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

Good afternoon, everybody. Welcome back to the afternoon session on the opening day of Citi's European TMT Conference. I'm Andrew Gardiner. I have the responsibility of covering the semiconductor sector here for Citi in Europe, and it's our pleasure this afternoon to have STMicroelectronics, in particular, Jean-Marc Chery, President and CEO; and we have Celine Berthier and Alessandra Fumagalli from Investor Relations with us as well. I've got a list of questions. Jean-Marc has kindly agreed to discuss with us. We'll have some time towards the end to open it up for Q&A if any of you in the room have questions as well. So first of all, thank you very much for coming, Jean-Marc.

Thank you, to invite me.

It's our pleasure. Perhaps we could start on the broader cycle, if we think of how you guys frame things with results back in January, you're acknowledging softness in certain parts of the market, continued strength in other parts of the market. Can you just give us a bit of a high-level update as to where things stand at the moment in terms of perhaps the broader cycle and the different end markets?

No, Mike, I can confirm exactly what I said at the end of January, early February. Now with some additional visibility because, of course, we visited customers, both in Asia and in U.S. It is clear that the automotive market and what we call the industrial B2B, so power energy, okay, robotics, automation, health care, this kind of stuff is very solid. And this is confirming by our backlog, which is still okay representing 6 to 8.25 of revenue and capacity. The consumer stuff, so personal electronics, personal computer is still weak. And there is a kind of consensus that you should restart in H2, having absorbed the excess of inventory, which has been built last year on personal computers and smartphone or accessories, depending definitively about the consumer behavior, both in America, Europe and China since China has reopened, let's say, early December. So this is this consensus. So far, for ST, that's the reason why, okay, we provided the indication for the year between $16.8 billion to $17.8 billion. I confirm that if the consumer market smoothly restart in H2, pushing the demand on smartphone and personal computers, we are exposed, okay, to this market, not significantly, but we are exposed. And Automotive and Industrial remain as we see today, clearly, we will be at the upper range of our indication. If the consumer market restart, is postponed next year, we could be at the midrange. So this is exactly the dynamic we see today.

And how would you characterize the, I suppose, like the push and pull within the market? I mean you've talked about capacity constraints. You talked about sort of where demand is. What's the bigger problem for you at the moment?

The complexity is maybe still higher compared to last year. Why? Because on some technology clusters and end market, we are still depending our capability to increase our own capacity. So it means we are dependent on our vendors. And especially, we are dependent on the equipment vendors, the process equipment vendors. I have to say that on assembly and test, the situation is improving. On process equipment, the situation is still under stretch. So we have to monitor very carefully the on-time delivery of the process equipment in order to enable ST to deliver the ramp-up in critical technology like the 14-nanometer, or the 28-nanometer, the silicon carbide, the IGBTs, MOSFET, the VIPower, also BCD and so on. So this is one kind of constraints. And here, we have a backlog, which is well above our capacity. Then on the other hand, on the consumer market, we have not, let's say, fully the backlog to deliver our forecast. And here, it's a matter of order entry. And we know that how many orders we are missing to fulfill our, let's say, high-end indication of our range for the year. So this year, okay, we have to manage still capacity constraints from our own supply chain. Implement on time our capacity increase and support our customer Automotive industrial and consumer staff, okay, it's a matter of order entry.

And you mentioned the full year outlook and sort of you touched on some of the swing factors there. I mean, you sounded particularly confident in recent presentations about how the year is looking, and I suppose, in particular on the bookings. Can you remind us of how you've addressed the level of backlog and concerns over double ordering that might be in there and sort of how you've cleaned that up to give you that confidence in the outlook for this year?

No, when we -- at the midpoint of our indication, so $17.3 billion, okay, so we will grow, let's say, 6%, 7%. H1, okay, with the midpoint of the indication of Q1, I have to say that H1, we will grow above 15% year-over-year. And H2 versus H2 last year will be at the midpoint, okay, with a slight growth. But here, that the reason why I communicated to the market that H2 for us will be special. Why? Because, okay, we will have a mix change in important, let's say, customer programs, where we will have less revenue, a significant decrease of revenue, but better gross margin, okay? And we absolutely do not lost any silicon content, but it is a mix change in the overall system. So that's the reason why at the midpoint, okay, the sequence will be average, we will grow 6%, 7%. H1, 15% year-over-year. H2 basically slight growth or flattish. And this will provide the $17.3 billion. Well, H1, basically, the backlog is there, okay? Order missing on consumer is quite small, okay? So again, what is important will be H2 for us in order to decide at which level we will run our manufacturing to be at the per to be at the midrange. Double ordering. Double ordering, I repeat, okay, what we monitor, okay, is the inventory at our distributor level. Yes, we have seen some inventory increasing when you look at the variation of inventory increase, okay, in a dynamic way, but they are still okay in inventory turn, which are totally, let's say, acceptable to run business in normal condition, okay? We are not detecting excess of inventory. This is something we are monitoring very carefully. And that's the reason why we manage our POP, POS to shift to distributors in order to not build excess of inventory. Well, then at OEM level on the car industry or industrial level, we have not detected excess inventories at all.

I suppose, with that combination of demand, supply, inventory, where does that leave us today in terms of the pricing discussions you're having and the kind of visibility you've got not only through this year, but into next year?

But here, we are very selective again, on Automotive and industrial B2B, there is price stable, sometimes, okay, through negotiation of the supply, there is still some price increase. Well, then there is some specific package or technology cluster end consumer market, where definitively, we see some price discussion and price pressure. But overall, we do believe that the price at company level will be neutral for our gross profit in 2023.

And from a longer-term perspective on pricing, do you think things have really changed in the industry over the last few years? Or are we going to go back to more normal dynamic in time in terms of the price erosion or given your end market exposure and the type of products you have, the increasing move towards longer-term supply arrangements, have things shifted?

For the market we address, Automotive and Industrial, I think there is many, let's say, points that change. But first of all, technology that are used to address embedded electronics. So power solution, our embedded processing solution or sensing solution. Now asking technology, either in 300-millimeter or for a new material like the wide bandgap in 200-millimeter. And this is calling for new investments. This will not be done by investment done in the past on a fully depreciated fab, because since 10 years the 300-millimeter fab are invested on FinFET and to enable microcontroller real-time or advanced analog and so on and so forth, you need 28-nanometer or 14-nanometer technology or 20-nanometer technology, calling for new investment. So nobody can say, I am enjoying a fully depreciated fab, except maybe to support the business. So this is the point number one. And clearly, the car industry and Industrial B2B, they need this technology to make their transformation. So decarbonization, automation, electrification and digitalization on top of the advanced processor. So this is when I expect that change a lot. In 10 years ago, in the past you have much of 8-inch available and you have some 12-inch available. Now it's no more the case. And it's for 200-millimeter fab is the same. So if you discuss with the equipment maker, they will explain to you that they are investing in 200-millimeter. But this brand new tool will cost investment, so you cannot leverage the depreciation too much. And in front of customers they cannot enjoy excess of capacity, making price pressure on semiconductor. But then the second element is inflation. This thing something looking, let's call it, structural for a while, because of the energy, because of some shortage in commodity and so on and so forth. So you will have inflation. More and then there is a labor cost. Labor costs for the time being will increase, maybe not at the level of inflation. It's our duty to contain it by the productivity improvement and automation and amortization. But labor costs will have to be decently increased. So you have driver of overall cost increase that the customer cannot blindly say we don't care. If not, they will have not, let's say, capacity allocated to them. So yes, of course, we will come back to the normal price discussion where you need to have a price road map, but price road map will improve with the substance of what you are doing. So technology improvement, process improvement, productivity improvement. Not price pressure related to offer versus demand, where basically you make some speculation about the capacity. The only exception I am seeing about this new condition is what is happening in China, because China is increasing capacity on the mature technology importantly, but it will be more to address the local market. So for sure, for a semiconductor company like us, certainly, this is something where we have to pay attention and we have to work with some Chinese player in order to leverage this situation.

Well, thank you. The -- so we've got a good idea at, I suppose, group level of how things are going. Maybe we can dive into some of the end markets. Automotive, in particular, clearly, in the news at the moment. Starting in terms of where we are in the cycle, I mean, how would you frame the robustness of demand from Automotive? Any change in order rates or behavior from those customers relative to what we've been seeing?

No, because the carmakers, they are executing their transformation. So more electrification is, let's say, well-known transformation. There is a progressively the change of architecture as software-defined vehicle, and the management by domain in order also to save some copper, let's say, in the car. So this is going on. Situation improving, most likely some carmaker that during the shortage, came back to some subsystem, old fashion. They will again introduce some subsystem improve, so it means with more semiconductor. And this is based on consensus of production vehicles about 83 million, 85 million vehicles. So today, what we see on the market is this happening. So production at this level, let's say, and the demand on our components to enable both transformation, change of architecture and coming back to new feature introduction. But of course, we monitor as everybody, the end market and we know that lead time for car -- for new car is improving, but this is basically not totally a surprise.

Okay. So I've given you 15 minutes or so to warm up. Let's get to the big topic from the last week or so. Your -- one of your big customers in Automotive, in silicon carbide, Tesla said they're going to reduce their need for silicon carbide in the next generation by 75%, a big headline number. I'm interested -- I know you can't say too much about what specific customers are doing, but perhaps you can frame it with your perspective on the road map for the industry in terms of how carmakers move from selling relatively high end, at least outside of China, should we say relatively high-end vehicles, electric vehicles and how do we get down to the mass market? What is the role for silicon carbide? What's the role for silicon IGBT still as we try and move down to lower price points?

Maybe the -- I can't answer specifically about this customer, but let's think about power overall. There is clearly a major acceleration of power solution, let's say, spread in the market, driven by this electrification of the mobility and the decarbonization of the industry, with energy renewable. And here I can list many, many applications. Clearly, inverters, onboard chargers, microinverter for residential, inverter for, let's say, base station for wind farm or, let's say, solar farm, motor control of the industry and so on. All this solutions, power solution will call for what? They will call for the best tradeoff between a device, which is capable to power switch high-power at a high frequency. And definitively, there is a cost of the solution. And that's the reason why if you remember well, I think we shared this chart at the Capital Market Day that, you have a curve where you have the power switch and you have the frequency and you have the landscape of the power solution, which is a mix between silicon carbide, GaN and IGBT and you have a MOSFET high voltage. This doesn't change. And it is based on this that we have built our $20 billion ambition, out of which, let's say, the power will represent, the number we share, but out of which the silicon carbide will have a presence more than $1 billion in 2023, $2 billion in 2025, achieving $20 million, so 10% of our revenue and growing, to what we say, around $5 billion in 2030 or below 2030. This doesn't change. It's not because -- last week, my friend, colleague, made this announcement that this changed our path. Well, this, I would like to repeat. So, this mix and match, let's say, strategy for an application is something we know perfectly. And our numbers remain exactly the same. Saying that, to be more specific on mobility, well, it is clear that I would like to confirm again that, when you want to run electrical engine on the battery at a higher voltage possible and switching the power at the higher frequency, the solution is a silicon carbide period. And then, you have also other implications about how you dissipate, let's say, the heart of the electronics and the overall cost of ownership of the solution. On top of that, the silicon carbide has many cost driver down. So wafer size conversion, new technology, whatever is a new technology. But for us, the SmartSiC as an example, the scaling, definitively, when you run fabs, at 10,000 or 20,000 wafer per week is not the same things when you run fab at 3,000 wafer per week. So scaling, wafer size conversion, technology improvement cost driver are none of thing. So more and more the silicon carbide will offer competitive cost of ownership versus IGBT, because the IGBT today is either on mature 8-inch or on 12-inch. So going more than some thought in terms of cost. However, you cannot prevent engineering to say, I can set up an optimum mixing and matching the technology. So today, on car, as an example, there is a car where you have the first inverter based on silicon carbide and the second inverter based on IGBT is not optimum because, the mileage is shorter. But, the overall cost of the solution is acceptable. I cannot prevent that engineering say, I want to make this mix as much inside inverter. Well, okay, fair enough. But after you have some technical challenge. That must be sold. As an example, the driver would drive IGBT based or a SiC-based is not the same. So here, you need to have a universal driver, which is not so easy technically. So you have other technical challenge definitively. So yes, certainly, this solution, will happen, but also, it's a flexible solution. Imagine in case of shortage, you have a shortage of silicon carbide MOSFET. You have a better supply on IGBT. You can make this mix and match with lower performances, but at least, your car is running. So more common than that, I have not -- I think it's an engineering activity that will lend certainly on the solution, but it is part of the overall power strategy where you will mix, GaN, SiC and IGBT in order to offer the best tradeoff when you need a best tradeoff in terms of performance, power switch, frequency and the cost of ownership.

So to hear you answer it that way. It feels to me like while we, in the financial community, may have been a bit surprised at some of the headline numbers from Tesla a couple of weeks ago. From your point of view, sitting within the industry, speaking to OEM partners, Tier 1s, this is something that has already been on the road map for some time, and it's all consistent from your perspective.

Exactly. Yes.

And when you talk about this idea of mix and match, is that something that's almost going to be, I don't know, for lack of a better word, sort of pin compatible, that sort of drop in, drop out or much more complicated than that?

No, it's -- after, you can drive it certainly by software, it is more complicated in the hardware, in the hardware, I repeat. You need to have a power driver, which is universal versus IGBT or SiC MOSFET. And if you find a way, after you can drive it by software. So that's the reason why this solution is maybe interesting for flexibility.

I suppose while we're on that topic, how are you feeling about gallium nitride now? It's clearly earlier in terms of its maturation and adoption curve, but it feels like there's perhaps a blurring as to where that solution may ultimately sit within the Automotive space?

GaN, again, it's interesting animal because, in GaN, you can address basically, again, the 3 markets we want to address is power management in personal electronics, like chargers, fast charger, but on any operated battery-based tool. If you want to have fast charger and a smaller charger, the GaN device is really good. And still today, we are already in this business. We are using TSMC technology. And we are in competition with a well-known company called GaN Systems. And we are a competitor of GaN system. And we tackle this market. More interesting for us is automotive and industrial market. Because the GaN as an interesting, let's say, capability either to switch power at higher frequencies at the silicon carbide, but lower, lower power. So -- and the second interest of the GaN could be bidirectional. And for onboard charger, it will be certainly very interesting. When you want to give back, the energy, you have stored your battery, whatever is a battery, could be a battery on the vehicle, but could be a battery that you use to store the energy you have from your solar panel as an example. So for the GaN, everywhere, you will have a charger of battery with the battery management system, where you need to load fast the battery and to be the bidirectional is an interesting solution. And ST to address this market, that's the reason why, we have done a few years ago, a small acquisition with Exagan where people have a deep know-how of the critical process step of this technology, the epitaxy. Good skill engineers in Southwest of France, Toulose. And we are developing our own technology, our own solution in combination with and our strategic plan, as we have always shared is based on the first industrialization line, we have set up in 2 for GaN. We are visaged to build 8-inch fab in tool to address this market. And this is our ambition and GaN will contribute to the revenue of ST significantly, but beyond the $20 billion ambition of '25.

And if we could go back to silicon carbide and sort of think about the competitive dynamics, you at ST have had an early lead, you've sort of accelerated away, $700 million of revenue last year. You're now saying more than $1 billion this year. There's lots of companies out there trying to chase you down. Some are closer than others in terms of that race. I'd just be interested in how you would frame the competitive dynamics at the moment and how you feel about your wins and losses as we look out to the future?

Well, we -- ST, we have decided to set up our communication based on KPI, which are very clear, is revenue. And now we give an indication, for the year, we give an indication at 2 years and then at 5 years, so revenue. And to sustain, we are communicating the number of programs and customers, the orders. And on Automotive and Industrial, because we will never, do on consumer, because generally, on the consumer, you have multisource, you don't know exactly what you will do. But on Automotive and Industrial, when you have your funnel of opportunities awarded, quite accurately how many revenue it will generate. So that's the reason why, first of all, we confirm and we have always said, we want to stick with a sustainable and profitable 30% market share. And looking at the market dynamic overall, this should translate in, above 1, 2 in 2025 and basically 5 before 2030. And we have many other opportunities. We have set up a manufacturing strategy consistent with this plan. Today, if we encompass Singapore as a wafer fab, Catania, existing Catania and the new Catania, which is integrated with our raw material factory that we are building, all this infrastructure and the related back end and module maker can enable this business plan. For us now, we start to question ourselves about -- to go beyond $5 million, between $5 million to $10 million. And here, I have to say there is one important consideration is China. Because China will be one of the main customer of silicon carbide-based solution. And for sure, because of the current, let's say, geopolitic situation, it will be challenging, to not produce something in China. And this is something that ST is considering very seriously, in order, to boost our business plan. Well, then after we have also, let's say, option and flexibility to continuously increase our capability in Catania. We have space available. It is our competence center. We have agreement with local university for talent and so on and so forth. So we are executing our plan. Well then, competition is healthy. It is a booster for us to introduce our road map of technology. So 3 in production, 4 in production this year. Then the 5 still planned up, but boosted by the 4 by SmartSiC. And then when we will introduce the 6th generation, we will change architecture, and this we are working on. Then we will convert in 200 millimeter, and we are installing, our own raw material substrate fabrication. So we are executing our plan, well, then after, again, the competition is healthy, is a good booster.

You mentioned the sort of the architectural road map. When you talk about Gen 6, you say it's a new architecture. You don't specifically say trench along the way that many in the industry are talking about tranches. Is it...

I don't want my competitors to understand what they are doing. What we are doing as a R&D and technology, but for sure, it will not be only tranche.

Okay. And you also mentioned China there where you are talking about making chips in China for the Chinese market. What about the local Chinese analog and power chip guys, who have -- they're relatively young, but they certainly have big ambition in this space? Do you see them as credible competitors in time?

No, we have -- Of course, we have to respect, as a basic [ principle ], China or any player also in America while you see now you have a microchip coming in, a certainly coiled is coming in as well. So -- but it's a difficult technology definitively because the reliability, the emission profile, automotive and industrial is quite challenging. And the failure mechanism are really different than the silicon. Today, in China, we see raw material orders, clearly. We tested them, and they are providing good quality substrate. We don't see yet device maker, except for diodes. And for MOSFET, most likely could come. That's the reason why, we consider adequate at a certain moment to make some strategic decisions to partner in China, and enable some production capacity to address the market, which will be, let's say, 40% of the total silicon carbide market. Today, China in terms of electrical car, electrical mobility, charging station, again, solar farm, solar panel is accelerating like a hell. So this is a market that ST, doesn't want to be out. We want to lead this market.

Okay. Perhaps if we can switch gears and look at some of the other divisions. Analog, MEMS and Sensors and imaging, in particular, in some of your opening comments referred to the change in business this year, some revenue pressure for a key customer in the second half, but not necessarily in terms of gross margin. Any additional sort of color you can provide as to sort of what is happening there in terms of the mix? And perhaps as we look beyond the 2023 cycle into '24 and '25, is there still innovation to come, in particular, for imaging? Do we -- are we finally getting to a point where we can drop the whole thing beneath the screen? What kind of things can you anticipate?

I repeat what I have always candidly said. We have a perfect visibility of what will happen. Linked to the volume of the phones sold for the next 3 years. And we are working in with our customers for the solution that will enable customer on differentiation and customer request. I guess, everybody understand that for this device, I cannot absolutely comment any detail, because it's very, very sensitive. But again, today, the -- really the relation we have with this customer is very well balanced between transaction and R&D. And I know exactly how to drive the imager. Then saying that, what is interesting for Imager is the diversification of course, we see some diversification in the Android comp. There is other smartphone maker asking for face ID, let's say, solution or any time of flight solution. But more and more, carmakers inside the car to monitor the driver, you need to have solution based on kind of face recognition, because you need to detect if the driver is in condition to drive. And this is an important, let's say, market for imaging for the next 3 years. Beyond that, I cannot disclose too much, because it is also related to the innovation we are doing with our main customers.

Okay. Understood. Just got a couple on sort of the financial side of things and then perhaps we can open it up to questions in the time that we've got left. You've given quite clear gross margin guidance for the first quarter. You still sounded reasonably optimistic into the second half of the year. What are the key puts and takes in terms of where we may end up in terms of gross margin for this year?

Well, we guide 48% for Q1. I am not specifically anxious. I think we will deliver it clearly. And I think I feel confident on the dynamic we share with you a few weeks ago that on Q2 will be very similar. And then on Q3, we will have to pay the ramp-up of Agrate. That will be no more in start-up cost will be, let's say, a manufacturing mode. And yes waiting for the critical mass, Agrate will not contribute positively to the gross margin at an average. So yes, in H2, we will have, let's say, lower gross margin versus H1 because of Agrate. Well then after, it's not a secret that the manufacturing input cost are under pressure with energy and some inflation. Thanks to coal growing in this year, so we'll improve the productivity. So we have also other opportunity improvement. We will increase our volume, because either at the midpoint of the range we have given or at the upper range, we will have some scaling effect that will mitigate the productivity. And then, we consider the price will be neutral. So that's the reason why, I confirm that connected to the range of revenue we have given, so $16.8 billion to $17.8 billion, we will deliver above 47% gross margin.

Okay. That's great. I've got a long list of questions. I could go for ages, but in the interest of being polite to our audience. Does anyone have any questions in the room?

Just on the AMS, you mentioned you have perfect visibility for the next 3 years. When you say visibility, are you talking about turnover, are you talking about what programs are on and you hope to sort of be designed in? And are we talking about proliferation beyond the 1 device that is 220 million devices a year? So are you looking to sort of move away or add outside the sort of main program?

Well -- sorry, the 3-year visibility is designed win and what we call engage programs, which can enable us to elaborate and accurate . This is what -- the reason why we have this 3-year visibility. Beyond that, there is some disruption on which we are working. And on which soon, when we will enter in the frame of 3 years, there is some milestone and when this milestone are positively passed, after it's granted. So this is the way we are working. So I confirm that I have a pretty decent sales plan and operating plan connected to the volume of phone that will be sold, but also PC and tablets and accessories that for the time being is, let's say, reliable. So this is what I am seeing.

Other questions in the room? Amit?

Jean-Marc, could you comment specifically around microcontrollers? And how do you see your confidence in levels of microcontrollers within distribution, inventory levels there? Because back in 2018, I think that was what surprised us with the changes in microcontrollers.

You have a good memory. No but today, microcontroller, there is many dynamics. But it is clear if we speak about mainstream microcontroller 8-bit or 32-bit addressing consumer application. Whatever is a pure consumer or what we call consumer industrial, so, let's say, white goods or battery-operated tools for consumers. But we are monitoring clearly the dynamic because, we see some inventory increase but we control it. Clearly, we control it. We control the POS. We control the POP. For sure, we are given more freedom in the way the distributor put orders. So with the flexibility to cancel it beyond the lead time of the manufacturing. That is business as usual. So I have to say on this part of the market, we are coming back to a pretty decently time inventory turn at distribution, which enables the distributor to make their business. Means flexibility, take short-term order and so on and so forth. Well, then there is microcontroller for Industrial. So high performance in the microcontroller, 40-nanometer. Here, the situation is very similar on Automotive. So still only time, still some capacity limitation and sometimes shortage and sometimes allocation. Then after you have ultra-low power microcontroller, a connected device with security and so on. So more, let's say, application-specific device, idea, the demand is solid. More it's secure microcontroller, there is no specific comment. So microcontroller is exactly the image of the market, so everywhere you have a microcontroller, addressing the Industrial market, B2B or addressing an application, which is application-specific with features, so security, artificial intelligence, connectivity, the market is solid and strong. And everywhere you are on consumer, you are coming back to a normal situation.

Anthony, do you have a question?

Yes, I was just wondering on 2 elements. Price, there is some carryover from simple calculation. So your guidance assumes some pricing decline in the second half. And then when we look at mix, last year was phenomenal in terms of mix driving revenue, so things like 13%. What drove that? And was there anything one-off in nature in that? And how do we think about when you talk about long-term guidance of top line, why it is mix contribute going forward? Is this any of that being brought forward in programs?

No. What is happening on the mix is consistent, sorry, with the strategy we have set up in 2018 and 2019 where we really want to focus on the smart, let's say, mobility and smart industry. And I agree with you that the post-COVID period has certainly accelerated this dynamic and strategy with us. And that's the reason why I mentioned that at the midpoint, of the indication of 2023, Automotive and Industrial market will represent 70% of this. This was anticipated only to arrive in 2025. But here, we have clearly an acceleration. And it is very beneficial for the mix, definitively. Then after it's inside the verticals itself. So inside the vertical itself, it is clear that where you are on, let's say, the electrification, moving forward mixing the customer we have versus a historical customer, it's positive and accretive for the gross margin of ST for silicon carbide and Industrial customers coming in as well. Expanding ourselves on different system like the battery management system, same. Going to more advanced technology on the legacy of automotive braking system, door control, body control, moving from old fashioned technology, at 0.5 micron or 0.25 micron. Going forward [ 110-nanometer ], it's improving also the margin. So there is 2 elements. Element is the weight of Automotive and Industrial versus total ST. In fact, we are 2 years in advance. And in each segment, you have also mix benefits related to the introduction of our new technology and a new application. This, of course, okay, we have taken benefits in a good price enrollment last year. But again, we consider that the price of the will remain pretty decent. Why? Because overall competition landscape and manufacturing landscape and supply landscape is totally different a few years ago. And now we are entering in the area where, again, I repeat, if you want to enable a car or industrial application, you have to invest in semiconductor wafer fab at 300-millimeter or wide bandgap wafer fab. So it's $15 for each dollar of additional revenue. So it's a cost. It's -- you have depreciation. Inflation is still there, even if the risk of a small relaxation, but inflation is there. Labor cost, in a certain extent, you do not increase the salary as you were increasing a few years ago. Productivity is still there, mitigating a lot, but you have ingredients that are pushing to continue to have a price environment, which is based on the substance of what you are doing and not a speculation or offer versus demand.

Well, with that, we are out of time. So Jean-Marc on behalf of Citi, thank you very much for joining us today. Very much appreciated your comments.

And thank you for my voice, it's allergy.

You made it.

Thank you.