STMicroelectronics N.V. (STMPA) Earnings Call Transcript & Summary

Good morning, everybody. Thank you for joining us today. My name is Andrew Gardiner and it's my pleasure on behalf of Citi to welcome you to the session on STMicroelectronics. We have with us today, Jean-Marc Chery, President and CEO; Lorenzo Grandi, CFO; and Selene and Cedric from the IR team. I've got a list of questions, as you might imagine. So I'll run through those, but I'll pause and make sure that we've got everyone's questions in the room answered as we go through. And we also have an audience participation question a little bit later on as well to just spice things up a little bit. But why don't we start, Jean-Marc, if you will? Well, firstly, welcome. Thank you.

Thank you.

Also, I should say just housekeeping, if you do want to ask a question, just press the button in front of you to fire up the microphone. Could we start then on your 2024 outlook? You've guided the market between $15.9 billion and $16.9 billion in revenue. Can you just walk us again through the different drivers of what's going to get you within that range?

Well, I think I have to come back by a vertical view. So first of all, clearly, Automotive that is representing about 41% of ST revenue, so we see in 2024, a growth about a mid-single digit. Well, we feel important that we have, let's say, to disclose a like-for-like growth of Automotive because if you remember what I said during our Q4 earnings announcement that in 2023, we have taken benefits for Automotive revenues, about 2 points. Point number one is what we call capacity reservation fees from a carmaker that touch as a peak in 2023. And the second, let's say, nonrecurring event was an important inventory buildup, agreed, because contractual inventory buildup from one important customer that will not occur again in 2024. So if we clean from these 2 elements, so capacity reservation fee and inventory buildups, very specific, the growth we expect on Automotive in 2024 is, I have to say, let's say, above 10% -- slightly above 10%. We have a good confidence level because the coverage we have in terms of backlog is, let's say, about 85% and for the time being, we don't see any massive, let's say, inventory corrected here and there, for sure. The customers are, let's say, tuning their inventory, depending the mix of the end customer demand, but we don't see a massive correction. So for Automotive we expect like-for-like slightly above 10%. Everything reported, let's say, mid-single-digit growth. Well, the second important vertical for ST is clearly industrial that is representing 30%. Well, clearly, here, at the midpoint of what we indicated, we see a mid-teens decrease. It is mainly related to a strong inventory correction, which is amplified by weakened demand. And at this stage, it is a mid-teens correction for the full year, decrease, but unbalance means as a correction is planned to happen mainly in H1, and we are expecting at the midpoint of what we have indicated to come back to, let's say, mid-single-digit growth year-over-year H2 '24 versus H2 2023 on this market. Today, there is no sign that this scenario is realistic but there is not yet sign that this scenario is fool-proof. So that's the reason why we have given this low head, let's say, forecast at 15.9%. And if it would happen, it would be mainly driven by a delay in the recovery of the industrial market and of the inventory correction. mid-teens correction for the full year, decrease, but on balance means as a correction is planned to happen mainly in H1, and we are expecting at the midpoint of what we have indicated to come back to, let's say, mid-single-digit growth year-over-year H2 24 versus H2 2023 on this market. Today, there is no sign that this scenario is realistic but there is not yet signed that this scenario is full proof. So that's the reason why we have given this low head, let's say, forecast at 15.9%. And if it would happen, it would be mainly driven by a delay in the recovery of the industrial market and of the inventory correction. But at this stage, this is a visibility we have. Well, on Personal Electronics, that is representing 19% of ST. But as reported, we should decrease about minus 5%. But it is here clearly driven by something I already warned the market since a while. The fact we are no more a pure optical module inside the application in 2024. So if we look like-for-like, Personal Electronics for ST will be more stable 2024 versus 2023. And here, again, with a good visibility from our main engaged customer program for the year. So here, we are confident that there is not too much uncertainty on this part. To finish with 10% of ST, which is computer, peripheral and communication equipment, here, we should be stable and clearly following 2 dynamics. So one key growth driver, which is the lower orbit satellite communication with an important customer where we continue to grow steadily quarter-over-quarter. And clearly, this is compensating the legacy exit we have in this field of, let's say, communication, entreprise communication, ASIC, purely digital. This kind of business that ST does engage step after step. Computer peripheral for us is mainly driven by printer and hard disc drive more and here on printer situation is certainly touch better than last year, but it is not a key growth for this year. And hard disc drive, yes, we expect H2 growth again, but not at a tremendous level. So at the end, at the midpoint of our guidance, is minus 5% for ST, I guess. So Automotive, about mid-single-digit growth like-for-like slightly above 10%, good confidence level because backlog coverage. Industrial, mid-teens decline, could be worse. But at this stage, the mid-teens decline, no sign that it will not happen. Is it fool-proof? Not yet because lead times are very short, so order will come late. Personal Electronics, about minus 5%, but we know exactly what like-for-like is flattish. Good confidence level. And communication equipment and computer peripheral flattish and good confidence level as well. So this is what all I can classify our outlook for '24.

Okay. Thank you. So I think that gives us a clearer view on 2024. In terms of the near term, I mean you've touched on the relative lack of visibility. Would you say that anything has really changed over the last 2 months since you first provided that guidance? Have things got any better or any worse?

Better, although, clearly -- I think no, we are entering, let's say, in a rather different market cycle than the past 3 years. I think it's a very simple, let's say, statement and very obvious statement. I think, yes, '24 is certainly a year of transition, mainly linked to the fact that there is this correction on the industrial market. Why? Because in 2022, 2023, certainly forecast of the end demand linked to the carbonation industry overall and consumer, let's say, spending were overestimated. On top of that, after the trauma of 2021 shortage for sure customer order a lot, and the industry has been capable to supply. So at the end a weakened demand versus adequate supply has created what we can classify a kind of bubble for 2024. That's the reason why 2024 is a transition year. When we see the activity on this industrial market from the 1 million developers we have on microcontroller. When we see the number of new products that we have introduced in 2023 that we will introduce in 2024 that are demanded by our customers, we don't see absolutely any structural issue that don't make us confident that ST as a broad range player on the industrial market capable to offer orbital processing solution capable to offer the full range of power device and power driver with the right analog sensing when we did, we don't see a structural issue, okay? Yes, of course, there is some activities in China, which are becoming very competitive. Yes, we see some effect of, let's say, what we can classify kind of decoupling between Western America and China. Yes, you see some consideration we have to take from the go-to-market approach or from our, let's say, manufacturing implementation approach. But I would classify tactical or strategy from a supply chain point of view, but nothing structural on our capability to address the megatrend of the industrial market, which is the combination of automation and more efficiency. On Automotive, more clearly, for the time being, we see China continue to execute their plan. So they should produce 10 million cars this year. Competition in China on car is very strong because the quality of car, the prices are very competitive. The features offered are very competitive. The ecosystem is very strong. There is many, many brands. So structurally, we could expect that mid long term, some consolidation will happen. On the other markets, so Europe and America, but yes, I am -- like you, we have heard that there is some, let's say, a period where the speed of the electrification of the mobility maybe will not run at the expected one. Maybe we will enter in some not turbulence but some sigmoid of the trend where some people are starting to say 2035 full ban of thermal combustion engine will not completely occur. Whatever is a percentage of electrical car we will have in the next 5 years. It is a trend. And clearly, we have to adapt ourself. Is it a radical change to compare what we saw 2, 3 months ago? No. Okay. I already communicated on that, that it is normal that electrical car will face period of, let's say, sigmoid up not down, but less up than expected because at the end it's a consumer that appetite will decide what happened. But then on the other market, Personal Electronics, honestly, for ST, as I always told you, I have -- we have 3 years visibility on the operation, 5 years on the R&D, on the main programs we have on custom design product addressing Personal Electronics. So it's going well. we want some sockets that were expected. We won some sockets that were expected. We won some sockets that were not expected. So this is a business that will continue to represent for ST between 18%, 19%, 20% and we consider this business good for also for the volume we can extract to run our 300-millimeter fab. About communication. Yes, clearly, we see an acceleration of the willingness of various region to equip themselves with lower hertz orbit satellite communication, for sure. We don't see major disruption on the deployment of 5G or 6G, except clearly in China, where it is clearly something that they are pushing like air but no measure was changed compared to months ago. Then server, are you aware that AI is very -- yes. I guess, yes. For ST, again, the important opportunity related to AI server is power stage because there is a change in the power stage architecture where it will provide important opportunity for silicon carbide. So this is what I can share about compare January or December as March 12 -- we are March 12 today.

And so you've spoken a lot about end demand. Where are we on the pricing side of things as well, sort of to complete the circle to revenue? Clearly, we've come through an unusual period of price rises and capacity reservation fees, et cetera. Where are we today? And what are your expectations embedded in that 2024 guidance?

So maybe if you don't mind, I give the opportunity to our CFO to answer.

Yes. Sure. orebody. But in terms of pricing, for sure, there are different dynamics when we look at the markets. In -- it's true that we come -- we came -- we are coming from 2023, in which pricing were substantially increasing not all across the board, but it was a positive year in terms of pricing. When we look what we have embedded and what we see in 2024, from Automotive, I would say that there is some price decline, but quite mild. I would say what we see is something in the range of 2%, 3% max price declining. And this is already substantially discussed with our customers. So it's not a significant impact in terms of pricing. I would say that it's even milder and lower than what was in the past. So some declining but not particularly strong. When we look at markets like the industrial, where you see there is this important inventory correction. At this stage, I would say that being the demand is so weak, to be honest, there are no major discussion on pricing. It means that at the end, even if you go to your customer and offering, let's say, low price, but considering that they have inventory on the shelves is not an argument that, that is going in the direction to boost the demand. But in any case, what we do expect that there will be some pressure on pricing. And this is what we have embedded in our, let's say, model for 2024 for this year. So this will be a little bit higher than what we had seen in Automotive. So we think that will be probably something in the range of mid-single digits or something in this range. We don't see any way any significant, let's say, collapse. This is not something that we are experiencing so far. We have not seen, let's say, requests from our customers to drop pricing at 10%, 20%, something like that. This is absolutely not something that is happening. So I would say that we are back in the normal price dynamic, maybe due to the condition of the demand in the industrial a little bit higher than the normal situation. This is what we have embedded but not yet experienced because today, let's say, as we were discussing before, it's more a matter of the fact that there is this, let's say, inventory correction that is substantially pressing the level of the demand. So we have no customer coming with a request of very big price decline. If you want, you can embed inside the pricing also what we call the capacity reservation fees. The capacity reservation fees that we are mainly addressing Automotive, I would say, substantially Automotive last year were quite material. Of course, this year, the capacity reservation fees declined but still remained quite material. It's not something that is going to disappear. There is a decline in the range of 50% of this -- of the value of the capacity reservation fees, but still is something that is positively impacting this year. This is normal, I think, because, of course, there is more balancing between demand and capacity available. Probably in the future, these capacity reservation fees will continue to decline. But this was embedded in our modeling for the company. We never thought that this could be a driver, a permanent driver for our, let's say, profitability when we were modeling our gross margin of 50% for the company, we were not embedding any material or significant impact from the capacity reservation fees.

Okay. I just -- before we dive into some of the end markets, I am just going to see if there were any questions around the room in terms of ST's higher-level financial guidance for the year.

It wasn't so much on the high-level financial guidance, but just more on the sort of environment. We've had one of your peers in the U.S. use the term green shoots a few days ago, which got people excited. So maybe you could sort of either corroborate or not corroborate find where you might be seeing or not be seeing green shoots in the business. That's the first question. And then secondly, another one of your U.S. peers kind of attempted to say, the bottoms end one quarter from now. Like how do you feel about that kind of statement as it pertains to your business?

On industrial market specifically on general-purpose microcontroller, we are not close to the correction. Is the backlog stabilized now and POS versus POP, so means from this orbital case who out, who in, stabilize? Yes, but speed of inventory correction accelerating? No. Inventory corrected, whatever is a channel, I repeat, whatever is a channel. So OEM, small OEM, EMS or distributor, not yet because, again, the end demand is not at the expected level compared with 6 months ago. So it will take more time. Yes, we can expect that end of Q2 early Q3. We see a restart of the growth sequentially. We can feel confident about that. The key question is what will be the magnitude of the acceleration? Is it mid-single-digit sequential growth? Is it low single digit? Honestly, it's too early to say. Why? For a single reason that the lead time offer by semiconductor are very short. And the inventory situation is still high. And then people, they have their own uncertainty still related to the economy, the interest rate and so on and so forth. So that's the reason why we have been not cautious, but we have been, let's say, pragmatic when we provided our guidance in January for the year, our indication for the year that at this stage we can confirm it. But okay, everything is possible on industrial market. Well, about Automotive, I think I have made my comments. On Personal Electronics, yes, there is a consensus worldwide that we touched the bottom last year. But now this smartphone business, accessories, let's say, are stable. There is not really yet a breakthrough in the future they can offer to boost the demand. In China, for sure, there is a guy coming back on high end, clearly, Huawei that was, let's say, out for a few years, but now it's coming back and is changing the competition landscape. Again, for the rest of stand-alone electronics, lower orbit satellite booming everybody want to have this capability everywhere in the world. For the rest 5G, 6G deployment is moving, but not at the speed of light. But the only application that clearly, we see some very strong traction. Yes, is AI and server AI and also because it has induced some change in the architecture. But our takeaway, again, we classify whatever will be the timing of the restart of the industrial market? Is it Q2, is it Q3, is it Q4 a transition year? More important is the substance and the structural trend. On the structural trend, the only things we have to pay a lot of attention is what has been induced by the geopolitics issues, what has been induced by the important investment in China on the mainstream node of technology. That's the reason why -- what is important is to adapt your tactics and strategy to pay attention this change in the landscape. But on the substance of the application and so on and so forth, there is nothing changed. So this is the way I classify the situation.

Could I ask? You dropped the hint there on winning some unexpected sockets in Personal Electronics. Could you share a little bit more color?

Of course, I cannot because it will impact other companies, so I cannot...

Is it sort of around -- is it tangential around sort of existing...

It is a Sensor.

Maybe just 2 high-level questions since you've touched on both. A number of semiconductor companies are trying to articulate how AI is going to impact them. In the power and analog space, we haven't really seen it and you have a combination of MCUs, which I suspect if edge AI gets rolled out might benefit in power. Can you remind us how you think about that because the take-up obviously has exploded. And everyone is talking about the power consumption being absurdly high and unsustainably high. And at the moment, it doesn't seem to be a major concern. So I wonder how that power equation gets resolved. And whether we're not hearing the power companies talk about it too much because it's not big enough to make a difference in your overall revenue. So if you could quantify this AI opportunity for ST, that would be really useful. And then I have another high-level one.

No, it is, let's say, pretty recent because now all the ecosystem especially the Asian one, Taiwanese one and Chinese one, which are basically the company that are working the most with the [ GAFA ] working of this kind of server or with NVIDIA and so on. All the feedback of the guy was please be prepared because the power stage architecture when you move from 48-volt to 15-volt and to lower voltage is changing. And this kind of architecture will call for different drivers and different power switch, including silicon carbide one. Now this is a dynamic, which is clearly moving fast, seeing the investment on this kind of server moving forward. To know also, do not forget that it has been self-limited by the chips of NVIDIA themselves, okay? So that's the reason why it was not so stressful. And everybody from power were more focused on automotive and industrial market. But now if you believe, if you trust me it's just what I have seen traveling in Asia during my past 3, 4 weeks. This is what the customer told me, and this is what they expect from ST to be capable to equip ourselves to support them in the future. So this is where, again, ST want to participate to the era of AI, but not as a provider, of course, of the graphic processor or the PMIC, but of course taking advantage of the peripheral device that we can provide to enable this application. Well, about edge AI. Again, here, I really want to be simply consistent. We consider that AI is an add-on enabler of the business. The 1 million developers we have on STM32 that up to now, let's say, were sensitive clearly to the cost. But after to the [indiscernible], the power consumption, connectivity feature, where needed, security feature where needed now to address the application to have a better operability to have better, let's say, feature offer to the handyman, they want to have this, let's say, AI capability available. But it is -- it will not boost completely as handyman, okay? What is the motor control is a motor control. A motor control where you have capability -- better capability with machine learning for predictive maintenance and blah, blah, optimization of energy consumption. Yes, but it is a motor control. So if you need one motor, you will have one motor. So you will have not a boost of the handyman because of edge AI capability. It will be something you need as an enabler, Why? Because the developer say, "I want to offer this capability to my customer and the field of potential feature improvement I can provide now with this capability is very important. But at the end you will not buy 3 washing machines if you have edge AI in the washing machine, you will buy one. So it is important we understand this. If you are not in, of course, will become a problem, okay? So that's the reason why, again, we believe that on edge AI, the winning company will be still the company having the same recipe. First of all, the wide product portfolio they can offer. The million of developers that are with you to develop application, then your supply chain capability, your capability to introduce new product and all the features that they need. So high connectivity protocol is the need, so high security protocol if they need and neural network embedded in the right ecosystem. This is what we see. Now to see this will weight this amount of dollar. Now we can see that the market of microcontroller or industrial microprocessor will continue to grow, offering improved features. That could maintain attractive price in terms of payback for semiconductor company because we offer more features. But on another way, we have to improve the productivity. So this is the [indiscernible] this market, not as something miraculous that will create end demand suddenly. Now it will improve the feature you can offer.

Understood. And the second point is on China, you mentioned it the level of investment they've made over the last couple of years is pretty staggering as a percentage of ASML, AMAT, Lam revenue. It's just now enormous. And it's all going to more mature nodes. I think the market is worried that you might be at the wrong end of that investment phase compared to more leading-edge providers. Run us through how you think about it? And what will your place be in China over the next 5 to 10 years as I suspect they continue to invest?

No, I mean, clearly, ST, we are making, let's say, since now 2, 3 years, ourselves evolving versus China to even if we remain global for more local activities. What we can say the glocalization. And everywhere we consider it is critical. Today, up to now, we consider it was critical to implement competence and application center in China. Why? Because on all the field of electrical mobility and digital power control energy -- renewable energy, the activity in China was by far the most advanced in the world. So to have our activity here was mandatory. Then we convince ourselves that with the implication related to the trade war and the embargo on advanced digital technology, yes, China will pay much more attention on, let's call it, mainstream not so much on mainstream node. And certainly, that capacity at a certain cost will be available. And if tension will continue to raise between U.S. and China, the willingness of Chinese end market to be self-sufficient in terms of production will increase year after year. That's the reason why we have decided on everything we consider critical consistently with our strategy to lead e-mobility and power energy transformation, conversion and decarbonation to implement activity in China for production of technology. SIC clearly, is the most visible part of the iceberg because we have make a press release, signed a joint venture and so on. But we are doing the same on microcontroller on a different way. It means we are using now more and more for our iron in China foundry in China, Chinese foundries. Our assembly plant of Changan, for sure, but Chinese [indiscernible] and the target is by '26, we will have 80% of our iron microcontroller produced in China, either using technology, we have transferred 14-nanometer technology or because we port the design of our product on Chinese technology using very efficient Chinese design house. So the second step of our glocalization strategy in China is okay on all what we consider critical to lead e-mobility and industry production in China. Design, we have already designed center in China. Now it will be a matter to size the magnitude of the resources we will put. Today, we have about 5,000 people in China. For sure, we expect to grow. Well, the benefits of ST in a certain extent will be to mitigate to our CapEx because clearly, with infrastructure, we have already in Europe efficient with still some, let's say, opportunity of growth like in [indiscernible] and of course, in Catania, but mainly we'll address the market of, let's say, America, Europe and APAC and still certainly time to time China. One is very specific example imaging. I don't want to implement something related to imaging in China. And the growth, we will have in China. So today, China for us is 15% of our revenue. We know that on some markets like silicon carbide China will be the fastest growth market. So our China penetration will increase. But we want to address more and more with the China production. So for us, it is a risk, for sure, to have seen this massive investment on mainstream technology, but it is also an opportunity. But it requests that we change our mind to say we want to be in China to compete in China with local competitor, and we consider them not as a threat, but as a competitor, as we consider TI NXP on Infineon as our competitor. So it's a change of mind and -- but we consider this is the right way to address this structural change that we are facing in the competitive landscape now.

Thank you, Jean-Marc. We're sort of -- we're down to the last 10 minutes. Why don't we dive a bit more into some of the end markets, starting with the biggest one, Automotive, as you said, just over 40% of sales now. And I think it's arguably the area where the market is most skeptical in terms of the near term. You've told us at a high level, you're expecting mid-single-digit growth. But on a like-for-like basis, given some of the adjustments for last year, you're talking about 10% or more -- slightly more than 10%. Can you walk us through some of the building blocks in terms of perhaps sort of different products, different technologies, mix within the market. What is it that's giving you that confidence in another year of growth in Automotive?

No, I mean, it's clear that everywhere, the demand is sustained by either -- but whatever are, let's say, the growth rate, okay? Clearly, if you take e-mobility, if you take, let's say, ADAS so the peripheral device not the vision processor. I've already spoken about Mobileye. And the legacy where now the carmaker again introduce a lot innovation in the legacy system. So body control, lighting, airbag control, cockpit and so on and so forth. Everywhere there is an innovation or change. Again, for sure, there is no inventory replenishment, but the demand is solid and the backlog is solid. And even for ST, there is some specific technology where we are most to most. We are not yet capable to have a big flexibility in our capability to supply. Everywhere clearly, where it is a strong legacy with, let's say -- and here, I use the word quite mature technology. Certainly, there is potentially some inventory, which has been put in place just after the shortage in order to be sure that nothing wrong will happen again. Our Tier 1 customer, EMS, are correcting the situation, which was not the case 6 months ago. But this we always said. We said, yes, we are pretty sure that at a certain moment, the automotive industry will change our behavior which was, I buy, I buy, I buy, whatever the mix because I am afraid that at certain moment, I am not capable to optimize. Now they change and they start to adjust. But nothing more than within a scenario where worldwide, they should produce about 90 million vehicles and out of which about 14 million vehicle electrical battery-based. Of course, with China that will produce 9 million to 10 million vehicle with the capability to produce 11 million if they want. So this is really where -- what we see. As a matter of fact we have a backlog in our hand. That is covering our sales and operating plan above 85%, I have to say, that is making us confident that to grow, let's say, about 10% in 2024 versus 2023. Again, like-for-like removing the effect of capacity reservation fee and Mobileye is a plan that is actionable. And this is the visibility we have today. So then what scenario could happen that in H2, the end customer demand completely collapsed and so on. So far, for the time being, this is not what customers said.

And within that, in terms of the silicon carbide expectation, you've given us guidance of moving to $1.5 billion to $1.6 billion of revenue this year. Can you speak to how that revenue stream is diversifying? You've clearly had a lead customer on silicon carbide, but it looks as though with the new models coming into the market, that should be diversifying. How is that happening? And also, how are you seeing the competitive dynamics on silicon carbide?

No, no, clearly, we confirm the $1.5 billion to $1.6 billion that we give, it's not an exact number. Yes, it is diversification because our main customer is not growing, let's say, tremendously. It was expected with the introduction of second source. And also, I repeat since July last year and updated okay, at the end of the year and they modified their forecast, okay? But in due time for us to adapt ourselves. And now we have taken advantage of the diversification, both in terms of customers, still mainly on the car industry, clearly. But now on different operating model because time to time, it is module. Time to time, it is packaged. But now we see more and more also demand on what we call a nongood dye and especially from Asia, where you have many competitive okay module makers that are associated either with industrial or with carmaker. They need a dye, okay. And really, there is many opportunity for us to grow our revenue. So the $1.5 billion, $1.6 billion, yes, the weight of our main customer will decrease, but as expected no more than that. We would have preferred to have our main customer, the original forecast they provide to us 2 years ago. But I am not so sure that we would have been capable to supply. For long term -- medium, long term, again, our ambition is to reach $2 billion in '25 and then to reach $5 billion before 2023, and we equipped ourselves to address this market. So with our 3 hub. So the hub of Catania, where you will have fully integrated from the raw material including to the module, the hub of Singapore and the hub of Chongqing. So with Sanan, our joint venture. And these 3 hubs will be, for sure, capable to deliver at least $5 billion revenue for ST within the next 5 to 7 years.

We're running low on time. Let's see if we can squeeze 1 or 2 more in. Anyone else in the audience have some questions they want to ask. I've always -- yes, [ Julia. ]

Real quick on silicon carbide. It's -- for the nonspecialist, it's hard to understand how we've gone from silicon carbide being almost impossible to make to now potentially oversupply as China seems to be making quite a lot of it, and it seems to be relatively decent. There's diverging views on that, but it's a very radical change in the space of 3 years. Do you agree with this idea that the quality of the wafers you're seeing in China is good and what has driven this incredible learning curve?

No, I mean, in terms of raw material, means the wafer of silicon carbide. Yes I confirm to you that the few suppliers, not tens of suppliers. A few suppliers we are assessing on engineering sample provide a good quality. First of all, it is engineering sample. It is not mass production. I guess you were aware that mass production still show some random surprise in the market, okay? Nobody is completely immune to have suddenly a major quality issue because the raw material that has been implemented we have only. So yes there is, let's say, this supply chain implemented. But again, in China, is not a surprise. Why? Because this ecosystem addressing the e-mobility or what they call the digital power, they have understood very fast that critical device are IGBT, SiC MOFET, microcontroller and they equip themselves. So I met a carmaker, a customer in China where the guy he has under his responsibility car factory, but he has a module maker. And he has a raw material provider, which is a subsidiary of course, which is a JV, but they want to control it. And again, that's the reason why at ST, we have taken this decision of JV with Sanan in China, okay? As traditionally our manufacturing strategy was to remain localized in Italy and in France and in Singapore. But will this situation create another capacity. Again, I repeat our position -- I repeat my strategic position. We do believe that -- to address this field of power electronics the recipe is to be broad range. Why? Because the customer they will change their mind. Some customers want to have a pure silicon carbide. Some customers now want to have mix and match, silicon carbide and IGBT on inverter. Some customers want to have silicon carbide an onboard charger because they don't believe GaN is a good technology. Some customers want GaN. Some customers want to have a very specific driver to make a differentiation on their power stage. So this is the point number 1. So the operating model to be vertically integrated. Then the second point, the market we address. We address e-mobility, we address the power energy market and motor control market. And now we address AI server market. These 3 markets should represent $15 billion to $18 billion by 2030. Between now and 2030, is at a certain moment of time some capacity implemented equal or above the demand. Why not? But it is business as usual, I have to say. Again, assuming that conversion to 200-millimeter will be done smoothly without quality issue, that conversion to shrink technology will be done smoothly without quality issue, means all improvement will be done smoothly. Up to now, it is not the case. I'm sorry. Nobody in the market can say on silicon carbide, the life is very long quiet river. Everybody has faced issue because at a certain moment the material behavior of [indiscernible] is different than silicon. So yes, '25, '26, when massively people will convert to 8-inch, maybe at a certain moment, you will have slightly overcapacities and demand. But what does it mean? We will come back to, let's say, shorter lead time first. And second point, customers will decide to change their architecture instead to remain cautiously on the power MOSFET or power IGBT, acknowledging that there is no capacity available, they will accelerate the conversion to silicon carbide because for the time being, there is still people thinking short edge will come. And that's the reason why they put a mix-and-match approach on inverter, not only driven by cost, driven also to secure and to have flexibility the case of shortage occur. So I expect simply that now silicon carbide in the next 10 years will come back in a normal path of semiconductor industry, but it's not a memory.

All right. Perfect way to end on silicon carbide not being like memory. So with that, we are out of time. Jean-Marc, Lorenzo, thank you very much for your comments. Thank you all for joining us.