Okay. Let's get started. So good morning, everyone. My name is Jérôme Ramel. I'm EVP Corporate Development and Integrated Communication of STMicroelectronics. So welcome, and thank you very much for joining us today in person at our 2024 Capital Markets Day in Paris. Thank you also to those joining us on our webcast. Attending the event today will be Jean-Marc Chéry, ST President and CEO. We also have all the members of the Executive Committee of ST presenting today as well. Before we begin, please note that we have issued earlier this morning a press release outlining our 2027-2028 financial model and path toward our 2030 ambition. Jean-Marc and the team will reference this trajectory during today's presentation, as you would expect. So let's move to the agenda. So here you have the agenda.
We're going to start with Jean-Marc and his opening remarks, followed by our product group's heads. To start with, Marco Cassis for APMS, where he's going to talk about smart mobility, power and energy, and intelligent things. Then Rémi El-Ouazzane for MDRF, where he's going to talk about what we do at the edge, in space, and in the cloud. They will both discuss their products, their IPs, technologies, and business opportunities. We'll then have a short break at around 10:45 A.M. for 20 minutes. During the break, for those in Paris, I invite you to go and see the videos that are running in the main room. I'm sure you will enjoy it. Then after the break, we'll cover two more topics, starting with technology and manufacturing with Fabio Gualandris, followed by Lorenzo Grandi, our CFO. We will present our financial models and targets.
Finally, Jean-Marc and the team will host a general Q&A session. We'll take questions from the room and from those connected on the webcast. Our webcast will then stop after the Q&A session. For the ones that are in Paris, there will be a networking lunch served in the main room where you went through, which would be an opportunity to meet with our management team that would be available for questions. This live website featuring the webcast presentation is accessible on ST's dedicated CMD website. The address is cmd.st.com. All presentations and materials can be downloaded and are available right now. As usual, a recording of the event will be available shortly after the end of the event. As usual, the presentation includes forward-looking statements that involve risk factors that could cause ST results to differ materially from management expectations and plans.
We encourage you to review the Safe Harbor statement presented there and also STMicroelectronics' recent regulatory filing for a full description of this risk. And now it's my pleasure to invite Jean-Marc Chéry, our President and Chief Executive Officer, to take the stage. Jean-Marc, the floor is yours.
Good morning to everyone, and thank you for attending ST 2024 Capital Markets Day. My thanks as well to Nicolas Dufourcq, our Chairman of the ST Supervisory Board, and thank you to Maurizio Tamagnini, our Vice Chairman of the ST Supervisory Board. I know that some of you today cannot be with us, and I thank them for joining our webcast. I am happy and determined to present this event with the current ST Executive Committee that I will introduce since I changed significantly the organization since last time we met in 2022. My colleagues will raise their hand. Marco Cassis, President in charge of analog power and discrete sensors. Rémi El-Ouazzane in charge of microcontroller digital ICs and RF products. Fabio Gualandris, in charge of manufacturing quality and technology R&D.
Lorenzo, President and CFO, recently joining as a member of the Managing Board. We have also in the room with us Rajita D'Souza, President, Human Resources and Corporate Social Responsibility. Jérôme Roux, President, Sales and Marketing. And the last but not least, Steven Rose, President, Legal Counsel. Since the current Managing Board and Executive Committee took their role, it has been committed to developing ST in order to reach a size and a scale that allow us to continue to compete with the leading global peers and leaders, as well as to protect the manufacturing and design of semiconductors as a strategic asset of Europe. This commitment has been supported since the beginning by a strategy which remains unchanged, based on a number of fundamentals that I would like to recall.
So first, a value proposal to stakeholders focused on sustainable and profitable growth, providing differentiating enablers to customers, and a strong commitment to corporate social responsibility and sustainability. Second, our global integrated device manufacturer operational model with a wide range of capabilities in proprietary technologies and innovative IP design, product and solution development, and with advanced manufacturing, both in-house but with a selected number of partners, this providing us with a strong supply chain, a strong competitive advantage, and a strong resilience for our customers. Third, three long-term enablers driving our strategy, so the smart mobility, power and energy, and last but not the least, cloud-connected autonomous things. These guide our market positioning on automotive, industrial, premium personal electronics, and computer peripheral and communication equipment.
These drive our customer engagements, the commitment of continuous and open innovation, product and IP design development, important technology R&D efforts, and of course, manufacturing investments, and this supports our conviction that with our customers and with our partners, we are a key actor of the transformation of all industries we are enabling toward a smarter, a safer, and a more sustainable future. Over the period 2018 to 2023, enriching the $17 billion revenue bar, ST has outperformed its peer in terms of revenue growth. We have invested more than most of them to sustain this growth and to capture the future opportunities, while reaching at the same moment a positive net financial position as opposed to most of our peers. We recognize that despite reaching 27% operating margin in 2023, this was structurally lower than our peers, but largely due to higher incidence on COGS.
However, our market performance and total shareholder returns have been broadly in line with our peers. Over the past three quarters of 2024, as the market demand significantly deviated from our expectations, our results have suffered more than our peers. Volume declines have not allowed us to fully deploy the capacity we have created, resulting in important underutilization costs. The revenue for the full year is now expected at about $13.3 billion and an operating margin of about 12%. I have to say that in such extreme market cycles, our pockets of weaknesses and opportunities for improvement are not offset significantly enough by our strategic strengths. We already acknowledged that the fact that the market we serve is becoming increasingly complex, with frequent crises and highly competitive.
This is driven by several factors and a lot of complex factors, including one, which is a strategic inflection in the automotive transformation path toward electrification and digitalization, and the evolving competitive landscape, such as the rise of Chinese carmakers, which is reshaping really the market dynamics. Another cause, even sometimes presenting advantages, another cause is government incentives spread out everywhere worldwide and trade restrictions. These have introduced distortion in the semiconductor industry landscape and led to unusual supply chain behavior, such as double booking, over-inventory buildup, and significant over-capacity investment. We have plans in place to adapt to this unprecedented situation, leveraging the acceleration of our manufacturing strategy and our innovative technology and product portfolio roadmap. You will soon, during the presentation, hear more from my colleagues, Marco, Rémi, and Fabio.
This is why we believe ST will come through this market cycle that I expect and we expect at 5% compound average growth rate, the market we sell. We expect that we will exit with ST with a stronger position, and we will reach about $18 billion of revenue and an operating margin from 22% to 24% in the horizon of 2027-2028. This prepares us to achieve our ambition of $20 billion plus and above 30% operating margin by the end of this decade. Following a severe 2024 revenue decline for ST, mainly driven by the industrial market weakness but severe correction, we will enter in 2025 with a Q1 below usual seasonality. We anticipate 2025 a transition year, both in terms of revenues and operating margin, still impacted by short-term visibility mainly on the industrial market and significant unused capacity charges.
We see some pockets of weaknesses and over-inventory also in automotive, amplified by significantly lower capacity reservation fees from the carmakers. However, the automotive demand weakness overall should be somehow mitigated by unexpected growth in the second half of the year by the industrial market and by our engaged customer program for premium personal electronics and communication equipment and computer peripherals where we have the appropriate visibility. Beyond 2025, so we anticipate market demand growth acceleration in both 2026 and 2027. Competing in the market we serve, which I repeat again, are expected to grow at 5% compound average growth rate by 2027-2028. We expect to grow roughly twice as fast as our SAM. This will be thanks to three major elements. First, application-specific analog products and sensors with, I repeat, our already awarded engaged customer programs in premium personal electronics and computer peripherals and communication equipment.
Second, our power and discrete products with a significant contribution of our silicon carbide power MOSFET and diodes driven by the automotive power train electrification, but as well with industrial power conversion application. Third, our General Purpose Microcontroller, for which we expect to come back to 23%-24% market share entering in 2027 once the current inventory correction in industrial and in mass market will be over. This growth will be driven and sustained and amplified by the continuous introduction of new products and features, including edge AI in our already market-leading broad-range STM32 portfolio. On the margin front, with the execution of our manufacturing reshaping program, cost-based resizing initiative, we would exit 2027 with a high triple-digit million-dollar savings compared to our current cost base. This would enable us to reach an operating margin between 22% and 24% at about $18 billion revenues in the horizon of 2027-2028.
Importantly, moving forward, we believe that our reshaped worldwide manufacturing and balanced footprint infrastructure spread within our current 14 manufacturing sites, with an acceleration of our capacity in 300 millimeters for silicon and 200 millimeters for silicon carbide. And on top of that, having set up a China for China value chain, this will continuously improve the competitiveness of our product portfolio and technology run and roadmap, and should result in our capability to deliver an operating margin above 30% by 2030. Today, we are reconfirming our $20 billion plus ambition that we expect now to be reached by 2030.
Our strong focus on a broad-range player and on high-growth applications such as in automotive, again, the electrification power train, the ADAS, the legacy system, but within the SDV architecture, in industrial with energy management at large from supply transportation and management at the point of use, all the battery-operated tools, so electrification everywhere, automation and robotics, whatever is process automation or discrete automation, on premium personal electronics, on data center power management and connectivity, and mainly the AI data center, and last but not the least, on low Earth orbit satellite communication, but again, this will be widely shared with you during the presentation of my colleague. All will be the key drivers of our revenue growth path toward the $20 billion plus ambition with the operating margin above 30%.
Embedded in our plan, we are committed to sustainability for all our stakeholders, so ESG, environment, social, and governance. Here, we are investing to further accelerate our effort by creating technology for a sustainable world in a sustainable way. Our commitment here goes back more than a quarter of a century, and we have the goal, I already shared with you many times, to achieve carbon neutrality on what we control, the Scope 1, the Scope 2, and partially the Scope 3 by 2027. So to conclude my opening remark, first, we are setting an intermediate financial model with revenues expected at $18 billion and an operating margin at 22%-24% in the horizon of 2027 to 2028. Our $20 billion plus revenue ambition and 30% operating margin remain intact, and we now expect to reach it by 2030.
To achieve this goal, we expect to grow about twice as fast as our SAM. We are reshaping our manufacturing, I repeat, within our existing 14 sites and resizing our cost structure to optimize our integrated device manufacturing model. During the next three years, we are anticipating a reduced capital expenditure intensity, and we will continue to rely on a unique manufacturing technology and product portfolio roadmap driven by innovation that gives ST a strong competitive advantage. Thank you, and I will now hand over to Marco Cassis.
Good morning to everyone. The one in the room and the one connected is really a pleasure to be here. Last time we met at ST Capital Markets Day in 2022, I was proud to be wearing a new hat and talking to you for the first time about our great ST analog MEMS and sensors technology and products.
Today, and since a few months, I'm even more proud to also represent the power products of ST and to discuss about what we do in APMS. I will start with a short introduction to APMS, our financials, market positions, and products. Then I will cover the three key trends that are driving growth for us, highlighting for each of them the most important applications and product areas. So let's begin. APMS is a group of around 4,800 engaged people delivering about 60% of ST's revenues in the first nine months of 2024. In APMS, our product families are grouped in two reportable segments: power and discrete products and analog MEMS and sensors. Our revenues were flat for the past two years with comparable nine-month earnings at around $7 billion, with different patterns for our two reportable segments due to different dynamics in the market that we serve.
On the first nine months of 2024, we reached approximately $6 billion revenues, out of which $3.6 billion in analog MEMS and sensors and about $2.4 billion in power and discrete. Overall, we are on a 10%-15% lower revenues base compared to previous years, impacted by continued difficult market conditions, especially in industrial. In APMS, thanks to our power, analog, and sensor portfolio, we strongly contribute to ST's goal of being a key actor of the transformation of all industries toward a smarter, safer, and more sustainable future. Our strategy is focused on the following portfolio: power and discrete from low to high voltage in silicon or in wide-band gap materials. In analog, we address both application-specific products, mostly driven by large customers, and general-purpose analog products for the broader market. In MEMS and sensor, we focus on MEMS sensors and actuators and on optical sensing solutions.
With more than 15,000 products that leverage the ST integrated device manufacturing model, we address a total market that is well above $100 billion, and we serve more than 100,000 customers. In terms of markets, automotive represented more than 50% of our revenues for the first nine months of 2024, and industrial was around 18%. Altogether, automotive and industrial, where we have a strategic broad approach, are about 70% of our revenues, while personal electronics and communication equipment, computer peripherals, where we have a more selective approach, are about 30% of our revenues. As you can see here at a quick glance, we have a very large product portfolio. Let me now go into more detail in each of our three product families. Power.
In APMS power products, we address all the technology and product categories from silicon to wide-bandgap materials, from diodes to IGBT and MOSFET, from low to high voltages, and from discrete packages to complex power modules. We leverage a large and diversified manufacturing footprint that we are expanding, especially in wide-bandgap with our new Catania Silicon Carbide Campus and the joint venture with Sanan. Our complete product portfolio and structured go-to-market approach has brought us leading market positions, such as in automotive, especially in the fast-growing electrification area, and we are overall among the top three players. Let's continue with analog, where we have a wide range of proprietary analog technologies, including many BCD flavors to support simple or complex smart power ICs tailored for specific applications and a large portfolio of general-purpose products.
We are recognized leaders in automotive and industrial application-specific products, and thanks to our strong system know-how, we develop a range of solutions that have solid market positions. Our strengths in power management allow us to hold a solid number one position on analyst estimates and to continuously win major designs in personal electronics and data storage applications. Last but not least, APMS sensors. In our digital world, sensors are increasingly present and important in a range of electronic devices, applications, and use cases, and they are an endless source of data for artificial intelligence, especially AI at the edge. We are serving the accelerations of sensor pervasion, and we are proud to bring unique competencies in mastering both MEMS and optical sensors. Our very specialized process technologies are the foundation for strongly differentiated products with the capabilities to offer either standard products or custom solutions.
In MEMS, we have a leading portfolio of sensors embedding more and more intelligence. In imaging, we are a recognized leader in time-of-flight and in specialized CMOS image sensors, and we also offer ambient light sensors and advanced optics. I would like now to focus on three key trends that APMS is benefiting from. As automotive is a major part of our business, I will start with smart mobility. Power and energy represent another key trend, and I will show you how our products contribute to greener generation of energy, smarter transmission, and more efficient utilization. Finally, we'll talk about intelligent things. Of course, here it's mainly about personal electronics, but also laptops, accessories, AR, VR, and so on. Let's start by looking at the automotive market. As indicated on the right side of the chart, if the chart comes, okay.
As indicated on the right side of this chart, the growth in the automotive market will continue to be driven by semiconductor content increase. On one end, in blue on the graph, even if the number of internal combustion engines cars will decrease, the semiconductor content keeps increasing at a moderate pace. On the other end, in yellow on the graph, with the number of electric vehicles steadily growing, either full electric or hybrid, we expect the semiconductor content to double by 2030 compared to what it is this year. This is a huge opportunity for ST, and we address this market starting from multiple very solid positions. So why is smart mobility key to our future growth?
As vehicles become more intelligent and autonomous, increasing safety-related norms and rules require redundancy for the most safety-relevant applications, such as ADAS or even braking here, where already today OEMs are duplicating the hardware. And e-mobility is driving silicon content per car, not only in the fully battery electric vehicle clearly driven by power, but also in hybrid cars, where the content of the drivetrain system is more than double that of a traditional combustion engine. On top of these two macro trends, the vehicle electronics architecture is now transitioning from a distributed ECU architecture to the domain zonal architectures, which clearly require more computing power and analog-wise smarter power management capabilities. On the safety trend, let me give you two specific examples. First, the evolution of braking systems represents opportunities for continued value in this legacy, but still transforming application.
We have historical market leadership in braking system application-specific analog smart power devices, and we lead in the electric parking brake market. Braking systems will transition from a pure hydraulic approach to electro-hydraulic and then to full brake-by-wire, which will ultimately double the silicon value. Leveraging our historical intimacy with top-tier suppliers, we continue to invest and innovate with new products to maintain leadership here. Safety is not only about smart power, but also about sensors. This market is growing, and we are playing on two dimensions. On one side, vehicle dynamics is a consolidated and large market where the migration from standalone accelerometers to advanced inertial motion units is a major opportunity for us. The second area, vehicle interior monitoring, is growing fast thanks to regulation, and here we have already a strong position and a solid pipeline of design wins.
We continue to introduce new differentiated products that are already winning on the market and with good market potentials. We aim to grow twice as fast as the market here. So safety is a major driver of innovation in smart mobility. The second driver is linked to the evolution of the car architecture. Let me give you a couple of examples. A first example is the move from today's distributed and domain-oriented architectures towards a zone-oriented electrical electronic architecture that uses only a few very powerful vehicle computers instead of many individual control units. One key domain is the door zone, where we lead with a 30% share of door zone integrated ICs. We took the strategic step to invest very early in the next generation, designing solutions in tight cooperations with key players to continue and to ensure our leadership.
Zonal architecture will also increase the market addressable for power E IC dedicated to new powerful automotive MCUs that you will hear more about in the MDRF presentation from Rémi. A second example of the impact of car architecture evolution is the market opportunity for e-fuses. With the increasing complexity of electrical and electronic systems, traditional mechanical fuses are inadequate for managing the power needs and safety requirements. The transitions to e-fuse technology are inevitable, and we are one of the early movers on this market. Our ST-i2Fuse products are already in production, leveraging our strengths in BCD and vertical intelligent power technologies, and our goal here is to achieve around a 40% share of the e-fuse market by 2027.
To finish with car architecture evolution, in the longer term, the automotive industry is expected to grow through another transition, which will open a new market opportunity: the transition from 12 to 48 volts. The wiring in modern vehicles is becoming increasingly complex, costly, and a source of power losses. So increasing the bus system voltage will allow the transmission of the same amount of power with lower current, allowing for thinner and lighter wiring. This will reduce the overall weight of the vehicle, improving the distance range and reducing cost. At ST, we already have several products available to support this new trend, and we are actively developing our roadmap to meet the needs of our most demanding customers. Really a great opportunity here. Let's now move to engine electrification. Hybrid cars are growing from around 12 million units this year to 20 million units per year in 2030.
As shown on the right side of the chart, the transition from the traditional internal combustion engine management system, where we have a high market share, to hybrid powertrain is doubling the smart power analog opportunities. Starting from our established leadership in powertrain, we are addressing this trend and have the right partnership to continue to lead with either existing or new products that we are developing with leading players. Now, about full electric vehicles. Here we lead thanks to technological leaderships, innovative solutions, and a strong market positioning, working with all the key actors. The opportunity is big, as you can see, with the growth rate and the addressable bill of materials. The key here is to have a comprehensive portfolio addressing traction inverters, which represent the biggest chunk of the market, as well as onboard chargers and battery management systems.
We are among the few semiconductor players that master all the key ingredients necessary for vehicle electrification, especially the combination of isolated gate driver together with power transistors, and here we are introducing innovative solutions to drive a combination of silicon carbide and IGBT in the same inverter, optimizing system performances versus cost. When we speak about car electrification, silicon carbide is an essential part, so let me spend some time on where we stand, what has changed, and what remains. So what has changed? While the adoption of silicon carbide is still growing, electric vehicles' growth pace, which has been a major driver for silicon carbide growth in the early years, has been reduced. The global battery electric vehicle production forecasts for 2024 have been revised down by approximately 15%.
Specifically, Chinese forecasts are down by around 3%, while forecasts for other regions have been revised down by about 20%. There is now an increased focus on hybrid vehicles, and we see several Chinese producers moving from full electric vehicles to hybrid vehicles. The result is lower semiconductor content, both silicon and silicon carbide. The customer landscape is also evolving. We will stay flattish with lead customers, while with the rest, we will grow. Third point, the industrial market that represents a second wave of silicon carbide-based opportunities is overall slower to recover than our initial expectation. Last, the intensity of the competition has increased. All in all, this leads us to adjust our short-term revenue goals and our midterm ambition to achieve above $2 billion now by 2027-2028. And we are also accelerating the 150 millimeter to 200 millimeter silicon carbide manufacturing transition. Now, what stays the same?
It's very important. First, the SiC market is still on a strong growth trajectory, and we target long-term market share of around 30%. Second, we reconfirm and accelerate our manufacturing strategy, as you will hear from Fabio later. On one side, we continue implementing our fully vertical integrated model from substrates to modules, ensuring supply stability for our customers. On the other hand, we have launched our China for China approach with the joint venture with Sanan that is progressing per plan. Third, we have a clear action plan to diversify our customer base by expanding into industrial markets, including power supply units for AI servers. This diversification is crucial for our sustained growth. Finally, we remain committed to continuous technology innovation, introducing new technologies and enhancing the existing one to maintain our competitive edge in the market.
In conclusion, while we face several changes in market dynamics, our strategic focus remains clear and robust. We are well positioned to navigate these changes and continue our growth path. So we have a clear positioning and strategy for silicon carbide as the market leader, and we are well placed to take advantage of the strong growth forecasted for this market. We have a long experience and already several generations of technology and products introduced in this industry. We have just announced the introduction of our fourth generation of products, and we continue to work on the generations to come, continuously improving power density, reducing conduction and switching losses, and many other parameters. More specifically in China, which is the fastest growing market for electric vehicles, we have a very strong position thanks to our design activity over the past years.
Today, we have more silicon carbide engagements with top Chinese EV car makers than any other supplier. This includes players like BYD, Geely, Great Wall Motors, Li Auto, and XPeng. We also have a strong pipeline with top Chinese industrial companies. Our China for China manufacturing model will further accelerate our growth in this important market. As a takeaway, why and how are we winning in smart mobility? Of course, long experience in automotive and relationship with key actors are crucial. We have numerous leading market positions, either in smart power analog or in discrete power technology. We provide the right manufacturing footprint and capacity, enabling strategic independence to our customers, including a well-defined China for China strategy.
We develop innovative technology and products tailored to support the high-quality standard automotive requirements, and we are among the early movers in deploying innovative solutions compatible with new car architectures, as well as anticipating future trends. Smart mobility is a huge and growing opportunity. We have outperformed the market in 2022 and 2023, and we target to grow twice as fast as the market we serve. Now, let's move to power energy trends. So to be successful in the evolution that we do see in power energy is led by basically three major needs: need for wider use of renewable energies, need for smarter transmission as it ensures that energy is delivered efficiently to where it is needed when it is needed. Finally, need for efficient utilization of energy to make better use of every watt in power-hungry data centers, in automated factories, or in every charging system.
I will now give you a few examples of how these needs represent major opportunities for us. Let's start with greener generations thanks to renewable energies. The increasing adoption of renewable energy sources such as solar and wind power is driving the demand for advanced semiconductor technologies. This is a big market that is growing at a steady pace. With our comprehensive portfolio of energy conversion and power management solutions, we address the various power levels needed, and we intend to grow at twice the market growth rate for this area. Now, moving to smarter energy transmission, the smart mobility trend that we described previously will not happen without a larger deployment of EV charging infrastructure. The combination of semiconductor content increase of more than 50% by 2027 and the increasing number of charging stations make this application very, very interesting.
Again here, power and analog semiconductors for AC to DC power conversion or DC to DC power management are essentials. And I'll let you read the long list of products that we are deploying for this application. We have already a strong position and aim to position ourselves as a leading provider of fast EV DC charging stations, capturing a 20% market share in this rapidly expanding sector. In scope of smarter energy transmission, smart metering is crucial for optimizing energy usage and sustainability worldwide. ST has been leading this market for more than 30 years with around 20% market share in electricity metering. To further enhance the efficiency of the global electricity grid, metering is evolving with wireless communication, for which we are deploying our module solution for Narrowband IoT.
This trend enlarges our addressable market, both with growing numbers of new connected meters, but also increasing the complexity of the solar system on chip and full module. We are already in production, and our module has been certified by leading telecom operators worldwide. Let's now move to efficient energy utilization. Sustainability is crucial when it comes to motion control. According to the International Energy Agency, almost 70% of industrial energy consumption is used by electric motors. More efficient motor control techniques, innovative algorithms, and pervasive use of wide-bandgap materials such as silicon carbide are key to improving energy efficiency. ST has a strong footprint in the motor control market thanks to a deep market and application know-how. Our complete product portfolio is a key asset to address this very fragmented and growing market.
Both bill of materials and volumes are projected to grow, and we target to grow twice the market in the years to come, driven by new innovative product introduction and wider adoption of silicon carbide in industrial servers. Now, where efficient energy utilization is needed, GaN transistors are game changers. The biggest market so far and for years to come is power conversion for fast chargers for personal electronics, and this is where we have focused our efforts for a number of years. The key here is to combine the best driver or controller with efficient power transistors in the same package. We are combining the best of the two worlds, and we are first to the market with our MasterGaN solution. And we are continuously optimizing the most difficult driving part. This enables our customers to build smaller form factors and more power-efficient chargers and adapters.
Of course, down the road, we will move to more integrated monolithic solutions. A last example of efficient utilization is the need to optimize the very high power required by AI servers to handle intensive computational tasks. The AI server market is experiencing very rapid growth, both in terms of units and in terms of semiconductor value. Thanks to our strong power and analog offering from silicon carbide to IGBT and from e-fuses to advanced multi-phase controllers and smart power stages, and our deep expertise in power conversion, we aim for a 10% market share by 2030. We are already engaged with design wins and are investing in localizing strong competence centers close to our customers to realize our ambition. In conclusion of this power and energy section, this is a growing and very fragmented market where our ambition to grow double-digit. What is key to win here is threefold.
First, you need a strong and broad technology roadmap to cover the large need for power and analog devices from low to high voltage, from silicon to wide-band gap. Second, a wide portfolio of products. Third, a solution-based approach. We deliver this through strong competence centers in all the regions, which are instrumental in supporting our go-to-market approach, addressing both key OEMs and distribution. Now, intelligent things. Intelligent things are not only personal electronics, but also all accessories, tools, computers, robots, and other electronic devices that are or will be connected to the cloud, and that embed intelligence to process real-time data and make autonomous decisions. This intelligence closer to the user reduces latency and offers personalized features. Localizing this intelligence at the edge improves energy efficiency and reduces environmental impact as the energy demand gets larger and larger.
All in all, electronic devices becoming more and more intelligent transform user and life experience on many aspects. So what is our approach? First, we are already a recognized leader in powering and enhancing intelligent devices of many types, especially personal electronics, premium personal electronics. Second, our products, we are talking here mainly about analog application-specific and sensor, are leveraging our broad analog, MEMS, and optical sensing proprietary technology portfolio to address selectively high-volume opportunities in the market. Third, we embed more and more intelligence in our products, for example, in advanced power management ICs or in our sensors, MEMS or optical sensor that serves as the eyes, ears, and the other sensory organ for AI. So now, talking about sensor, what is our vision for tomorrow? We see that sensors are becoming smarter and represent the main source of data for AI and are key in human-AI interaction.
They gather raw data and pre-process information that are essential for any system to understand and interact with their surroundings. For developing high-performance, accurate, low-power, and context-aware solutions, it is important to first, own, develop, and manufacture proprietary technologies. Second, master several types of sensors to be able to combine and optimize solutions. Third, embed algorithms, machine learning, and processing or computing capabilities to improve context awareness and offload the rest of the system. And you need to provide the tools to develop so they can effectively exploit these capabilities. We do that in three steps. Step one, we provide intelligent MEMS sensors for higher efficiency in high-volume applications. With already two generations and more than 20 products deployed, ST MEMS sensors with embedded edge AI are growing with applications where the energy and computational efficiency have a key role. But in sensor AI, it's not enough.
We have also deployed a complete ecosystem fully compatible with STM32 Cube. AI environment to support the developers along all the design cycle. We are a leader in this market, and we are investing in a new generation of sensors to deploy tiny edge processing in more and more high-volume applications. Step two, we leverage our CMOS image sensor and digital technologies to create a range of smart optical devices. Our solutions are capable of merging several types of sensing to enable more use cases. For example, in the car, fusion of RGB and infrared sensing increases the market we can address and can also open to additional opportunities outside the car in the future. We can also put together advanced image sensors with image signal processors that embed AI algorithms to further enhance features in intelligent devices.
With more than 2 billion units shipped, our time-of-flight system solution, where we leverage our sensing, processing, optics, and packaging capabilities, generates limitless applications and possibilities. Step three, as the world of sensors is very technology-driven, we invest for the future and continue building competitive advantages in sensors technologies. In MEMS, we improve accuracy, reduce the size of our components, and integrate more sustainable materials in our supply chain. Our THELMA motion sensing and PETRA piezoactuation platforms are market references here. In imaging, it is all about increasing sensitivity, reducing noise, and offering more integrated solutions. But it is also about micro-optics, where we continue innovating, for example, with meta surfaces that we are the first to release to the market.
All in all, we are relentlessly pushing innovation on many fields: silicon process, packaging, pixels, new materials, optics, embedded AI, to enable our future differentiation, opening to more high-volume opportunities. So what does it take to win and why are we winning in intelligent things? We win because we develop innovative technology roadmaps. We win because we own and can scale our capacity in line with customer volumes, expectations, and market windows. We win because our products, the widest industry sensor portfolio, are embedding more and more intelligence, which enables countless new applications and use cases. We win because we have a very strong expertise and are supporting innovation of multiple market shapers with whom we partner. Thanks to these strengths, we have a clear roadmap of engaged customer programs and opportunities, and we target to grow twice as fast as the market.
To conclude, we are investing in proprietary technologies and leveraging our integrated device manufacturing model to continue to address the key trends. Our three product areas are strongly technology-driven, and all our efforts are focused on consolidating where we are already a leader, while fueling innovation in technologies that bring differentiation to benefit from the new trends. We have a clear manufacturing strategy and optimization ongoing that are supporting our product strategy, and we are aiming at further expanding our portfolio to gain share across the markets that we serve. Thank you, and I now hand over to Rémi.
It's good to be back in front of you to give you an update on our microcontroller, digital IC, and RF products group. In the next 40 minutes, I will give you an overview on MDRF strategy, our financials to date, and the trends we intend to benefit from.
I will then give you a detailed update on microcontrollers with a focus on general purpose and automotive MCU. Then, an update on what we are up to in low Earth orbit satellites, and I will wrap things up with a brand new initiative we started more than 18 months ago in the field of cloud optical interconnect. I'm hopeful to keep things interesting throughout the course of the presentation. Let's start with the vision. The MDRF vision remains the same. We believe in a world where things can sense, autonomously infer, and intelligently actuate. The vision relies on three pillars. The first one is edge AI, where we expect AI compute performance at the tiny edge to grow by a factor of 100, 100X in the next coming five years.
The next pillar is security to protect an ever-expanding set of devices, roughly 40 billion of them by 2030, against a growing surface of cyberattacks. And the last is cloud, but not only to backhaul data to and from the cloud, but also to complement compute at the tiny edge in an edge-to-cloud continuum. This is the vision that unified all the activities of ST on digital, RF, and optical. As you've noticed, we are finishing a quite difficult year at MDRF, where we've seen both our industrial SAM decrease by 8% and the general purpose SAM decrease by 31% as per WSTS. Our situation on general purpose MCU was compounded by an inventory build-up over 2022 and 2023 that we are still resorbing. I'm going to come back on this.
The general purpose MCU situation was not compensated by the better automotive MCU market dynamics simply because we are the very early innings of our Stellar family. I will tell you more about that soon. The year was also the witness of a significant inventory correction at one of our large automotive engagement programs. Now, on a brighter spot, we saw more than 50% year-on-year increase in revenue on our LEO business, thanks to our current engagement program in that space. During the course of my presentation, what I'm going to do is explain to you why my faith in our general purpose MCU remained as strong as ever, how we are rebuilding a strong automotive MCU franchise, demonstrate to you the health of our LEO business, and last but not least, explain a new class of data center-centric business that is opening up for the company.
Now, before I get into product specifics, let me talk to you about the four mega trends that we intend to benefit from. Those trends are, to our belief, structuring a new era of growth. First in the trend is AI at the tiny edge. I will show you subsequently why and how we at ST are actually making the AI at the tiny edge market a reality. The forecasts for this market are quite numerous, but assuming a sum of $2.5-$3 billion of AI hardware accelerated MCU in 2030 is, as of today, no more far-fetched. This will represent roughly 10% of the overall MCU sum at that time. Second is the electrified software-defined vehicle. Marco spoke about that a little bit. Successful new energy vehicle must address two challenges. One, to reduce the cost of the electrification, and second, deliver the new software-defined architecture, a.k.a. SDV.
This is, for us, for all intents and purposes, a true inflection point. I will explain how we will help our customers and how we intend to leverage this transition. Third is the low Earth orbit, a.k.a. LEO market. We believe for a long time in a world of satellite constellation, bringing connectivity to homes, businesses, governments, ships, planes, vehicles, and mobiles. We believe this will become a new internet modality and will morph into an essential part of the global communication infrastructure. Last in the trend is a cloud optical interconnect. We are entering the era of the accelerated AI infrastructure. The fundamental part and shift of this accelerated AI infrastructure is one of workload requiring many, many connected XPU, X being GPUs, CPUs, or ASIC.
This is going to lead to a network of hypergrowth, hypergrowth that we believe will be fueled and based on silicon photonics. Okay, so now I'm going to enter the MCU section of my presentation. We last met, excuse me, in May 2022, and since then, many things have changed, and clearly not for the better. Globally, we've been at the intersection of an inflationary economy, growing geopolitical tension, and overall a weaker economy than expected. This translated into a weaker environment for general purpose MCU that got hit hard by inventory build-up over 2022 and 2023 and a sluggish industrial market. As you can see from the graph, the effect was quite drastic since the forecasted general purpose MCU WSTS sum for 2024 got reduced by 37%, or more or less more than $5 billion when comparing the 2022 versus the 2024 projections.
To calibrate you all, the GPMCU sum in 2024 will be in the same zip code in size as what it was in 2020. We essentially went four years backward. The GPMCU sum dynamic, however, was partially compensated by a better-than-expected automotive and secure microcontroller market. At this stage, what we expect our general purpose MCU to do is a slow recovery leading to 2027 sum more or less clipping the 2021 levels. Some will call this WSTS forecast now conservative, and we hope 2025 will start to expose signals demonstrating that we are on a much better trajectory than this one, so it's again this backdrop that we should look at the global MCU market and the assumptions we have taken in our three-year plan.
Our assumptions are based on a conservative 5% CAGR from 2024 to 2027, with industrial being the main driver, simply due to the fact that the volume and revenue drop in the industrial market were quite significant in 2024. The driver for that market, you know them: factory automation, robotics, power and energy to name a few, and industrial markets. And on automotive, and as previously mentioned, it's a combination of electrification, be it full battery EV or plug-in hybrid EVs, together with the migration to an SDV architecture that is creating this inflection point in the automotive MCU market that will, we believe, accelerate dollar content in the car. Specific to MCU, AI acceleration will be a critical aspect of how MCU will evolve over the coming decade, but not only. Growingly, features that were the privilege of MPU will step by step land in the MCU world.
Before we go into specific MCU categories, I wanted to come back on what I think are the key fundamentals to be a winner in the MCU market. There is no silver bullet. It's a string of pearls that are required to be long-term competitive. You can see the success factors on the slide: manufacturing capacity and resilience, controlling your NVM technology, AI hardware and software stack, an approach to developers that has to be unique, your security features, and system-level solutions. How we in ST fit against the success factors are what you can see also on that slide. We obviously have our in-house technology and manufacturing footprint through our IDM model, and I will describe later on what we're going to do for China. We have unique NVM technology. I will get back on this as well.
We are, at this stage, the leader when it comes to tiny edge AI. I will demonstrate that to you. We have the largest developer footprint today on the market. We have a long security lineage thanks to our smart card history, and obviously, we have the complete portfolio. Now, I want to double-click on the technology aspect. It's important. One of the secret sauce behind the GPMCU leadership is our unique and highly innovative technology roadmap and manufacturing capabilities. By showing you this graph, comparing ST to its competitors, we wanted to expose one of the strengths of the ST model. And that strength is our unwavering commitment to push forward the eNVM CMOS logic process technology, a key differentiator versus our peers. As of right now, we are the only player, together with another company, although lagging, that owns its eNVM CMOS manufacturing roadmap below 40 nano.
We don't do that in a vacuum. We do that because we believe this gives us a true competitive advantage from both a differentiation standpoint at product level and from a cost standpoint. And we complement it with a continuous partnership with TSMC on complementary technology. Now, our 18-nanometer FDSOI with phase change memory is a perfect example of that. We call this process P18, and we have announced it in March 2024 to support next-generation embedded processing devices. This new process technology was co-developed between us and Samsung, and it delivers a cost-competitive, leap in performance cost and power consumption for next-generation microcontrollers. The first-generation STM32, based on the new technologies, is already sampling to customers as we speak with production planned on the second half of 2025. As a matter of fact, if you look at the upper left side of the chart, you will see that die.
Today, I have some news for you. I'm glad to announce that Samsung Foundry and ST have signed an agreement for Samsung Foundry to provide ST with foundry services from their depreciated fab in Korea to complement the P18 capacity we are currently deploying in Crolles, France. This is an important milestone in securing the future success of P18. Now, it's all good, but before I leave this slide, it's important for you to put it in perspective. And for that, I need you to understand that in the industry at large, more than 60% of the MCU that will ship in 2030 will be using a node at or below 40 nano. This makes our strategy and investment even more relevant. Okay, I'm moving on now to the general purpose MCU part of my presentation.
It has not been an easy year by any stretch of our imagination for the STM32 franchise as our sales materially decreased between 2023 and 2024. So it's with some humility that I'm telling you that we have, in 2024, underperformed the market. Now, let's dig into the factor behind our 2024 revenue decline. First, a main factor to rationalize this revenue decline is the inventory correction in the supply chain. During 2022 and 2023, our customers forecasted a strong growth for their sales and placed firm purchase orders to cover the needs. Not only the growth expected never materialized, but the GPMCU market is also facing a strong drop in 2024. I'll come back on this. This change, coupled with the fact that the purchase orders were non-cancelable and non-reschedulable, generated a significant over-inventory in the channel that is now being resolved.
We expect this to continue through part of 2025. Now, we estimate this inventory resorption to be responsible for 60% of our sales decline between 2023 and 2024. The second factor is the industrial market that decreased, excuse me, by around 21% in 2024 versus 2023 as per Omdia. Given our commanding position in the industrial market, this one market dynamics did hurt us and represent 30% of the drop in our sales. Finally, during the allocation period of 2022-2023, we lost some share due to the reduced focus on our 8-bit microcontrollers family in favor of our STM32 products, as well as a reduced focus on some consumer application. All of this got tangled with some of our mass market customers in China adopting local solutions that were in high availability. We estimate that market share lost to represent around 10% of our 2024 revenue decline.
Now, with that said, we are seeing signs of recovery, and our book-to-bill ratio is one of those weak signals that we track. The index, excuse me, varies with the market dynamics and has dropped in 2023 as customers place their orders for 2023 and of 2022. We are seeing a progressive rebound of the index, indicating the order intakes are resuming, and we have crossed the one value in the third quarter of this year. The impact of this dynamic in our GPMCU market share is illustrated through the blue dots that represent the mathematical share, essentially our sales divided by the market. The dashed red line represents our share, excluding the inventory effect. Now, you can see in the red curve that we have lost some share due to the allocation effect that I described earlier, which we are currently recovering.
However, this loss was not seen in the reported market share as it was masked by other effects, be it the large inventory build-up that I've just explained or even some gain share elsewhere in the market that we've achieved. We are now in a period of inventory correction, and as a result, the reported share will show a material decline. However, and that's an important however, the bottom of our share was reached in Q2 2024, and we are now in a recovery phase. I can guarantee you, on behalf of the team, it has not been an easy year for STM32, but I'm confident that we know that we are now on a resurgence path, and I'm going to give you some more about that. So, to back up the confidence, let me show you now some market facts.
We are systematically and repetitively reviewing our market share status with a set of hundreds of customers. This allows us to keep ourselves honest and keep a close pulse on the market positioning. Things are looking quite solid, with close to 80% of our recently surveyed customers maintaining or expanding the adoption of STM32 in the portfolio. To be clear, these hundreds of customers are representing a worldwide sampling. Another metric of importance is what they're telling us when it comes to the value they attach to reuse. More than 70% prefer to reuse MCU platforms that have already been adopted. Okay, clearly, there is a limit to their indulgence if the pricing gaps become too large, but such statistics demonstrate the power of incumbency. To be even more valid when you take into account that we have shipped more than 13 billion of STM32 since inception.
This is indeed a big install base. Now, talking about reuse offers a good transition to talk about something we consider mission-critical: our software developers. We've been vocal about our developers' first mindset powered by the STM32Cube, simply because our revenue growth is somehow linear to the size of our ecosystem. As a reminder, our STM32Cube is a comprehensive developer framework providing all the necessary software assets to ease and accelerate embedded design. The framework is consistent across the entire STM32 portfolio, and it does leverage ST worldwide support channels and communities. Since we last talked in May 2022, we have broken records, growing our unique active developer base by 50% over a little more than two years, with close to 30% or one-third of that growth coming from China.
We're also enjoying a pretty vivid technical community with more than 500,000 visitors every single month and a developer satisfaction index that is now reaching 80%. This is quite a high performance, and next year, we'll be launching our biggest ever upgrade in the lifetime of STM32, a brand new version of STM32Cube, dramatically improving efficiency and productivity for both our developers and for ST moving forward. This is something really we can't wait to show our developers. AI is something else we've been pretty vocal over the past three years. Marco touched base on this a little bit. I will tell you in all humility, in AI, we are unapologetic. We are making the tiny edge AI market, and we intend to continue to do so.
According to ABI, the market for Tiny ML MCUs will grow with a CAGR of 113%, with applications like battery management, arc fault detection, on-device face or object detection and recognition driving the growth. This is exactly what we observe in our tool usage. We announced in December 2023 the ST Edge AI Suite, which is bringing all the different tools from ST into a single user interface and a set of partnerships with market leaders such as NVIDIA or AWS to link their tools to our toolset. In the first nine months of 2024, I'm repeating, the first nine months, we have seen more than 51,000 enterprise projects initiated on our tools. This is actually almost three times what we had the previous year, and that's confirming the rapid ramp of edge AI projects in the market.
Our vision in order to make the tiny edge AI a reality is to combine a very software toolset, what I've just described, together with the needed hardware acceleration that will enable the execution of neural network algorithms that were in the past too demanding to run on small microcontroller devices. To achieve this, it's important for all of you to understand that a few years ago, we made a strategic move to develop our own accelerator capability. We brand that Neural-ART, and it's leveraging all ST key technology assets. The first generation of Neural-ART is coming to market, and I will give you more on this right after. And more importantly, we have the roadmap to go further.
With the generation two, we'll introduce digital in-memory computing, and we'll bring a 4x performance over Gen 1, and we foresee a Gen 3 that will bring another 10x performance jump over Gen 2. The first product with Neural-ART is the STM32N6. Now, if you are a student of history, you will remember two seminal moments for AI hardware in the industry. The first one was NVIDIA Pascal in 2016 that opened the path for GPU to be used for AI, and the second one was Apple A11 Bionic, the first application processor with AI acceleration in 2017. STM32N6 is the first-ever MCU with AI hardware acceleration, and we do hope that it will be the third seminal moment in the AI hardware history.
Today, I'm thrilled to share with you that it's already in development, adopted in more than 50 customers worldwide with super good feedback. This leads us to believe that STM32 N6 will be one of the fastest products to reach $100 million of revenue in the history of STM32. Now that we've established our quest for AI leadership at the tiny edge, you can more easily put in perspective the collaboration we announced with Qualcomm Technologies in October of this year. It's an enabler to make our edge AI vision a reality. We sense, we autonomously infer, we intelligently actuate, and we automatically connect. In this collaboration, we bring the best of both worlds: the best-in-class MCU platform, STM32, together with the best-in-class wireless connectivity technology, the one of Qualcomm.
It's important to understand that Wi-Fi is a hard market, with standards being defined and driven by both the mobile and networking industry, and where the next decade velocity of feature upgrade is going to be quite steep. This is what we in ST count on Qualcomm for. The initial product offering resulting from this collaboration is expected to be available to OEMs in Q1 of next year, 2025, with broader availability to follow. It's the first step in a collaboration that envisions a roadmap of Wi-Fi, Bluetooth, Thread combo products, as well as an intention to expand into cellular connectivity with one goal, only one goal, to deliver on our vision of edge AI-powered industrial application. It will be hard to conclude my section on STM32 without talking about China. The world is STM32's oyster, and we compete globally. Globally includes China.
Like it or not, we have been, we are, and we'll continue to see an acceleration of the supply chain decoupling over the years to come. And it's our job at ST to adjust to those new market dynamics. And this is why, close to 18 months ago, we took the strategic decision to enable our eNVM 40-nanometer generation MCU to be manufactured in China end-to-end. Okay, I need to go and tell you more. We are not talking about designing our products twice against our own PDK and against the Chinese foundry PDK. Okay? That will be woefully inefficient. We are talking about the exact same process that is running in our 300-millimeter fab in Europe, now running in China.
As a matter of fact, we are using the exact same mask set to produce the wafer in Crolles or moving forward in China at HH Grace, the partner we have selected for this initiative. You can look at this two ways. From an ST perspective, it's all about efficiency. There are no double efforts of designing or industrializing the same product on two different technologies. It's the same good old ST eNVM 40-nanometer process technology. From a customer perspective, well, it's a seamless second source qualification, and they get instant benefit of our process stability. This is a big, and this is a unique step forward that we are officially seeing here today, and this will be enabled on all the ST eNVM 40-nanometer MCU, starting with STM32 late 2025. Okay, obviously, this is being made possible because we do control our technology. Remember five slides ago.
Let me give you some interesting use cases of all of this. One is a European or a U.S. OEM, for instance, who has designed a worldwide platform with STM32, but wants the ability to claim that the STM32 they are using for products based on that platform that they assemble and sell in China are actually produced in China. Another use case is a Chinese OEM with international exposure who can make their operation efficient by producing their products in China with STM32 made in China while leveraging ST fab outside of China for the products they sell outside of China. Or even a Chinese customer domestically focused but wanting to get access to STM32 technology and feeling comfortable to know it's produced locally end-to-end. It's one thing for me to say it.
It's another for you to see just a tiny sliver of the customers who have encouraged us to go down that path. In that list, you can see many Chinese customers' names like Growatt or HH Grace, but also industrial heavyweights such as Danfoss, Schneider, or Siemens. At the end, our goal is simple. We want ST to be perceived as the most Chinese of the non-Chinese MCU vendor. And this is why, by doing so, we are confident we can grow our local customer base in China by 50% over the next five years. Okay, we are done with the STM32 section. I'm now moving into the auto MCU part of my presentation. On automotive, we are not confused. We are going to double down on the market that is the most congruent with our IDM model, and that market is MCU.
It's close to $17.7 billion sum in 2030, composed of higher-end products for X-in-1, meaning integration of multi-domain ECUs into a single one, and gateway products, meaning zonal application. Those higher-end MCUs are represented by the dark blue rectangles in the picture you're currently looking at. The light blue rectangles do represent ASIL B only, actuation and sensing MCUs, and I've got something special to share with you on that front in a few slides from now. For all the reasons I just mentioned, it's now time to show our ambition in automotive MCU. We intend to double our revenue between 2024 and 2030 in automotive MCU. Let me tell you, we acknowledge this is a tall order challenge, and we intend to do that because we believe we've got the recipe to do that, and the recipe is the one that you're looking at.
It's more or less the same recipe as we have on STM32, our ideal model, our ability to harness best-in-class eNVM technology, new class of products I will describe in a second, and a pretty steep release of products that will come on the market over the next three years. One of the key foundations to double our revenue in MCU is our Stellar MCU family. We have built the ARM-based Stellar architecture with two sets of principles in mind. First, the principle that integration of ECU and sensor fusion is the main direction that the industry will take, and second, an understanding that software-defined vehicle introduces new constraints that will become more and more visible soon.
By that, I mean an increase, an expected increase of over-the-air instances, as well as software as a value creation over the lifetime of the vehicle lifetime, excuse me, becoming more and more important, leading to flexibility and scalability of memory resources becoming more and more of a strategic requirement. It's a combination of both aspects that inspire the way we architect at Stellar. Stellar P is more tuned to X-in-1 architecture where multiple virtual ECUs are integrated within a single ECU. Stellar G is the ultimate gateway and zonal products. Few of those products have already or are, as we speak, reaching PPAP status, meaning ready for production in an automotive supply chain, and as such, they are right now ramping in production at Li Auto OEM in Tier 1. I have more for you on that later on.
Behind Stellar is ST's phase change memory technology, also known as PCM. PCM is recognized as the industry's smallest memory cell for automotive MCU and is pioneering a transformative breakthrough in automotive, set to redefine what's possible when it comes to OTA. As you can see from the table, it's a one-of-a-kind technology, now totally proven, and I insist, shipping in the automotive space. As I said before, vehicles are becoming increasingly software-defined, which requires the ability from the vendor to continuously introduce and upgrade new features, functions, and enhancements. PCM groundbreaking technology will support this shift toward dynamic, flexible, future-proof software development and deployment, giving the automakers what they need to refine experience as vehicles continue to evolve and adapt over its lifetime. In short, PCM is the linchpin of our auto MCU strategy. It took time to convince customers of that.
It took time to convince customers that PCM was simply superior, but as we ran the technology in production, we are beyond this stage now. And it's with a certain amount of pride on behalf of the team that I can share this quote from Axel Aue at Bosch. Axel is a world expert in automotive NVM, and his word comparing PCM against ReRAM or MRAM means actually the world to us. Now, we are at the beginning of our Stellar journey, the early innings, like we say in baseball, but I'm happy to report that we are making steadfast progress in our market engagements. You can see on the left side of that chart an extract of early wins we have at car OEMs. I'm proud to share with you today that both BYD and BMW are in that list and have adopted Stellar in their next-generation car architecture.
On the right side, an extract of the early wins we've got at Tier 1s. Here, I'm proud to announce that Bosch, Continental, and Sunwoda in China are among the first Tier 1s who have selected Stellar for the next-generation ECUs. Like I said, it's a beginning of the journey, but a very solid start as we ship our first 500,000 units to our lead customer. Now, if you allow me, I've got one last thing to share with you on Stellar. Over the past 12 months, we've been hard at work to make Stellar relevant in ADAS and autonomous vehicles. In general, we've never really cracked that space with automotive MCU. As you know, we've been working with Mobileye for close to two decades. The partnership was defined around Mobileye EyeQ products.
Today, we're proud to announce a new and different type of partnership between Mobileye and STMicroelectronics, where Mobileye will adopt Stellar as a functional safety MCU in their EyeQ6L and EyeQ6H reference designs. This is opening a new class of businesses where we are not present to date and is giving us the chance to work with the ultimate leader in ADAS and autonomous vehicle systems with shipped close to 200 million systems to date. So many thanks to Amnon and the Mobileye team for the trust that they are putting in us. I have one last thing to share on automotive MCU. That will be the last. If you've paid attention, I've mentioned earlier that I have something special to share with you with regard to ASIL B technology for actuation and sensing MCU. It was the light blue rectangles of my first slide on automotive MCU.
So here you go. It's been requested by many of our customers for some time, and we are delighted to share with you today that we're going to bring our STM32 microcontrollers to the automotive world. It's leveraging our STM32 architecture, IP technologies, and ecosystem. We're going to develop several STM32 automotive series targeting ASIL B safety level, focusing on actuation for car body, convenience, and onboard charging. It will include a wide range of processing power, IP set, memory sizes, and packages based on our cost-effective solution, the one that we have demonstrated in the industrial market. And it will for sure benefit from our leading software development ecosystem that we will augment with safety and AUTOSAR solution. It's a brand new market that we have never touched to date, and the feedback we've received from our customers, as recently as in Electronica, is really super positive.
We are quite excited about that new venture. This is going to wrap up the automotive MCU section of my pitch. We are going to be patient. We are going to be committed to the journey, and we intend to make it a stimulating one for ourselves and our competition. I'm switching now to a different class of business. As anticipated in the introduction, we firmly believe in the satellite constellation bringing connectivity to the world. There are today two wideband internet constellations existing and in service. One is OneWeb, and the other one is Starlink. Starlink is dominating the market and has recently announced more than 4 million subscribers connected to their network.
Based on phased array antennas, the LEO user terminals are built off a multitude of radiative elements, each of which needs power amplifiers, low-noise amplifiers, and they get associated with beamformers ICs, modem ICs, GNSS receivers, power management, sensors, and more. We have sized the market of electronics for user terminal to be in the zip code of $600 million in 2024 alone. And actually, ST delivers very important technology for this, which I will detail in a minute. Moving on, we see other constellations coming up. China appears with a couple of distinct constellations, and we expect Amazon Kuiper to launch services by the end of 2025. We believe the numbers of user terminals will continue to increase, driven by the ubiquitous needs of connectivity in underserved areas, new applications for corporate, government services, mobility use cases such as the planes, ships, or vehicles.
Finally, looking at the longer horizon, 2029, 2030, it appeared clear to us that China's constellation will represent a relevant size of that business, and as recently confirmed by the European Commission, excuse me, IRIS², the independent European constellation will appear. All in all, we intend to pervade in all of these constellations with differentiated technology, especially adapted to these markets, so what are those technologies? The first is BiCMOS, which is essentially providing the LEO user terminal manufacturer the perfect technology paradigm for flat phased array antennas. Its electrical performance is allowing actually use of the technology in both Ku or Ka band. Its unbeatable noise figures characteristics help the network available bandwidth to be used efficiently, leading to more individual users per infrastructure unit of capacity.
And last, its linearity and relatively high gain allow to reduce the numbers of antenna elements on the user terminal panel, essentially saving cost. The second we have not talked too much about. It's called panel level packaging, aka PLP. And as you've understood, the numbers of IP to deliver to those user terminals is simply huge. Just to better comprehend that number, simply multiply the numbers of user terminal, say five million, by the numbers of radiative elements per antenna, say 1,000. That means that the world, that there is a need for five billion LNAs, PAs to be delivered each year. To achieve this, believe me, an efficient packaging technology is as useful as a good way for process. And this is what PLP is all about, that it provides ideal RF parameter performance at a very high production rate.
Fabio will touch base on this subsequently in his presentation, so great, there is nothing like a customer testimony of this technology, and for SpaceX, it is clear. Differential technology from ST have helped them create what is today an undisputed engineering fit that connects people all over the world, and so this is what we have achieved. Together with Starlink, we have delivered a stunning 4 billion cumulated ICs to date with an impressive increase over the last couple of years as the Starlink constellation deployment was going on, and we are planning for an even bigger year, 2025, but I know it's difficult to figure out what this means, so here is a tentative visualization of that amazing amount.
First, as discussed, this is essentially Bi CMOS, but not only, as we have also been manufacturing several beamforming and modem devices using our FDSOI process. You may be wondering, what is he doing with this rectangle? Well, what I'm trying to tell you is that actually what we've shipped physically is the equivalent area of two American football pitches, more than 10,000 square meters, almost three acres of silicon proudly farmed from ST. That is indeed a big number. I'm entering the last part of my presentation to talk about the last mega trend that I've talked about at the beginning of my pitch, the cloud optical interconnect. As you know, we are entering the era of the accelerated AI infrastructure. The fundamental paradigm shift is one of a workload requiring many, many, many connections.
It's a combination of the numbers of ports required and the speed at which those ports need to function, 800 gigabit per second today, 3.2 terabit per second tomorrow, that is leading to a decade of network hypergrowth. All these interconnections are materialized by optical fibers. At both ends of each of those fibers, there is a need for what's called an electro-optical transceiver, whose job is basically to transform a signal modulated on light onto an electrical signal transmitted then on a wire and vice versa. These transceivers are most of the time pluggable objects that you plug onto a switch, plug onto a server, allowing you to build a flexible interconnect network.
They are essentially made of three things: an MCU, we make that, which controls the transceiver operation, an electronic IC, also called IC, driving the optical source, in that case, a laser, and amplifying the signals, and more importantly, a photonic IC, also known as PIC, that is doing the actual conversion of light into electric signal and vice versa. I will not cover it in detail today, but I want you to trust me when I tell you that in ST, we have the best BiCMOS technology to build the EIC. We are seeing an acceleration of our design wins. To date, about 20 large-scale customers designing EIC on our BiCMOS technology, an example of them being MACOM.
Maybe more importantly, we believe that most of these transceivers are moving the photonics IC, the PIC, to a technology called silicon photonics, aka SiPho, that is bringing clear advantages over VCSEL or electro-absorption modulated laser, i.e., EML, and that makes this technology, SiPho, the fastest growing technology in the fast-growing market of cloud interconnect. New paradigm, new connecting paradigms such as co-packaged optics, where the fiber is brought a lot closer to the actual XPU, are even going to be manufactured from the get-go using silicon photonics only. But that is a thing for the future as most of the volume in the data centers, optical cloud interconnect moving forward, is still going to be based on pluggable optics simply because of ease of use, serviceability, and cost. In all, we believe the total market for both BiCMOS and silicon photonics foundry in 2030 will top $2 billion.
It's a brand new opportunity for ST, fully congruent with our IDM model. Let's come back to why the silicon photonics is the right technology for photonics IC moving forward. When comparing to existing VCSEL or EML, one can clearly see a few definitive points of advantage: ease of integration, flexibility in range or reach, better losses, and performance supporting up to 200 gigabits per second per lane today, which VCSEL are not able to do in all conditions. This is why in our mind, the transition to silicon photonics is a one-way door, and it's with that backdrop that today I'm proud to pre-announce to you ST's brand new silicon photonics technology platform. ST was one of the pioneers in silicon photonics 10 years ago, but decided to hold as the market was not ready for volume, but as described previously, it now is.
We have decided to come back with a new technology called PIC100, which we believe stands as the sole silicon-only 300-millimeter technology able to support up to 200 gigabits per second per connection lane. PIC100 will benefit from our 300-millimeter fab in Crolles and from its advanced lithography, ensuring a yield in the same category of digital CMOS yield. PIC100 will ramp in 2025 and will be announcing publicly more technical detail very soon. Clearly, the largest hyperscaler agrees with us. We have signed a collaboration agreement with AWS, who has been intimately involved in the definition and the development of PIC100, and who will deploy the technology in their infrastructure as it reaches production stage. Honestly, we could not have dreamt of a better partner to work with.
We have also ongoing collaboration with leading optical solution providers such as InnoLight, excuse me for that. We'll be using actually InnoLight will be using PIC100 in their next-generation 1.6T transceiver for data center interconnect and AI. This is a big step for ST, fully congruent with our IDM model and making ourselves highly relevant in next-generation AI-powered data centers. I'm quite excited about that. As promised, this was the last section of my presentation. So to conclude, you can count on us to focus our R&D dollar on markets where we can lead or co-lead, on markets consistent with our IDM model and where we know we can deliver over time financials accretive to ST gross margin and operating margin targets. Clearly, MCU is the largest of those markets. And on these markets, we intend to leverage ST differential technology and manufacturing platform.
On GPMCU, if we really want to measure our performance, then we need to look at the decade from 2017 to 2027 to normalize diverse effects that you've seen recently. On that period, you have our commitment to grow 50% faster than the market. On automotive MCU, we can compete. We can materially increase our market share in the next five years thanks to Stellar MCU and our newly introduced STM32 automotive platform. As importantly, with both now general purpose and automotive MCU under the same leadership, yours truly, we'll be able to drive hardware and software platform synergies. Regarding what we'll be doing on both the LEO and cloud optical interconnect market. I hope I've given you enough colors to understand why we are excited about what those two opportunities will bring us. This marks the end of my presentation. Thank you for listening.
Next is a break before Fabio takes us over to cover our technology and manufacturing activities. Thank you very much.
Thank you very much, Marco and Rémi, for this very insightful presentation. So now we're going to have a break. I am the timekeeper. We are running a little bit late, so let's make the break only 15 minutes. So please come back at 11:15 A.M. Thank you. Okay, thank you for coming back. It's my pleasure to introduce a short video on manufacturing, and then Fabio will make the presentation. Thank you.
Good morning. Welcome back. Today, I want to emphasize the role of process technology and manufacturing that, as you understand from the presentation of my colleague, is one of our pillars for our future success, and I want to walk you through how we are structured and how we are working.
We are deeply convinced that our mastery of technology and manufacturing is essential to guarantee our competitive advantage. The vertical integrated approach allowed us to align our operations to the customer needs and allowed us to be integrated, to be faster, to interject the development needs, and to create not only product of value, but also reliable operations and supply chain. The strategy is not just about producing at scale. It's about producing with excellence. We combine proprietary technology with the state-of-the-art R&D facility to cater to the needs of the end market we serve. At the same time, we are strongly committed to sustainability, ensuring that our operations are both efficient and environmentally sensitive.
Here is, as for my colleague, I want to divide my speech of describing first the manufacturing strategy, then we go to the technology strategy, then we will describe how we are going to reshape our manufacturing, and then we go with the takeaway. So here you can see our ST technology vision for manufacturing strategy as a key enabler is articulated on a specific understanding of how to be agile, how to be fast to the market, and how to bring differentiated product. To do that, we make use of our installed base. You see here where our main sites are located in the world. We use the usual color code to distinguish the front-end. These are the factories that we use to run silicon, with the back-end where we do the assembly test.
The front end is reporting in the blue, the back end is reporting in yellow. Our strategy here and our manufacturing roadmap is articulated on a few key points. The first one is the acceleration of the silicon capacity transition from 200 millimeters toward the 300 millimeters. The second step is run a testing operation consolidation in two hubs. That's mainly to guarantee efficiency. The third is on silicon carbide, that, as you understood from Marco's presentation, is one of our major pillars, is to accelerate also here the transition from 150 millimeters to 200 millimeters operations. We do it making use of two major infrastructure and plants. The biggest one in Catania, that I will comment later on, and the second one is with our joint venture in China and with Sanan. The last point of our strategy is China for China.
Remy already gave you an explanation of what do we mean by China for China. I will articulate a bit more on how we are going to execute it. Here, to give you a clear message of how our manufacturing is conceived, we start first put our focus on the major high-volume manufacturing pillars. Here we divide the pillars into categories: the front-end pillars and the back-end pillars. The front-end hubs are number one is the factory in Agrate, 300 millimeters. Number two, we have the factory in Crolles, 300 millimeters. And then we have the mega fab vertically integrated in Catania, 200 millimeters. And the last one is the operation that we are building up in China, Chongqing city.
Moving to the back end, we identified among our high-volume manufacturing factories three operations: one in Malaysia, where we also host the panel-level packaging line that Rémi already introduced to you. We have Shenzhen, that is our biggest operation in the back end. Lastly, we have Malta, which is a totally new building where we make use of the advanced Industry 4.0 technology. Complementing our manufacturing hub pillars, we have also focused on centers of competence. Here we split again to help you to read through the complexity of technology between front-end and back-end. In front-end, we identify as areas of competence Agrate, where we develop the BCD, mostly the BCD technology and advanced analog. We have Catania, where we concentrate the power technology capability and development. We have Crolles, where we have advanced CMOS, where we run the embedded non-volatile memory and imaging technologies.
We have Singapore, where we are recently opened a technology hub to support our activity on silicon carbide. And lastly, we have Tours, where we keep our MOSFET on GaN, and we have installed a one-of-a-kind pilot line. Moving to back end, we identify Agrate, Grenoble, Shenzhen, and Singapore, where we develop the packages and the technology to allow us to run the packages operation. We have Muar, where we have the high reliability for space and the sense products. We have then Shenzhen, that is, as I said, the biggest plant, but also hosts one of the R&D centers for power devices. And finally, in Shenzhen, we also open what we call a scouting team to take advantage of the network of a supplier we have in China. So, to conclude our manufacturing view, it's important that we also spend a couple of words on our manufacturing models.
As we said, we concentrate our proprietary technology in-house manufacturing. Within the many plants we have, we have focused on the major hubs. We also complement the activity we do on the major hub, taking advantages of partnerships with other suppliers. Here, we give you a split of how we divided our total manufacturing operations between internal and outsourced technology, with the split between the front end, where we make use of foundry, and the back end, where we make use of OSAT. Important to notice that in the foundry, we are using specific partners to complement our technology offers. We already mentioned two key suppliers and partners for us in the front end. There is TSMC, where we work with FinFET technology, and there is Samsung Foundry, where we co-develop the embedded nonvolatile memory on P18.
On the OSAT, we are using a set of OSAT inside and outside of China to support the fragmentation and geo-localization. Moving now from the global picture of manufacturing and having a closer view of our technology, we here report again, just to realign your memories on what are the major buckets in the technology we have. Here you see we go from the sensor MEMS that Marco has already described what are inside, FD-SOI CMOS technology. We have analog and RF CMOS, embedded non-volatile memory, smart power for BCD, discrete power, and overall all the technology for silicon carbide and gallium nitride, the GaN power technology, and optical sensors. All these need to be complemented with the suite of products and technology for back-end. So this is the scope of our technology R&D.
If you go more in detail, I want to drive you through what are the distinctive capabilities of our technology R&D. Talking on smart power first, we are mastering the advanced BCD that gives us leadership in many segments, followed by the monolithic GaN and BCD and the galvanic isolation. In the world of power, we are also covering the entire suite of technology that offers us to support our product power offering that is going from the silicon carbide with the multigeneration that goes from the Gen 2, 3, 4, as Marco described, the new one, including the future multi-drains. And here we are complementing with our technology of assembly and testing, where we can offer all these products in different versions that go from the non-coded for some specific activity to the discrete packages up to the modules.
Concluding the powers, we have what the technology that we call power GaN and PMOS. PMOS is very important for us in the two versions, the low voltage and high voltage, two different technologies, both very important for the market. And here we are mastering a distinctive technology on the superjunction. Last, we go to the mechanical sensor, where we have, as a distinctive capability, the piezo technology, the multi-generation of mass and six-axis gyroscopes, and finally the pressure sensors. Moving from the mechanical to the optical sensor, very important to underline the unique capability and capacity we have on 3D stacked sensors and our RGB and infrared sensors. These are complementing where the planar optical element that we have been the first in the world to announce and to use. Moving to CMOS and BiCMOS, Remy already said about the unique performance of this technology.
Today, we are supporting the RF CMOS and the BiCMOS with our 55-nanometer BiCMOS technology and with our 28-nanometer radio-frequency FD-SOI CMOS, complemented by the PIC100 that Remy already introduced to you. On the analog and mixed-signal, we are focusing mostly, but not only, on 40-nanometer analog with the embedded PCM, and then in advanced digital, we have two technologies. We have the 18-nanometer, the P18 FD-SOI with embedded nonvolatile memory, and the FinFET that we use out of the TSMC Alliance. Front end is important. It's complex. To master it all together is a challenge, but we do not stop at the front end. In this chart, quite complex, I want to give you an idea of how we master the needed technology for back end. Without entering too much into detail, you can read the slide.
You can see that we can go in from the left to the right, taking the technology of back-end on the left, moving versus the area of application on the right. So the blue box gives you which technology, the yellow light, the yellow box, sorry, where do you land. In between, you see the different examples of the packages that we develop and master. We divide it in three columns. Column one is what we have in production. The central is the one that is in startup production, early qualification, or that is just start, and the last is the next-generation node. Important to understand here is that the unique ability that we have to master both front-end and back-end technology allows us to open up the technology in a sort of a combination.
Among these very many, I want to focus your attention on the technology that we call PLP, or panel level packaging. That is a combination of substrate technology, front-end technology, and back-end technologies. This is a unique feature that, combined with some design methodology of fan-out, allowed us to combine in the same low cost, better thermal dissipation and thermal control and miniaturization. At this point, after you get the point of the complexity of front-end, complexity on back-end, ability to mix the technology, we open up our methodology to a brand new world for our product. So rather than going and integrating in a complex package, already existing silicon, we move adapting an architecture and design technology that is called system on chip and already established in the pure digital world.
So we are adapting this methodology of design into our product family, allowing our colleagues in the future to combine IP in a scalable, very efficient, highly reliable, fast, and with a very important time to market. Here on the right, on the top part of the slide, you see an example of specific designs that are not the combination of already existing chip that will be the traditional system in package technology, but it's a specific design on a specific technology to maximize the performance, then integrate it in the subframe. To do that, we are committed, sorry, to develop an entire ecosystem where we put together all this design and technology capability, and we also decide to open up a pilot line that will be operative for 2026. One word before to move out of the technology is what is PCM and nonvolatile memory.
In this slide, I don't want to go back to the same information that you already took from Remy's presentation. We took on the upper left of the slide the same picture of the real devices that we are doing. But I want to stress not only on the competitive advantage that PCM technology, also on the cost-effective and also on the fact that we are doing a resilient supply chain. This technology, as we said, has been co-developed between us and Samsung Foundry and will be manufactured in our plant in Crolles, 300mm, and in Samsung. And just to give you an idea of what it is, this is a technology that we took many, many years for us to develop to the point to convince the customer that is called what the people call phase change memory.
It's a memory that we can create in our devices utilizing the capability of design, mastering the manufacturing, and using specific materials that give us this unique performance in the memory. This is proprietary to ST, and we are convinced will generate a lot of advantages. Now, after you get the front-end memory, sorry, the front-end technology, the back-end technology, the combinations, the chiplet, we talk about panel package line. In this picture, in this slide, sorry, I took a few pictures of our running line. It's a mass production panel package line. This line today is the first of the kind and the first in the world. We opened up last year. We are already producing, as we speak, more than 4 million units per day, and we are in the ramp to 6 million units per day.
You can notice that we will produce on a large panel. You see the picture, the top on the right, 700 mm by 700 mm. One of the panels, one of those panels can be up to 70,000 devices. It's fully automated. There are no humans inside. Make use of the advanced process control, working process control, tools control based on the AI and the machine learning. We are particularly proud of this machine, of these operations, not only because it was one of the enablers, as Remy already mentioned, to take some business, but also because it represents really the first operation of this kind in the world. Now, I want to drive your attention on the strategic manufacturing program and give you an update. We go back to the four main front-end high-volume manufacturing hubs. Starting from Crolles, today in Crolles, we have a 300mm line.
It is our high-volume manufacturing line. It is well established. It is used to run advanced CMOS node and optical products, but not only, and have a capacity at the full build-out of an average of 20,000 wafers per week, depending then on the specific mix. Going to Agrate, 300 millimeter operation here is another high-volume manufacturing infrastructure. It is devoted to digital and advanced BCD with a capacity of full build-out of 9,000 wafers per week. Catania is the first of the kind in the world. I will come in later on. It's a 200 millimeter operation. It's fully vertically integrated. That means in that operation, we start from the powder, we do the ingot, we refine, and then we go through all the other steps of manufacturing, and the capacity of full build-out will be at 15,000 wafers per week.
Last, we have in Chongqing an operation that we are building together, our partner in China. This is a front-end manufacturing joint venture. We are co-designed. We are working with our colleagues in China to build it. We make use of the Chinese infrastructure and Chinese ecosystem, but we design and we move there our own product and our own technologies, and in this infrastructure, we are building the first capacity for the phase one of 5,000 wafers per week that can go up to 10,000 wafers per week on the second phases. I'm talking here about the potential capacity. Of course, the real capacity will be tuned to the demands. Now, going more into the details, among those four hubs, this is what is going to happen here. We report in these pie charts the total wafer capacity that we have.
If you see on the left side of the chart, we are reflecting what is going to happen in our silicon base capacity. Here, we don't plan to expand the capacity, but we are planning to reshape the average weight of the capacity among the three corridors: 150 nanometer, millimeter, sorry, 200 millimeter, and 300 millimeter. The numbers here are calculated using 8-inch equivalent wafers, normalized at 25 masking level. This is a normal technique used because when you have to put together capacity coming from very many different products and processes, it will be otherwise impossible to have unique metrics. So that's the metrics, and those are the results of the plan that we intend to execute up from now up to 2027, 2028. Moving on the right side, see what happened on silicon carbide.
Basically, today, we are producing almost 100% of our devices on 150 millimeter wafers, and we are moving to produce 100% of our production on 200 millimeter wafers. Now, we already qualified our 8-inch, and I will comment later on this status. Now, going back to 300 millimeter to give you a better understanding of what's inside of 300 millimeter and why the 300 millimeter, Agrate 300 millimeter, will be one of our bases of this transition. Today, the line is fully qualified and is moving on high-volume production. Capacity is always tailored to the demand, and today we are running BCD and HCMOS. In Agrate, in combination with Crolles 300, we also create what we call a center of competence on 300 millimeter that allows us to have the total interoperability between the factory in France and the factory in Agrate.
That has been one of our advantages in starting up new products for one of our biggest customers. The roadmap here is focused on mixed signal, smart power, advanced BCD, HV power, and HCMOS. Complementing what we are doing in Agrate, we have our 300 millimeter fab in Crolles. This is, as we said, the state of the art production line that has been designed and built with a modular approach, both for facility and capacity, and today represents the reference of our manufacturing capability. We plan to have a growth in our output of 20% within 2026, exploiting investment that we already manage. Advanced devices, technology, and imaging will be the main product that we will bring in that factory, along with the already mentioned P18 technology with FD-SOI and embedded nonvolatile memory.
Complementing what I said on Crolles, on Agrate, sorry, also in Crolles, we are working with the seamless interoperability, where the two factories are capable to exchange not only recipe, product, but also reticles. This is important, and this is what we call a concept of fabs twins that is different from digital twins. Digital twins usually are referred to as a model to run planning and simulation. Digital fabs are something real. We also make use of digital twins, but this is not the same. On Crolles, of course, we are focusing on advanced technology, CMOS, and embedded volatile memory. Catania, now, let's spend a few words on Catania. Catania is a one-of-a-kind, first in the world, and this technology, this infrastructure allows us to run from the raw substrate. That means we start using graphite and silicon powder. We create out of this what we call ingot.
We wafer it. Then we do the process in front-end. Then we do the testing assembly. Finally, we go to the front-end and back-end. Now, the entire ecosystem is made in Catania, first, the big one, and in Chongqing. In general, complementing Catania and Chongqing, we have also the factories that we report here. The wafer processing, Catania and Chongqing, is a testing assembly that is in Shenzhen and Catania. Ingot making that is run in Catania with the R&D in our facility in Norrköping, Sweden. Then we have front-end in Catania SiC campus and back-end in our Shenzhen campus and Catania. Here are the few distinctive features of this Catania, first-of-its-kind 200 millimeter campus. As I said many, many times, it's a fully vertical integrated that goes from the ingot out to the wafering, epitaxial growing, wafer fabrication, test, packaging in very different scales.
It's a scalable manufacturing plan, produced at scale with a lot of synergy that enables innovation in high volume. We use in this facility a lot of robotization, a lot of automations, and we use working process control, tool controls, tool management with a system that is powered by artificial intelligence. This is to secure efficiency and guarantee the quality. It's important also to notice that in this operation, we make use of a unique mix of substrate. So we use here what the people refer to as the silicon carbide monocrystalline and sandwiches composed by silicon carbide monocrystalline and polysilicon. We use this not to guarantee specific device performance on specific devices, to optimize our loading and to open the way to the future generation of technology. Here, a few pictures inside of Catania. You see the construction in progress.
We start. We just announced the opening of this operation back in May this year. The construction is going as per plan, same as in Chongqing. On the other picture, you see inside the portion of the factory devoted to the ingot creation. Even here is a fully automated line. This is very important, not only for the quality, but also for safety. So, all in all, this operation will be in production starting Q4 2025. That for us is an anticipation versus the previous announced starting date, perfectly aligned with the needs that my colleagues reported before. Once again, important to underline here the massive use of robotics, automation, and artificial intelligence algorithms to control the entire operation that includes even the facility. A word on Singapore Technology Hub.
This is a team of mix between manufacturing and R&D technology that has been instrumental to speed up the scaling up of the technology from 150 millimeter wafer to 200 millimeter wafer. And because of geographical location, it's also useful in supporting our joint venture in China. So, it's important now to elaborate a bit more on our model China for China. Rémi already mentioned the advantage for a product and business, Marco the same. So, what we are doing in the operation for China for China. So, we established, as we already said, a front-end manufacturing for silicon carbide in a joint venture with Sanan. We create captive capacity corridor in partnership with HH Grace, where we run the 40 nanometer technology. Rémi already commented on that. But we do also the OFT that is a specific technology used for low voltage PMOS technology, BCD, and IGBT.
This is to have a totally controlled front-end capacity in China running ST technology. Complementing the front-end, we also open up back-end. It's not that we open up, but we expand what we already have in back-end. So, as I said already, in Shenzhen, we have the largest back-end operations. So, we are expanding this campus to host manufacturing assembly, test R&D, and the manufacturing will be not only for the production that we do as today for all the world, but also for guaranteed internal production of 100% of the wafer that we produce in China, either with our joint venture or with our partner. Last, in Shenzhen, we have also established a world-class failure analysis and reliability labs, coupled with the application center in Shanghai run by our colleagues of sales and marketing.
That makes a unique network in China to cope and compete in China with the Chinese supplier. This gives us the advantage of cost-effective operations, resilient versus the local supply chain, but opens up access to the entire network of suppliers for materials and tools. So, at the end, what are our major takeaways? So, as we said, we master front-end and back-end technology. We are able to cope in a flexible way, our distinctive capability. We are focusing on manufacturing, accelerating the reshaping of our silicon footprint, moving focus on our Crolles and Agrate 300 mm factories, rebalancing the 200 mm capacity. We are moving the 150 mm technology in a single unit concentrated into Singapore.
On silicon carbide, as we said, we are using as two main strategic assets, our mega fab in Catania, where we are migrating most of the productions, complementing with the front-end operation in Chongqing. We have opened up a 200 mm technology hub in Singapore, and we run silicon, ASIC, and test in Shenzhen. China for China is, as we already just said, one of our major manufacturing pillars. And the last is keeping mastering the state-of-the-art technology in front-end, BCD, VA powers, MEMS, imaging, HCMOS, BiCMOS, embedded nonvolatile memory, and digital with a standard and comprehensive assembly and testing roadmap. Last but not least, we are also starting the development activity to bring into our product portfolio the advantage of integration through the chiplet technology. Thank you very much.
Okay. Good morning. Still on time to say good morning, and welcome to the last presentation of today.
What we will do in the next half an hour, we will review the ST financial performance and the financial expectation for mid-term and longer term. Last time we met in this place, I was illustrating the company financial model targeting to achieve $20 billion plus revenues by year 2025-2027. Let's see how the company performed over the last years and where we stand today in respect to our financial model ambition. Since our financial model definition in May 2022 targeted revenues at $20 billion plus, gross margin around 50%, operating margin above 30%, we experienced many changes in our reference market and in the competitive environment that ultimately resulted in a slowdown of our path toward our financial model. Since last time we met in May 2022, we experienced an acceleration in some geopolitical and competitive dynamics.
More trade restrictions, increased state incentives for semiconductors, especially addressing manufacturing, resulting in current and future expansion of capacity, increased economic decoupling, as well as increased competition, especially in China. Additionally, we experienced a slowdown of the economy with China's late recovery from the pandemic and Western economies suffering from higher interest rates and increased debts. Two of our main reference markets, automotive and industrial, saw a significant reduction in their expected size after a period of inflated demand and capacity shortage. Carmakers recently delayed their path growth toward electrification and, to some extent, toward digitalization with the mix moving from high-end to mid-low-end cars. In industrial, the entire value chain suffered from significant inventory correction, and the market started to experience some increased price pressure and fierce competition, mainly in China.
However, the secular trends supporting the electrification and digitalization of mobility, the pervasion of cloud-connected devices in industrial, and the decarbonization of society are still there and intact. So, there are no fundamental elements that will not allow ST to achieve its ambition of the $20 billion-plus revenues. The path will be longer than previously expected. And while we are convinced our $20 billion-plus financial model is still actionable, an intermediate and medium-term step is defined, as well as considering the changes that occurred, the relevant action needed to secure the achievement of our $20 billion-plus revenue financial model ambition. Over the last eight years, ST revenues grew substantially in line with our reference market. But when we look at the period 2017-2019 pre-COVID and 2019-2022, the company outperformed the market, respectively by 1.4 and 1.2 times.
Also, in 2023, reaching $17.3 billion revenue with a sequential growth of 7.2%, ST was outperforming in a market slightly growing. Surely, 2024 has been a year of correction impacted mainly in industrial by over-inventory in the channel, built by many of our customers placing during 2022 and 2023 a very high level of orders as a reaction to the capacity allocation with the intent to protect their business. Business that unfortunately materialized much lower than what they had anticipated. For ST, the correction has been strong, resulting in a decline of 23% of revenues, much larger than the market decline. Anyway, in the years between 2017 and 2024, the company in average was growing at the same pace of the reference market and much higher than the market, excluding 2024. Over the last years, ST has significantly improved its profitability.
Our operating margin has moved from 13% in the 2017-2020 period to 21% over the period 2021-2024, and excluding 2024, a year of strong correction to a level of almost 25%. The operating profit and operating margin improvement has been coupled with a strong increase in the company's ability to generate a positive free cash flow, totaling a positive $6.6 billion over the period 2017-2024, after investing in the same period $16.5 billion in CapEx. These investments set the basis for the current manufacturing infrastructure, enabling us to access two world-class 300 millimeter fabs for silicon and two silicon carbide fabs suitable for a fast transition to the 200 millimeter, as well as to a state-of-the-art back-end infrastructure. This has been achieved without leveraging the company, exiting 2024 with a very solid net financial position and significant cash on hand.
As just shown by Fabio, today, ST can count on a strong and unique technological and manufacturing infrastructure, a key competitive differentiator factor. After a period of consolidation and selective capacity increase initiatives, we have in these last two to three years accelerated the transformation of our manufacturing infrastructure. Crolles 300 at scale is today delivering one of the best wafer costs of the industry, and Agrate 300 will be gradually out of its initial ramp-up phase and soon start to contribute to the margin increase of the company. On silicon carbide, where ST is leading, the foundation to secure our current leadership position also for the future has been set with the creation of the new integrated silicon carbide campus in Catania and the JV fab with Sanan Optoelectronics in China, which both will start 200 mm production.
While the company was preparing the basis to secure profitable growth to achieve our ambitions, we faced an evolution in our market environments and, to a certain extent, increased headwinds in the overall economic landscape, which resulted in a slowdown in the pace to achieve our $20 billion plus revenue model. Compared to the expectation at the time we set our model in 2022, our reference market for 2025 has been significantly revised down. All end markets gradually entered a correction post-shortage at various times and with various intensity. Electric vehicles' growth pace has been reduced, and the industrial market entered in a long and deep inventory correction. Compared to the 2022 WSTS expectation for year 2025, the market has now been resized by around $50 billion, and probably this reduction is still somehow underestimated. But the key long-term secular growth trends have not changed and are still there intact.
Smart mobility, power and energy, cloud-connect autonomous things. Marco and me have shown that ST has in place all the ingredients to benefit from these trends. This reference market decline compared to our original 2022 expectation led us to delay over the time the achievement of our $20 billion plus financial model ambition. That remained, yes, delayed, but still there. An intermediate financial model is set over the next three, four years horizon, with revenues at around $18 billion, gross margin in the range of 44%-46%, operating margin between 22%-24%, and free cash flow generation of about 20% of revenues. Okay, let me now go more in details on the ingredients at the base of this intermediate financial model.
To achieve in the next three, four years revenues of around $18 billion, assuming a serviceable addressable market with a compound annual growth rate of around 5%, and this is based on WSTS forecast, ST must grow around two times faster than the market. We saw that already in the past, ST has been able to outperform the market. Outperform by around the double of the market growth is quite an ambitious target, but as illustrated by Marco and Rémi's presentations, we believe we have solid elements to support such an ambition. A first strong growth driver will come from our engaged customer programs. With these programs, the visibility is quite high, and the expected growth is higher than the market average. These programs will mainly address selected applications on premium personal electronic devices and low Earth orbit satellite communication.
Additionally, we do expect to outperform the market in three further areas: general-purpose microcontrollers, where we started from a quite low revenue level in 2024 due to the strong inventory correction of the industrial market, automotive with silicon carbide, and for a lower extent in ADAS. The strong potential of growth of general-purpose MCU product line from the revenues of 2024 is comforted by the strong activity of developers around our STM32 ecosystem, as described by Rémi and as an example, the over 50,000 active development projects on ST's artificial intelligence tools, where ST was the first company to introduce an MCU featuring artificial intelligence hardware accelerator. On silicon carbide, we lead the market, and we will continue to lead, targeting by the timeframe of our intermediate model above $2 billion revenues.
These accelerated growth drivers will represent around three-quarters of the around $5 billion expected revenues growth in the next three, four years. In the industrial market, the combination of accelerated growth drivers and the evolution of the other business will translate in a revenue growth in the mid to high teens in a market growing mid-single digit. General-purpose microcontroller products, analog application-specific and general-purpose, and silicon carbide power MOSFET will be the ST's main revenue drivers. In automotive, ST's revenue will grow high single digit in a market expected to grow mid to high single digit, mainly thanks to silicon carbide driven by car electrification and in car digitalization by ADAS. These two end markets will contribute by around 75% of our total delta revenues 2014 to 2017-2018, giving also a positive product mix contribution in terms of profitability.
Personal electronic and communication equipment computer peripheral markets are expected to grow low- to mid-single-digit, while ST will, in both markets, increase revenues mid- to high-single-digit, outperforming mainly thanks to our engaged customer programs. These two end markets are expected to contribute for the remaining 25% of our delta revenues. All reportable segments will contribute to our top-line growth, with power and discrete leading with a low-double-digit to mid-teens compound average growth rate over the period, followed by microcontroller with high-single-digit to low-double-digit compound average growth rate. Analog, MEMS, sensor, and digital ICs and RF product segments are both expected to grow mid- to high-single-digit. The portfolio will remain quite balanced, with microcontrollers, analog, MEMS, and sensor representing slightly above 50% of the delta revenue between 2024 and the $18 billion revenues target.
Digital ICs and RF product slightly above 10%, and power and discrete around 36% of the dollar revenues increase over the period. The medium-term market growth expectation change and the competitive environment evolution resulted in a delay of our $20 billion plus financial model achievement. We are therefore taking actions to adapt the company to a lower revenue base than initially expected for the next three years, preserving the ability to deliver profitability and cash generation in such muted condition, and at the same time, securing our ability to innovate. Three vectors have been identified to strengthen our growth capability and to improve operating efficiency: synergize and enhance product development, reduce time to market, and increase customer focus through a more focused and efficient organization.
Reshape our manufacturing footprint with an accelerated path to the 300 millimeter in silicon and 200 millimeter in silicon carbide, leveraging the infrastructure put in place over the last few years. Resize our global expenses, taking advantage of a more efficient and more focused organization, thanks to the digitalization of many support processes. Since the beginning of this year, ST has made significant changes in the way it structures and operates, including the reorganization of its product groups. Now, our product groups are organized, combining design expertise for similar technology and products, boosting company knowledge sharing, mitigating overlap, and as an ultimate result, enhancing product innovation, design efficiency, and time to market. First results are already visible in our ability to bring to the market new products and improve and accelerate our product development roadmaps.
At the same time, our sales and marketing organization, on top of the traditional organization by region, has introduced an application marketing organization by end market, boosting our ability to complement our product offering with an end-to-end system solution approach, exploiting more and more the complementarity of our wide product portfolio with the aim to extract more value from our activities. Well, our competitiveness cannot rely only on an innovative and wide product portfolio, but needs to be coupled with a competitive cost structure. The evolution of our market requires significantly accelerating our transition to the 300 millimeter for our silicon-based product and the 200 millimeter for our silicon carbide-based product. We will leverage on the manufacturing infrastructure we set up in the last years for the 300 millimeter, saturating our 300 millimeter in Crolles today underutilized and accelerating the ramp-up of the 300 millimeter in Agrate.
These will be also achieved on top of the contribution coming from revenue increase by selectively reducing our 200 millimeter legacy fabs capacity, transitioning, whenever possible, such production to 300 millimeter, while keeping our overall silicon capacity broadly stable or slightly growing. As a reminder, the benefit we can expect transitioning from 200 millimeter to 300 millimeter represents at least a 20%-25% productivity increase in front end. Silicon carbide needs to transition rapidly to 200 millimeter from the current 150 millimeter. Everything is ready with our Catania campus and JV fab in China so that we can expect over the mid-term horizon, 150 millimeter technology in production substantially disappears. Our plan was already contemplating such a transition for the silicon-based products, but now we will reach a 300 millimeter production value weight higher than originally expected, also thanks to the early rationalization of our 200 millimeter footprint.
Such a transition will be a key enabler to secure improved profitability in our mid-term financial model and toward our $20 billion plus model ambition. With the slowdown of our path to the $20 billion revenues, we need, on top of reshaping our manufacturing footprint, to accelerate our efficiency on R&D and SG&A. The new organization will reduce the need of additional resources to keep high pace of innovation. In addition, thanks to the automation digitalization of many of our G&A processes, we will need a lower level of resources to accomplish the same support tasks. This will allow us to start immediately a program to resize our expenses base, mainly leveraging on attrition, only partially replaced, and early retirement.
These expenses containment and the reduction, coupled with the comeback of the company to a revenue growth trajectory, will strongly improve our expense-to-sales ratio, bringing it back in the intermediate model to the low 20s. The accelerated transition to 300 millimeter, the 200 millimeter manufacturing footprint optimization for silicon, the fast moving of silicon carbide from 150 millimeter to 200 millimeter, and the resizing of our expenses will bring a high triple-digit annual saving exiting 2027 compared to the current cost base, saving including a mix of cash and non-cash items. While the saving in our expenses is expected to start to yield some benefit already next year, COGS savings will start to be visible in 2026 with a strong acceleration in 2027.
In addition to the well-known positive impact connected to the diameter change, the use of more automated and modern fabs and the expected full saturation of our 300 millimeter will significantly improve our manufacturing efficiency, contributing to driving down our overall cost. Operating margin will significantly improve from the expected just above 12% of this year, 2024, to the 22%-24% range in 2027, 2028. This profitability improvement over the next three years will stem from three main drivers. First, the full reabsorption of the unused capacity charges impacting about 300 basis points in 2024, which is mainly related to the current inventory correction and the continued productivity and efficiency gains in our manufacturing. Second, the reduction of our overall cost base thanks to the manufacturing reshaping and the resizing of our operating expenses.
Third, the positive contribution of the net operating leverage resulting from the higher revenue base moving from $13 billion in 2024 to about $18 billion in 2027-2028. These positive drivers will more than offset two negative impacts, which are the expected shrinking of our capacity reservation fees in automotive, which in a sense were related to the shortage and are not expected to represent a significant dollar amount exiting the period, and a slightly negative effect from the combination of negative pure price effect only partially offset by positive mix. With the operating margin up more than 10 percentage points at the end of the period, we will be then well positioned on our path toward the higher than 30% operating margin of the $20 billion plus revenue ambition. Our capital intensity over the last few years was accelerating compared to the past.
Over the period 2022-2024, our CapEx to sales ratio was above 20%. This level of CapEx, above what we consider a long-term sustainable ratio, significantly improved the modernization, the efficiency, and the readiness for the future challenges of our manufacturing footprint. All this was achieved without leveraging the company and, on the contrary, improving our net financial position. We have now in place the infrastructure that will allow us to make an acceleration on the next steps, transitioning a significant portion of our production to more advanced fabs to be ready for the challenges of the current market and the competitive environment. After the past effort, the capital intensity in the next three years is expected to decrease.
We do expect in the coming years to reduce our CapEx to sales ratio by at least 5-6 percentage points, allowing us to continue to generate a free cash flow at the intermediate company model around 20% of the revenues. Summarizing, the company will continue to invest in innovation with a more effective and efficient organization. Our relentless commitment to innovation will remain and will be reinforced. The basis to be back after a couple of years of transition to a path of growth will stem on top of our innovative and valuable product portfolio from a competitive cost structure thanks to a modernized manufacturing infrastructure with a reduced area of inefficiency. Our internal manufacturing will be complemented by partnership, allowing not only to mitigate our capital intensity, but also to strategically address reference market with a local integrated supply chain.
We intend to complement our organic growth with a fast bolt-on specialized acquisition aimed to reinforce our portfolio and boost innovation. The company was showing the ability over the past years to generate cash, and we will continue, on top of the fundamental value increase for our shareholder, to have a dividend policy consistent with our cash generation. Conclusion: The $20 billion plus ambition and related financial model is still valid, but considering the market evolution postponed. The company has set an intermediate model with the revenues expected at around $18 billion in the period 2027-2028, as well as the right innovation capabilities, product portfolio, and the market exposure to achieve this target. Gross margin is expected in a 44%-46% range in 2027-28, supported by the acceleration of our manufacturing efficiency thanks to reshaping and improved efficiency of our manufacturing infrastructure and should reach about 50% by 2030.
Operating margin is expected in a 22%-24% range in 2027-2028, supported by the resizing of our global cost base and on the path to reach more than 30% by 2030. Free cash flow to revenues ratio will improve, supported by increased profitability and lower capital intensity. It is expected to reach around 20% in the period 2027-2028 and above 25% by 2030. Thank you for your attention, and I will now hand back to Jérôme Ramel. Thank you.
Thank you, Fabio, and thank you, Lorenzo. And thank you, everyone, for your attention. So it's now time for the Q&A. As a reminder, we're going to take questions from the room and from the webcast. So everyone with online, please feel free to send your question in the dedicated box. So now I will invite our president to come on stage. All right.
We got the whole ST management team with us. So that's a great opportunity for all of you to ask a question. We're going to have a Q&A of roughly one hour, fully webcasted. So as always, limit yourself to one question and one follow-up, and please introduce yourself. So thank you, and we can start to ask the question. So maybe for the first question, let's see. Maybe on your right.
Yeah. Hi, Sara Russo from Bernstein. One of the things you've talked about in the past is insourcing your silicon carbide wafer production, and the target, I think, was 40% this year. Has anything changed on that?
So I suggest to Fabio to answer. It is about the sourcing, the 40% target.
So I assume you are referring to the sourcing, so in our internal sourcing.
So today, as we said, we are using our needs of silicon carbide. Today's 150 nanometer, sorry, 150 millimeter, moving to 200 millimeters. Today, on 150 millimeter, we're using mostly the external sourcing. Moving to 200 millimeter, we will move between a mix between internal and external. The mix that will start with the 40% internal, that will move gradually into the future up to the level that we will define according to the demand.
Great. And maybe as a follow-up on SmartSiC, I know you've talked in the past about an agreement with Soitec on that. Do you have any update on what you're doing with SmartSiC?
Okay. So please again.
Today, we are making our product not using the SmartSiC technologies. We are using the SmartSiC technologies for the R&D.
We plan to use it, as I report in my presentation, in a mix that means we do not envision it today to go from 0% SmartSiC to 100% SmartSiC. We will have a blend between SmartSiC and non- SmartSiC . The plan will be the mix, sorry, ratio will be determined by the product mix because there are some products that will take advantages or some applications that take advantage of SmartSiC versus the others.
Thank you.
So in a nutshell, I would like to confirm that this mix and match strategy to be completely vertically integrated or not is protecting our full freedom to innovate and our full freedom to purchase any material at the best price we want. It is protecting our freedom and agility.
Thank you. Maybe Lucie, we keep on the right with François-Xavier.
Thank you very much, François-Xavier Bouvignies from UBS.
My first question is on maybe the manufacturing footprint. I mean, you showed all the sites that you have. And one thing nowadays, you know what is striking when I look at this map is the U.S. is kind of missing in terms of footprint. And obviously, we live in a geopolitical environment where there is a threat around tariffs, and manufacturing footprint in the U.S. could be important strategically going forward. So I was wondering if you are thinking as well about increasing your footprint in the U.S. if it looks like not with this presentation, but I was surprised about this, for example.
No, but at this stage, I guess you have seen we are engaged in a pretty heavy manufacturing project.
As you have detected, ST is certainly one of the last semiconductors capable to manage to build a plant, whatever is a 300 millimeter or 200 millimeter. So in the short medium term, we have no other intent to set up from scratch any project. However, we have always said that we are open through partnership as far as our customers are clearly requesting this to make partnership in the U.S. about manufacturing. But at this moment, it is not on the table.
Thank you very much. And my follow-up is on silicon carbide as well. The move to 150 to 200 millimeter in two years' time or three years' time, I mean, would be remarkable. I mean, the migration usually 150 to 200, 200 to 300 millimeter, especially for SiC, which is not mature and quite complex in the manufacturing.
I think you can speak better than me on that. It seems a very aggressive timeline. It seems that you are really accelerating quite aggressively your roadmap on migration to 200 millimeter. Obviously, the market for substrate is challenging, and you have the China as well investing a lot. I assume it's because of that you're accelerating the roadmap. But I have the feeling there is no margin for error, if you see what I mean. So how do you think about this complexity and how fast you target to move to 200 millimeter? And how is it important for you to do that?
I will let Fabio to answer, of course, but I would like to repeat one thing. Semiconductor is not a job for beginners. It's a job for professionals. So that is why we are aggressive and we are taking risks. But Fabio will explain.
Yeah.
I think that your comment is absolutely correct. We are facing a very complex transition to move in such a short time a technology from 150mm to 200mm. It's complex, yes. We have created specific teams with a specific charter to decouple activity we can do in parallel. So we have a super dedicated team on design of the facility. We standardize the facility. So we didn't redesign from scratch any kind of cleanroom infrastructure for hookup, pickup of the tool. So we are copy-exact across the board. We do the same with the pool of suppliers. We do the same with product technology. We do what we call parallel engineering to qualify the product, qualify the facility, qualify the modules.
Finally, we are making a lot of use of what we call machine learning algorithm to speed up the qualification of devices and qualification of the tools. Nevertheless, it's complex, yes. It's a challenge, yes. That's why we are paying for. We are confident. We have the technical capability and the tool to do it. This represents one of our biggest projects and is within my first three top priorities, me and my team, of course.
I repeat answering to the first question, our mix and match strategy, okay, being vertically integrated, but keeping the freedom to not be one technology, let's say, lock or one or two supplier lock is offering all the options to us to deal with this complexity because, yes, it is an increasing complexity of the competition landscape, okay, between Europe, America, and China.
But the way we organize ourselves with a China for China strategy with Chongqing, our JV with Sanan, Shenzhen, and our Catania campus is offering for ST the maximum flexibility and options to maximize the market we serve.
Thank you. Maybe Lucie, one more question on the right with Janardan.
Hi, it's Janardan Menon from Jefferies. Two quick questions for me. One is on the outlook from 2027-2028 to 2030. I just want to know what you think is going to drive the gross margin from that 45% midpoint to the 50% midpoint. Is that going to be more a function of further migration to 300 millimeter, as you showed in some of your charts?
And also within that, I mean, given that a lot of your competitors are doing pretty much exactly the same thing when I look at Infineon or TI, or they're all accelerating towards 300 millimeter or 200 millimeter and silicon carbide. So how do you expect the price pressure as well in the industry and even some of the Chinese are doing that? And can you put that also in the context of that increase in gross margin from 45% to 50%? And the second question is there was one of your charts, Lorenzo, where you showed that the SAM is increasing in 2025.
I was just wondering, when you look at your performance into 2025, and I'm not asking for guidance here, but do you think you can stop that outperformance versus the SAM in 2025 itself, or are you still going to see some inventory correction headwind in the first half of next year, which may reduce that level of outperformance versus SAM in 2025, and you will start seeing that outperformance versus the SAM or doubling of your growth versus the SAM more from 26, 27?
So Lorenzo will answer about the gross margin. I will answer about 25.
Okay, I start with the gross margin. Of course, when we will be, let's say, $18 billion, our path, let's say, on the drivers to improve the gross margin is not over. There are, I would say, two elements.
If you remember the model when we were discussing in 2022, there was a difference in respect to what I was presenting today. In the next, let's say, three years, four years, you see that we do not play a positive impact when we look at pricing and mix, while in our model was there, let's say. I think that moving forward, our mix will be more and more enriched, and one of the elements will be definitely, let's say, moving from a slight low level of red, let's say, to something that is more green, and then it means that we will think that we will have a positive impact in this respect, and then there will be definitely also a positive impact on continuous improvement in our manufacturing efficiency.
We have, let's say, we will not be at, let's say, in 2017, 2018, exploiting all the potential that we have for our 300 mm scale. We are not there still, let's say. So it means that we have still opportunity to improve in terms of efficiency and to improve in terms of, let's say, ability to deliver a positive impact on our gross margin. I would say that these two elements are the most important one in order to achieve our target of 50% gross margin.
The second question is, first of all, I would like to say that we confirm that we will enter in 2025 with a Q1, which will be under the usual seasonality that I recall is about, generally speaking, -11%, and this year will be amplified by a Q1, which is really shorter than usual, so impacting -6%.
And then I have qualified during my opening remark, 2025 still a year of transition because we face in Q4 in 2024, sorry, a kind of business quake, an earthquake. And now 2025 will be more a transition year. We do believe that we will be still impacted by some inventory correction that creates two effects. First of all, a really low visibility on the market, especially still on industrial. Some distortion between what is the reported market share and the clean market share of inventory distortion. And there is something specific to ST on automotive, which was the benefits in 2023 and 2024 of capacity, let's say, reservation fees. So if I would like to classify 2025, I would like to say that we are pretty confident to be about what I have seen the consensus from the analysts. Would it be better than the market?
Will depend verticals and will depend the sub-application. Clearly, on premium personal electronics, thanks to our engaged program, we will perform better. In some pocket of communication equipment, we will perform better as well. On industrial, depend of the product. Certainly, general purpose will be still, let's say, impacted by the inventory correction. On silicon carbide, okay, we will continue our leadership position. And on the other automotive legacy system, we will perform better or equal the market. So this 2025 year as a transition, as a takeaway, is a year we have to look, okay, around the consensus, and we have to deep dive inside the subsystem and sub-application to understand because too much distorted by the inventory correction that would last at least on the first half of the year.
Thanks.
Thank you. Maybe Claudia, a question on the left-hand side. Is there a Sandy?
Hi, Simon from Barclays. Thanks for taking the questions. You mentioned on general purpose MCU specifically about the non-cancelable contracts. I was just wondering if you can give us some more color if there are any of those contracts still running in 2024, or are they completely out of the revenue base now, just so that we have a clearer picture on a clean revenue number going forwards?
So it is capacity reservation fees, or it is non-cancelable order that we add?
So you've given the capacity reservation fees in the slides. I think that's clear. It's the non-cancelable contracts. So those customers still taking product because they signed those contracts. How much of that is still in the revenue in 2024, whether it's in general purpose MCU or in other divisions as well on a total group basis would be very helpful.
Basically, we manage the shortage of semiconductors using two tools. To manage the car industry, okay, we have put in place what we call capacity reservation corridor fees that the car maker agreed to pay to be sure they will have either directly from us or through the Tier 1, the quantity of pieces they want. So this was one tool. The peak of these tools was reached in 2023, decreased this year, and will materially significantly decrease next year.
Capacity reservation fees, yes. When we look at the capacity reservation fees, definitely there is a significant drop when we look at next year, 2025. And then, frankly speaking, we have no model any longer capacity reservation fees moving in 2026, a very, very low level of dollars. So this will be something that is going to disappear.
And then the second tool we use in 2022 and up to March 2023 was for the industrial market, which is, I recall, super fragmented, both in terms of customers and in terms of, let's say, operating model. So EMS, distribution, and so on. And here, the tools have been non-cancelable order warranting volume at a certain price. And this policy has been canceled in March 2023 for all new orders coming starting April 2023. Up to March 2023 from 2022, all the backlog on the industrial market was protected by this non-cancelable order. So this was basically where we are. So no, we have no more.
Very clear. Thank you. And then as a follow-up on gallium nitride, you obviously have a very strong position in silicon carbide. You've hinted at gallium nitride offerings that you have mainly in consumer on smartphone charging.
I was just wondering how you see gallium nitride impacting the power market in the future. Many of your competitors are talking up its opportunity, potentially even into the inverter as well. So how do you see that mix and that competitive dynamic developing over time? Thank you,
So I will pass the answer to Marco. I would like just to repeat before Marco is answering that, okay, our power and discrete strategy in terms of technology to enable is to be broad range. Okay, so ST is convinced that the winner and the future winner addressing this market will be the company that are capable to control and manage the full set of technology from GaN, from silicon carbide, from IGBT, MOSFET, maybe in the future diamond, and to be, okay, either totally or partially integrated. I think that I let Marco answer. Hello.
As I have shown in my chart for the next, let's say, few years, the biggest portion of what is related with GaN will be, let's call it in the personal electronics, in the power adapters, and so on. And there, I was trying to pass the message that, of course, the transistor is important, but a key part of the system is related to the way you are going to drive this transistor, which is extremely complex. We are speaking about digital power there because the loop in order to control and to achieve the efficiencies is complex, and you need to do it embedding in the same silicon package because you want to work at very high frequency. Very high efficiency is coming from that. So putting, as I was saying, the two worlds together is what it is very important.
But no doubt, GaN is important, and we are working on GaN. Now, if you think about GaN moving to automotive, you need to consider what is the voltage of the batteries. I am speaking about traction inverter, the voltage of the batteries, and where the batteries are moving. We know very well that from 400 volts, there is a trend to move to 800 volts. So to cover that, you need to have GaNs that will be well above the 1,000 volts bidirectional and so on. So extremely complex. Now, Fabio will go a little bit more in detail, maybe in terms of technology, but it is a barrier that you need to go over in order to enter in automotive. And clearly, we are looking there at what we can do, but it's not something which is coming in the next few months.
Thank you, Marco.
So when you talk about power, you need to understand that there is only one way to master power if you read from a technology perspective. So you need to have a capability on PMOS, both low voltage and voltage, IGBT, silicon GaN. And you need to have the ability to integrate all this capability depending on the application. So the market is very fragmented, as Marco said. Different market segments have different product specifications or mission profiles, if you wish, of the product. So the only way to be future-ready in this arena is to have the full suite of products, of technologies to combine tailored to the product in a scalable way. That is our strategy. Don't forget that it's not only a matter of GaN, SiC, IGBT, but it's also there is a huge know-how in the way to test.
And that's where we have an advantage versus many of our competitors. And then, as I explained in my presentation, the ability to create packages that are small, flexible, can carry a lot of current, and can manage a dissipation. So this is the way to attack this market, so fragmented and so diversified.
Thank you. Maybe Sandy?
Yeah, hi. This is Sandeep Deshpande at JP Morgan. I want to go back to the 2025 indication that you gave. When you look at where you start off in 2025, to have even what you call a transition year, you need to see a better second half in the year. And the critical thing to that is the industrial market and the microcontroller business of ST. I mean, there are two parts to this. So firstly, on the microcontroller side, what visibility do you have about inventory at your customer base?
In terms of microcontroller itself, how do you see underutilization going away next year so that the margin can improve given that your microcontroller is one of your most profitable businesses?
I will let Rémi answer on the microcontroller. Yes, again, okay, we will start with a Q1 that we consider to be the bottom, lower than the usual seasonality, and then to start to grow sequentially. It is clear that the sales and operating plan, okay, we have built is based on, as I said during my opening remark, clearly some pocket of weakness and inventory correction in automotive across the year. Okay, so no uplift or backloaded across the year, except again Q1. Somehow, this will be mitigated by two things. I repeat clearly, analog and sensor in already awarded engaged customer program in high premium personal electronics, which is a really well-known profile of business.
The second point is industrial market accelerating, okay, starting Q2 and H2. And this is what we see. Visibility on industrial is pretty short because inventory level is still above what is normal. But here, I let Rémi to comment specifically on microcontroller.
Hi. Clearly, in a situation where you have a lot of accumulation that is not translating into POS, and it's not an ST pattern, it has been actually a pattern across the board by a few players. Clearly, we've been helping our distributors to go and move their inventory as efficiently as possible. And this indeed adds a burden on our margin. So that's point number one in terms of the margin recovery. The second part is we have actually a slew of new products coming in where we're migrating away from what used to be TSMC 19-nanometer technology to ST 40-nanometer technology.
It's a pretty steep migration because a bulk of those new products are covering the entire gamut of product regime STM32 that is also going to go and help our margin increase. And the last point is we do have, and I touched base on that a little bit during the presentation, we do have products that are carrying a much higher ASP just to go and calibrate you. When you look at this STM32N6, it's between in average, let's say, it's a 5x increase on the nominal ASP of an STM32 that does help as well in the complete mix.
And my follow-up question is on.
I want to comment also about inventory. And Jérôme will complement if needed. I think the inventory situation, okay, through the go-to-market approach, so distributor or OEM must be carefully assessed by sub-segment.
What is related, let's say, power and energy infrastructure, so supply and transportation, and at point of use, it's pretty clean of over-inventory, and the demand is pretty solid, even if recently, okay, we see a little bit less incentive on this market, but it's pretty clean. As soon as you speak about residential and buildings, here, the situation is really still very challenging. Over-inventory and the demand is weak. And why? Because, okay, this part of the business has been put under stretch during the shortage, and when they had the opportunity to build inventory, they have done it.
Then moving forward, if you go to robotics and automation, and it is consistent because I am pretty sure you heard the recent announcement of Siemens and Schneider, it is clear that whatever is related to discrete automation or process automation, which is linked to the overall economy, especially China with the implication of Germany, here, the demand is weak. Inventory is not super, super high, but of course, the inventory has not been built to face a weak demand. So temporarily, you will have over-inventory. Then what is related to consumer industrial, so white goods, battery-operated tools, and so on, demand is weak, inventory has been clean. Okay, on this view, confirm, Jérôme? Yes. Okay. And inventory has been clean. Well, military, everything is good for reason you understand. And healthcare and medical, also the situation is pretty healthy. So this is the situation on the industrial market.
Thank you. That's very illustrative. Just one quick question on the.
Sorry, Sandy, we have to move on to next. Maybe we're going to take some questions from the webcast. So Cédric.
Thank you, Jérôme. So two questions from the webcast. First question on MCU. You mentioned China for China strategy with likely some technology transfers. How do you protect ST IP in this context? So that will be the first question.
So I address the question to Rémi El-Ouazzane. Fabio will complement if needed.
We've been quite careful to deploy our 14-nanometer technology. And the way we've done that, I'm not going to go through all the processes, but there was a lot of encrypted data files in that process. What is important for people to understand is that actually China is already having a 14-nanometer eFlash process being deployed in that case, HH Grace already have.
So we are not accelerating something they're already having. It just so happened that actually our process technology, we believe, is way more efficient for many reasons, one of them being our eNVM architecture, which is in itself something that would be difficult to go and reverse back. The idea we are having here, and by the way, we've repeated, we step repeated this approach on BCD process technology here. The idea, I insist, is anchored into efficiency and productivity. You will hear a lot of people out there talking about the China for China strategy. But what they're really telling you is that actually for any given product, they design it twice. They design it on a given process development kit outside of China, and they redesign it in a process in China.
What they want to do, our competitors with their resources, is obviously their own concern, but more importantly for me is what it means for the end customers. The way it works in real life, take one of the industrial customers I was talking about earlier. They build a system, say, for example, a PFC, a power factor correction subsystem for a totem pole, 300 kilowatts, let's say. When they build a system, they qualify the system. It goes through a very rigorous validation and verification process with all the components on the board. When you change a component, and believe me, designing something with a different PDK is a change of a component, they have to go and requalify the entire system. And by the way, they will have to go and also requalify the entire software running on that system.
This is where I'm saying that there is China for China and China for China, in the sense that actually what we are enabling at this stage is for the same totem pole maker I was describing as an example to essentially have seamless dual sourcing. And when they produce that system in China, they can actually use, in that case, STM32 made in China, or where they can rebuild that system outside of China, they can actually get STM32 sourced from, in that case, for example, Crolles. And this is actually, it's a very subtle point, but I'm insisting on that point because I do believe there is a lack of education on what China for China truly means in the industry.
Maybe I ask Fabio to comment on one point, especially about test, which is a critical step in the value chain.
So when you go to China, it's always a complex matter. There are advantages and risks for sure. So the advantages are clear, very dynamic market, a lot of suppliers, interesting network, those are all advantages, big market size. So today, what we have on our manufacturing strategy, we also maximize the protection of our IP using three specific approaches. Approach number one is what Rémi already commented. We do something that, to run, needs the software, so it's not easy to clone the software. Protection number one. Protection number two, we have the testing. So whatever you want to make in terms of product, you have to test, especially on power, especially on silicon carbide. The huge know-how stays on the testing. So we segregate the location where we do manufacturing for front-end with testing, and we do not design testing programs in China.
We just bring our testing program in China in a different location where we have the total control. And the testing is a software. As I said before, software is not easy to clone. So this is protection level number two. Protection level number three, I already mentioned, is we keep the facility segregated. This is a cost in terms of logistics, yes, but it's a cost that is worth to protect the IP.
Cédric, what was the next question?
So another question from the webcast on silicon carbide competition in China. How do you see competition today and into the future from the local Chinese players for silicon carbide devices?
So I pass to Marco, but take your micro and don't give my.
Okay. We see China in this moment as a supplier of substrates. Clearly, they will move step by step at final device.
If you remember the chart that I was showing about where we are present in terms of design wins among all the carmakers, we have the one with the highest numbers of sockets that has been awarded or probably will be awarded to us. And without making names, on the competition list, there were basically no Chinese there. There were some Chinese module makers that are taking, let's say, known good dies from, let's say, Western suppliers. But at this stage, it is not really a strong competition there. While, again, on substrate, they are extremely competitive and extremely low price. Now, looking forward, clearly, they will come. So what is the recipe for us to make sure that we keep the competitive edge? Is to keep pushing on innovation, is to leverage on the economies of scale that we are creating.
The fact that we are vertically integrated will allow us to improve cost and quality because let's not forget that where you use silicon carbide typically is in situations that are extreme. High voltage, high temperatures, huge peak of currents. This kind of know-how you build through time and through experience. We have learned a lot. We are still learning. So again, the recipe is simple. As everywhere, competition will come. We need to stay ahead of them. And we have the manufacturing structure, the R&D resources, the reach to the customers, and the wide portfolio that allow us to be competitive overall, not forgetting that we are probably the best position in terms of driving silicon carbide, which is, again, not simple.
As I was saying in the presentation, driving not only silicon carbide, but a combination of silicon carbide and IGBT, which will extract the maximum value in terms of cost and performance. Again, not an easy task to do for competition. So I hope this answers your question.
And also, one point important. When we talk about China SiC, as Marco said, mostly we're talking about a substrate today. And we have access to the SiCs made in China for our China operation. So with our presence there, in a way, we compensate this potential handicap. We can have access in China to silicon carbide substrate at China cost.
Maybe back to the room. Maybe with Sebastien.
Thank you. Sebastien Sztabowicz, Kepler Cheuvreux. I've got one question on the automotive market.
Could you please elaborate a little bit on the level of inventory in this specific market, where we are today standing versus a normal level of inventory, both as your OEM customer and also as a Tier 1? And the second question will be on silicon carbide because you target more than $2 billion of revenue. This year will be a little bit below $1.2 billion. Or should we think about the trajectory of your revenue in silicon carbide moving into 2025? Should we expect something very back-end loaded, or it will be more linear with an acceleration already visible in 2025? Thank you.
Alors, to make Marco's life simple, he will not comment on 2025 silicon carbide because 2025 we will comment in January. I gave already some color about 2025.
Inventory level, I think we have to, on the automotive market, and I will let Marco on analog and power and Rémi to complement. I think there are points which are specific. There are really points which are specific to one or two customers where, because of the change of the, let's say, path of growing of electrical battery-based vehicles in H1, these customers, they have to adjust their inventory. Point number one. Point number two, also because the change of the mix within the various fleets the carmakers are selling, some specific customers which are across the board, they will also have to adjust their inventory. Unfortunately, okay, these customers are with ST. So that's the reason why we commented that Q1 will be below the usual seasonality because they will have to adjust this inventory.
Then, when we have spoken more about across the board automotive industry, again, I repeat, we have already seen in Q1 some inventory adjustment by the Tier 1. That's the reason why they started to decrease what we call their delivery forecast. Now, at this period of the year, when we see the delivery forecast for next year, it is already with inventory correction embedded. But we don't think the inventory are like the industrial market spread everywhere and over. It's more driven by mixed adjustment where because we follow us pretty frequently. If you remember well, still one year ago, the carmakers, the Tier 1, they were purchasing all the components across the board without optimization. And starting Q1 this year, they start to manage their mix. So we are still facing this situation, but not a bulk of over-inventory like industrial market.
So Marco, you want to comment more? And you want to comment more?
No, I just want to comment on our engaged ADAS customer. Clearly, they still, at the end of 2024, some accumulated inventory, both for safety stock and far less with their Tier 1s, as a matter of fact. I think it will be a continued process in 2025, but pretty consistent with what they've announced themselves.
We have time for final question. Maybe Johannes?
Yeah, Johannes Schaller from Deutsche Bank. I would say two of the major concerns you're hearing from investors are pricing and the future of EVs generally outside of China. I was hoping you could shed a bit more light on that. Maybe starting on pricing, Lorenzo, what kind of pricing assumptions have you made in your 2027, 2028 targets? You were showing us price and mix combined.
Maybe you could give us a bit more color and maybe more interestingly, how did you go about forecasting pricing given where we are in the cycle and how low visibility is at the moment?
In terms of pricing, what we have embedded in our model, let's say, is a little bit of modulation in terms of pricing. Next year, in 2025, even when we talk about pricing, it's always a little bit generic because, of course, it depends when you look at the various market dynamics. For instance, when you look at personal electronics, we know about the trajectory of our pricing in our engaged customer program, as well as, for instance, when we look at the low Earth orbit satellite engagement that we have. In this case, the pricings are quite well-defined.
When you look at some other markets, let's say, general purpose in industrial, you need, let's say, to make some kind of assumption. What are the assumptions? But we have taken, let's say, some kind of, let's say, assumption for 2025 consider still, let's say, a year of transition in the substantially mid-single digit. When moving in 2026 to 2027, we think it will be more, let's say, low single digit, something like that. This is the assumption that we have taken in terms of pricing. Is 2025 too much pessimistic? We will see. Let's say, at this stage, it's difficult to say. There are still negotiations of the contracts. We know that we are facing a market that still could be difficult. So this is the assumption that we have embedded in our model in terms of pricing.
And the other question, can you repeat, please?
Yeah, sorry.
The other question is just on the future of EVs. I mean, Marco, you've given a lot of detail about how the growth is composed in automotive, and overall, you're aiming for 9%. If we were to assume that EVs are not seeing growth outside of China in your planning horizon, which is, I think, some possibility, how much do you think you can grow automotive in that period of time?
Okay, first of all, I'm not as pessimistic as you are because let's not forget that there are regulations that are going to push to accelerate the production, maybe calling for incentive at the end, but are pushing to accelerate the production of battery-operated vehicles. The CAFE, otherwise, the result of that is all they will not produce even internal combustion engine if we do not produce battery-operated electric vehicles.
So I tend to agree with you, and I think it is embedded that there is a slowdown in terms of growth of battery-operated vehicles, but I do not see these super dark scenarios coming in. By the way, in case there is a reduction there and assume the number of cars is going to be still there, we are extremely well-positioned in the rest of the automotive market. So we will grow more somewhere else, maybe less in silicon carbide, and this will mitigate. And by the way, we are also at the same time accelerating our penetration towards non-automotive application with silicon carbide. Industrial is a good opportunity. AI data servers taking a little bit more time will be there. So overall, the growth of silicon carbide is going to be there.
I think we already took in our numbers relatively conservative numbers, but I think for what we see and the discussion we have with carmakers, clearly there is a delay, but it's going to be a delay and just a different slope in terms of growth.
Thank you,
And this is basically, if I may, what was the testimony of all the carmakers we met in Electronica. Basically, we met all the European carmakers in Electronica. Nobody questioned the medium-long-term vision of new energy vehicles, and by the way, nobody spoke about over-inventory. There were more concerns clearly about the CAFE implication for 2025. They were a bit concerned by Taiwan and, of course, by the rise of the Chinese competition, but not about to question the strategy on electric car. Absolutely no carmaker.
All right. I know there are many other questions, but unfortunately, time is running.
So for those who are in the room, you can continue our conversation with the management during the networking lunch. For the others, feel free to contact the IR team. And once again, thank you. Thank you, everyone, for joining us today. This is concluding our Capital Markets Day. Thank you.