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

That's right, yes. Well. Hi, everyone. I'm Alex Duval, delighted to see you all. I head up the European tech hardware and semis research team here at Goldman in London. Delighted to be joined today by Jean-Marc Chery, CEO of STMicro. And I'm sure you'll see all the disclosure slides, but this conversation is off the record and not intended for media.So welcome again, thank you so much. Perhaps we can kick off with 2 or 3 minutes recap just on your latest results, and in particular, the high-level demand trends that you're seeing across the business at the moment.

Okay, so maybe I will speak only about 2022. I don't come back on Q4. So 2022, in fact, okay, we delivered so $16.1 billion, so it's an additional growth of 26% year-over-year. And so as we are looking at our benchmark, it is clear looking at our peers. So TI, NXP, ADI on microchips and Renesas, they all communicated their results. I have to say that ST overperformed the average of the peer. Infineon -- I forget the Infineon, sorry. Our SAM, okay, grew 16.5%, so this year of 2022 has demonstrated our capability to overperform the market we sell and the average of our peers. Interestingly, I would like to give some color about the breakdown by verticals. All the product group have contributed to the growth, all the product group, but interesting is to look at the growth by verticals. So on Automotive, we grew slightly above 50%, so 5-0. Industrial market, we grew well above 30%, around 35%. Communication Equipment and Computer Peripheral, okay, we grew about 20%, 19%. And Personal Electronics, we have been basically flattish. We have been flattish in Personal Electronics for basically 2 reasons. Reason number 1 is the market in H2 that started to be softening, and reason number 2 is because it has been allocation decision. We have squeezed, okay, some customer, part of this consumer market, by decision to allocate our capacity on Automotive and Industrial. So all in all, okay, we are -- 2022, if I want to take -- give the takeaway in terms of revenue, it's just consistent with our strategy to have a leadership position on Automotive and Industrial market and being very selective on Personal Electronics and Computer and Communication. And when we address Personal Electronics, with general purpose device and so on and so forth, facing allocation decision, of course, we allocate to Automotive and the Industrial market. So at the end, these 2 verticals altogether, in 2022, represented 62% to 65%, and Personal Electronics and Computer and Communication, 38% to 35%. Well, what about the dynamic? Well, after I can speak about the gross margin and we finished at 47.3% and operating margin above 29%, more into, let's say, improving performance, of course, versus '21. How about the dynamic? So entering in 2023, I have to say that we have a higher backlog compared the same period 1 year ago, so entering in 2022. So basically, we have a backlog of about $28 billion total. Out of which, okay, $20 billion are requested by customers on 2023. It is not the peak, and the peak has been 30%, 31%. But in Q4 and end of Q3, we have cleaned the backlog mainly in distribution or the orders, which were still on -- not scheduled to be advised or too much in delay, okay? We have decided in cooperation with distributor to really clean the backlog, so now, our backlog is clean, I have to say, of old order. Okay, what sometimes you ask double ordering this kind of stuff. So now the backlog is really, really clean. I have to say this is basically a bit of balance because, okay, it is clear that this backlog is still very high, representing a 6 to 8 quarters of running revenue for Automotive and Industrial, what we call the B2B. So energy, power conversion, power storage, transportation, robotics, automation, this kind of B2B business. And on consumer overall market, of course, we came back to a normal -- more normal situation where you have a lead time around 6 months and a backlog coverage representing, okay, 3 to 4 quarter maximum of revenues. So our situation, a bit unbalanced, okay? However, it's important we speak about what we provided as an indication for the year 2023. And of course, I can speak again on Q1, but Q1 is done. I have nothing to say more. On 2023 overall, so we have given an indication of -- at the midpoint of $17.3 billion, so which is representing a growth of 7%. But here again, the dynamic by vertical will be different. At the midpoint of this indication, Automotive and Industrial market will continue to grow about 20%, driven by the acceleration of the electrification and the ADAS, definitively, and the content of semiconductor still increasing in legacy automotive. And this is based on 83 million, 85 million of vehicle produced. It's not based on 92 million, 93 million. It's based on 83 million, 85 million of vehicle produced. Of course, of which, okay, electrical car will be about 10 million cars. It is based on the visibility we have on our engaged customer program in automotive and mainly the one related to ADAS and mainly the one related to the electrical car. Personal Electronics will significantly decrease in 2023 at ST for 2 reasons in H1: Because the overall volume weakness and demand weakness clearly on, let's say, the Android-based operating system for phone. And the second reason why ST will have a revenue decrease in 2023 versus 2022 is because of the change of mix in the second half of the year in one of our major engaged customer program. And this change of mix, I have to say, is significant in terms of revenue losses, but is accretive in terms of gross margin. But as it's quite material, that the reason why I have to explain it. On the dynamic, let's say, H1 and H2, it's a bit complex that the reason why I have to comment. Well, you see that at the midpoint of our guidance, Q1, we will grow year-over-year 18.5%. Well, I have to say that H1 2022 and '23 versus '22, we can grow about 15%, okay? And then in H2, it will be more flattish at the midpoint of the indication. Why? Because of the implication related to the product mix I have mentioned to you a few minutes ago. If you look this dynamic without imaging, which is impacted division by this product change, the growth in H1 is basically the same, so about 15%. Why? Because as I have explained during my various, let's say, address in H1 2023, we are enjoying this famous, let's say, module. We have not 1 year ago, so that's the reason why the growth year-over-year is similar. But in H2, if we remove imaging, the growth H2 2023 versus H2 2022 will be 6% at the midpoint. If we are at the upper range, the growth will be 12% -- 12.5%. And clearly, it is driven again by Automotive and Industrial, and the difference between the midpoint and the upper range will be most likely driven by what will happen in China. And of course, under the assumption that we will have no collapse of the market in Automotive and Industrial in the Western countries. But if China recover smoothly as expected, the company feels confident that can deliver the high range of the indication we have provided to the market. With the current visibility we have, we provided this midpoint. So this is the dynamic in terms of market revenue. So my take away is, at the end, 2023 should be another step consistent with ambition at $20 billion between '25 to '27. And more important, will be a strong step to demonstrate our strategy to have Automotive and Industrial representing 70% of ST, so [indiscernible] leadership and being very selective on the other market, representing about 30%.

Thank you very much, Jean-Marc. I'd like to talk next a bit about technology and specifically silicon carbide, which I think is probably one of the drivers on the way to that long-term intention. Your traction there continues to be very strong. It'd be great to investigate a bit the competitive differentiators you have on the automotive side for silicon carbide, whether we're talking about performance or about cost. And could you give us a bit of a sense of what you're shooting for in terms of 2030? If we think long term, where do you want to get to in terms of your share or your revenues, and what needs to fall into place externally or internally for you to do that?

Yes. Let's try to be, let's say, structure. I can say, again, it's a sales and operating plan. It's not a guidance. But we indicated this year that we are comfortable to deliver above $1 billion revenue. We have capacity, we have the demand, so it's I think -- and starting April and July, of course, I will disclose the number in a more accurate way. Then we are really confident to achieve in 2025 above $2 billion and going to a plan today to achieve $5 billion by 2030. Maybe a bit in advance, but certainly, by 2030. Why? Because you know that every quarter, I provide our over 100 project. So now it's 115, I guess, projects spread between Industrial market and Automotive. Well, I can disclose one number but -- today, only today, thanks to the fact I am here. What is important to be aware about this program on is what we call the project value opportunities. So what is project value opportunities? When you are awarded in a project for Automotive, you know that you are participating to a platform, okay? And you know the revenue that will be extracted in terms of this platform overall. So you have [ now ] the warranty of the market share, okay? You will have, because I'm speaking, the carmaker or the industrial OEM, they are spreading the share between 2 sometimes resource. But you have a good ideas of what we call the project value opportunities. Basically, the project value opportunities ST are because have been awarded between '23 up to '27 is $25 billion. So if you apply a decent but realist market share of ST in silicon carbide, which is 30%. Today, you know that we are much higher. We are basically at 50% market share. So if you apply a 30% market share average, which is pretty realistic of $25 billion, so it's $8 billion. So cumulatively, between '23 up to '27, we will have with this new program, $8 billion of cumulative revenue. So this makes us comfortable to have established this sales plan of 1, above 2 and 5 or we will support it in terms of manufacturing strategy. So clearly, today, we run 150-millimeter with 2 sources, which is Singapore and Catania. And we are sourcing the substrate from mainly 2 source, okay, which is one American and one Japanese source, and very marginally from one Chinese one and from our own source in Sweden, so it is 150-millimeter. What will happen is in the next 3 years? We will have 2 important cost down driver and capacity increase driver. It will be the 200-millimeter wafer size conversion. And it will be the internal source growing up to 40% as a run rate in Q4 2024 and onward. And of course, we will work as well on 200-millimeter with one of the major external source we have today. So we will introduce to support our business plan in the next 3 years the 200-millimeter conversion and the internal source. And that's the reason why we have built and we are building this fab in Catania, be capable to deliver the monocrystal silicon carbide ingot and the epitaxy and then the wafer processing. What we will introduce as well in the next 3 years not in mass production, but ready to support mass production beyond 2024 and 2025, is a SmartSiC. So the SmartSiC technology is a technology that, to make it simple, today, when you use a substrate. The substrate is often slicing an ingot of monocrystalline silicon carbide. So drawback of this approach is that the ingot of silicon carbide monocrystalline is quite seen, it's not an ingot of silicon. And you slice so you can obtain a limited number, and it's got -- the cost is quite high. The SmartSiC process to make it simple is the following. So you remove the layer of monocrystalline of SiC, and you use a polycrystalline silicon carbide substrate, which is easier to have and much easier to produce. And on top of that, you can [ drop ] this polycrystalline. So at the end, you put the layer of the monocrystalline onto polycrystalline, drop it so you have a lower resistivity, so you enable better performance. And overall, the cost of ownership is lower. So we will introduce this technology in 2024 on our Generation 4 planar technology, and we will qualify it. And this will be another cost decrease and performance improvement. Well, so this is basically the 3 main manufacturing action that will support our ambition to go to $5 billion. So 8-inch conversion, internal sourcing, SmartSiC introduction. Then, of course, we have the technology road map to continue to enable the performance of the device. So now, we have our Generation 3 planar. We will introduce mass production Generation 4 in Q4. We will introduce Generation 5 on 200-millimeter boosted with the SmartSiC, still in planar. And then we will introduce a new technology that I cannot disclose publicly for secret reason of our IP and technology development. And honestly, ST, we have always a discipline to never disclose information which has not been proven as mature enough, okay? So that's the reason why we have this discipline. But for sure, it is in our road map. So all in all, we are very convinced that on silicon carbide, we will support this EUR 5 million with adequate cost structure, 200-millimeter SiC internal sourcing vertically integrated. And by the way, introducing the technology road map at the right speed to follow the demand of customer in terms of performance. So this is our strategy and plan on SiC.

Can I also just follow up on 2 things there, and that was super helpful. Firstly, on margins, how should we think about the kind of profitability versus the group or ADG? And obviously, if we think about the structure of the market, there are more people trying to do this, and perhaps on IGBT for auto inverters. So how do you think about that in the long term given that actually, the value theoretically you can deliver to a customer is very high? And secondly, just thinking about the tech road map. You talked about planar, high level for the industry. How important is trench technology in your view?

Trench for the time being is volume margin I have to say after it is an architecture that has been decided by our competition. We respect it. What we will introduce will be more complex, and -- but for the time being we don't need this tranche. Coming back to the product group, it is clear that with our 50% gross margin ambition, taking into account the portfolio when we deploy this ambition by product group, you do not extract the same gross margin when you have a power device versus a microcontroller or a microprocessor for industrial or an analog. In analog, depends if it is a power switch or if it is a consumer analog or high precision analog. So we have a clear breakdown of this objective by product group categories and technology cluster. What I can confirm to you that the silicon carbide will play the contribution of the 50% gross margin according to the model we have deployed. Thanks to the initiative I have disclosed to you, which I repeat is a 200-millimeter wafer size conversion is internal sourcing at 40% or beyond, but at this stage, 40%, and the SmartSiC introduction. And the SmartSiC introduction is a buffer because when we disclose the 50% gross margin, we were not yet sure that the SmartSiC will contribute. Now, it's an additional buffer we have in our weapons to improve our profitability. So the SiC, which has been an investment up to let's say, last year or 1 year ago. Now, it's really a contributor according to the model we have fixed on the ambition of ST in terms of gross margin.

Super helpful. If we sort of stay on the Automotive side in ADG, it's been interesting to contrast some of the guidance given by STMicro and some of your European peers versus some of the American peers, and you've been quite bullish near term and others have been a bit more subdued near term. I wondered to what degree that's a function of the different areas that you're serving within Automotive? That's my first question. Second question, there seems to be a high degree of tightness in certain areas of Automotive. And I wondered if you could talk about the extent to which you think in the long run, auto companies and then those working with them are more willing perhaps to reserve capacity for longer or perhaps price a little bit more favorably for companies like ST?

So I think here, it's complex because the Automotive is really facing one of the most challenging and biggest transformation of its history. And basically, no industry faces such a transformation. You -- to, let's say, try to assess the complexity, we have to break down in 3 parts, okay? It is what is related to the electrical cars, so the electrification. So 10 million vehicles in 2023, and you know that up to 2030, it will grow close to 40 million cars. Then at the other range, it is the ADAS. And ADAS, you know that for security reason, norms reason and so on, so you will have a permeation of Level 2, Level 3. And then after you will have maybe in some business case -- sorry, mission profile at the Level 4 or Level 4 plus. But the main -- the big volume will be Level 2, Level 2 plus or Level 3. In 2023, both -- this both electrification and ADAS, will grow a lot, and ST is totally exposed to this market. That's the reason why I disclosed above $1 billion revenue in silicon carbide compared to the $700 million we have done this year. So it's another, most likely, 70%, 80% growth. But the ADAS as well. We will more than double our revenue in ADAS, thanks to the partnership we have with the guy owning 75% of the market share of ADAS. So this is the 2 big growth driver of ST in 2023 on Automotive. Then you have the car, because whatever is a car on the ADAS or car electrical, you have all the legacy of the subsystem. So the body, braking, lighting, airbags, all this stuff. And here basically, there is a different dynamic. There is one dynamic which is still very active today is that semiconductor are increasing in the legacy without still main architectural change of the car. Today, we are not yet in a domain architecture car or zone, particularly will be more zone than domain. Domain certainly something that will not fly. But in the pure architecture of the car, the permeation of semiconductor is very high, okay? You have much more device on braking system, you have much more device on door control, much more device on internal monitoring system of the driver. But the permeation of the semiconductor is huge, and it is making us immune to the fact that in 2023, the production of vehicle will be 83 million, 84 million. We don't care if it is 1 million car or 2 million car or less, we don't care, okay? And they are doing it, other technology which are still strongly in terms of capacity. It is a 14-nanometer for microcontroller, it is a 28 or 16 for standard microprocessor, it is advanced BCD 110-nanometer technology for power switch -- for power driver, I have to say. It's IGBT, it's VIPower, this kind of technology are still struggling a lot. Of course, everywhere, you have general purpose analog diodes is performing so the situation is now clean. But on all the critical devices, the situation is still under stretch. And here, by the prices and volume are warranted for 2023, and we are working for '24 and '25 because they want to have warranty and volume. So this is for the next 2, 3 years. Well, then we know that the architecture of the car will change. So going forward, let's say, software defined and certainly zone architecture, and following 2 important paths. One path is to have the capability to update the car on the over-the-top, because the complexity to update the car always is a nightmare. And also is for service purpose. It means you can improve the consumer experience of the car, changing the software. We know it. It has been already done. Some carmakers are offering a booster of performance depending of what we pay as a service. It's point number one. Point number two also, the carmaker, they want to make their car lighter, especially on the electrical car. Because they want to look for an electrical car for many things. First of all, the critical point is a range of kilometer between to charge. And of course, here is a matter of battery, is a matter of investor and is a matter of weight. So lighter is your car more extended will be the range, so to help as much as you can and decrease the weight of a car is critical, and especially the weight of the wiring. So that's the reason why the zone architecture will be certainly the winning one because you will minimize the wiring instead of the domain. Because the domain you have still, unfortunately, you need to wire everywhere. This change of architecture for some side of the product portfolio will reduce the number of ECUs, so maybe will reduce the MPUs and MCUs. But the computing power, you will need real-time or let's say, not real time, will increase tremendously. So in terms of what we call DMIPS, Digital Million of Instructions per Second on the car, it will increase continuously and will be breakdown between real time and not the real time. Of course, with lower number of either microcontroller and more high-performance microcontroller and microprocessor. ST, we play both. We play high-performance microcontroller, what we call as Stellar on 28 FD-SOI with PCM. But we play the Stellar up which is a microprocessor we are codesigning with Volkswagen and CARIAD. But okay, we will also offer other products timely. Then you will have other implications, wiring and fuse remove using semiconductor. So some technology like the vertical integrated power which is key to replace an electromechanical fuse and replace all this wiring, will do in the next few years. Power management ICs, more computing power you have, more power management ICs you have. More feature you have with processors, more power management ICs you have. So in fact, the semiconductor content of the software-defined vehicle by zone will continue to increase in term of dollar, maybe less in term of numbers, because the architecture will be more structured and concentrated. So this is the trend that we will see in the next 5 to 10 years. And this is everywhere, okay? So electrification, ADAS, and the change of architecture and consumer experience increased. And of course this period will cover now for 2030 definitely.

Super interesting. Jean-Marc, I think some other semis players have talked about potential for a cyclical inventory correction sometime this year, perhaps in the second half of the calendar year. Just wondered if you could give an update on what kind of visibility you see on the second half? How do you get comfortable it won't be weaker than expected, and how do you bake that into your guidance? And then perhaps related to that, if you can talk a bit about inventory levels, what you see in the industry, what you see in the channel, any sort of signs of double ordering? Things like that would be really helpful.

I mean in the channel, for sure, we are going to a normalization, which basically is a copy paste of the image of the market dynamic I disclosed to you. When you have a microcontroller, high performance, including some connectivity or security feature for industrial market, inventory is very lean, return, very high. If at the extreme, you have a mainstream general purpose microcontroller addressing consumer devices wearable, this kind of stuff, yes. Now you have inventory at the right level, lead times are quite short, definitively, we are entering the phase of more term business. It's business as usual. I can never forget that. When you look what kind of components, the normal situation entering in a year is basically to have a coverage of 60% to 65% of our revenue expected for the full year. When you enter in the year with the backlog, this is a normal situation. Also, when you enter in a quarter, you have 75%, 80% of backlog coverage. The rest is term business and that's the reason why, you have some go-to-market approach with distributor. And that's the reason why a distributor, you need to have inventory term at least 3 to 4 because there is this flexibility because a bit of material is very complex. Here, we are again in a total different situation for the reason I explained a few minutes ago. For consumer, we see for us something coming back to normal. We are not [indiscernible] maker. We are not [indiscernible] maker. Here, we know that the price are collapsing 30%, and there is a big excess of capacity, and that some company will struggle in the next few quarters. Here, we are in a different situation. For ST, I repeat, we have the visibility on Automotive. We have the visibility on the B2B Industrial market. We have the visibility on Personal Electronics from our main customer, that is updating us every 2 months, and those visibility is really clear. We have the visibility of the main engage customer program on communication, the low earth orbit satellite. Then the rest, on Personal Electronic, Android base, it is what it is. If China recover and there is stimulus on consumer market in China, this will increase. And then it will increase at the speed, we don't know. What we have taken is a cautious approach which is embedded in our guidance. Then on consumer, industrial, appliances, battery-operated tools and so on and so forth, same. Today, yes, there is the right inventory level in the channel. There is the right inventory level at our customers. The voice of our customer is H2 should start to recover because of China as a main driver plus they expect that the inventory will be fixed, okay, in H1. At the end, the result is what we told to the market, it's 7.3 at the midpoint. If something worst happen, we do believe that we can be resilient at 16.8. If something positive, China restart and the market in the Western countries remain quite stable, we will be at the upper range.

Great. Well, I think we're coming into the last -- just under 3 minutes of very illuminating discussion. Just to squeeze one more in, we had some questions about reshoring. And obviously, we see more and more headlines about what players like TSMC are doing, more production perhaps happening in Europe. Could you help us understand and put into context what that could mean for ST in the coming years?

Today, we acknowledge what is granted, 100% sure. Today, we have received 2 important information. TI, I think, they communicated 1 week ago their capital allocation. Clearly, as they announced $15 billion of CapEx during the next 3 years to set up 300-millimeter fab that will address analog mix signal and embedded Flash technology, including the 28-nanometer. TI is already a competitor. They are using [ fully ] for this kind of technology. Now, they will use their internal fab that they will have to go, they will have to manage and so on. For us, for sure, we expect to see TI more aggressive on consumer market because to load a 300-millimeter wafer fab, you need to have volume. So this is one element. Second element is our friend Infineon. Well, the announce Infineon has done for power electronics and smart power is not a surprise. They have the ambition to go. The power market, if my memory -- only power should be a market above $30 billion in 2030. So Infineon is one of the leader. They need capacity, so this is the right time, they implement this kind of fab. For us, again, it's not a surprise. Today, we have capacity, more spread on the 8-inch and on the silicon carbide. But in our mind, we have a roadmap of manufacturing and we know that what we will do in the next 5, 6 years. But we did not announce because we did not yet decided or validated. So it's -- for us it's not a surprise. Then the third element you did not mention is microchips, on silicon carbide. It's not a surprise they have the silicon carbide, Semikron. so we know they are developing this kind of activity. And why? Because they want to take the Industrial market. And when you want to take the Industrial market, like micro-inverter for renewable energy and so on. If you go and you offer the full solution, it's much better. This is what we have in the position. So for us, it's not a surprise. Then all the rest. TSMC in Europe that, by the way, has just been postponed as a decision. U.S.A., which is more Samsung SMC for very advanced technology. It's not our business. Intel in Germany is not our business. Then TSMC in Japan, that has been already announced for Sony, yes, okay? And Renesas, yes, it's something important. What we are monitoring very carefully is what is happening in China. Because in China they are building capacity massively at SMIC or other foundry on 28, 40, 65. And this is something we are monitoring very well because we want to have access of this capacity to address the China market. So all in all, yes -- but this is something we take into account, but it's not disruptive for us.

Great. Well, on that note, I really like to thank you very much, Jean-Marc, for a very interesting dialogue. We look forward to continuing the discussion, and thanks everyone for joining.

To play a good match, we have to be 2.