOI customer inventories, it is going in the right direction. If you remember, a year and a half ago, they peaked at a much higher level, around 18 months.
At the end of December, they were sitting around 14 months. These inventories are coming down step by step. If the smartphone market sustains its ongoing recovery, they should continue to go in the right direction, probably around 11 months at the end of December 2025. Moving to POI, on its way to becoming the industry standard for advanced SOI filters. Momentum is strong. POI is our fourth product to reach around $100 million in annual revenue in fiscal year 2025. We now have 10 customers in production and 13 in qualification. Adoption continues across high-end smartphones, Chinese OEMs, and also wearables, including notably the iPhone 16e, Google Pixel 9, Samsung Galaxy Z Flip 5, and even smart glasses now with the Ray-Ban Meta model. We continue to see strong engagement from all key US fabless. We are also working on next-generation POI to address ultra-high-band performance requirements.
Finally, FD-SOI adoption continues to build up. We have secured strategic design wins for millimeter wave and envelope trackers across several flagship devices, and we are now expanding into new verticals beyond smartphones. Turning now to automotive and industrial, the market environment was tougher than expected, but our long-term roadmap remains robust. Starting with Power- SOI, volumes were lower in f iscal year 2025 due to market weakness. Interest from existing customers and new prospects continues to grow, especially as we are developing roadmaps aligned with a growing China-to-China trend. With a strong outlook for battery management systems, we are preparing the transition of Power- SOI to 300 mm. FD-SOI is gaining traction in radar and connectivity, with adoption expanding across leading players, including GlobalFoundries 22FDX. That said, in fiscal year 2025, sales were impacted by two key customers who put on hold orders.
Despite this temporary pause, the fundamentals remain strong, and we see clear momentum building up. Finally, on SmartSiC , despite the previously communicated delay in ramp-up, commercial interest remains high. We have engaged in a sixth customer qualification, and 35 prospects are evaluating our product, many of them in China, a strong testimony of our efforts to remain competitive. Customers continue to validate our value proposition: superior device performance, 10x mono SiC substrates' reusability, and lower CO2 emissions. Beyond EVs, we are seeing rising interest from sectors like renewables and data centers. Let's move on to edge and cloud AI, where we are seeing some of the most exciting dynamics across our portfolio. Photonics- SOI is another strong example of our product diversification and should become our fifth product to reach $100 million in annual revenue in fiscal year 2027.
We are now leading suppliers of Photonics- SOI for optical transceivers in AI data centers, enabling greater performance, lower latency, and lower energy consumption. Photonics- SOI is embedded in 100% of next-generation AI data centers. Silicon Photonics is a leading architecture for co-package optics, being adopted by industry leaders like NVIDIA, Broadcom, and Marvell. Together with AI players, we are accelerating the Photonics product roadmap, keeping Soitec at the core of next-generation AI architecture. FD-SOI is also powering AI at the edge, in smart devices, wearables, and industrial applications. Adoption is growing fast, and we are collaborating with multiple foundries and IP providers to bring the next wave of products to market. Finally, in Imager- SOI, the phase-out of our first generation is completed. We are already developing the next-generation platform with improved performance for advanced near-infrared sensing and imaging. A few words now on our key financials.
Despite the top-line pressure, EBITDA margin remained resilient, a clear reflection of our operating discipline. Operating cash flow improved, supported by tighter control of working capital, and our balance sheet remains very solid, giving us both flexibility and stability. All in all, we maintained strong cost control, optimized fab utilization wherever possible, and continue to prioritize investment in R&D and sustainability, even in a more constrained environment. We also made strong progress on our sustainability roadmap, which remains a core pillar of our strategy and a key differentiator for us and our customers. It enhances our attractiveness to talent and partners and drives industrial efficiency. On the environmental front, we reached our carbon emission reduction target two years ahead of plan, as validated by the SBTi. We have also made significant progress on water reuse and reduced overall water intake.
On the people side, we continue to lead on diversity in our industry while strengthening our efforts to make Soitec a place where people can grow and thrive. In governance, we are raising the bar. We are rolling out training and awareness programs across the company and ensuring that ESG performance is embedded in incentives at all levels. As we just spoke about people, another key highlight of this year is the appointment of Ruth Hernandez as Chief Commercial Officer. With a strong engineering background, Ruth is a veteran in the semiconductor industry, where she has held several executive positions in Texas Instruments, Maxim Integrated, and GlobalFoundries. During her career, she has driven exponential revenue growth in all different segments, working in several countries across the globe. After only four months with us, she's already a very strategic part of the executive team, contributing to take Soitec to the next growth chapter.
Also, you heard yesterday about a new addition to our management team, and I'm very glad to introduce you to Albin Jacquemont, who was appointed yesterday. Albin brings over 30 years of international experience in financial leadership, strategic planning, and corporate governance. His career spans listed and private equity-backed industrial and technology companies, including Inetum, SOH, Altron Technologies, Darty, and Carrefour. Throughout his tenure in this organization, he has led major financial transformation and delivered significant value through operational performance improvement, cash flow optimization, and M&A execution. In his role, Albin Jacquemont will be responsible for all finance-related matters at group level.
He will play a pivotal role in reinforcing Soitec's financial and operational foundations and supporting the company's next phase of sustainable growth and value creation. He succeeds Léa Alzingre, who will step down to pursue new professional opportunities, having supported Soitec's growth over the past six years. I would like to warmly thank Léa for her strong commitment and valuable contributions to Soitec's development during her tenure. Looking ahead, given the ongoing uncertain environment, we are withdrawing any guidance given previously and will guide revenue on a quarterly basis. Q1 2026, we reflect the phase-out of Imager- SOI and ongoing inventory corrections, with revenue down approximately 20% year on year.
This number reflects a further significant correction in our FD-SOI inventories, a continued weak automotive market, and a strong edge and cloud AI dynamic, partly offset by the phase-out of Imager- SOI, which contributed $25 million in Q1 fiscal year 2025. We are adjusting our investment pace. CapEx will decrease from EUR 230 million in fiscal year 2025 to around EUR 150 million in fiscal year 2026, reflecting our agility and disciplined capital deployment. Our fiscal year 2026 profitability model is designed to preserve margin resilience by reinforcing cost control while continuing to invest in R&D with discipline to secure future growth. The macro environment continues to evolve from geopolitical tensions and tariffs to ethics, volatility, and climate-related risks. We operate in a world of growing complexity. That is why we remain absolutely focused on what we can control.
We are strengthening our foundation and accelerating diversification, continuing to invest in R&D, deploying capacity with agility across both SOI and compound platforms, driving diversification across product, customers, geographies, and suppliers, building strategic partnerships with key players in the ecosystem. Just as important, we are investing in our culture, our tools, and our processes, all of this while maintaining cost under control as we navigate a tougher environment. Let's talk about product diversification, one of the clearest signs of our momentum. Just three years ago, only our FD-SOI generated revenue well above $100 million. Today, Power- SOI, FD-SOI, and POI are now around or well above $100 million revenue annually. We expect Photonics to join them this year or next as adoption continues to accelerate. These products are now becoming established industry standards.
We are focused on the next wave of innovation: Smart SiC gaining traction beyond EV into renewables and data centers, LNOI emerging as a key enabler for optical transceivers in AI infrastructures and crucial to expand the range of our POI solution for higher-frequency SOI filters, GaN and SmartGaN targeting both RF and automotive segments, and next-generation Imager SOI expanding into a broader range of advanced sensing applications. We are not stopping here, of course. We are developing new materials to extend SOI roadmaps within our existing market and markets that we do not address yet, and new materials beyond SOI, new layer transfer technologies to strengthen our position across markets. There is not a market that we cannot serve today or tomorrow with our technologies. We are also accelerating diversification on other strategic fronts.
On the customer side, we are expanding our base and reducing dependency on our top five customers, a sign of broader market traction. Geographically, we are building a truly global footprint. Take China, for instance. Our direct exposure to China has more than tripled in the last three years. On the supply side, we are actively diversifying our supply base, securing long-term growth and strengthening our ability to protect our margin. While the short term remains challenging and uncertain, our fundamentals are strong, and we continue to strengthen them as we accelerate diversification. As end markets normalize, we are getting ready to fully deploy our operating model. Our addressable market is expected to grow from around 5 million wafers in 2024 to 12 million wafers by 2030. This supports our potential two-times revenue opportunity.
This timing will depend partly on what we control: R&D, execution, and agile capacity deployment, and partly on external factors like macro trends and customer adoption. When we get there, our cost structure is designed for operating leverage. With tight OpEx control, improved EBIT margin, and disciplined working capital, we are positioned to significantly enhance cash generation over time. To conclude, fiscal year 2025 was a year of transition, but also a year of progress. Headwinds will persist into fiscal year 2026, but we are getting ready to emerge from this cycle more diversified, more resilient, and more agile. We have continued to invest in our innovation, our people, and our long-term impact. We are now ready to deploy our operating model with strong upside across revenue margin and cash generation. Today, we are building Soitec for the decade ahead. Thank you very much for your attention.
It is now my pleasure to hand over to Steve Babureck, our EVP Strategy. Steve.
Thank you very much, Pierre. Hello, everyone in the room and on the webcast. Let's take a moment now to walk through our strategic outlook. First, we are seeing powerful megatrends reshaping the global economy. AI, mobile connectivity, and electrification are all fueling semiconductor demand. These are long-cycle shifts, and they continue to drive strong, sustainable growth for Soitec. Second, our differentiation is clear. Engineered substrates are not just materials; they are enablers. They unlock features that make chips smarter, more connected, and dramatically more energy efficient. This is the core of our value proposition. Third, our addressable market is scaling fast. We expect it to grow at around 15% CAGR through 2030, rising from around 5 million wafers today to 12 million wafers by 2030. That is a massive opportunity for us.
Yes, as Pierre said, we are managing short-term headwinds, notably excess SOI inventories, but we also see upside from our innovation incubators and emerging markets. Put together, these three drivers give us a clear, compelling path for long-term growth, and we are executing with speed and precision to capture it. From the PC era to mobile, and now, of course, AI, semiconductors have powered every major technology shift. Today, we are laying the groundwork for the AI era. In 2025, global semiconductor sales are expected to hit around $710 billion, and the industry is on track to reach $1 trillion sales by 2030, with AI, of course, driving most of that growth. Now, looking at the history of semiconductors, it has not all been smooth sailing. Geopolitical tensions, macro uncertainty, and inventory corrections have tested the industry again and again.
Each time, we have seen the same thing: a pretty strong rebound. Because beneath the noise, the long-term fundamentals of the industry remain rock solid for semiconductors. On one side, we have the analog chips, the semiconductor chips that capture and communicate real-world data. By 2025, global data creation is expected to hit 180 zettabytes and more than double to 400 zettabytes by 2028. This growth is being fueled by AI, IoT, and a massive wave of connected devices. To put this in perspective, one zettabyte could store the entire contents of every book ever written, not just once, but over 100 million times. This is the volume that we're building for. On the other hand, you have digital chips, the chips that process and make sense of all that data.
As the data volumes explode, the ability to process it fast and intelligently becomes just as critical. We're not just following Moore's Law anymore. We're going beyond it. Since 2010, the computing power required to train an AI model has been doubling every five to six months. That is not incremental. This is exponential. This acceleration is powering real-time AI, autonomous systems, and the next wave of digital experiences where data becomes insight instantly. To put these enormous numbers in perspective again, the human brain is an astonishing machine, but it would take a single brain over 30 million years to match the amount of thinking that went into training GPT-4o. Data is exploding. Compute is accelerating exponentially. Together, we're setting the stage for a major shift. This is what is driving the AI revolution from smart assistants and robotaxis to education, healthcare, and beyond.
Of course, this is only the beginning. AI solutions and services are projected to generate nearly $22 trillion in cumulative economic value by 2030. This creates a powerful self-reinforcing cycle, putting semiconductors at the core of the next wave of global economic growth. What is the flip side? AI and electric vehicles are driving global electricity demand to new heights, faster than anything we have seen before. This is creating real pressure on energy infrastructure, and this is why we need scalable, energy-efficient technologies that can meet this demand without compromising sustainability or reliability. What does that mean for us, for Soitec? Today, Soitec engineered substrates matter more than ever. Let's zoom out for a second and look at the global wafer demand.
Our semiconductor industry consumes on an annual basis around 300 million wafers a year, and the vast majority of these wafers are made on bulk silicon. Only a small share of that volume today falls within our addressable market, which is focused on engineered substrate. What's the difference? A bulk wafer is only the starting point. An engineered substrate is a bulk wafer that has been enhanced with additional layers that add physical functionality, improving performance, power efficiency, both at the chip, at the semiconductor level, but also at the system level, at the smartphone, automotive, edge, and cloud AI. By design, engineered substrates deliver value where it matters most, at the device and at the system level.
Engineered substrates unlock better performance and greater energy efficiency across a wide range of applications, whether it is RF chips delivering faster connectivity and longer battery life in your smartphones, optical interconnects powering the backbone of AI data centers, or maybe even quantum computing in the near future. In short, engineered substrates are what makes these breakthroughs in semiconductors possible. Given the role, the critical role that engineered substrates play in enabling high-performance, energy-efficient semiconductors, this is no surprise that the demand for our product is accelerating across our three end markets. In mobile communications, supported by the recovery of the smartphone market, the ongoing adoption of 5G and Wi-Fi 6, 6E, 7 technologies, and the emergence of new use cases such as satellite communications or fixed wireless access.
In automotive and industrial applications also, growth is being driven by the adoption of electric vehicles, the increasing penetration of ADAS systems, and also new industrial applications for data centers or renewable energy systems. Finally, in edge and cloud AI, momentum continues thanks to sustained investment in the hyperscale AI infrastructure, as well as the expansion of the edge AI ecosystem. Taken all together, we estimate Soitec's addressable market to grow at a robust 15% CAGR through the end of this decade. In terms of products, we can split our addressable market into two main categories: SOI first, which remains the core of our portfolio. This market is expected to grow at more than 10% per year. Second, compound materials: silicon carbide, gallium nitride, POI, LNOI, now for silicon photonics.
This is scaling even faster, and the growth in this segment is projected at around 30% per year until the end of this decade. Overall, we anticipate 15% CAGR for our addressable market, and this is why diversification across materials is such a strategic advantage for Soitec. Now, in the short term, clearly, there's still some uncertainty, mainly due to the persisting excess SOI inventories at some of our customers. Even as Pierre said in RF-SOI, it's clearly going in the right direction. This could continue to weigh on near-term demand for SOI before we see a clearer rebound. That said, the fundamentals are intact, and we are focused on navigating this phase with discipline. On the other hand, looking further ahead, supported by our innovation, we do see strong potential upside beyond the 12 million wafers projected for 2030.
This would come in part from our incubators, which are exploring new products in RF, power, image sensing, and even advanced computing. These programs are still early and at different stages of maturity, but they reflect our long-term ambition and commitment to keep expanding the scope of engineered substrates and to capture new growth opportunities over time. In summary, global megatrends, from AI to energy efficiency, are driving a major shift in semiconductor demand. Our engineered substrates are central to that major shift, and we are accelerating our diversification to meet that demand. Engineered substrates enable smarter, more connected, and more energy-efficient semiconductors, delivering value right at the heart of the device and at the system level. As a result, our addressable market is expected to grow at 15% CAGR, with potential upside from our new technology incubators.
This is a perfect transition to introduce our next speaker, Christophe Maleville, our Chief Technology Officer. Thank you.
Thank you, Steve, and hello everyone in the room, in Paris, and in the webcast. If strategy sets the direction, technology is the engine that gets us there. I am very pleased today to walk you through how we are advancing innovation at Soitec and why it is key to everything we do. At Soitec, our innovation model is scalable, focused, and built for impact. We continue to invest steadily in R&D, not just to maintain our competitiveness, but to accelerate diversification and drive growth. This model has proven its ability to deliver efficiency and speed, taking innovation from lab to fab fast. Innovation is not just about what we are building today. It is about what we are unlocking next. It is about value.
Our unique and versatile material expertise enables our customers to deliver devices with superior performance and energy efficiency. It all starts at the substrate level. We innovate at the substrate level and create value where it matters most, directly at the chip and device level. This is how we differentiate. Finally, we have launched dedicated incubators to push the frontier even further on SOI and beyond. Incubators help us explore potentially disruptive projects, assess different revenue streams while focusing on high differentiation and optimum resource allocation. Customer engagement is central, ensuring alignment with market needs and accelerating the path to impact. We do not innovate in silos. We are constantly working to push the boundaries of what our technology can do. We do that with our own capabilities and also by staying fully connected to a rich and global innovation ecosystem.
We partner with leading players around the world: foundries, fabless, OEM, research institutes to amplify our innovation power and stay ahead of the curve. We're deeply involved in driving global innovation roadmaps, ensuring we do not just follow technology trends. We help shape them. We keep our eyes wide open, constantly scanning the technology landscape to watch for changes and spot emerging trends early and turn them into real product opportunities. If a disruption happens, we want to make sure we are not reacting to it and we are leading it. When we talk about semiconductor innovation, the industry typically follows three main paths. First, there is Moore's Law, and I will not spend time on that one. You all know the story. Second, more than more, where innovation comes from new materials that unlock new functionalities: GCRF, power, photonics. Sorry. And more.
This is where Soitec is fully active and where we continue to lead and differentiate. Third, 3D integration, combining different technologies in a tighter footprint, more aggressive form factors to deliver higher performance and energy efficiency. Here, we are developing a unique and promising platform, transistor layer transfer, or TLT. TLT enables 3D heterogeneous integration of any substrates. Wow. Combining more functionalities into a single package and doing it with higher throughput, lower material usage, and greater design flexibility. What do we bring to the table specifically? We bring three things: high performance, high-quality data, and energy efficiency. On performance, we enable more value per square millimeter and a lower cost of a chip, with products like POI and Photonics- SOI delivering real system-level gains. On data, our substrate drives better connectivity and faster transfer speeds, from RF-SOI in 5G to Photonics- SOI in data centers.
On efficiency, we help reduce power and greenhouse gas emissions with up to 70% lower CO2 footprint on SmartSiC and lower energy use across the board. It is this combination, performance plus efficiency, that defines our impact. Let's now take a first concrete example on how we push device performance, POI for high-frequency filters. We are now developing next-generation POI substrates to get to higher frequency, higher performance. These higher performance filters are for advanced RF front ends. At the heart of this advancement is our Connect POI stack. This unique structure combines a piezoelectric layer with trap-rich silicon engineered for high performance. What does that mean? Sorry, some technical problem. What that means, first, new materials are being introduced to support wider bandwidths and higher frequencies. We are also optimizing the top layer thickness.
We're also optimizing the top layer thickness to extend performance beyond 30 Hz while ensuring strong manufacturing yield and multiplexing capability. The result is a better integration of filters for both transmit and receive paths, a critical requirement for the next generation of 5G and 6G devices. Among the new materials, we are introducing a lithium niobate, which brings more performance in the POI family. We are leveraging this platform as thin-film lithium niobate for high-frequency modulators in silicon photonics. Talking about silicon photonics, let's look at Photonics- SOI, a key enabler for high-performance optical transceiver in all data centers. Our substrate is designed to minimize optical losses and maximize signal quality. What makes the difference? Precise control of top silicon thickness, improving modulator performance, cross-talk, and yield. Optimal surface roughness, reducing optical losses at the interface and boosting photodiode responsivity.
Mechanical stability to ensure robustness in high-volume wafer handling and load effectivity, which is critical for through silicon via driver integration and chip-to-fiber attachment. All of this enables the performance needed for next-generation pluggable transceivers and co-package optics, which lower latency, faster data rates, and better energy efficiency. This brings me to our anything and anything matrix. This is our toolbox, a toolbox that enables us to deliver these next-generation products. It shows you our strengths, our ability to combine different materials to create differentiating substrates. All of these wafers are a combination we have demonstrated, combining the best active layer with the range of functional substrates we have used so far. Behind each of these wafers, there are multiple years of R&D, numerous patents, and secret recipes.
With all of these assets in the bank, we can innovate faster, we can innovate better, and bring unique added value engineered material to market. Let's now talk about how we stay creative, focused, and enable upside. That's exactly what our incubator strategy is designed for. We've launched four incubators on SmartGaN, image sensor, advanced SOI, and next-generation FD-SOI. Each one is a focused internal venture program built to explore high potential opportunities at low cost and high speed. The model is simple. We start small, invest progressively, and track progress constantly. We work hand in hand with customers and ecosystem partners from day one, and we focus only on segments where we see true differentiation potential and strong return on investment. It's about expanding our product portfolio with agility, with cautious investment, and sustained execution excellence.
To wrap up, at Soitec, we've built a scalable and efficient innovation model, one that delivers today and prepares us for what is next. Our technology roadmap is driven by deep material expertise, unlocking performance and efficiency right at the device level. With our incubators, we are expanding into new frontiers with focus, agility, and a clear eye on where we can truly differentiate. Innovation is not just part of what we do. It's the foundation on how we grow. Execution is what turns it into results. To show how we are scaling our operations with agility, efficiency, and sustainability, I'll now hand it over to Cyril Menon, our Chief Operating Officer. Thank you very much.
Thank you, Christophe. Good morning, good afternoon, and good evening to all of you. Let me now take you through how we are managing our industrial operation at Soitec, especially in this uncertain and rapidly changing environment. We are navigating short-term volatility with discipline while building the foundation for long-term. Our operational focus is centered on three pillars. First, deploying a global industrial model that scales efficiently without overbuilding. Second, keeping this model agile and capital-efficient, enabling us to adapt to demand fluctuations. Third, continuous sustainability in every aspect of how we build and run our fabs. Let me start with how we drive our operation at Soitec. Our model is built to scale efficiently and adapt, even in volatile environments. We focus on three levers. First, maximize operating leverage. In the current uncertain environment, we have strong levers to protect our margins. Fab allocation, including SIMGUI and flexible workforce deployment.
Bulk optimization to align our sourcing with market trends and in-sourcing steps like SOI refresh to lower cost, reduce emission, and absorb fixed costs ahead and earlier. Second, minimize cash consumption. We scrutinize every cost on the opposite side, from materials to logistics. We also make sure that we delay fixed costs like building or hiring as much as possible, finding creative ways to optimize our current assets. Third, reallocate assets and cut idle capacity. Speed matters. We use common tools across fabs, move teams where needed, and shift production lines to meet demand, all without new infrastructure. This flexible, lean model keeps us responsive, protects margin, and preserves cash no matter the global context. What this means for our industrial operation footprint is our ability to keep delivering in line with a dynamic customer demand using the same fab year after year. That is by design.
We build for resilience and adaptability, not just for scale. In downturn, we protect without destroying, preserving talent, infrastructure, and readiness to rebuild fast, especially in mobile and auto. In expansion, we are ready to ramp quickly and efficiently. That is why our footprint is scalable, robust, and agile, able to absorb both slowdowns and surge with minimal disruption. Tech POI demand is strong, and we are scaling without building a new fab. Leverage existing capacity at Bernin 3. At the same time, the SOI market is shifting to 300 mm where we are more competitive. If 200 mm SOI demand softens, we can relocate that capacity to POI where growth continues. This ability to switch allocation lets us respond in real time and boost output with a limited investment, a core principle of our capital discipline. We also drive efficiency through synergies.
Tool commonality across compounds means fewer assets and shared maintenance, and internalizing SOI refresh lowers cost, reduces emissions, and helps absorb fixed costs earlier. These are the levers: flexibility, reallocation, and shared infrastructure that make our model resilient, efficient, and built for profitable growth. In terms of capacity, we are deploying our operating model to grow from around 3.4 million wafers per year to around 4.3 million wafers per year. The biggest driver of that increase is POI, where we now see a potential to reach 1 million wafers annually. What is important is that we can achieve this by raising part of our Bernin 1 fab, which drastically reduces the C apEx required. In parallel, we continue to invest in 300 mm SOI to support growing demand and stay ready for new opportunities. We are advancing Smart SiC deployment steadily in line with the customer program.
In short, we're scaling, but we're doing it with discipline and by making the most of what we already have. This approach is also reflected in our CapEx strategy. To deploy our operating model and materialize to roughly double the size of the company, we anticipate a total investment of around EUR 770 million. Importantly, that's not for building new fabs. It's about making smarter use of what is already in place. For 300mm RF-SOI, we're investing mainly in tools, notably for new products and also for the completion of the Singapore extension. For POI, we are scaling beyond the initial plan by reusing capacity, especially in Bernin, which drastically reduces the investment needed. Smart SiC continues to progress with investment aligned to confirm customer demand. Beyond those, we're also funding enablers like automation, sustainability, and IT, which help us scale efficiently and future-proof our operations.
Overall, we are investing, but we are doing so selectively, step by step, and with a strong capital discipline. In this unstable macro environment, resilience matters more than ever. We are focused on what we can control, and we've made meaningful progress in three key areas. One, yield improvement. We've implemented tighter process control across fab, integrated fab teams with engineering, and continued our global yield management strategy. The result? Year-on-year yield improvement, which directly translates into margin protection. Second, automation. Robotics deployment continues across our production line, improving direct productivity and consistency. Our goal is not just to automate for the sake of it, but to support scale with fewer fixed costs and lower variability. Three, sourcing. We've secured long-term strategy agreements with key suppliers. We've also diversified, especially on critical materials, to reduce concentration risk and improve negotiation leverage.
This operating backbone is what allows us to stay competitive through the cycle. Now let's talk sustainability, a topic that is increasingly central to our operational strategy. Scope one and two emissions are down 37% compared to 2020, and we reach our 2026 SBTi target two years ahead. Water reuse. We're now reusing 44% of our water, and we're on track for 50% by fiscal year 2030. Energy. In France, we operate on 100% low carbon energy, and in Singapore, we are now above 50% and still rising. Transport. We shift to sea freight as our default mode, reducing internal logistic emissions by 75% over the past two years. These efforts are not just about compliance. They are about operational resilience, cost saving, and strengthening our partnership with customers, who increasingly factor ESG in their sourcing decisions. Let me close with three messages.
First, our operating model is scalable, efficient, and ready to support the next phase of Soitec's growth without compromising capital discipline. Second, we are building agility in our DNA, reallocating capacity, aligning with demand, and driving productivity with automation and shared assets. Third, we are executing a best-in-class sustainability roadmap, reducing our environmental impact while enhancing competitiveness. Thank you very much. I will now hand it over to Steve for the last part related to finance.
Thank you very much, Cyril. As you understand, during this CFO transition, I have the privilege to present the finance part, warming up for Albin for the next time. Key messages for this finance section. Fiscal year 2025 has obviously tested our ability to manage through uncertainty, both in terms of visibility and performance. Before diving into the numbers, three takeaways.
First, we have accelerated the diversification of our business model across products, geographies, and customers to better absorb short-term market volatility and prepare the group for recovery. Second, we have strengthened the flexibility and agility of our operating framework. This is what enabled us to preserve profitability and manage cash effectively in a volatile environment. Third message, we are now positioned to capture the full potential of our operating model. This means significant revenue upside, improved operating leverage, and stronger free cash flow generation. With that in mind, let's take a closer look at the fiscal year 2025 results. Let's start with the highlights. We deliver EUR 891 million in revenues, down 9% year on year at constant forex and scope, in line with our revised guidance back in February this year. Despite this revenue decline, we maintain a resilient EBITDA margin of 33.5%.
This performance reflects strong cost discipline and our ability to continue investing in innovation. Also, importantly, we generated positive free cash flow of EUR 26 million. This was driven by a reduction in CapEx and better working capital management, two areas where our agility made a difference this year. Pierre already commented on the revenue performance, so let's move directly to our gross margin performance. Our gross margin in fiscal year 2025 came in at 32.1%, down 2 points versus last year. This reflects two main headwinds: higher depreciation expenses, as expected with recent capacity investment, and lower volumes, particularly in mobile and automotive. That said, our model once again demonstrated its resilience. We were able to partly offset these pressures thanks to tight cost control, strong industrial execution as outlined by Cyril, and to some extent, support from public grants such as IPCEI.
Let's continue on the P&L. We delivered EUR 136 million of current operating income, or 15.2% of revenues, a six-point decrease compared to the previous year. This performance reflects three drivers: first, a drop in gross margin, as previously discussed; second, a 39% increase in net R&D expenses. We continued to invest in talent, strengthen our innovation partnerships, and reduced R&D capitalization. Third, SG&A expenses were up 4% year on year. While we kept a strong handle on cost discipline, we were again a difficult comparison with EUR 4 million of non-recurring tailwinds last year. Despite all this, our EBITDA margin held steady at 33.5%, just half a percentage point below fiscal year 2024. A few additional remarks on the P&L. Other operating expenses totaled EUR 16 million, mostly related to the disposal of Dolphin Design, a non-core business for us going forward.
Net financial result came in at EUR 9 million lost, reflecting interest from recent financings and some foreign exchange losses. Our effective tax rate increased from 11.3% to 17.4%, mainly due to non-recurring items this year. All in, we delivered a net income of EUR 92 million, representing 10.3% of revenue. Now, let's turn to cash flow. Operating cash flow reached EUR 202 million, up from EUR 165 million last year. Working capital was more moderate than last year. The EUR 79 million change is mainly due to higher inventories driven by a shift in customer demand and two customers putting deliveries on hold, higher receivables reflecting a less favorable customer mix and specific commercial agreements.
Regarding CapEx, which totaled EUR 230 million before leases and interest, in line with the guidance, they included additional POI capacity in Bernin, expanded RF-SOI and Photonics- SOI production in Singapore, refresh investment for 300 mm SOI in Bernin, completion of the Smart SiC 200 mm pilot line, and ongoing investment in IT and sustainability. Thanks to tight cash discipline, we delivered positive free cash flow of EUR 26 million. Looking at the working capital, which ended the year at EUR 488 million, up EUR 95 million compared to last year. This increase in working capital reflects two main factors. First, receivables remain high, with a stable DSO at around 113 days. This was largely due to a strong Q4 revenue contribution and diverse payment terms across our customer base.
Second, inventories remain elevated as we navigated an adjustment phase in demand mix, a consequence of two customers requesting to put some deliveries on hold. Turning to net debt, we closed fiscal year 2025 with a net debt of EUR 94 million, up EUR 55 million from last year. This net debt increase was primarily due to EUR 65 million in new lease liabilities, mainly related to lease-back arrangements for the Bernin fab building and tools related to POI and Smart SiC. Overall, this increase in net debt is moderate, and we continue to operate with a strong and flexible capital structure, fully aligned with our investment strategy and long-term goals. To conclude on fiscal year 2025 numbers, our balance sheet remains healthy, with strong equity and moderate net debt. Now let's talk about the outlook.
As Pierre mentioned, we have moved to a quarterly guidance approach to reflect our uncertain environment and ongoing volatility. Pierre already covered the outlook for Q1, so I'll focus on the rest. First, on CapEx. We are moderating our investment in fiscal year 2026, with CapEx expected at around EUR 150 million, down from EUR 230 million in fiscal year 2025. As Cyril highlighted, we're operating with very strict capital discipline, and we prioritize flexibility and maximize asset utilization. Regarding financing, as you know, we have EUR 325 million OCEANE convertible bond, with a maturity in October 2025. We're planning a partial refinancing of this convertible bond, with about 2/3 through non-dilutive instruments, and the rest covered with available cash. Regarding profitability, we will continue to manage the cost tightly, and we will maintain a high level of R&D to support our long-term differentiation.
Now, to give you a sense of our operating leverage potential, in fiscal year 2025, if you look at our cost of goods sold structure, it was made up of roughly 70%-75% variable cost and around 25%-30% fixed cost. This setup enables margin expansion as volumes recover. Regarding the revenue potential, you remember the slide that Pierre described earlier. We are deploying our operating model and see significant potential to double our revenues. The timing of this subside will depend on a mix of a few factors, some within our control, like how we execute R&D, deploy capacity, and move with speed. Some of these parameters are outside of our control: macro trends, end market cycles, customer adoption, etc. When we reach that next level of revenues, there is significant upside on profitability.
Our current EBIT margin, which today stands at around 15%, has the potential to move toward around 25% thanks to three key drivers. First, operating leverage. Remember our cost of goods sold breakdown. It gives us strong leverage in growth cycles. That is going to be a big part of our EBIT margin improvement. Second, as we scale, we will see better absorption of SG&A and R&D with a larger revenue base. This is a structural driver of our margin expansion. Third, we will continue our cost discipline, a cornerstone of our model, even in a growth mode. At the same time, we will remain committed to sustain investment in R&D. Just to note, this model assumes a EUR 1.10 to $1.00 exchange rate.
Let's now turn to cash generation and the potential we unlock as we scale our model. As we deploy our operating model, we see clear upside in free cash flow driven by three core levers. First, profitability improvement, as we just talked about. Second, working capital efficiency. We are targeting a reduction from 55% of sales in fiscal year 2025 to around 40% over time. We will reduce working capital with a better inventory and receivables management as we return to a more normalized production run rate. Third, lower capital expenditure intensity. We expect our annual CapEx-to-sales ratio to decline from 22% in fiscal year 2025 to roughly 15% as we better leverage our existing industrial footprint. Together, these levers create a clear path to stronger free cash flow and support a step-up in our post-tax return on capital employed from around 7% today to around 20% through the deployment of our operating model.
To conclude, one, we accelerate the diversification of our model to mitigate current market volatility. By doing so, we are preparing the group for our upcoming recovery. We have a flexible and agile model, which should help us secure margin in a quite uncertain environment. Finally, our operating model enables 2x revenue upside with strong operating leverage. Thank you. I will now hand it back to Alex to open the Q&A session.
Thank you, Steve. Before we start the Q&A, I just would like to remind you that you can find our presentation on the homepage of our website. For those of you who are joining us online, you have the ability to ask questions by clicking on the orange button to the right of the player. We've got on stage Pierre, Cyril, Christophe, and Steve, but the rest of the management team is in the room as well, so please ask your questions. We will start with the questions in the room.
Hi, this is Alex from Bernstein. I have three questions for me. The first one is, you pointed out in your presentation that three years ago, only RF was well above $100 million, and now you have three businesses in this category, soon four. Yet your revenue base is basically the same as it was three years ago. Obviously, you know, great execution on what's growing, but the decline of RF is stronger. I'd like to understand, you know, at what point you will see this erosion in RF stop and stabilize, and then we can enjoy growth. That will be my first one. I'm going to have a follow-up then.
Thank you, Alex. Maybe I will ask Jean-Marc, of course, to complement my answer. We have started three years ago with a very, very high level of inventories, very high. We moved to 18% the year after, then 14%. We expect to have another, let's say, a depletion of a few months additional for the coming year. Of course, this level of inventories is based, the calculation is based on a consumption rate that remained modest. If the consumption rate is increasing by 15%, that is possible. You gain two months in the calculation. For the moment, we prefer to be cautious because you understand that looking at the volatility and lack of visibility, we are cautious and we prefer to calculate on a modest consumption rate.
We believe that we are reaching now a kind of plateau that is going to normalize the revenue of RF we are going to generate, not perhaps this year, but the year after. The drivers of this market remain for us because after this correction is done, the content increase, even if it is a few percent, is going to continue. The new applications to come, we were talking about new products, 9SW, 3DIS, 3DIC with UMC, Wi-Fi 7 extensions and so on, are going to continue. We have, even if it is modest, a growth in volume by a few percent per year. Everything cumulated will make RF-SOI pick up again. We need this correction now to go to a plateau, and we are now very close to it. Perhaps Jean-Marc, if you can add some elements.
Yeah, just to complement, thank you for the question. He's right. In the past year, we did correct the inventory by about three months. It is right that the market today still sees some growth. I mean, the front-end module growth for calendar 2025 is expected to be 5%, including the smartphone sales at around 2%-3%. Which means that there is extra content, extra content coming right from the 5G penetration, which is still not at 100%. As Pierre mentioned, we are winning share also in the Wi-Fi, Wi-Fi 7, with the integration of the PA. We see also the premiumization of the smartphones. The share of the premium phones is increasing. On the drawback, on the headwind, we see there is more and more shrink.
The fact that the fabless, the foundry are designing today in 300 mm helps them to do more shrinks to address advanced nodes and also to optimize the footprint within the modules, which was not the case in the past. All in all, I mean, we see a content growth of about 3.5% in the coming years. I mean, before the 6G penetration, which is a different story. Again, I think it was well mentioned in the slides, we have to face these inventory corrections. We believe we still have to correct about three months this year to bring the level below 12 months overall. Then we will be close to what we call a new normal because, I mean, again, before COVID, the normal was more in the range of six to nine months.
Today, we see a just-in-case scenario is still in place, especially we've seen some restocking in the latest FQ4 because probably of the tariff threat. Yeah, I mean, we still have to correct a little bit, I mean, before seeing the real demand of the market.
Okay. Yeah. Thanks. Thank you very much for this. Now, this brings me to my second question. Could you quantify the revenue so that they could realize if you didn't have all these moves in inventory? You had a bit of destocking, you had a bit of restocking towards the end of the year, apparently as well. It would be helpful to understand where your revenue base is if your customers were just ordering what they consume in their production. Just a final one for Cyril. What was the average capacity utilization?
Is my assumption that it's about 50%-55% correct? Or is there a mistake there? Also, on the passive risk expansion, is that now shelved? I didn't see that in the slides. Obviously, the market doesn't really support that right now. If you could give us a perspective on that. Thanks.
I propose to start with capacity loading. For sure, this is not as straight as giving you a standard figure for all the capacity. This is really heterogeneous in between our different plans. When we look for sure at POI, POI was 100% full last year. We'll be 100% full this year as well. Here the demand is strong.
We have to, for sure, the 200 mm loading is much lower than 300 mm. That's the reason why we revise our industrial strategy, transforming, I mean, cutting some capacity in Bernin to transform this capacity into POI and reuse assets, which by the way are fully depreciated since it has been invested two decades ago or almost. Basically, we're getting 300 mm. I think that this one is a bit tricky as you have, we have seen in fiscal year 2025, as it's fiscal year 2024, we face up a high seasonality in between H1 and H2, especially. In H2, 300 mm loading was 100%. We will see in H1 it was half of it almost, which gives you a bit of an idea what was the average loading in between in 300 mm for the fiscal year 2025. Back to your question regarding Pasir Ris.
Obviously, in Pasir Ris, just to recap for everyone, we have already one plant with 1 million capacity. The shell is fully equipped, will reach 1 million. Today, we have deployed 80% of the capacity, meaning that we have a capacity in Pasir Ris deployed at 800,000 a year. We do not use yet the total industrial footprint in Pasir Ris 1. That is the reason why even if the shell of Pasir Ris expansion is ready, we will not equip today and in the coming quarters a clean room to not generate any fixed cost in Pasir Ris extension. Waiting for, I mean, some tangible information traction from our current core business or potential upside to deploy a clean room and equip and obviously generate fixed cost only when we are 100% sure that we will use this capacity.
Yeah. Just maybe regarding the first question, we are guiding on the next quarter, so we do not want to play too much what-if scenarios. If you remember, we posted a presentation at Mobile World Congress this year where we showed you the RF-SOI revenues against the smartphone market in volumes. Pre-COVID, which was the year ending in 2020, RF-SOI was in the range of $450 million. You can see that it jumped pretty high despite and against the smartphone market lower performance. If you extrapolate from that $450 million and plug some 10%-ish every year until now, you see roughly where we could be in RF-SOI if we did not have the inventories disruption in the supply chain.
Emmanuel Matot from ODDO BHF. First, have you seen a further deterioration of some of your markets recently due in particular to the trade war? Do you see more customers asking for orders to be put on hold? Second, do you think Soitec could capture some of Wolfspeed's business for silicon carbide wafer manufacturing? Because you know this company is facing major financial risk and still doing $400 million of in revenue in silicon carbide wafers. Last question, just to clarify the euro/dollar exposure, because you mentioned $0.05 evolution as an impact of 150 basis points on your EBITDA margin. It was 100 a few years ago, so it has increased. I'm quite surprised knowing that the Singapore platform has also increased at Soitec. Do you know why this exposure is even higher than a few years ago? Thank you.
Okay. I'm going to give the last question, Emmanuel, to Steve. For the two first questions, then we don't see deterioration. The term is very strong. Due to volatility and uncertainties, we see customers taking time before making decisions. We see customers thinking yearly, not multi-yearly. We see customers, in some cases, asking for delivery in advance in order to avoid some tariffs impact. At the end of the day, we see a bit more of nervosity regarding the overall geopolitics environment. No deteriorations in, first of all, the segment shares we have today, on the contrary. Secondly, in the dynamic of penetrations we have in the different markets where we diversified so rapidly. Regarding, and one of it is, of course, Smart SiC, even if, of course, we are still in the adoption phase and we see through the Wolfspeed difficulties that the market is taking time and is a kind of new, is looking for a new shuffle.
The SiC market is now, the big Paris center is in China, where the rising stars in mono- SiC are settled. We see that some key players are now working with mono- SiC suppliers in China. Wolfspeed is still, of course, let's say, one of the non-Chinese suppliers. The difficulties of Wolfspeed, but of course, I don't wish any bad things to this company, but in case of weaknesses, persistent weaknesses, it could be an opportunity for us to be the kind of non-Chinese independent SiC suppliers, with on top of it the ability to enhance the product because Smart SiC is bringing additional advantages towards mono- SiC. It is recognized by the six under-qualification customers we have today, plus the 35 under evaluation. This could be an opportunity, but of course, we don't wish any bad things to Wolfspeed. That's something we have to contemplate. Steve, for the third question.
First of all, the cost base in euro has continued to expand. We spoke about the expansion of POI, Smart SiC in Bernin. Remember that what we mentioned for this fiscal year is that 75% of the net exposure is already hedged. In the previous years, we had a higher net hedging exposure, which could limit maybe the impact of the forex change on the EBIT margin.
Hello, Sébastien from Kepler Cheuvreux. One question for Christophe. On silicon photonic, you have already a very strong market share on Photonics- SOI, almost 100% market share. We now see the market shifting to LNOI and ultimately InP. Do you see any competitors coming to those markets? What could be your market share with this silicon photonic market beyond Photonics- SOI? Attached to that, what kind of market opportunity do you see by the end of the decade? We know the photonic integrated circuit market is about to grow 40%-50%. What kind of growth do you see for your own Photonics- SOI business? That would be my first question. Thank you.
Okay. Indeed, we are quite successful with our silicon photonics platform. The important point is that the others you are mentioning with the lithium niobate for the high-frequency modulators and the indium phosphide for the light generation are others to this market. The silicon photonic space remains, and the lithium niobate modulator is coming on top, and the indium phosphide is coming on top as well. This is going to further improve our position into this market. Okay.
The second point, as you understand from the discussion today, is that we are developing a global lithium niobate platform that will serve the filters and the silicon photonics, which is super interesting for us in terms of the synergies. As far as the growth of the silicon photonics is huge in front of us due to the data centers needs, the AI, and so on. How much are we going to get? More. You'll see that we will be growing with the silicon photonics platform. As I said, again, we'll add new material on top as early as possible. That is why we are really accelerating on lithium niobate. I believe this will bring our position even better. I think it's difficult today to clearly identify where we're going to be in terms of this business, but the pluggable transceiver is going to grow. On top of that, the co-package optics is going to accelerate that.
Okay. Maybe only one question. If you want to, yes.
Maybe I can. You covered most of it. Maybe one element that he did not really mention, right now, if you look at silicon photonics, the application is really in the data center. This is where it is happening. Christophe already commented on what is the trajectory when it comes to modulation speeds and efficiency, etc. We have new materials coming in. Obviously, if you think about silicon photonics, it can also be used in different kinds of applications. I do believe that moving forward, we see new opportunities for the same technology. Right now, and we have, of course, the market share you talked about, I am looking into other kinds of applications.
Think of healthcare, think of biosensing, think of lighter applications, think of quantum computing. I mean, the technology as such is not limited to the data center. I believe that further down the road, there are a lot more applications that we can serve with our silicon photonics solution.
Okay. Thank you. One question for Emmanuel behind. On Smart SiC, you announced a sixth customer in qualification for your technology. Can you comment a little bit on the qualification process? We know you are progressing step by step, but it has been a long process since you started to qualify the first one that was ST. Can you update a little bit on the progress you have made on the qualification process? Thank you.
Okay. You have seen the silicon carbide market evolving because at the beginning, it was the early development of the silicon carbide just after Tesla. There were more shortcuts in the qualification, especially for the auto. Now, the business is back on track on the normal auto qualification. Roughly, evaluation, it's between 12-18 months. Qualification could be up to three years. It's a long one from the automotives. The good thing is that when you are in, you are in for a long time. As of today, what we notice is that the qualifications are doing okay for both 150 and 200. We are focusing more on the 200. What we also notice, Pierre underlined it, is that the demand of the customers is a little bit different also for the data center. With the question of the increase of the need of energy into the data centers, there's more and more centralized now data centers starting in DC, 800 volt, going in DC, 400 volt.
After that, you touch a bay. For us, that is a sweet spot for silicon carbide first. We are also thinking about GaN that could play a role. That is the kind of thing we are looking for.
Hey, guys. Hi. It is Daniel Schaffer from CITI. Thank you for taking my question. I just wanted to come back to RF-SOI inventories, given that you mentioned your goal is to go to 11 months roughly by the end of 2025 calendar year. If I understand correctly, the first quarter, we will see the, again, kind of big hit. Given that you are targeting 11 months, so three months down, plus the effects and potential tariff headwinds, can we then expect the second quarter to be more muted in comparison to last year where we saw a bigger spike?
We will only talk about the first quarter. On Q1, clearly, as a kind of classical pattern, we have been, so it took for a long period of time on RF. Our customers are getting big orders on their Q1 calendar, that is our Q4, in order to be sure that they are going to be on time to deliver the big two periods for smartphone sales that are back to school and Christmas times and so on. This is always the same cycle. Clearly, we still have some inventories in excess. The fact we got big orders and deliveries in Q4 is making Q1 quite weak. Whatever we forecast, a Q1 quite stable for mobile because POI is offsetting a part of some softness on RF-SOI, including a bit of also FD-SOI activities. This is really the frame we see for this Q1.
Okay. Thank you. The second just follow-up in terms of operating model. Before, you had a medium-term target of EUR 2 billion. You kind of now scrapped it. You now have an operating model. In your presentation, most of the time, you mentioned 2030. Can we imagine that this operating model is something to achieve by 2030? What conditions have to happen in order for this to be achievable?
No date, but clearly, as soon as we have wins in the sales again, and we have many sales now, that's good news, and many healths, it is clear that today we have the structure, the people, the partners provided EUR 770 million additional CapEx investment. We are in a trajectory able to double our revenue. Again, in this world of volatility and uncertainties, the strengths and the coming of the wins in our sales propel the multi-health Soitec that we have today.
Perfect. Thank you.
Oliver Wong, BofA. Thank you, guys, for the presentation today. A couple of questions. First is, could you talk a bit about your customer inventories for not RF-SOI, but maybe Power- SOI, FD-SOI? What's the situation there? My second question is, you mentioned the EUR 770 million long-term CapEx investment plan. 30% of that will be for POI. 5%-10% of that will be for Smart SiC. Is that roughly correlated to what you expect for the long-term revenues for these product lines? Thank you.
I will let Cyril take the second question. On your first question, is the breakdown of customers per other SOIs and RF products. That was your question? Okay. Customers' inventories. Okay. Overall, inventories are fortunately lower than for RF in other areas. We are more in normality. Everything depends also on the demand. If we take automotive, inventories are in normality, but the demand today is very low. The automotive industry is suffering from low demand because the market today is very weak and very shy. That is the point. For the rest, if we look at the different pictures we have today, we are in a normal mode. In some areas, like photonics, there are quasi very small inventories because the demand in that case is very high.
Regarding CapEx and capacity expansion, I mean, I guess that you noticed that in our operating model, we consider that we have to increase our capacity from 3.4 million wafer to 4.3 million wafer. This is the capacity install. It doesn't mean that today we have 100% fab loading. That's for sure. This will be necessary to extend capacity and load our fab to reach the operating model at EUR 2 billion. Out of it, this plus 900, basically 900,000 wafer, 25% of it is for Smart SiC. Back to your question. You have the indication how much this + 25% capacity increase means in terms of CapEx.
Thanks. Jakob Bluestone from BNP Paribas Exan. Just to get back to the inventories, I mean, you mentioned a few times that you expect to get to 11 months of inventory. Can you maybe just help us understand what gives you the confidence that's the right number that we're going to end up in? Why don't we go back to the 6-9? Is this what your customers are telling you? Does that take into consideration maybe a slightly weaker economic environment? Just if you can maybe explain your confidence around that that is genuinely the plateau for inventories.
Thank you for the question. I will add a few things to my previous comment. I mean, pre-COVID, the pattern was 6-9 months. I mean, as you see, during COVID, they were short in terms of, let's say, inventory. We had to put most of the customers in a location mode. Today, that's why we see the confidence level of our customers to increase this level to something which is close to 10-12 months as the security, I mean, for them, especially with the current uncertainties today, geographic, I mean, in terms of geopolitics. That's why, I mean, today, I believe it will be the new normal will be 10, 12, definitely for RF-SOI. Thank you.
If I can ask a question just around sort of margins and, I guess, things you can do to offset revenue pressure. I appreciate you're not guiding for this year's revenues, but you obviously have a fair bit of operating leverage within your business model. I guess the question is just, what can you do to offset some of that? I mean, in last year, in FY 2025, your revenues were down 9%, and your operating margins ended up down 6 percentage points. If we saw a similar revenue decline, do you think margins could fall similarly? If you can maybe just help us understand what are some of the cost-cutting measures that you can take to offset some of that operating leverage, which obviously works against you in a declining revenue environment.
Sure. I mean, definitely, this operating leverage is great in an environment where we grow. Unfortunately, last year, we were not in that situation. Anything that we can play on the P&L is at stake. Basically, we can start from the first item, one of our major costs, which is our supply of bulk. I think that definitely we have reorganized totally our supply chain, diversification, and leverage that situation. Two things that we emphasize on that one, yield improvement, quality improvement, which drive customer satisfaction. This is important even more in such an environment. At the same time, it does improve our yield and does improve our bottom line.
This is really important as well. Three, in terms of structure, being able to manage the fab allocation. Two examples, which is pretty simple. One, the fact that we have managed allocation in between SIMGUI, our partner, 200 mm, and our fab in France, where competitiveness obviously is better thanks to synergy on that side. Plus a second item, which is pretty important. You notice that we have a new fab, Bernin 4. Fortunately, Pasir Ris is idle, I mean, the extension, the project. In Bernin 4, we have a new fab with new fixed cost. We have been able to offset that by having a strategy, an agile strategy, implementing our refresh, which is basically something which was subcontracting to some supplier. We internalized that one, which is in a way to absorb all the fixed cost from this new factory.
Obviously, you see the benefit, especially in ZBID day.
Thanks. Alex, a follow-up, and then maybe we take the online questions.
Thank you. Thank you. I just have a follow-up. You've mentioned shrink a few times, and I'm just looking for reasons why. We've seen the smartphone market basically stabilizing for the better part of the past two years, yet your revenue is still going in reverse quite strongly. I'm looking for explanations here. Is shrink an enemy here? On that note, when Apple brings the modem in-house away from Qualcomm, what's that going to do to the SOI content? Thanks.
Okay. I will start with the question. It's right that the foundry moving from 18 to 12 inch brings them to advanced nodes. This is obviously an opportunity for them to benefit from the tighter node and to shrink the die. We see, I mean, there is a trend also to add more features on the same die so we have more compact die. This is also a demand from the smartphone makers to optimize the modules. A few years ago, I mean, the modules were not optimized in terms of footprint. Today, there are more and more modules. The smartphone's size and thickness is still the same or even shrinking when you look at ultra-thin smartphones. Overall, this is the headwind in the model. I mean, yeah, the shrink is impacting the overall growth. Again, it is compensated by other factors. Other factors include what we mentioned, I mean, in smartphone, but also diversification to adjacent markets of front-end modules like automotive, like IoT, like fixed wireless access, like infrastructure, etc., where there are front-end modules which are also integrated.
On the second question, today, I mean, Apple is reporting or Qualcomm is reporting that they are going to keep 70% of market share in the modem of the next iPhone generation, at least for calendar 2025. It might be less in calendar 2026, etc. This could have some consequence on our side that we have some FD-SOI today which is utilized for envelope tracker. On the other side, we have strong, let's say, expectation that Qualcomm will be able to bring their next generation of millimeter wave chipset with FD-SOI. Today, which is integrated already in the Galaxy S24 and S25. Also, I mean, the Apple modem designed by Apple will bring more content to the front-end module makers, the legacy ones like Skyworks, Qorvo, where today we are penetrating also with new products like the POI. We do not believe that we are going to lose anything, but we should win, I mean, in terms of content overall by this new Apple strategy.
Okay. We have a few—sorry, we have a few questions online. Maybe Steve will start with your questions on Imager- SOI. Can we quantify the revenue in fiscal 2025 for Imager- SOI and the breakdown, especially in Q1, Q2, and the rest of the year?
Okay. Regarding Imager- SOI in the edge and cloud AI division, the revenues in fiscal 2025 were in the range of 4% of the total revenues. They were in large part in H1 2025, so very front-end loaded. This is why the impact of the phase-out of this product line completely in fiscal 2026 will mostly impact our H1 revenues. As we discussed yesterday, it impacted for $25 million in the first quarter ending in June.
Thank you, Steve. The other item, I'll package two questions that would be for Cyril. Inventories and working capital. Working capital on sales at around 55% today. You mentioned a target to go towards 40%. Can you help us to get a sense of how to get there? The second one is on inventories. Inventories on the balance sheet went higher. What is the right level of stock that we should expect in the future, and is this sustainable?
Definitely, this year, we are definitely hit by what we communicated beginning of February with two customers putting on all deliveries. Basically, these deliveries were already either in inventory in our WIP or at least in bulk. That is a direct impact in our inventory and then on our working cap. We expect this to be a one-off situation and to improve over the years.
Yes, Damien Antani, DNCA. Could you elaborate a bit on Smart SiC with the collapse of the silicon carbide price? The bulk, I understand you bring more qualities and properties to the product, but what impact in the end? Because one of the advantages is to cut in more than 10. Work on that. We were talking about three years qualification. I think end of the year, you would be at three years with ST. I know you do not give guidance anymore, but when should we anticipate the start of some real deliveries for Smart SiC?
First question is about monocrystal. It is true that the prices are really going down, especially from China. The Chinese government allowed 6,060 licenses for 150, and each region wanted to have their own fab. The result of that is that in 150 mm, the price really went down much faster than what we thought at the beginning. The 200 mm is more reasonable in terms of, first, the barrier of entry. It is more complex to make it, and there is much less license in China that has been given, less than 10. Probably the price will remain a little bit higher. What does it tell us today? First, with this level of price for silicon carbide, that has opened a new domain that was not touchable by the silicon carbide because of the price. Now, having this price much closer to the IGBTs, to the silicon, probably there are much more options for silicon carbide to be touched, especially where there is more power density needed.
For us, because we buy monosilicon carbide and we buy polysilicon carbide and we do the combination of both, we are working very hard to reduce the cost of the poly at the same or not exactly the same trend, but in the trend similar to the monosilicon carbide. For the monosilicon carbide, we are benefiting from the low price. We are working with the production to reduce our production costs. We are very good at challenging our production costs when the volume will be in. We know how to leverage that. Now we have delayed the ramp, so we cannot have right now the price we want. In the future, if we follow the volume we are expecting, I am pretty confident we will remain very competitive in terms of what we have with our solution of Smart SiC.
On top of that, we are creating the value we are creating through this solution. That is still being to the market. Regarding the qualification, it is a long process. Started with the GDA. After that, we have EVAL. Now we are in a long process. We have now six qualifications ongoing. Some of them, all of them, they are in different regions, which is great. Pierre underlined it. We are also qualifying in China. Just telling you how we are probably seen as a real value creation leader in terms of silicon carbide because we are able to donate in China. It will take time. I cannot give the end of the story because that is a long story. In parallel to that, silicon carbide at the first was for the car industry.
You know the car industry today is really a little bit shaky for plenty of reasons. We are patient. We are controlling what we are doing. We are sharp on our engineering. We are pretty confident that we will see the ramp of silicon carbide very soon.
Thank you, Emmanuel. To make sure we finish on time, I suggest we leave the couple of last minutes to Pierre for his closing remarks. If you have any follow-up questions, the IR team is here to answer them. I remind you again that the presentation is available on our website. Thank you very much.
Okay. Thank you. Thank you, everyone, for your attention, attendance, and also your questions. Thank you, of course, for the whole management team, including the newcomers, for being there. Of course, we are living the longest crisis ever in the semiconductor industry. Okay, that's a fact. Longer than ever. In this difficult time with probably volatility, uncertainties, and so on, as I mentioned and as it has been mentioned by Steve, Cyril, Christophe. Soitec is getting stronger with a stronger mindset. Stronger mindset. I hope that during the lunchtime, you had occasion to discuss with the management team, also the key leaders who are here, and of course, right after these meetings. A very strong mindset because at the end of the day, we passed two crises, and we are passing a third one. The first crisis was a governance crisis. It has been the biggest clash on the market, Parisian place, the title of Les Echos. We passed this crisis.
I can tell you that the relationship we have today with the board members is very, very strong, very good, with a lot of interaction, a lot of synergies. We passed two days on Monday and Tuesday with all boards and all the management meetings, exchanging a lot of information as a team. We passed a second crisis at the same time. That was a structure crisis. Soitec, in 2022, was depending on one customer, more or less, one product, one supplier with huge stock, huge inventories on this product. The market was getting down, collapsing. In three years' time, we managed and the team, I'm very proud of what the team did everywhere. In three years' time, we managed to shift 50% of our revenue from one product to three other products, making Soitec more stable and stronger and others coming and joining the $100 million, let's say, entity.
We are passing now the third crisis, business market crisis, with volatility and so on. We will pass it. We will pass it. No doubt. Every morning when I wake up somewhere on the planet and it is in a different place, what makes me confident, resilient, super motivated? I think about five things. First of all, our financials. We are keeping control. We know how to keep control. We have demonstrated two years in a row by generating a good level of profitability, improving our cash, managing in difficult times the different metrics of the company with a balance sheet that is very solid. I am thinking about the fact that we are stabilizing, if not increasing, our segment shares, even in some areas where we are already very strong. We have a very, very inventive R&D.
Yeah, we are spending money. No, we are not spending money in R&D. We are investing money in R&D for our future. Christophe did not say anything. We are working on many, many interesting elements, key elements for our future. Our plants and factories are extremely agile and flexible because shifting 50% of our activities after more or less eight years of monoproduct pattern in such a short period of time is demonstrating a lot of flexibility and agility by Cyril and the team. The fourth element is our customers. We have excellent relationships with our customers. I met personally over the last three weeks with more than 11 of our key CEOs of our customers because of meetings, exposition, booths, and also calls and whatever, review. I can tell you that today, our relations with our customers are in a solidarity mode.
Yes, we look at quarter as everyone. We are working also on the foundation for the futures. We are developing common products more and more. The level of quality we are delivering to them is praised and is a reference more and more. We are entering into the top three of some of our main customers for each of the products we are delivering to them, delivering more and more product because the diversification is now in our DNA and is everywhere. The fifth thing that makes my morning happy to lead this company is my people. What has been made, what has been achieved in this so difficult environment is really amazing. I am very proud for my team, my management team, and all the guys who are leading this company with me for this achievement.
It makes me confident because when the winds will come back into the sails on this, again, multi-earth boats we built as a Soitec group, I can tell you we'll have to fasten our seatbelt. Thank you very much for your time and your listening, and see you very soon.