Shall we start? Okay, so good morning, everyone. This is day three of Morgan Stanley's TMT Conference 2024. And I'm very pleased to say that we have both Jean-Marc Chery and Lorenzo Grandi, the CEO and CFO, respectively, of STMicro on the stage.
Good morning.
Welcome to Barcelona.
Thank you.
Thank you very much. I think maybe just to start off, we've clearly had the Capital Markets Day earlier this week, and there was some clear messaging coming out of the event. So maybe you could share with us what that messaging was and maybe some of the key assumptions behind it. Thanks.
I will start, and Lorenzo will give complementary input. We have defined at the Capital Markets Day. We share an actionable target of revenue at $18 billion and with an EBIT margin at 22%-24% within the planning horizon of 2027-2028. The main revenue driver, addressing a market that is supposed to grow at 5% compound average growth rate, the main growth driver for us will be the analog and the sensors that are mainly related to our high premium personal electronic business with one of our key customers. The second main driver will be power electronics. Inside power electronics, mainly the power MOSFET and the diode on silicon carbide, and here driven by clearly the electrification of the mobility and the industrial market.
The third growth driver will be the general purpose microcontrollers after 2024 and 2025 correction year, but I will come back on it. The main success factor of the EBIT margin target will be, on one side, reshaping of our manufacturing footprint and infrastructure. For silicon technology, will be some capacity, let's say, removal on mature technology, 6-inch or 8-inch, and accelerating the 300 mm or the 12-inch. On silicon carbide, will be the conversion to 200 mm, doubling the capacity here, and removing the 6-inch capacity. Of course, the other side will be some cost-based initiative resizing to leverage our growth and to leverage our productivity improvement. 18 billion, 22%, 24% EBIT, growth driver based on the three main points I mentioned, and critical success factor is the execution of our plan.
It is clear that we consider 2025 as still a transition year, and we will see [six] and [seven] as an acceleration. 2025 overall, we have good visibility on the automotive market and on personal electronics and computer and communication. Indeed, on automotive, some risk still exists if there is, let's say, the competition with the Chinese car maker. Maybe in Europe, what is related, the famous CAFE, so the fine on non-electrical vehicles, if the car makers are not at a minimum point. Basically, what we have as a visibility is really appropriate. Where we will operate still in 2025 with a low visibility is what is related to the industrial market. I would like to recall very simply that for semiconductor, the industrial market is power energy, is automation, robotics. It is, let's say, white goods. It is a battery-operated tool.
It is healthcare and medical. It is space and military. Here, the visibility is pretty low. There are two possibilities for 2025. There is a bear approach where we see the semi market flattish, and there is a more bull view where moving forward in Q2 and H2, we see an acceleration of the growth, resuming a situation where inventory has been corrected and the end demand is growing smoothly, so this is 2025 as a transition year, acceleration. To close the main message, we believe that beyond 2027-2028, we will have a booster, mainly driven by the megatrend, so electrification of the mobility, decarbonation of the industry, but AI, and AI, basically two ways.
This is AI related to the servers, where ST wants to be a key player in the power supply of AI servers and wants to be a key player in the connectivity, communication with silicon photonics. We believe that beyond 2027-2028, the Edge AI, [ST] Edge AI, will start to be boosted because application now will fly. For each of blocks, so connectivity, Edge AI, and let's say power, we believe we can extract $1.5 billion-$1.7 billion additional revenue. That's the reason why exiting 2027-2028 with $18 billion. We do believe we can reach $20 billion plus by the end of the decade. I guess this is the main message.
Yeah. So maybe there's a lot we went through there, a lot of touch points. So maybe we'll go through some of those individually. Maybe Lorenzo, if I turn to you next, just talking to the cost restructuring that Jean-Marc pointed out, some discussion in the investor community around how that splits between the cost-based management, the treatment of 300- mm acceleration, but also looking at headcount reductions and how much each of those play in that high triple-digit million savings. So maybe you could help us understand the mix.
On our plan, I would say there are three main vectors that we can consider, let's say, when we look at the plan. The first one is something that we have already implemented, and this is the reorganization that we have done in ST starting at the beginning of this year. We have, let's say, split our organization substantially in two groups. These two groups are characterized by similar technology and product development. It means that we have somehow synergized our effort in R&D. We have eliminated some duplication that in the previous organization was existing, and we have somehow boosted our ability, let's say, to have more aggressive roadmaps for our products. And definitely, this transforming in time to market. This, at the end, what does it mean? It means that at the end, to do and to follow our innovation, to improve our innovation, we need less resources.
I mean, we need to add, if you want, less resources. This is the first element. So at the end, what we intend to do is to protect our R&D, but at the same time, let's say, not to increase significantly in terms of our R&D cost, as we think that this organization is much more effective in this respect. Then what is said by Jean-Marc, let's say, the acceleration to the transformation of our manufacturing footprint toward the 300 mm. This will be done leveraging on the infrastructure that we have created over the last years, in which our expense, to say, our CapEx- to- sales ratio was increasing significantly. It means we were in the range of 20% CapEx- to- sales ratio in the last years, level that we believe is not sustainable.
But in any case, this has, let's say, allowed us to create an infrastructure that now we can leverage in order to have, let's say, something that is more modern, more efficient, and more automated. The CapEx- to- sales ratio will decline, and we will move significant production from the 200 mm to the 300 mm for the silicon, from the 150 millimeter to the 200 millimeter for the silicon carbide. For the silicon, substantially, the overall capacity will not increase significantly. We remain substantially flat, but with a cost, let's say, much lower, as we will for sure reduce the activity in 200 mm, especially for where the 200 mm is not efficient. The silicon carbide, moving from 150 mm- 200 mm, will represent also here a cost saving, even if in silicon carbide we are growing and the capacity will grow.
Then the third element is on expenses, especially when we look at our G&A. Here, we intend to reduce our level of people in our G&A overall worldwide. This is thanks to the fact that over the last years, we have done investment in terms of digitalization of our process, rationalization of our process. Now we are in the position, let's say, to do same things, even more things with less headcount. But this will be done mainly, let's say, through attrition. It means not replacing people, early retirement, maybe in some areas, some layoff. All in all, we will achieve this triple-digit, high triple-digit, let's say, overall savings in respect to the current cost structure, the current cost structure that we have in 2024. But this will be, let's say, a combination of cash item and non-cash item.
I think these are the three main actions that we have taken. For the expenses, we think that you will start to see some benefit already in 2025. Of course, for the COGS, it will take a little bit longer, as we here need, let's say, to transfer process, transfer product, requalify product. But I think starting 2026 and with an acceleration in 2027, there will be a clear impact of saving in our COGS.
Gotcha. Pretty clear. So Jean-Marc, it seemed to us, at least as you guide out to next year on sales, Q1 does seem to have a below seasonality profile. And clearly, there are inventory corrections out there, both industrial and potentially in autos as well. Can you walk us through the dynamics that you're currently seeing, maybe reference distribution and perhaps visibility to end market as well?
2025. Yes. First of all, on Q1, the usual seasonality is -11%. Then we have, especially in 2025, a mechanical one, which is 6% of calendar day less, which is impacting as well. You see one of our competitors share - 18% of their Q1 forecast. So clearly, we will have a pretty significant Q1 below seasonality. Then moving forward, moving forward, I repeat, entering in 2025 on automotive, personal electronics, and let's say communication equipment and computer peripheral, we have a visibility, which is usual one. So it means between 70%- 80%. Again, we know that on automotive is mainly on frame orders, and that frame order could be adjusted. Here, I repeat, the risk is maybe in Europe. If the car makers, they have to adjust their production to avoid the fine. We expect that the story of [CAFE] will be fixed sooner or later.
On personal electronics and computers and communications, honestly, the visibility is appropriate, so we are confident. Here, it's clear that it will be backloaded. It is backloaded as usual, and it is more backloaded than as usual. Why? Because for us, H1 with our main customer will be released at the lowest point. And in H2, we have a new device [per] version with our first customer that will accelerate the H2 revenue. Then the critical point is, again, clearly industrial. Industrial, as I said during my summary of the Capital Markets Day, there are two scenarios. One scenario is a bear scenario, means 2025 flattish versus 2024, and basically all the quarters very similar, except Q1 because Chinese New Year, because less calendar day, and then sequential growth in Q2, and then very, very light in H2. This is a bear scenario.
A bull scenario is, yes, acceleration in Q2 and acceleration in H2, showing somehow a significant growth of the industrial market in 2025 versus 2024. But I repeat, for us, 2024 was - 40%, -45%, and the overall industrial market - 30% of semiconductors. So this is the two ways we have. It is clear that the current consensus of analysts versus ST for 2025 is in between a bull and a bear scenario, I have to say. More close to a bear scenario than to a bull scenario, I have to say. This is the way I can today classify. But of course, we will communicate more visibility, most likely in April, as usual, maybe in January. But we have to understand that for industrial, the visibility is very short term. And it is valid for our customer and the customer of our customer.
Gotcha. So maybe just touching on that short-term visibility and the difficulties, Lorenzo, in industrial, how should we think about the margin as we go through this next year, particularly referencing underutilization potentially and such like? So maybe help us with that.
Of course, let's say this visibility, this short visibility is not helping also in modeling where we will position ourselves in terms of margin. Let's say that there will be some drivers, some headwinds and some tailwinds. Clearly, let's say while in 2024, we were still benefiting from a significant amount of capacity reservation fees, lower than 2023, but still something that is material for us. In 2025, there will be still some of that, not disappearing completely, declining, significantly declining. This, of course, capacity reservation fees are impacting 100% of our gross margin. This will not be any longer there.
Then, on the other side, what we may think, especially considering, let's say, the evolution over the year of the revenues, assuming that in the second part of the year, we will have some recovery, as was explained by Jean-Marc, we see a different pattern in terms of unloading charges. For sure, H1, still we will have material unloading charges in our gross margin. But moving in the second part of the year, of course, based on the strength of possible recovery, especially in the industrial market, we may see start significantly reduce this level of unloading charges. Not disappear. I don't think that we will be in the position to have zero unloading charges, but most likely, let's say, we will start to see a reduction. This year, unloading charges are more than 300 basis points impacting our, let's say, gross margin.
Here, where we see today is that it should decline at this level of a basis point next year. But on the rest, I would say that mix and price should be substantially offset each other. So I don't think we will have a strong impact in this respect. And the fact that we will start to see some benefit in the loading of our fab on top of the underloading charges, we should have also an improvement in terms of efficiency. Because clearly, let's say, the fact that when you run a fab, let's say, at a very low level of production, the efficiency is very low, not on top of the fact that you have impacted by underloading charges. These are the effects.
Now, to quantify and say it will be equal, will be a slightly lower, slightly above of this, is very difficult to say at this stage because we need to enter a little bit more in the year to understand the evolution of the volumes at the end, to understand where we will position ourselves in terms of volumes.
Makes sense. Makes sense. I'll ask one more question before I open it to the floor. And I wanted to just maybe go a little bit off-piste with this question, if you don't mind. China- for- China was an interesting development at the Capital Markets Day, Jean-Marc, and I noticed a tie-up with Hua Hong this week, which makes sense. You want local production. But tell us a little bit about your vision for China- for- China, how this will impact, what you see this will do towards getting to the $18 billion number, how it all fits in.
Here, I would like to repeat what is the substance of the strategy. First of all, the semiconductor market in China will be one of the main ones. You like it, you don't like it, it is the reality. Then for us, it's vital as a diversified semiconductor company playing in the field of automotive and industrial market to be present on this Chinese market and to compete. Because today, what is the situation against the local semiconductor companies? I have not spoken about the Western company competing against us in China. I have spoken about the local. If we look at what is happening on other industries, industrial power and so on, it's always the same scheme. They compete on the low-end and middle-end part of the business. Then they learn very fast, and then they compete against on the high-end.
And then they attack you on your domestic market. When I say domestic, Western market. So this, we don't want to face this scenario. We want to compete in China in a pragmatic way because our customers are super pragmatic. If you offer quality, cost, productivity, and lead time, they will work with you. China- for- China is a way to compete against local players with the same competitive advantage: cost base, speed of implementation, speed of innovation, because they are super fast in innovation, and a local supply chain which could be protected against constraints on importation, exportation, and so on. So it's very simple. We want to compete on this market in order to not let growing the semiconductor in China that will compete later on our domestic market because they are super good. So this is our strategy.
Then what we will do to be more accurate, we will implement in China the critical technology that are the critical enabler for the automotive and industrial. So it means silicon carbide, IGBT, low-voltage MOSFET, BCD technology, and 40-nm microcontroller technology. So in any year, we have either our JV with Sanan in Chongqing, or we have now a partnership with [Hua Hong Semiconductor]. So this is the two key partners with whom we will work together for a long strategic period of time, I have to say, in order to enable ST to be the most competitive Western company in China.
Very good. Very good. So I did promise to open it. Any burning questions from the floor?
No.
Maybe I'll continue. So I think that was a pretty full answer on the strategy for China. But maybe staying with just specifically autos, we're seeing quite a change in the mix for automotive as a demand structure next year. EVs seem to be growing a lot more slowly now, hybrids growing in the mix, particularly in the West. But there does seem to be this reliance on growth in China. So maybe if you can give us a sense for your exposure on these end markets and how you see the mix effects in automotive?
In China, we were not, let's say, the top leader provider of semiconductor to Chinese car makers now. Why? Because three, four, five years ago, our focus were more to give priority on Western companies like our main customer, Tesla. Now, what we shared at our Capital Markets Day is everywhere where we want socket with the main Chinese player versus our competition. And today, by fact, we are the one for silicon carbide that are one of the most important number of socket. In China, what is the dynamic today? Today, they have basically a domestic market of 21, 22 million vehicles. 60% is a new energy vehicle. And what they call new energy vehicle is battery electric vehicle, plug-in hybrid, and battery-based extended range means they have a small turbo combustion engine that is loading the battery.
It is clear today that the most growing part is hybrid and extended range. Of course, for these two kinds of vehicles, the content, and mainly in terms of silicon carbide or power electronics, is a bit less than the two other ones. However, the car makers, they are not crazy. They know that they will have to manage this kind of mix. And more and more, they are in fact designing platforms that are flexible. They can adapt their manufacturing process to offer the platform that is growing versus the one that is not growing. So here, we see, of course, some fluctuation in our capability to grow at a certain speed depending on the mix. So far, what is important for us on SiC specifically is to lead 30%, 33% of the market.
We are super well positioned from our existing base today, where we estimate to have 40% market share. Taking into account the socket we won, both in China and in the Western world, plus the growing industrial market, and later on what we could do in AI server, we are confident to keep a sustainable and profitable 30% market share on SiC.
Perfect. You just touched on AI server at the end there. Maybe if we look at AI more generally. So you've got the opportunities with the server. It looks as though there's maybe interconnect opportunities as well. And then Edge AI, of course, TinyML is coming into view. Maybe help us with the last one, TinyML, because it did come out in the Capital Markets Day as an opportunity set. Where do you see your strengths? I mean, you've got the whole tool chain, but where's the real strength for STMicro?
Today, you have, as usual, you will have two, let's say, staggered dynamics. You have one dynamic that will be on the existing use case, whatever are the application industrial, consumer industrial, or pure consumer, where to have an hardware accelerator and then to have Edge AI capability running a neural network will be an enabler. People say, no, I need this because I want, if it is an example, motor control, I want to have a system that is capable to analyze some weak signal and according a neural network to make some predictive maintenance or something like that. So enabler. But then after, later on, you will have what we call in high-tech, the use case expansion. Because of the capability of this microcontroller, which will be much more powerful, it will self-create demand. And this is a systematic, let's say, trend on high-tech.
That's the reason why we do believe that on top of to be an enabler, the Edge AI will have its own market. We estimate that for ST that it should generate about $600 million by 2030, specific market link to our Edge AI microcontroller. On top of, but to be a booster of the growth of microcontroller, addressing the usual, let's say, point of use.
Yeah. So it's just the microcontroller market?
Yeah. Then after for ST, clearly, silicon photonics will be also beyond 2027, 2026, 2027. And it will be mandatory because the usual, let's say, connectivity technology will be no more capable because of the footprint, I have to say. And for us, it could represent a $500 million business by the end of the decade. Last but not the least, today, clearly, we are an incumbent competitor of the leader on AI server for the power stage. But we have all the technology blocks and the product know-how. And we want to achieve 10% of market share, means roughly also $500 million by the end of the decade. But the reason why I repeat, the potential of growth of ST on the next three years with a market growing at 5% is to achieve $18 billion.
If we achieve $18 billion in 2027, means we grow twice the market, so 10%-11%. If we achieve in 2028, means we grow 7%, so 1.2, 1.3. And then boosted by what I just shared with you, that the reason why the $20 billion plus is really actionable.
Gotcha. Sounds pretty clear, actually. Maybe when we think about all this and put it together, clear innovation around 300- mm silicon carbide, you've got your rationalized cost base, you've got new market entry in AI, both cloud and Edge as well. And you've got the margin or the cost-saving program at least as well. So walk us through where we think the sustainable margin structure should be, Lorenzo, for this business. I think you called it out at the Capital Markets Day. But how should we think about that, just putting it all together?
At the end, if you want, is what a little bit we discussed the other day on our Capital Markets Day, because when we will be at the level of $18 billion, we do expect it to be in the range of [44%, 46%] gross margin and [22%, 24%] operating margin. Here, of course, we are not at the end of the journey. We think that it's still valid that the model that we have set in 2022, and we think we may have the opportunity to accelerate moving forward. Why? Because at the end, we will reach the right scale for our 300 mm. At $18 billion, we are not yet fully exploited the potential that this infrastructure, this manufacturing can do. So it means that we believe that both in silicon carbide and in silicon, we will have, let's say, a very competitive cost structure.
We believe that on what the initiatives that have been shown by the groups, by Remi, Marco, during the Capital Markets Day, we have opportunity really, let's say, to extract value from what we do together with the fact that more and more we move to our, let's say, a system approach for our marketing. It means, let's say, giving the possibility for us to extract more value than the single, let's say, semiconductor part. At the end, we see that the potential for the company to achieve the 50% gross margin, the $20 billion revenues, and to be in terms of operating margin at 30% is there. So I think that the trajectory is still there, is what we believe the company will be, let's say, in the medium-long term horizon.
I see we're in red time, but I wanted to squeeze one last one in, specifically for both of you, really. All of this seems to be coming from organic growth, but there is, and I think you entertain it yourself, the possibility that we may need bolt-ons or perhaps not transformative acquisitions, but perhaps bigger than you've seen in the past. Can you give us a sense for the appetite for that, perhaps, the availability of cash, perhaps, for this? And where do you think there might be a relevant area to focus on?
First of all, I would like also to repeat why we are still strong believers of our capability to grow organically. I give you another example. Thanks to what Lorenzo mentioned a few minutes ago about the way we are reorganized the company, putting under the same umbrella our embedded processing solution business. Immediately, we have taken decision on some high-end standard microprocessor for industrial to rebalance the resources and to redefine very fast a more aggressive roadmap on automotive microcontroller. Now, the company is equipped and will be equipped with three families of microcontroller. One microcontroller will be super efficient on 28 nm with PCM to address what we call the X- in- one concept for power train. X- in- one for the most sophisticated is eight- in- one, where you have the full power train in one system. And here you need a really efficient real-time micro.
It's done. And it will grow between 2027 to 2030. The second one is a microcontroller that is completely complementary and suitable for software-defined vehicle architecture. It was not existing. It has been developed, and we will sample very soon. Done. And the last one is to put a 32-bit microcontroller based on STM32, not a real-time micro, but it's suitable for many applications in the car. We call STM32A. This will boost the growth of ST beyond 2027, for sure. So you see, we are strongly believing that on organic paths, we can continue to grow. Then saying that, yes, if Lorenzo allocates to me some capital, I would be ready to make some acquisition more at a bigger scale than what we have done up to now as far as it is accretive on our EBIT or and accretive on our market share on automotive or industrial market.
And yes, there is some target. So it is in our radar screen, but we are working on it. And if we need to move, we will move.
Perfect. Gentlemen, thanks very much. Welcome to Barcelona.
Thank you very much.
Thank you.
Thank you.
Thank you.
Have a nice weekend.