Infineon Technologies AG (IFX) Earnings Call Transcript & Summary

Yes. Hello, everyone, and a very warm welcome here to the GIP update call at the PCIM in Nuremberg. With me are Peter Wawer, Division President, GIP; and Peter Friedrichs, Vice President, SiC. And after the presentations, we are going to have a Q&A and everyone is invited to put your questions just by e-mail directly to me. I will read it aloud, and this will then be part of the Q&A. And with that, the floor is yours. Please, Peter.

Yes. Thank you very much, Alexander. Warm welcome here to the few ones who made the way here, but of course, to all of you on the screen at the video interface. I'm happy to be here and share with you now the recent advancements of GIP, formerly known as IPC. So green industrial power, driving decarbonization that is the motto and the agenda of today and as already introduced by Alexander, I will share this presentation with our top silicon carbide expert, Peter Friedrichs. So that's the outline of the agenda. Maybe a few words about Green Industrial Power. So what's it all about? With the renaming, there is no reorganization or something like this ongoing. That's the question which I'm always asked multiple times. It's basically to adapt a bit to the things that changed. And you will see it in the following slides. It simply emphasizes our contribution to the energy transition. While we are totally committed to our industrial backbone and to our industrial customers, it reflects the reality that renewables and all these kinds of applications, which drive decarbonization have massively contributed and are massively contributing to our revenue growth and, of course, also profitability. So regarding today's scarcity on resources, meaning highly skilled talent and employees, of course, it shall also foster pride to the team, which is already there, but it shall be also help to attract new talents to support our growth and send a message to also external shareholders like you. The paradigm shift I already mentioned on and it simply demonstrates a bit the transition that we are undergoing as a society, but of course, also in our business. And to give some meat to these bones, we prepared for you the following slide where you see how the split on CO2 emission is being shared among the applications. So the majority, 40% globally stems from the energy that you need to -- that you burn to provide the electricity and the traditional fossil-powered power plants, and of course, producing heat, simply for warming our homes and, of course, also in the industrial space. But electricity and heat producers account to 40%, then transportation, meaning combustion engines, majority is road-bound traffic, but of course, also airplanes and ship is contributing to that one, then the whole industry sector, followed by buildings and remaining others. And while we know that the planet cannot afford something like 37 -- close to 37 gigatons, which were emitted in 2022, again, the highest level in history, we think -- we know that energy consumption continuously will increase because simply, population continues to grow globally. And of course, therefore, energy being consumed continues to grow, even outperforming the growth of population to a certain extent. And here are, again, some numbers. EJ are Exajoules, that's a unit not too well known. Anyhow, we know these are huge numbers, and they simply continue to grow. And now the good message for a semiconductor manufacturer related to the growth is that to have a chance to achieve our CO2 emission targets globally, the world has to electrify. That's the message on this pie chart on the right-hand side. And to what extent we will succeed in 2050, that's the long shot here, remains to be seen. But the estimation goes here regarding on the IEA scenarios, depending on how aggressive the policies will kick in, that up to 30 -- between 30% and up to 50% -- maybe up to 50-plus-percent energy consumption will become electric. And so the need for semiconductors to support this transition is, of course, reflecting our growth looking forward. And then now translating it for GIP division, green industrial power at Infineon basically looks like this. So we have the topic of growth in general. We have the specific role of silicon carbide. The transition is accompanied now by the conversion from traditional silicon-based technologies to wide-bandgap, in our case, specifically silicon carbide. And of course, you see here that the traditional business -- the industrial business continues to grow with a substantial single-digit growth rate. That's the gray area below. But the new market opportunities out of renewable energies, power infrastructure related to EV charging and then, of course, transmission and distribution because the electricity -- renewably generated electricity has to be stored and distributed accordingly. That contributes to more than 20% growth CAGR for the years to come. This is now assumed for the 5-year time frame and overall totaling to well above 10% growth. How far above 10%, we will see, somewhere between 10% and 20%. And of course, I will give you some numbers for now the ending first half year in a second. Maybe some selective examples to where we see these kind of trends. The topic of heating -- decarbonization of heating has been now a big political topic, especially in Germany because there's now a new ruling upcoming still heavily debated in the politics arena, where simply it's prohibited to replace oil-fired heating system and gas-fired heating systems by new ones. But if you replace the system, you have to go for a CO2-free solution, for example, the heat pump. That is already seen and that translates into tremendous demand growth on the heat pump side, of course, also certain amount of M&A activities. Maybe you are aware of the recent announced takeover of a privately held company, Viessmann, in Germany. A very traditional midsized company is privately owned now by the U.S. brand, Carrier. And there is a lot of activity here on growth and, of course, also on efficiency, right? One thing is that you have tremendous growth opportunities, now more than 20% CAGR being here estimated, but of course, also the requirements for high efficiency are increasing. Therefore, this is also a playground for semiconductors and of course, also very advanced semiconductors like our recent to-be-now-launched generation IGBT8 and also silicon carbide. Another example, which we now also announce here at the occasion of PCIM are our high-end high-power modules, and we are really proud to now expand the -- in the community already well-known.XT technology also into the arena of silicon carbide. So high-power 3.3 kV is here the first beauty, I would say, out of our product family. So the well-known.XT technology really leverages on silicon carbide, where the traditional connection technologies show certain limitations if we talk about soft solar, high temperatures and some cycling wear out, .XT provides superior properties, which are already known from the silicon time and are now being transferred and used here for silicon carbide. And of course, Peter can comment a bit more on the details. Still key message, you reduce the losses. The losses equals to energy consumption, electricity consumption for, in this case, public transport trains. Another important topic is also that while reducing the losses, you even improve other topics as for example, noise. Also the noise of the rolling stock of the inverter is a significant environmental contribution, right? You do not want to have loud trains, trams or high-speed trains and therefore, also certain requirements regarding quietness, enjoy the silence, you see the motto, can be fulfilled with applying this kind of high-end technology. Okay. These were 2 selected examples. And now let's talk about numbers a bit, business update. So as you already have seen in the Q2 numbers, we really enjoy extreme strong first half of this fiscal year. So with a record breaking segment result of slightly more than 30%, having achieved more than EUR 1 billion of revenue for the first half of this fiscal year, we're clearly targeting to go for more than EUR 2 billion for the full fiscal year. If you do the math, more than 20% growth for the first half of the year, and we have a high confidence that we will sustain this momentum also looking forward. I think this is particularly remarkable while we know that certain areas of the semiconductor market are in turmoil, deep trouble. For example, the computing arena and memory players and also the consumer goods are weak. They continue to be weak. Of course, certain portfolio of Infineon and also GIP is affected. But now again, talking for GIP only, this weakness is completely overcompensated by the strength of the new energy topics. And to be a bit more precise, on the subsequent slide here on this one, you see now the splits by applications. That is how we monitor and typically assess the market development. The percentage points you see on the left-hand side are the share, how our revenue is being split to this application. This is not precise math because we, of course, do not know exactly what our customers are doing with our discrete products that we are selling to them, but to a certain extent, of course, regarding the customers, we can cluster those topics. So in aggregate, roughly speaking, this fits nicely. And you see here, majority of the arrows is green, that's the good message. And also, if you look into automation and drives, which still accounts to a bit more than 1/3 of our overall revenue, that is a yellow, meaning stability, but a stability on a high level. So it's not further growing, but it's flying high. So if it stays like this, I'm also very happy if it remains like this, I'm even happier also for '24. So very clearly for the remaining calendar year, we are -- we see it overall quite positive. The yellow arrow in the area of home appliance, this one I would like to highlight as something where we stay stable on a low level. So this market is down. That relates to what I said. Consumer exposure also that we have in the area of close to 15%, close to 20% of our overall revenue, here, we are in a weak situation already for the last 3 quarters. So also here, we rather expect a bit of a rebound towards the end of the year. Nevertheless, the evidence is not strong enough. So therefore, we keep the arrow simply yellow, flat. All the other topics, I think maybe very short comments on the green ones. PV area of renewable generation, extremely strong. We are completely sold out. That's also a huge momentum coming out of the demand from solar customers for IGBT, but of course, also SiC. We are here completely capacity limited. Power infrastructure, especially the area of EV charging, exactly the same, being sold out for '23 and also huge order backlog for '24. And the area of transportation, we see also very positive momentum regarding on-board charger, but of course, also the electrification of the larger vehicles, not passenger cars. That is, of course, predominantly served by our automotive colleagues, business is also booming there, as you know, but also the area of CAV, so larger trucks, buses, that continuously gains momentum, which is quite positive. Others, I think I don't need to comment too much. With that, let's have a short look into the wide-bandgap strategy topic. And there, of course, then also Peter will take over. That's a picture which I like very much. That's our application landscape, and that's also how we knew this is the environment -- the industrial environment we live in. And maybe it's, of course, already a bit biasing because the emphasis here on the renewable topics, you see in the upper left corner, wind generation, PV modules and then, of course, the distribution network. And here, of course, just the topics to explain what is meant here. CAV, of course, the world becomes electrified more and more. You see this tower on the right-hand side, grid-scale battery storage, not to underestimate grid scale battery storage is a big thing. So now containers of batteries to stabilize the grid, which is required when we step-by-step shut off the fossilized fuel or in case of Germany, even the nuclear electricity power plants. You nevertheless, of course, need ways to stabilize the grid. And therefore, decentralized storage or like grid scale storage technologies have also a huge momentum. And of course, you might also be aware that Tesla is very much investing into these kind of applications. The heat pump you see over there, rail, CAV, et cetera, as already mentioned. Now comes the interesting thing. What is the technology being used? So far in semiconductors, that was served by silicon-made power devices. And we very simply now added here where we see -- how we see from Infineon perspective, silicon carbide and also gallium nitride will penetrate this market. And if you look into it from a simple perspective, you say, okay, it will more or less penetrate each and every way. Well, I would say this is true, but of course, there are certain areas where the value proposition is very clear and others where -- yes, let's see. It depends then as always from a cost performance perspective and also from the customer history and there are customers who are very conservative and say, yes, I'm looking into wide-bandgap, but give me a break, it will take another 5 years, and until then, please supply me with the latest and greatest IGBT and MOSFET -- silicon-based MOSFET technologies, and other customers being much, much more aggressive. And that's, of course a topic, being a broad liner and the market leader in power, we need to take into account, and we want to serve all these customers. Going into some specific examples, which are not all mainstream, but where you see and we indicate certain trends are now applications for very innovative concepts. Here, for example, for the famous PCB motor, as we call it, where we see an opportunity that much lesser copper is being used. So there's cost reduction on the motor design side by its own. This enables you to apply significantly higher switching frequencies and by using then silicon carbide for the inverter, you achieve energy savings, less size, less weight and of course, overall, less material, less consumption, less CO2 reduction -- less CO2 generation, meaning CO2 reduction. That's also an interesting example, the on-board charger for the electric car where we see, of course, an IGBT-powered system. And there you see now as a KPI, so to say, the kilowatt per liter, so the power density. And here you see consequent improvement, while going from silicon to silicon carbide, you double. And then again, you increase by 50% from 4 to 6 kilowatt per liter looking forward, being expected for the future, where already now with the silicon carbide, we are able to double the power density, which is quite impressive. And last but not least, the perfect fit for gallium nitride. So lower power, switch more power supplies, the blocking voltage on gallium nitride is still being limited to the area of 600, 650 volts of course, also here innovation goes to higher voltages, but it's well below 1,000 as we speak of today. And therefore, the area of switch more power supplies is a very nice fit. Again, the same logic applies, energy savings, less size, less weight and overall lower system costs. And that is something we always have to keep in mind. The semiconductor -- the wide-bandgap semiconductor typically is more expensive than silicon. So from the traditional view on how to procure from our customers' perspective, how to procure the devices, you need to understand the value proposition on the systems side. I dare to say, silicon carbide will remain, on the component level, more expensive, more pricey for the foreseeable future. But the value proposition kicks in, if, for example, for an EV charging system, you increase the power you can use, the rated power by 30%. That's massive, right? Same enclosure, 30% more power or the other way around, same power, you shrink the whole system by 30%, saving again material, being more efficient. Okay. Most recent automotive SiC design wins, I think that's, of course, kudos to our colleagues on the automotive front. Very prominent names, Stellantis, then also some U.S. OEMs where we're not allowed yet to disclose the names, the hidden car in red. And then Genesis, for example, overall, 20 OEMs and roughly 10 Tier 1s are already won. And it's now, of course, a huge challenge for manufacturing to ramp up capacities, but Peter will comment on it in a minute. And then SiC design wins on the industrial side, that is typically the infrastructure topic. Here, you see also very well-known names ChargePoint, Bloom Energy, Delta, SolarEdge, where I would like to highlight also the topic of Bloom Energy. The next big thing which we already see, it's not yet there, but everybody is preparing for it is, of course, then since you cannot electrify everything, you still need high-power density fuels, and that is, for example, hydrogen. So to replace other CO2 emitting energy sources, the hydrogen industry, the hydrogen environment will become the big thing for the remainder of this decade but of course, then also for the years to come thereafter. And the topic of electrolyzers -- megawatts up to gigawatt electrolyzers for realizing very cost-efficient hydrogen is a topic where I expect also the market will grow with a high exponent looking forward. To be a bit more specific here, an example how it looks for EV charger, example is here. Thanks to charge point, you see on the left-hand side also the CAGR, and these are not our market numbers, but from the estimation of Yole for DC Charging for Automotive. If you ask me, I would say, the light green area being the SiC MOSFET is here slightly underestimated, but also due to this huge growth globally, actually, each and every semiconductor is needed to fulfill here the demand forward looking. Very nice example, and you see here, from a block diagram perspective, the different stages. And you see, of course, our product portfolio, which simply fits very nicely to this application. So depending on the power rating, a couple of these EASY modules are inside such a power wall. Then let me briefly comment on gallium nitride. The gallium nitride topic is not the priority one for GIP, but it's an extremely important topic for Infineon. And we -- while we are engaged already there since almost a decade, we thought it would be really good to add competency and to complement our knowledge and our team. And therefore, we were very happy now a couple of weeks ago to announce the agreement between GaN Systems and Infineon. It's, of course, still subject to the final closure and the granting of the political authorities. But we are very positive and optimistic that this will be successfully executed in the second half of this calendar year. So again, we are addressing the fast-growing applications. I think it's not a need to reiterate this topic. And it provides us with a unique know-how on top of the know-how we already had and have, simply increasing our knowledge base and the economy of scale. So we completely focus here, not only on the switch, but on the whole semiconductor value chain. Based on system understanding that we derive together with our customers, we want to provide a total solution to the customer, where software algorithms are included, thanks to control, drive and then, of course, also switch. And that is true, by the way, not only for gallium nitride, but for silicon, silicon carbide as well. So also here to quantify it a bit in a graphical way, where we stand today and how we see the evolution until tomorrow, we realize this chart. So that's now the outlook from last year where we have the numbers, '23 not yet fully completed, and how we see the growth momentum looking forward. And not only we, but here, the sources are given on the slide in the lower right corner. And you see, of course, the big area, which is silicon. And it's very important also to realize and recognize that silicon continues to grow. That's the clear assumption. And that's the solid green area. And with the assumed CAGR of 4%, it increases into this shaded green area for the next 6 years. Then you have silicon carbide and gallium nitride. Of course, here, the starting points are significantly lower from where the market stands today. But of course, you see the well-known double-digit CAGRs estimated for silicon carbide here for 30% and gallium nitride, from a lower base, even more than 50%. And you see how this will evolve until end of this decade '28. So the message here is, of course, huge growth potential. That is nothing new, but also important to say and to see how silicon still will dominate the power arena until the end of this decade because while the growth numbers are remarkable, it's not so easy to catch up 20 or 30 years of silicon power history. With that, I come to my very last slide. We are rather frequently, of course, asked about our strategy on wafer sourcing and silicon carbide, now focusing on silicon carbide. And you know that we had a missed attempt a couple of years ago to acquire Wolfspeed. But we then decided, okay, what are the alternatives? And we thought it would be a good alternative to, of course, generate know-how and competence because the material topic plays a very crucial role regarding yields, performance of the devices. So we want to understand and know how material quality interacts with device manufacturing. And of course, we want to also understand levers to reduce cost. Therefore, we acquired Siltectra a couple of years ago. But regarding crystal growth, like in the silicon area, we then finally came to the conclusion that it's also a good idea to follow here a multi-sourcing strategy on the wafer supply side. As it's publicly known, a very important supplier of ours is Wolfspeed and Coherent, formerly known as II-VI from the U.S. We announced the agreement of long-term supply with Resonac, the Japanese supplier, formerly known as SDK. And then very recently, we added 2 very important suppliers -- new suppliers for silicon carbide, raw wafers, namely SICC and TanKeBlue from China. And with that, we are very well set regarding supply security globally, but also locally. So basically from all major regions in the world, we have now secured significant amounts of silicon carbide raw material looking forward. And what's also interesting to see is, of course, that the environment gets more competitive. Of course, at the moment, I wouldn't say that there is already an oversupply on the silicon carbide wafer side. I would consider this a bit balanced. But you see here, normalized, of course, the variations between suppliers and supplier pricing. And here you see that the competitive pressure really increases, as you can guess from the numbers, since supplier A is normalized to 100% but the delta between supplier A and C in price is more than 30%. So that gives a competitive environment. And for us, of course, a comfortable situation since we have very ambitious growth plans, we count on the strategic players that are named on the slides, maybe more to come looking forward, let's see. And also an important topic I would like to highlight, we only qualify suppliers if they reach a certain performance level, Prerequisite for qualification and for appearing here on this slide, the names, left, and of course, the numbers A,B, C are that yield and performance targets are met. So it's not that, for example, supplier A and C differ very much in price but also in performance, that's not the case. If we compare these suppliers and if we qualify these suppliers, they are plus/minus on a par level regarding performance, meaning defect density and overall reliability, the relevant KPIs that usually apply for suppliers. And that is a topic which is, of course, very positive from our perspective. With that, I think I'm done, and I'm going to hand over to you, Peter, going a bit into the meat on the technical side.

Yes. Thank you very much, Peter. Yes. As we have shown, basically, the big question of substrate material, we believe, is on a very good way. And therefore, we can concentrate on the following steps, what is our core know-how at Infineon, basically, that's a device technology plus I will add also a few information about that, the ecosystem around the actual chip. But let's start with the device technology. There is this magic question in silicon carbide still not completely solved, whether the planar or trench-based concept is the one to go for if it comes into a MOSFET. We have decided at Infineon to focus on the trench. And I would like to give you a few explanations which hopefully also convince you that this is basically the right path, especially the one which has future outlook for further improvements. While we believe that the classical planar MOSFET has some limitations, It's actually indicated by the errors here on the left picture. So you have to reserve some space on the surface of the wafer or the chip for critical parameters like channel lengths, also, we need to turn the current which starts to flow at the surface and down to the bottom -- back side of the device. So all those things require a certain space. So we are limited basically in the ability to make chips smaller. And this is a magic basically in semiconductors, not just in power. We know it from memories, we talk about smaller nodes and the similar thing is taking place in power semiconductors, we always want to shrink our devices in the form of the cell size, and this is basically somehow limited in case of the planar concept. And for that reason, also driven by the fact that we have a lot of experience in silicon at Infineon with respect to trench-based devices from low voltages up to very high voltages, we decided from the beginning to address our MOSFET product by a trench-based technology. Here you see those dimensions, which define more or less a limit regarding shrink and the planar concepts now are in the vertical direction. So we need much less space to generate what we call unit cell. And that means, in the end, the valuable silicon carbide material can be more effectively utilized for current flow. For the customer, it means you have a smaller die giving you the same performance in the application. So that helps us to reduce cost to basically reduce the number of required wafers. And eventually, then it turns out also to be a technical advantage. If you look to the benchmark studies done by external groups here, always the trench-based concepts are the top ones with respect to performance. And eventually, also, we are very proud to say this trench-based concept is a very reliable one. So all the rumors that the planar is favored because of reliability aspects, we can definitely not confirm up to now, and we are in the field for more than 6 years, meanwhile, we have not a single device returned from any of our customers showing field failure or reliability problem. And therefore, we believe we are here on the right path. What is crucial to achieve such a very high reliability? One of the most critical parts in the device in the power MOSFET is the sensitive gate oxide. That's a very thin layer being more or less the key element to turn on and off a semiconductor-based power switch. And this very thin insulating layer sees, permanently in operation, certain electrical stress. As long as the layer is perfect, what is indicated here on the left-hand side under the name bulk, so that is basically a well-designed, well-structured gate oxide, we are able to design the device in a way that it has a very high lifetime, which is indicated by here in the lifetime curve, the last part, there is a steep increase, which is always more or less the intrinsic life a device can achieve. However, especially in the case of silicon carbide, also triggered due to the still higher defect density in the material compared to silicon. We have to take into account that there might be distortions in this very thin insulating layer. And those distortions basically electrically act as a so-called thinning, that means with an applied electric field for the bulk material, typically at such an imperfection, the electric field is much higher. So every electric field triggered destruction is accelerated. And that's basically also then depicted here in this lifetime curve under the label unscreened. Unscreened means you take the devices as they are basically coming out of the production. And the problem you're facing is that initially, that's the Y-axis here, initially, you have an extremely high failure rate, much higher than people are able to accept today for applications. And we have to address this topic. And we do that at Infineon by a process which allows us to detect those small imperfections and to sort out those devices with what then statistically means, basically, you drop down the failure rate, which is very high initially to a low level. So this is a couple of orders of magnitude what we can achieve here with respect to reduction of a failure rate, and this allows us to give -- even with the danger of such an imperfection device to the field, which is basically reliable as the silicon technology is. You can imagine this is a lot about understanding mechanisms, statistics, models to extrapolate from accelerated testings. So here, this graph just shows you what is necessary to do in order to develop those models in order to predict the lifetime under certain mission profiles. We did hear the so-called marathon test. This is a methodology also being applied historically in silicon. So large groups with thousands of devices are stressed under relatively high electric stress. So in this case here, it's a 30-volt for a device, which typically is operated with 18 or 15 volt. And based on the failure statistics you achieve with that, you can develop an extrapolation model, which then gives you a lifetime under the operating conditions by this linear E model. So basically, that helps us to extrapolate from our lab tests with thousands of devices to the behavior of our population with millions of devices in the field. The key aspect now in order to achieve low FiT rates is the fact that the screening process I have described needs to be very effective. So we come from a relatively high failure level, you have to drop it down by some orders of magnitude in order to be comparable with silicon. And here, again, the beauty of the trench technology steps in. The trench technology allows us to have a screening procedure with an extreme high failure reduction factor. So we can achieve here close to a factor of 1,000, so 3 orders of magnitude reduction by screening compared to the best follower, which is a planar technology, which is only close to 100. So basically, order of magnitude higher failure rate reduction is possible by using the trench technology. Most of the other technologies are even worse. So what does it mean now for the real device? If there is a 10x higher FiT rate because you cannot drop down the failure rate as good as we can do it with a planar, that means on device level, your failure rate is also 10x higher. And in the final product, the same story. So therefore, it's an enabling factor here to get a very high reliability as people expect in -- especially also automotive applications, but also in industrial applications. It's enabled by the fact that the screening capability of trench devices with a [ signal ] gate oxide is substantially higher than in planar technologies. Regarding an outlook for the chip technology, so we just recently introduced our Generation 2 technology where we are able to shrink the area specific [ resistance ] by 25%, and we are already in an advanced stage with upcoming generations. So concept phases are nearly finished. So we can foresee next substantial steps down in the area specific [ resistance ], all enabled by certain shrink combined with more refined sales structures. And even what we have today more or less already sketched is not yet the end of the road map. So the material itself allows for even better performance, which is indicated here by Gen 5. It's still dashed. That is something where we are looking into various concept how to achieve it, but should just indicate we are not yet at the end of the road map. So that's about the chip technology. However, with all those shrinks, we are confronted with a challenge, which we also have to address in order to enable our customers to use this technology. We have to combine the chip with a very powerful package in order to connect the performance from the semiconductor to the outside world. And for that, we need, again, smart technologies, which address the challenges we have with wide-bandgap technologies. One aspect is thermal performance. The chips are very small. Yes, we can reduce the losses in absolute numbers, but the remaining losses are concentrated on a very small area only. So we have to handle very high power densities, we have to remove heat from a very small area. And for that, we have invented a couple of nice technologies on the packaging side. So one is shown here. We call it also.XT. It's for discrete device, a little bit different approach like in the power modules, but giving you a similar result. What we are doing is, we replace the so-called solder, which typically connects a chip to its package on the back side by a diffusion solder process. So you see already directly here, the transition layer is much smaller. And it comes along with very nice additional aspect, and this is a significant reduction in the thermal resistance. That means we can allow to shrink our chips, but still due to the improved thermal performance, the customer can fully utilize also this shrink potential in his application. And a similar story, we do also have available for power modules. So in power modules, it's very important to understand the total picture for the thermal parts from the chip to the heat sink. And if you make a breakdown, this is shown here on the right-hand side, of the contributions to the total thermal resistance, so you will see that the so-called thermal paste, which you have to apply between the heat sink and the power module, has by far the biggest contribution. So it's not so much about the ceramics itself or the chip, et cetera, but the dominating part is the thermal paste. And with this picture in mind, a smart solution is, of course, to work on both, an improved ceramic, but taking also into account improved mechanical properties. And that is shown here on the left-hand side. You see in a colored way more or less a flatness of a ceramic on its back side. The standard ceramics has a relatively wide variation in total thickness, which you have to compensate then in the final assembly by the thermal grease. That means you have different thicknesses of thermal grease to more or less generate a smooth and fixed connection between the ceramics and the heat sink. On the left-hand side, our Infineon ceramics comes with a better thermal connectivity, but also with a significantly better flatness. So we address both here, this contribution from the ceramics, but to a large extent, also the contribution from a thermal paste, because we can now shrink the thickness of the thermal paste significantly. So in total, we can offer now again, especially important for silicon carbide, a very smart thermal solution in combination with wide-bandgap chips. Yes. And of course, there's always a question after our acquisition in 2018 where we are with our cold-split technology, which gives us the opportunity to significantly reduce the waste in the production of semiconductor or silicon carbide wafers from boules, which is done typically in most of the fabs today by wire sawing, very high material loss of 50%. We can do here a significant reduction down to 20%, but also then gives us more material from the same original boule, of course, so more wafers out of a given boule, but also, of course, it's a cost factor. Yes, and we are very proud to say we made it. So we went out of the start-up phase, and we're able to generate a production-ready technology, which is ramping right now. And for that, I would like to share with you a very brand-new video from our new pilot site in Dresden. [Presentation]

Okay. I hope you enjoyed this short clip. As mentioned, of course, we have to continue our path of innovation of driving up the capabilities enabled by wide-bandgap. And of course, we have to cover all the question of supply and volume. Just to give you an update where we are here. At the moment, we are transitioning our existing power fab in Austria and [ Villach ] in the 150-millimeter and 200-millimeter part completely to wide-bandgap, which should give us end of 2025 revenue potential of about EUR 1 billion, which is approximately 3x where we are today. And the next step is already in construction. You see here on the upper picture, our site in Kulim. And on the right, this is the third module of our Kulim fab, being completely dedicated to wide-bandgap production, enabling further 2 billion of production capacities. With this move, basically, in '27, we have a 10x higher capacity installed compared to what we have basically right now today. There is still some upside taking into account potential coming from 200-millimeter. This is something we will finalize and fine-tune during the upcoming calendar year. And therefore, we're very happy to share with you the message, we rock the ramp at Infineon. We keep our promises. And with that, thank you very much. And we are, I think, both ready for question and answers, Peter.

Okay. First question comes from Aleksander Peterc from SocGen. We heard yesterday from Coherent, formerly II-VI, they are exploring strategic options for the silicon wafer substrate in the [ EV ] business. Given the roughly 20% SiC wafer market share and the partnerships with both yourselves and with Resonac, that is also your supplier partner and, given the substantial move to deeper vertical integration in the SiC industry with your key SiC device competitors, including STM, Wolfspeed, [indiscernible], isn't this a perfect opportunity for Infineon to join this industry verticalization trend by striking a deal with Coherent?

Is this a perfect opportunity? I truly don't know. Since I was yesterday also at PCIM, this news did not reach me so far. And as I presented or we presented, I think we're pretty happy with the current setup. So if the opportunities come across, we definitely will have a look at it. But from today's perspective, we feel confident with the setup that we have. And so that's our plan A.

Okay. Different question from Gianmarco Bonacina. What will be the impact for Infineon from opening the new facility in Kulim in '24 in terms of revenues and in terms of costs?

Yes. Overall, that is now the very long awaited upside -- capacity upside that we will be able to provide to our customers. So the demand is huge. For GIP, we were growing now with a 50% CAGR for the last 5 years, of course, starting on the lower level revenue-wise. Now we talk about substantial 3-digit revenues. You know that we announced that we intend to achieve the EUR 1 billion overall for Infineon, EUR 1 billion revenue of silicon carbide around '25. And I think Kulim -- output of Kulim is prerequisite to achieve this target. And of course, if we execute as ambitious as we now plan, then maybe there's a chance to even pull in the time line a bit, it's not a question of demand. The demand is there. It's simply a question of supply and we would love to have this output 1 year earlier. But I think already next year and also with the ongoing ramp that still comes out of Villach, we are nicely set up to support this growth, massive double-digit growth for the years to come.

Yes. His second question is on what is the contribution of GaN Systems in 2024? Well, I think since we haven't closed yet that acquisition, question is a bit premature. And I'm not sure whether you want to comment on that any further.

Yes. I don't know how many questions are still to come, but a very brief comment. I think, of course, assuming now that the deal closes in the next couple of months as expected, the contribution will be know-how, right? So not to be underestimated, they provide a lot of know-how, also a very knowledgeable experienced team. And so the primary contribution for us will be know-how on the application, on the technology and also on the product side, and it's up then to the both teams combined to make something more out of it, where I'm pretty sure that this will be the case. And maybe also to comment on it from a GIP perspective, of course, we have a certain overlap in the area of 650 volt devices and, of course, applications. Also for us, it's very important and interesting to learn how the GaN Systems perspective looks on the industrial market. Our focus, as said, is IGBT and silicon carbide, but we have already prototypes, module prototypes based on gallium nitride at our customers. We have a couple of R&D projects, nothing productive yet, but extremely relevant and extremely interesting for us.

Okay. I have another couple of questions from Stephane Houri from ODDO. Can you please summarize all the SiC wafer supply contracts that you have right now? I think we did already on the slide. And how much they cover of the forecasted chip demand you see by 2027? I think the second part is the new topic.

That's a good one. I think I would have to ask procurement, sorry, for the second part of the question, but the ones are on the slide.

The question came in before you presented the slide.

Okay. Okay. Yes. But maybe just to repeat, right? It's Wolfspeed, it's Resonac, it's Coherent, it's SICC, it's TanKeBlue. Did I forget somebody, Peter?

These are the ones which are publicly known, there's a few more. Maybe to comment on the reach of those contracts. Of course, all of them have a different term. Some of them end before '27. That means we go into the arena, negotiate the next phase, this is ongoing. But the discussions today indicate that we are well covered. So we don't expect really that the material is going to slow us down.

Okay. About the 2 SiC supply contracts you announced, they are going to cover only Chinese client demand? Or is it also then being useful for other customers?

Well, first of all, front-up is that we don't distinguish the material selection according to a customer, except the customer requires that. We can do that, but it's not the front-up solution. And we are pretty sure that there's also, outside China, a willingness to accept Chinese material. Again, we can react. If this is not the case, if there are any restrictions coming up, for that, we have this multi-sourcing opportunity. But it's not just China for China, I would say.

Okay. One question that is -- I'm not sure whether -- can you do an update on the ramp of the different 300-millimeter plans you ramped in Villach, Kulim, Dresden? Maybe I can very briefly step in. Well, yes, Kulim, I think we just discussed, and that is what we are ramping there, the 200-millimeter or the wide-bandgap fab.

Kulim, 200.

Yes, absolutely. And interesting, the first 300-millimeter plant is, I think 2/3 or even I think even more 3/4 full. So the remaining clean room will be used in the next 1 or 2 years. In Villach, there is more room for buildup of capacity. I think here, we are at around maybe 30% of the clean room capacity, so there is more space to fill it. But at the moment, all fabs are running in a high utilization rate.

And we ramp with a higher speed.

And we ramp with a higher speed, of course. Now a question more back to you. Can you update us on the pace of the ramp of SiC revenues from '23 to '24 already for the next year? I mean we gave a guidance for this fiscal year, but...

Yes. So the challenge for us is very clear to simply keep the ramp rate. So as said, looking back now in average 50%, speaking for GIP per year, the bigger the number gets, the harder it becomes to keep this growth. But definitely, if you have looked at the slides, if we now assume overall market growth of 30%, definitely, our ramp rate will be higher, assuming that these 30% are somehow realistic, then we will continue to gain market share because we significantly, for this period, considered '23, '24, we will be able to increase our output at a higher rate, so higher CAGR.

Okay. And since we are here on the power fab, what kind of revenues do you expect in GaN in the coming 3 years?

Okay. That's a good one because, to be honest, these are -- I would have to ask my [indiscernible], so to say, the PSS guys. It's definitely something in the range of double-digit euro. Do you have any numbers?

Yes. I think they go into the double-digit revenue this year and the CAGRs we have shown. So it's still not yet at the level of silicon carbide, but especially the last year was extremely successful also for our gallium nitride guys. So this is starting to make a lot of fun.

Okay. So the -- yes, and then from our official perspective, we didn't hand out a guidance on that material. But for sure, it's ramping fast. And yes, I think the whole technology is just a couple of years behind silicon carbide and in some ways it's going to be the same story, I think.

Yes. And important to understand that it's not competing against each other, right? So we have an overlap at the -- many times already mentioned [ 600, 650 ] thing, but it's always a competition silicon or gallium nitride, silicon or silicon carbide. So there's a good argument to have silicon carbide and gallium nitride in the portfolio.

Okay. So next question comes from Andrew Gardiner from Citi. First one on -- again on SiC substrate sourcing. You have highlighted the cost benefit of the newer low-priced SiC substrate suppliers. How does Infineon plan on using that substrate cost advantage? Will you be more aggressive on device pricing to gain share more quickly? Or will you try to hold pricing and drive higher margins?

Guess what? I mean, first of all, of course, we are in a competitive environment. So the pricing is set via competition via demand -- by demand/supply. That's the usual game in semiconductors. So we have to adapt our pricing strategy to the competitive nature. But of course, if you're able now with the help of lower pricing on the material side, then you have more freedom to operate. First of all, we have to achieve the target operating model, which is communicated from the Infineon Board. And if we have more room to improve profitability, while we, of course, continue to grow with or above the market, then we go for more profitability for sure.

Okay. I'm switching back to that slide because the next question is concerning that. So 10% to 20% GIP growth range, from Slide 6, your assumption seems to be that you will grow at the same rate of the market or to go on to roughly 30% market share. In SiC, however, you're targeting share gain from your current level through the end of the decade. Are you therefore assuming any decline of share in silicon? Putting the question another way, it feels like the competitive dynamics are more intense in the new growth areas of SiC and GaN, while it is less in SiC, if competitors are less focused on SiC -- in silicon, sorry, if competitors are less focused on silicon, while Infineon continues to innovate. Why shouldn't you gain share over time? Very long question.

Yes. Why shouldn't you gain share over time in which direction? In silicon or silicon carbide?

I think in silicon. One is the assumption, yes, we improved our share in silicon carbide. Yes, today, we are around, give or take, 20% and then the target is to become 30%. So I think that is communicated. We didn't communicate a target for silicon-based stuff.

Yes. I mean, from the market research, we see that for IGBT in the industrial area, our market share is in the range of 30-something-percent, depending on the precision of these market research numbers, but 30-something is somehow reasonable, I would say. Since the silicon market continues to grow, we simply want to keep our market share. That is meaning at least growing with the market, if there's the opportunity regarding supply/demand and also performance. Why not also grabbing one or the other percentage point on market share? I will not refuse for sure, if the profitability and the margin is accordingly. And also the silicon innovation road map is not yet completely at the end. I mean, it definitely is flattening. But as shown today, the IGBT7 is now into the market, very well received, we must say, also here, we could do even more revenue. Therefore, we need to ramp the already mentioned facilities with the highest speed, while we still have also the IGBT8 now approaching the market, being in the final phase of development. And so definitely, we want to keep or if there's the opportunity, even expand the share in silicon, while at the same time, of course, we have more room for growth and also more room for market share on wide-bandgap.

Okay. I have another one from Janardan Menon from Jefferies. Infineon seems to be many generations ahead of its competition in SiC technology, especially on trench. Yet this does not seem to be resulting in any clear market share advantage in the automotive market in terms of design wins and revenues, even excluding Tesla. Why is that customers are not recognizing the significant cost and/or quality advantages of Infineon's SiC by awarding a higher design win market share? Is this mainly because of capacity constraints? Will you expect your market share to increase once capacity constraints reduce?

To put the answer simply, I would say, yes. Indeed, at the moment, the design wins rolling in, in this one slide I've shown are very relevant, and we are internally fighting for capacity. So we, of course, have an agreement how to allocate the volume, but also in the automotive area, we are supply-limited, meaning if now with the activities operations undertaking, if we are able to increase the ramp speed substantially, I expect increasing market share, not only for industrial space but also for automotive.

Okay. Within the overall longer-term growth of 10% to 20% for GIP that you are now guiding to, what will be the growth rate of only the renewable energy portion?

Here, I would say, ballpark number, of course, it's a bit the crystal ball, but I would expect something like 20%. Of course, now you have to differentiate a bit again, there's competition out there, which you take very serious. PV is really booming, and it will continue to boom. That's very clear, what I hear also from the customers here during the fair and elsewhere, while wind has, of course, always longer lead times, right? So the investments for wind, especially then also the -- granting the permission to build the wind power park onshore, but even more complicated and challenging technically also offshore. That takes time. Typically, we discuss about years. We realized in Europe, especially also in Germany, that these lead times for getting those permissions, of course, massively need to be reduced. There's a lot of activity coming. And there are other areas in the world where these lead times are anyhow very shorter, and we see this kind of a demand really getting prepared, so to say. It's continuously growing, but there will be more boom also coming from wind that is clearly visible in the next 1 to 3 years. And we are preparing for it the best case we can -- the best way we can.

Okay. There's one directly fitting to you from Sebastien Sztabowicz from Kepler Cheuvreux. Do you see any opportunity for silicon carbide within electric drives? Or do you think IGBTs will remain the main technology there in the coming years?

I suppose it's about the traditional motor drives -- industrial motor drives, yes, definitely. We do have already a substantial revenue contribution today from certain segments in industrial drives. So Peter showed this very innovative new motor solution. So there's a couple of very nice projects stepping in. Yes, of course, if you look to the -- what we call general purpose drive, so the absolutely low-cost standard performing things, it will take some more time. But in the high-end side, a lot of designs are going in or taking place right now. And not to forget with silicon carbide, we are enabling the [ reinverterization ] of [ noninverterized ] motors. That's a huge potential for future energy reduction for operators of such a system. And this is something what also slowly starts to develop right now. So therefore, clear yes. We expect also drives to be penetrated by silicon carbide.

Okay. And being a layman, I think robotics and server drives is an area where silicon carbide is fitting especially well.

Absolutely. So that's where we have today already our very nice first revenue contributions. Again, the value proposition coming along with a MOSFET compared to an IGBT is so substantial in those applications that it's more or less a no brainer to go for SiC. And again, we can shrink the system, we can save on cooling, save materials. So there's a lot of very strong arguments, yes.

Okay. Great. And I have a follow-up from Aleksander Peterc. What is your assessment of Soitec's Smart Cut technology? Are you still evaluating it? Do you think it has a place? Can it be game-changing if it delivers on its promises?

And before Peter takes over, I would say that should be asked to Soitec. But as we are an innovative company, of course, we look into all kind of interesting innovations, right?

Absolutely. It's still ongoing, to be very honest. So still there's a couple of things to look into because it's very important. It's a different material. It's not like every other silicon carbide and therefore, we have to check various additional things. It's not like another supplier, which we usually can [indiscernible]. So this approach has a lot of potential, but we have to be sure that we don't bring in any, let's say, hidden problem for the rollout in high volume. And that's -- we want to be sure on that side before we really comment on this as a final assessment.

Okay. Yes, that's again a bit difficult for us. Could you outline what are the difficulties at the 200-millimeter SiC that the industry is experiencing and what is Infineon's plan in terms of timing for the migration from 150 to 200? I think the second part is easier to answer.

Yes. So we are starting the conversion projects as we speak. We have 8-inch wafers, we meanwhile also get 8-inch wafers from more than one supplier. That is very good news. Of course, it's in the very early stage being also cost or price per square millimeter, still innovation style. And so the time line, maybe I already mentioned this, we assume that this conversion will take roughly 2 years, meaning we expect production readiness in around '25. Regarding the technical challenges?

Yes. I think we see the similar thing what we all experience with each and every transition from a smaller diameter to a larger one with silicon carbide. So at Infineon, we experienced it starting from 2-inch and exactly the same problems we see. Of course, it was always at a smaller scale compared to now where, of course, the total volumes are much higher. But this is a typical transition phase where initially, yields in crystal growth are not where they are with a smaller diameter. So it's simply a learning curve, which is also a little bit bound to the technology itself. So it's not a surprise for us, I would say. And what also steps in here is that, of course, at the moment, the focus is on ramping 150 millimeter. That's where the volume is today, where also the suppliers have to deliver. And yes, once more effort is spent to 200-millimeter, those problems will also go away.

I like the recent quote of Jean-Marc Chery from STM very much. He said 200 millimeter is not a piece of cake.

We agree.

Absolutely.

And with that, I have no further open questions. So yes, thank you very much for joining us in this call. Any follow-ups, always happy to take it in Investor Relations. So Peter and Peter, thank you very much for your contribution and for the insightful presentation. And -- yes.

Thanks, Alexander, for the moderation, preparation, and thanks also to the team and the audience. Thank you.

Thank you very much.

Thank you.