🎁 Black Friday: 50% off Unlimited and 30% off Pro annual plans
STMicroelectronics N.V. (EPA:STMPA)
France flag France · Delayed Price · Currency is EUR
40.53
+0.72 (1.80%)
Nov 18, 2026, 5:35 PM CET
← View all transcripts

Investor update

Nov 18, 2026

Summary

STM32 targets doubled revenue from 2025 to 2028, backed by early market-share recovery, its developer and AI ecosystem, and expansion across automotive, industrial, robotics and data centers. Growth is described as broad-based, with automotive MCU revenue expected to become meaningful in 2029.

Operator

The STM32 Pioneering the Convergence Conference Call and Live Webcast. I am Sandra, the Chorus Call operator. I would like to remind you that all participants are in listen-only mode and the conference is being recorded. The presentation will be followed by a Q&A session. You can register for questions at any time by pressing star and one on your telephone. For operator assistance, please press star and 0. The conference must not be recorded for publication or broadcast. At this time, it is my pleasure to hand over to Jérôme Ramel, EVP Corporate Development and Integrated External Communications. Please go ahead, sir.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

Thank you. Thank you, everyone, for joining STM32 Pioneering the Convergence Era Conference Call. Hosting the call today is Remi El-Ouazzane, President Microcontroller Digital ICs and RF Products Group. This live webcast and presentation material can be accessed on ST Investor Relations website. A replay will be available shortly after the conclusion of this call. This call will include 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 contained in ST most recent regulatory filing for a full description of these risk factors. Also, to ensure all participants have an opportunity to ask questions during the Q&A session, please limit yourself to one question and a brief follow-up. I now would like to turn the call over to Remi El-Ouazzane. Remi?

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Thank you, Jérôme, and welcome everyone, and thank you for joining us today. I want to explain today how our STM32 microcontrollers is pioneering the convergence era, as application across industries increasingly share the same technology foundations. We see this technological convergence as a major growth opportunity for STM32, expanding the markets we can address, increasing the value we can bring to customers and developers, and creating new avenue for growth over the coming years, like I will be showing you today. I will start with our ambition, and our ambition is clear. We want to double STM32 revenues between 2025 and 2028.

In the presentation today, what I am going to do is I will start with a business update, then explain what differentiates STM32 with our unique value prop, how our unified microcontroller platform supports the technology convergence era across various market segments, and talk a little bit about the growth opportunities it does create. Without any further ado, let me begin with where we stand today. An important chart for you to get the state of the union on our business. First, I am glad to report that our market share recovery is progressing well. It is technically 12 months, one year, ahead of schedule. Revenue is strongly growing, and we are in a much better position than we were in the second quarter of 2023 when we reached our peak in sales. Conditions are vastly different today. Let me explain to you why.

Revenue at that time in Q2 2023 benefited from an exceptional inventory accumulation and significant price increases, while the book-to-bill was already low throughout 2023, as you can see from the lower right side of this chart. Today, our distribution inventory are now below our standard targets, which you refer to the upper right side of the chart, and the book-to-bill has remained well above one, excuse me, over the last four quarters. It's actually taken together that these indicators support the strength of the recovery already underway. I shall say also put under perspective the comment on the lower left side of this chart, where we explain that our Q3 2026 revenue is 25% below the peak of Q2 2023. But as you can see, with dramatically different inventory and book-to-bill dynamics.

Okay, I'll come back later to our longer-term growth ambition after walking you through the capabilities and opportunities behind it. Let's now start with what makes STM32 unique, and it's a slide that I had the chance to present to you back in the day. ST microcontrollers' success rely on multiple key factors, spanning from manufacturing capacity, technology, software developer productivity, including AI-augmented solution, security, and system-level approach. What makes it unique is that we bring that together, all those capabilities together, and I'm going to go through them with you throughout this entire presentation. Now, one of the foundation of our GP MCU leadership is our technology roadmap and our manufacturing capability. By showing you this graph comparing ST to its competitors, we wanted to expose one of the strengths of the ST model.

Our unwavering commitment to advance eNVM and CMOS logic process technology is actually a key differentiator versus our peers. ST is actually the only player ramping 18 nanometer in-house, the most advanced MCU node, not in development, but in production, and in production today. We don't do that in a vacuum. This specific process technology, which we call P18, is co-developed by ST and Samsung Foundry. We do that because we believe it gives us a true competitive advantage from both a resiliency and co-design standpoint. We complemented it with our longstanding TSMC partnerships for complementary technology. Now, I think it was only natural to move from that chart and move to something that gives you more context about our China for China strategy. The first message, or the bottom line of this chart is quite simple, is that actually we are expanding STM32 market share in China.

I repeat, we are expanding our market share in China when it comes to STM32. Our China for China strategy is quite simple and is a committed one, which back in the days was a little bit of a contrarian view. It's actually becoming actually the reference approach to the market. It is simply spelled out by saying it's a one product with two supply chains strategy. Using the same design and the same technology, we offer one product over two supply chains. At 14 nanometers, customers can access a non-China supply chain based on ST and its partners, alongside with a fully China localized STM32 supply chain. In China, the way we do it is that we combine wafer manufacturing capabilities of HH Grace with ST Shenzhen backend plant, and also some local assembly and test partners.

We delivered the first fully made in China STM32 products in the first quarter of 2026, end-to-end. We are gradually extending this approach to additional family. The chart on the right tracks, which I preface this chart with, the chart on the right tracks, on the right, excuse me, tracks our revenue and market share with Chinese end customers. 2026, as you can see, put us above the previous peaks at this stage. We can only conclude that localization is therefore supporting our position in China with market share gains alongside with revenue recovery. Very happy about that performance. Let me switch gear and dig into a more technology-centric perspective related to our P18 technology. At the heart of our roadmap is our ability to continue pushing the boundary of MCU innovation. The V8 series of our STM32 franchise is a very good example.

It builds on the manufacturing strengths we just discussed with in-house production at 18 nanometers, the most advanced MCU node in production today. It brings MPU-class hosting performance to the STM32 offer, combining high performance with robustness in harsh environments. Just by naming SpaceX as an example, our customers are already benefiting from the STM32V8 and this unique 18-nanometer technology. It significantly expands what can be done with the microcontrollers, addressing demanding application ranging from space to automation and robotics to energy management and transportation. We support this technology with a dual-source strategy through ST Crolles in France and our foundry partner, Samsung. As usual, I will tell you that the hardware is not the only key pillar of our value proposition. Since the very beginning, we have deployed a developer-first mindset. This is where STM32Cube comes in.

STM32Cube is a critical source of differentiation for MCUs, and we continuously invest in it to give developers the best tools available. Let me give you three concrete examples. First, we are making STM32Cube ready for agentic AI. This brings STM32 expertise directly into the coding agent workflow to improve developer productivity, and I will come back on this with a dedicated slide. Second, we are enriching the ecosystem with more than 14,000 validated embedded software examples to help customers with their designs. Third, we launched STM32 Sidekick, our AI-powered technical support assistant, available to all developers 24/7. For our customers, this means a more reusable software and a better support to move from development to working design to production. Okay, it is all great, but the most important is the following statement. This is that the ecosystem continues to scale.

STM32Cube is now approaching 2 million active developers. This represents approximately 25% compound annual growth rate since 2023. STM32Cube is the gold standard of the software ecosystem for the MCU industry. Let me now come back on the importance of the agentification of our STM32 software ecosystem. Faster development and design success. That is, in short, the goal of bringing Agentic AI into STM32Cube. We give coding agents a trusted, reliable source of STM32 knowledge, maintained by us. Through skills and MCP servers, they can access our resources and tools to help developers make better use of a microcontroller. The quote from Ben at OpenAI illustrates the benefit: greater developer efficiency and better design optimization. I will let you read it. This goes beyond just OpenAI.

We support various platforms, from Claude to Gemini to GitHub Copilot and others, to give customers the freedom to choose their preferred environment. We've seen AI at this stage, and through this chart, helping developers build products. Now, let me turn to another part of AI, which is edge AI, which is to bring intelligence into the products themselves. I will start by re-describing to you how we in ST have defined edge AI. It refers to the deployment of machine and deep learning models directly into resource-constrained physical devices, such as microcontrollers, rather than relying on cloud servers' inference. Edge AI is a model running where a signal becomes data. Our edge AI solutions are adopted at scale, and as you can see, with already 60% of our top 100 STM32 customers using them.

Beyond this adoption, we have spent years and a lot of time to build the most complete edge AI software ecosystem. As an example, with more than 400 AI models for MCUs, our library is by far the largest in the industry. Now, usage of our AI tools is also surging. More than 250,000 active projects over the past 12 months. This is 150% up versus 2025. Finally, AI is also enabled across our entire portfolio with multiple series embedding our proprietary Neural-ART hardware accelerator. By the end of 2026, we'll have delivered more than 10,000 development boards for our highest performance version, the one you know, the STM32N6. It's customer adoption, it's development activity, and it's hardware capability, and all of them together that are actually advancing as one unit.

That allows me to switch gears and to bring me to our next topic, which is how technology foundations are converging across markets. It's fundamental for you to understand, because it's a linchpin of what we have decided to do. We see today that automotive, industrial, and robotic systems increasingly share the same foundation: edge AI, safety, security, sensing, and connectivity. For sure, certification and qualification remain market-specific, but more of the underlying technology architecture and know-how can be actually reused. Together with software lifecycle management, these foundations enable physical AI, systems that perceive, that can decide, and act in the real world, from smart meters to industrial automation to autonomous vehicles. Different markets are increasingly calling for the same requirements, and our answer to that trend is convergence. Convergence, bringing them together in one unified STM32 platform.

It helps customers and developers reuse their work, scale across applications, and get to market faster. What does a unified STM32 platform mean in practice? It actually rests on three pillars. The first is portfolio consistency and continuity. Customers are familiar with Arm Cortex-M architecture, and they're familiar with shared technology blocks and IP across our products. This makes their design easier because it leads to software reuse. Remember the scale of that foundation. Over 150 billion, this is nine zero, Cortex-M cores shipped, supported by the biggest software ecosystem that developers already know and trust, yours truly, STM32Cube. Our partnership with Arm stands since more than 20 years, okay? With STM32 leading Cortex-M adoption. It's also going to shape what comes next. We are the lead partner working with Arm to define the roadmap for future Cortex-M cores.

For developers, this means advancing towards physical AI while enabling long-term scalability and protecting their software and system investments. Now, let's hear from Drew, Executive Vice President of Arm Physical AI Business Unit, a dear colleague who shared his perspective at our recent STM32 Summit.

Drew Henry
EVP of Physical AI Business Unit, Arm

Physical AI will be one of the largest, if not the largest, computing market in the history of computing. We're right on the cusp of it. The fastest way to move is to bring our ecosystem together to work on it. That's why we introduced the Arm Total Design Program for Physical AI in September, where we bring our ecosystem together to accelerate this growth. STMicroelectronics and Cadence are key partners in defining a core part of that platform, which is our Cortex-M roadmap.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Thank you, Drew. Not here, but I thank him nonetheless. Let me now keep on describing the other pillar of the STM32 unified platform. Our second pillar is a common ecosystem foundation, STM32Cube. We talked about that, including as well the ST Edge AI Suite. It helps close to 2 million developers to reuse tools, software assets, and development know-how across projects. And third, our shared process technology, be it 40nm flash, 18nm FD-SOI with our embedded phase change memory, or TSMC eNVM technology. They provide a full manufacturing approach supporting both industrial and high safety and automotive roadmapping needs. So how does this translate into our product portfolio? Here, you can see our STM32 portfolio today. We have built over time the biggest MCU portfolio on the planet, with more than 4,000 part numbers spanning high performance, mainstream, mixed signal, ultra-low power, and wireless MCUs.

Now, with STM32A, customer gain high safety and automotive-grade capabilities within the STM32 platform. It also support industrial and robotics application as they expand into new market, notably physical AI. The value of the unified MCU platform is that customers can build on a familiar software ecosystem and architecture while moving across domain and handling different level of safety complexity. Let me now explain the strengths we bring to this extension. As I said earlier, similar trends are appearing in industrial automation, robotics, and emerging physical AI system. STM32A brings the automotive-grade capabilities and higher safety standards to these application, building on the convergence we have discussed previously. It leverage on the same core strengths that make the STM32 successful.

That is, one, Arm Cortex-M core foundation, one, long-lasting expertise in designing for industrial and automotive markets, one, common software ecosystem foundation, one, longstanding expertise in Edge AI security and safety, and last, one, set of proven process technology. At the end of the day, it's all about scalability and continuity for developers. STM32A is the extension of the STM32 platform strategy that everybody in the industry was waiting for. As an example, let me show you how the STM32A portfolio works together in the zonal architecture of a next-gen electric vehicle. STM32A4 addresses smart edge nodes, including local actuation, distributed sensors, motor control, and basic body function. The STM32A5 extends this coverage into body and comfort control, including lighting, doors, seats, and climate systems. STM32A9 addresses zonal and real-time control, including I/O aggregation, power distribution, and actuator control.

From distributed function to integrated architecture, this family broaden STM32A's opportunity throughout the vehicle on a shared platform. Believe me, we have more STM32A to announce. Volkswagen adoption of STM32A is an important customer proof point for the franchise. Who better to explain this than Carsten, Member of the Extended Executive Committee at Volkswagen? Let's hear what he shared at the recent STM32 Summit.

Speaker 5

As the automotive industry continues its transformation, we at the Volkswagen Group are shaping the mobility of tomorrow as a global automotive tech driver. We highly value our strategic partnership with STMicroelectronics. The STM32A family is an important step towards future automotive electronics, as offering us more choice with a unified platform and software ecosystem.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Thank you, Carsten. Now, I'm going to turn to humanoid robotics, another demanding application of physical AI. Let me start by saying that STM32 is today the leading MCU franchise in humanoid robotics, with a portfolio addressing all six MCU application classes. This range from hand and body actuation to gateways and connectivity, battery management system, body control, and sensing. The breadth of our portfolio allows us to address these different functions within the same robot with the same STM32 software ecosystem. We already have strong customer engagement working with all leading players in the industry, as you can see on that chart. Let me remind you that a humanoid robot typically includes around 100 and above microcontrollers, with MCU representing approximately 20% of the estimated $800 of current addressable by ST bill of materials in a humanoid.

The potential market is significant too, and Goldman Sachs forecasts global annual humanoid shipments of around 6.5 million units by 2035. It is an opportunity of a lifetime. STM32 also integrates with our partner, NVIDIA, in robotics ecosystem, connecting our control capabilities to the wider physical AI development environment. I invite you, if you have time, to also see what we have announced together with Deepu in the STM32 Summit. With that said, let us hear from someone building these robots. Jason gave his view at the STM32 Summit. Jason is the Head of International Market Development at a very famous company called Unitree.

Speaker 6

Humanoid robots combine high-level intelligence with many distributed control functions for motion, sensing, power, and safety. Extending the familiar STM32 ecosystem toward higher safety applications while preserving the scalability, software continuity, and development efficiency is what we need to build more capable humanoid robots.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Thank you, Jason. We will finally close the application, or end application discussion, if you wish, with another mega trend ongoing in our industry, which is cloud AI, where microcontrollers already represent a significant business for ST. Where do our MCUs fit when it comes to cloud AI? In three essential areas: optical connectivity, compute, and power conversion and delivery. In optical transceivers, our high-end performance and ultra-low power family support the control function around high-speed connectivity. Our STM32 portfolio also addresses control function in compute board and in power conversion systems within AI data centers. We serve a diversified customer base from top-tier hyperscalers to OEMs to ODMs, and we also hold a commanding share of the optical transceiver MCU market.

For 2026, we expect MCU to account for more than 30% of ST's data centers-related revenue, and we actually expect our data center MCU revenue to double in 2027. Let me wrap this up with our ambition and some key takeaways for you. Our ambition is clear. I gave it to you at the very beginning of this speech. It is to double the STM revenues between 2025 and 2028, and we have a clear path to get there. Our market share recovery is 12 months ahead of schedule. We have a differentiated value proposition from resilient sourcing and advanced process technology to our AI-ready ecosystem and edge AI expertise. At the same time, our unified STM32 platform is pioneering a new ara of convergence, bringing greater software reuse and faster time to market across application.

It is important for me to re-insist on the fact that STM32 is the only platform in the world enabling this level of convergence in the industry. We take great pride in that. Beyond this, cloud AI and physical AI create new opportunities to extend the growth potential of STM32 even further. This is it. Thank you. I hope you found this interesting, and we are now ready to answer questions.

Operator

We will now begin the question and answer session. Anyone who wishes to ask a question or make a comment may press star and one on the telephone. You will hear a tone to confirm that you have entered the queue. If you wish to remove yourself from the question queue, you may press star and two. Participants are requested to disable the loudspeaker mode while asking a question. In the interest of time, please limit yourself to one question only. Anyone with a question or a comment, may press star and one at this time. Our first question comes from Sébastien Sztabowicz from Kepler Cheuvreux. Please go ahead.

Sébastien Sztabowicz
Analyst, Kepler Cheuvreux

Yeah. Hi, everyone, and thanks for taking my question. The first one is on the automotive market, STM32A. Could you comment a little bit on the design activity with the STM32A, and when do you expect the revenue to start picking up for automotive MCUs, and what are your ambition for automotive MCU going forward?

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

The first answer to your question, Sébastien, is we expect revenue to start to kick in in 2028. It will be modest and start to be meaningful more in 2029. It takes time, as you know, to go and do that. In term of our ambition in auto MCU, I will answer it two ways. A few ways, excuse me, Sébastien. First is, what we presented at the CMD in 2024 remains valid, although it slightly will take longer than initially anticipated. Why? Because we have been facing a still subdued automotive market backdrop that did not help. Now, we believe that the MCU convergence strategy is going to be a big help and a big strength for us actually to gain share in the automotive market over time. Why? Simply because we see a convergence happening in automotive makers.

If you take all of them, you will all see them actually diversifying their activities and becoming more active either in industrial or in robotics activities. As they do that, the need for them to go and leverage a converged platform will actually play in our favor when it comes to actually the momentum we will be building when it comes to the automotive part of the market with the STM32A. What I'm saying now is not a theoretical statement. We have actually described what we are doing with Volkswagen and the announcement that you just saw. We can also tell you that actually, prominent makers involved in both cars and humanoid robots is actually enjoying leveraging STM32A over the long run and be able actually to exploit that convergence strategy that I've described.

Sébastien Sztabowicz
Analyst, Kepler Cheuvreux

Okay, and a follow-up. Could you comment a bit on the pricing environment today on the MCUs, and how do you see the prices trending 2027, if possible?

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Okay. Here, I have obviously learned a lot from my first year in ST when it comes to the pricing policy we have deployed during the 2022 and 2023 era, where clearly, I believe that actually we were potentially overly aggressive in the price increase that we deployed. I will tell you that actually, I am hypervigilant in making sure that actually the inflationary market that we are living in right now is not something we exploit above and beyond what is actually related to the increase of cost that we have to induce, that we have to absorb ourselves.

The reason we are doing that is because actually, we are on a path to go and double the revenue, 25, 28, like we just explained, which by the way, sorry for the aside, but obviously, we'll make 2028 a market share that we have never reached ever in the lifetime of STM32, okay? Because of that, we have this pricing policy, which is essentially the one I've described. We will pass through our cost increase, but we will actually be there to support the growth of our customers without exploiting the inflationary world that we're living in right now.

Sébastien Sztabowicz
Analyst, Kepler Cheuvreux

Very clear. Thank you.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

Thank you, Sébastien. Next question, please.

Operator

The next question comes from Stéphane Houri from Oddo BHF. Please, go ahead.

Stéphane Houri
Analyst, Oddo BHF

Yes, hello. Good afternoon. You've just alluded to that, but can you maybe specify now that you're back at your previous market share level to what level you're targeting, and for instance, for 2028? Thank you.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Yes, Stéphane. I do confirm again that we are 12 months ahead in the context of getting back to our nominal market share, which we have always said was around 23%, if you remember what either myself or Jean-Marc have communicated. As we move and migrate toward doubling our revenue between 2025, 2028, you have to assume that this market share will be between 25% and 30%. I would say more on the second part of this range than the first part.

Stéphane Houri
Analyst, Oddo BHF

Okay, very clear. The follow-up is about the Chinese market, where now you have a manufacturing footprint. Can you maybe comment on the level of pricing there and the margin that you can extract from this market, and maybe make a difference between gross margin and EBIT margin? Thank you.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

I will tell you that at this stage, the Chinese market is not gross margin dilutive for me. That is the first statement I want to make. It is also because of the mix of our products there. Clearly, we have gained the market share that was described in the chart earlier, not by playing with pricing, but because of the strength of the software ecosystem. I can tell you that a non-negligible amount of our growth in unique active developers is coming from China. We have not played with pricing to get there. We have played with the strength of our product portfolio and the fact that we have offered to our customers this dual supply chain. It is a very powerful value proposition, which, by the way, come also with the benefit that we can scale up capacity very quickly in China.

Sorry, Stéphane, I think that you had a second question on gross and operating margin, or maybe I misunderstood something.

Stéphane Houri
Analyst, Oddo BHF

No, I wanted to understand if you make a difference between the dilutive or relative or accretive aspect of the gross and the EBIT margin.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

You mean us versus group or China versus the rest of STM32?

Stéphane Houri
Analyst, Oddo BHF

Versus the group.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Yeah. We are, let's say, a very good profile. We contribute highly accretively to both gross and operating margin of the company.

Stéphane Houri
Analyst, Oddo BHF

Thank you very much.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

Thank you. Thank you, Stéphane. Next question, please.

Operator

The next question comes from Christopher Rolland from Susquehanna. Please, go ahead.

Christopher Rolland
Analyst, Susquehanna

Hi. Thank you for the question. I guess my first one is around optical. You guys have pretty high market share here on the MCU side. Perhaps talk to why that is, what your capabilities are, and your expectations for share going forward. That would be great.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

I think it's a great question for many reasons that I will explain in my answer. The reason why we have this oversized market share in optical is actually only explainable by the size of our developer ecosystem. To be completely intellectually honest, we have understood our presence in optical after the fact more than before the fact. One, STM32, especially in China, which is right now where the bulk of the optical transceivers are coming from, and by the way, the same in the U.S., I have to say, has been the gold standard for many mass market applications. It just so happened that actually 10 years ago, that was actually the same thing. 10 years ago, optical was a mass market marginal application.

Because of the strength of the software ecosystem, because of everything that is available on STM32, they started to engage on that platform. When you engage and develop code for optical, and what does it do? It control often the laser, and it also controls temperature of the ecosystem, which has an impact itself on the laser and the frequency of the laser, and do a lot of housekeeping functions that are super tricky to manage. When you have invested like this and you know very well software games, you're not really happy if you have to go and change platform. So you want to continue down that journey. Obviously, all our MCUs, 4,000 of them, are all software compatible. That's reason number one, Christopher. The second reason is attached to the 4,000 partners.

Because when you go into this optical market, you realize that actually all those transceiver makers, they are doing it different ways. They are in need of different I/Os, different packages, different mix between DACs or ADCs. They are actually very customized in the way they approach. There is not one standard way to do a transceiver. Let's say that actually, the world of transceivers is 15 guys that are actually controlling that market in terms of transceiver makers. I can tell you that those 15 guys are having 15 different architectures. The good thing about that is actually they all started in a mass market fashion, and they have grown with us because now they have 4,000 products that they can play with to go and mix and match features, memory size, I/Os, and analog interfaces.

It's a combination of all of that that is making us right now the gold standard of this industry. I insist, okay, it's not like we realized it this year, okay? But we realized after the fact what was our presence there. It's actually the beauty of STM32. It is the most efficient fishing net of the industry.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

Another question.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Thanks, Chris.

Christopher Rolland
Analyst, Susquehanna

Yeah.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

You have any follow-up?

Christopher Rolland
Analyst, Susquehanna

Yeah, just a follow-up. If you are in the micro control, if you are a developer or an engineer, there are development environments, there are libraries, there are other development tools. I think other companies are now starting to use AI to kind of vibe code some of these MCUs, as opposed to hardware languages. I guess, do you have those environments, those capabilities? Secondly, what is different about the ST environment that people like, that helps create stickiness, something that is different than using one of your competitors, for example? How would you describe your differentiation and your draw?

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Yeah. It is a little bit of the GitHub syndrome. I will start backward and answer your agentic coding AI question at the end. Look, if I look at the amount of people that are actually active part in the community every month, I have got anywhere between 200,000 to 300,000 developers actively engaged in the community. As a matter of fact, I have to tell you that actually a bunch of support on STM32 is not even provided by us. It is provided by the community. The community augment our code. The learning cycle of the community allows us to go and strengthen our tools. It is a combination of all of that which make the environment that we deliver, you get used to it, and you get engaged in it, and you baseline on it.

When a software developer, as you know, is as a way to go and do, to program a HAL, or has a way to go and configure an MCU, or has been taught and used to go and use a certain type of software example with a certain type of documentation, or has actually been accustomed to a certain type of compiler or an IDE, a software engineer is by design, if he is a good engineer, must be lazy. Laziness means that actually it is yours to lose. Often the yours to lose is not because of the software ecosystem that we keep going. It is because you are not feeding the guy with hardware underneath it that he can actually continue to use to build other products and stuff. That is actually the flywheel works like this, and that is the way we make it work.

Can someone over time try to reproduce what we do? Yeah, everybody can. No problem. With enough time, with enough dollars, you can do it. But actually, it's a combination of many things that makes it very sticky. We realize, to your first question, that our software environment can be consumed directly by us, but will eventually be consumed because an agent will be calling on it.

What we are doing now is kind of similar thing, except that actually we are using Model Context Protocol, MCP protocol, to go and wrap our software assets, either directly or through RAGs, to expose them properly to an agentic coding platform so that actually an agent or even a human will have access to the exact same infrastructure, except actually it will be abstracted through RAGs and MCPs protocol connected to your Cursor, your Codex, or your favorite tool, favorite coding agent tool. At the end of the day, it will most likely, my view on this, it will make our software ecosystem, I would say, even more important because it will be, we manage the single source of truth.

We manage our errata, we manage our architectural document, we know how our products work, we know how to design and optimize software data bricks to get maximum hardware performance. We know all of that. Whether we expose it to you directly or we expose it to you through RAGs and MCPs, same old, same old. You still need to have the same infrastructure. At the end of the day, it's the quality of this infrastructure that will make the difference because it will eventually translate into time to production.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

Thank you. Thanks, Chris. Next question, please.

Operator

The next question comes from Gianmarco Bonacina from Banca Akros. Please go ahead.

Gianmarco Bonacina
Analyst, Banca Akros

Yes, good afternoon. A very question for me, sorry. The 25% compounded growth between 2025, 2028, can you give us a ballpark of volume and price? The other one is also here, a ballpark of the absolute amount in dollars of the revenue that you expect in 2028, and the breakdown by main applications. Thank you.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Okay. We do not give absolute amount. I believe we only share them at EMP level, right? I will try to help you with some data points maybe. I would say that STM32 represented in 2025 between a third and half of the EMP revenue.

Gianmarco Bonacina
Analyst, Banca Akros

Okay.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

You can take it from there to go and do your math. In term of growth volume, I would say that we see our high performance portfolio growing quite a lot, as well as ultra-low power. Because if you look at the underneath the growth that we see, I would say non-negligible amount of that growth is related to data center. And if you double-click, it is related to optical. And if you double-click, you will quickly get to either high performance or ultra-low power MCUs. So, it gives you a little bit of an idea. They are often, as you can imagine, a bit more expensive than mainstream MCUs. That is about what I can share with you at this stage.

Gianmarco Bonacina
Analyst, Banca Akros

Okay. Okay, thank you.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

My pleasure.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

Thank you. Next question, please.

Operator

The next question comes from Adithya Metuku from HSBC. Please go ahead.

Adithya Metuku
Analyst, HSBC

Yeah, good afternoon, guys. Thank you for taking my questions. Firstly, just on the share excluding data centers, to me, it feels like a lot of the growth over the coming three years in that doubling will be driven by data centers. If you were to exclude data centers, is your share still growing? If it is, then what are the key drivers here? Any color there would be helpful, and I've got a follow-up.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Yes. Like I said, I completely agree with your statement, even though it's not all the growth. It's a non-negligible high double-digit amount, but it's not funding all the growth, that I can tell you. What's growing is actually industrial. We see industrial booming. You know there is a part of it, maybe related to also what's happening in AI infrastructure because, we have fantastic customers, without naming them, Schneider or ABB or Eaton, who are actually super active as well in the power delivery architecture, except they're not in the rack, right? They're outside, building SSTs, building cooling chambers, building stuff of that nature. We see a lot of growth happening there. We see a lot of growth happening with cooling vendors as well. We also see a lot of growth in industry automation happening.

This year was a big growth when it came to solar in some industries. We see a lot of growth in industrial, most likely because the pendulum swung so hard over 2023, 2024, 2025 in term of inventory correction that I would say, in 2026, 2027 minimally, you're going to see a very healthy industrial market. Okay? That would be what will also bring a lot of growth for us on top of the AI data centers around compute, optical, or power conversion that I was mentioning earlier.

Adithya Metuku
Analyst, HSBC

Got it. But would it be fair to say that you are also gaining share in the MCU market, excluding the data center opportunity, maybe excluding the optical opportunity, I should say? Can you hear me?

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

I would say so.

Adithya Metuku
Analyst, HSBC

Yeah.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Yeah, no, I was thinking of your question. You know what? I would say so because as good business people, we are also not Our share in optical is so high that when we modelize things, we are assuming share loss just to be on the safe side, and we are able to achieve what we are achieving. So, I would say so, yeah.

Adithya Metuku
Analyst, HSBC

Got it. Just as a follow-up.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

I would certainly not want the audience to. It's a great question because I would certainly not want the audience report out to be that actually we are getting there because of optical and data centers. That would be a little bit of a deceiving outcome to be direct. Your question is great, and I can confirm that actually the growth is across the board.

Adithya Metuku
Analyst, HSBC

Got it. Just as a follow-up, on the optical side, it looks like a lot of your growth is coming from China and the developer ecosystem there. There's a lot of talk about potentially banning the Chinese optical transceiver makers into supplying into Western data centers. I do not know if you had any thoughts around how your share compares to within the non-Chinese optical transceiver makers. How does that compare to your share with the Chinese optical transceiver makers? If this ban were to get implemented, how should we think about the implications?

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Well, if the ban gets implemented, I think you will have to call 1-800-NVIDIA first because they may be in trouble with their deployment. Pun aside, my view on this is that actually, we are used as a microcontroller across all top players in the world, be it Chinese or non-Chinese. There is only very few exception to my statement as I speak. I am talking here to your question, which was talking about MCU utilization in optical transceiver. Here, I can tell you that actually our market share is strong in China, but it's also very strong outside of China. Okay?

I have nothing to add on top of that because we have never shared our exhaustive customer list, but I really want you to trust us when I said that actually, for example, in the U.S., the largest of all is 100% relying on ST microcontroller. I will not name names because we never actually disclose that, but I can confirm that to you with confidence.

Adithya Metuku
Analyst, HSBC

Got it. Can I ask a follow-up or should I join the queue? Sorry.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

Sorry, Adithya, you already had your two follow-ups.

Adithya Metuku
Analyst, HSBC

Absolutely.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

You can get back to the queue. Thank you. Next question, please.

Operator

The next question comes from Alberto Gegra from Equita SIM S.p.A. Please go ahead.

Alberto Gegra
Analyst, Equita SIM S.p.A.

Hi, good afternoon. Hope you can hear me. I have a first question on the manufacturing strategy. Since few days ago, we read rumors about Samsung that will increase capacity also on the collaboration with you on chips for LEO application. More generally, how is the split today between internal and outsourced capacity? How should we expect this mix to evolve in the next years? Maybe also, what does it mean for you on the margin perspective?

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Sorry, could you repeat again the first part of your statement because we want to make sure we understood it clearly, if you don't mind.

Alberto Gegra
Analyst, Equita SIM S.p.A.

Yeah, sure. We read some rumors about Samsung increasing capacity in the U.S. to increase the output dedicated for STM32 for the LEO application. Just wondering how is the mix today between internal and outsourced manufacturing?

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Okay. I won't comment on rumors. I will just tell you for sure that our collaboration with Samsung Foundry is strong, and it should not come as a shock that eventually for microcontrollers, we also propose for some of our technology corridor the equivalent of what we're doing in China, meaning a U.S. for U.S. That could make sense at the right point in time, but we have nothing to announce today about that. In term of the foundry utilization of the business, it's actually not the same as nominal ST. We are, in microcontrollers, much more agile in using foundry businesses. Some of it is historical reason, some partnership with TSMC, but some of it is also self-induced, our China for China strategy that I described earlier. We, as an IDM, we control our process technology.

When we deploy our process technology, for example, in HH Grace as a foundry, we are not using their technology. We deploy our own process technology so that actually we can deliver the one product to supply chain that I was mentioning earlier. The business and the business discussion are slightly different than if you buy a wafer with their own technology, for which they have done R&D and for which they deserve to get most likely a different price point. I will let you connect those dots so that actually you understand it. But clearly, this is, at the end of the day, all factored in our gross margin, and it's something that we can completely deal with.

Alberto Gegra
Analyst, Equita SIM S.p.A.

Thank you. If I may, a quick follow-up. Do you see opportunities in this space for particular technological investments, maybe also some bolt-on M&A to strengthen the positioning?

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Yeah. Straight question. Yes, clearly possible. We are looking at that through anything that can make our software ecosystem stronger. We are game, and we are always constantly investigating things. Also, actually, what we could do to secure even more our presence in physical AI and in humanoid in particular. That's something of high interest to us. That's actually where our focus is. It's more on bolt-on than anything else, if I can put it this way.

Alberto Gegra
Analyst, Equita SIM S.p.A.

Okay. Thanks.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

Thank you, Alberto. Next question, please.

Operator

The next question comes from Johannes Schaller from Deutsche Bank. Please go ahead.

Johannes Schaller
Analyst, Deutsche Bank

Yeah. Thanks for taking my question. I want to zoom in a little bit on the auto microcontroller business. Quite a few of your competitors, including the market leader, are going to market now with RISC-V microcontrollers, specifically targeting software-defined vehicle and other use cases. Maybe can you give us an update just what's your latest thinking on that? I mean, you have this very, very strong Arm partnership, but are you open to RISC-V? And also, how does that impact your ability to compete if a lot of the other players are going to RISC-V and that kind of becomes maybe more the standard for certain applications? And then I have a follow-up as well.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Okay. That is a great question, Johannes. First, we are obviously keeping a very close eye on RISC-V. You know we were part of Quintauris, and we clearly acknowledge that RISC-V ISA, Instruction Set Architecture, has interesting feature. As a matter of fact, you need to know, we use RISC-V a lot, except it is not exposed to developers. It is for some fixed function deeply hidden in our process. Look, Johannes, I tell you, I would say I meet roughly 200 STM32 customers every year. Roughly that number. And I have been doing that now for five years, that make it 1,000. If I can remember of more than five discussion where the RISC-V ISA topic versus Arm came up, I would be stretching that. Said otherwise, no one give a damn.

But where it is important, and this is where, as much as you are holding me accountable, I think I will invite you to do that globally, is you need to have consistency. Me, I have no regrets. We have defined our convergence strategy on Cortex-M and Arm. We are working with Arm on this roadmap. More will be announced in the coming months on that topic. And I am consistent. Please, if you believe in convergence, and whether it is Arm or RISC-V or a third ISA, please make sure your entire portfolio is converging to the same ISA, because the fundamental of what I have just explained will remain. Developers will want to have a single architecture that they can scale across, and this architecture is not only the fixed function accelerator, it is also embed the scalar processor, also called RISC-V or Arm.

So, for as long as you have logical approach where you do that across the board of your portfolio, general purpose, automotive, and robotics, me, I am all in. I can only respect the logic. I am all good whether it is Arm or RISC-V. What I think is a bit of a misplaced statement, if you allow me, is if those statements are made when your portfolio is completely heterogeneous, where on industrial you do X, and then for part of your automotive you do Y, and then for part of automotive you do W. Then, I have to admit that this is losing a lot of credibility in my eyes. So, I repeat, we have a conviction. The conviction is that all the industries will converge and that our customers will be multimodal in the market they go after.

That conviction leads you to what type of microcontroller they can have for maximum reuse. That conviction leads you to convergence. That conviction must lead you to a single ISA across your portfolio. And that single ISA can be RISC-V or Arm. This, I have no problem with it, and everybody will select and decide what they want to do, and I will respect it, because RISC-V is a good architecture. But what is very important is it is being deployed exhaustively with a conviction where your portfolio is completely uniform.

Johannes Schaller
Analyst, Deutsche Bank

That is clear.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

I think you have follow-up question.

Johannes Schaller
Analyst, Deutsche Bank

Yeah, just a quick clarification on the capacity side for your STM32 target in 2028. Maybe just can you give us a bit of a feeling where you are capacity-wise to get to that target and where the capacity growth will come from here? There were a few comments here and there, but maybe just summarize that again for us.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Yeah, yeah. No, no. This is a great question. Actually, let me answer it this way. Capacity unbounded, the numbers 2x 25 to 28 will be bigger. The 2x is without constraint in capacity.

Johannes Schaller
Analyst, Deutsche Bank

Got it.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

So, you have to hold me accountable to 2x, because if I look now at my complete demand, it will be, I would not say dramatically, but it will be a non-negligible X factor on top.

Johannes Schaller
Analyst, Deutsche Bank

Great. Thank you.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Thank you.

Jérôme Ramel
EVP of Corporate Development and Integrated External Communications, STMicroelectronics

Thank you. I think this was the last question. That concludes our session today. If you have any further questions, please reach out to the investor relations team. Thank you everyone for joining today.

Remi El-Ouazzane
President, Microcontrollers, Digital ICs and RF products Group, STMicroelectronics

Thank you.