Good afternoon, and welcome to our Investor Day 2023. I am Victor Bareño, Head of Investor Relations at ASM. It's great to see so many of you here in person today. Many thanks for joining us, and also a warm welcome to everyone on the webcast. Before we get started on the presentations, I'd first like to go over a few housekeeping items and introduce the program for today. First, an instruction on safety for our guests here in the room. The exit is on this side of the room and also in the back. And in case of an emergency, please follow the instructions of the hotel staff. We also kindly request you to keep your phones on silent during the presentations and the Q&A. Let me now walk you through the program.
We'll start with Benjamin Loh, our CEO, who will talk about the long-term trends in our industry, our growth through innovation strategy, and how ASM remains well-placed for continued outperformance. Benjamin will also provide more details about our commitment to sustainability. Next, Han Westendorp, our Corporate VP, Global Marketing, will discuss the outlook for the wafer fab equipment market and opportunities for us for continued growth, such as in the transition to Gate-All-Around. Han will be followed by Hichem M'Saad, our CTO and member of the Management Board, and he will talk more about the technology roadmap, innovation in our key products, and also highlight a number of new, exciting applications.
After these presentations, we will have a break of 25 minutes, and after the break, we will continue with Kent Rossman, Senior VP, Global Operations, and a member of the Executive Committee, who will discuss how we continue to deliver on our customers' growth needs by building a resilient supply chain and expanding our manufacturing capacity. Kent will be followed by Hichem, with his second presentation, this time about ASM product sustainability, the actions that we are taking to reduce GHG emissions and the environmental impact of our products. Finally, our CFO, Paul Verhagen, will give a financial update and present our financial targets. Paul will also talk about sustainability initiatives, specifically our net- zero by 2035 target.
After the wrap-up by our CEO, we will have a Q&A session of 50 minutes, in which we will be taking questions from the audience here in the room. We will have one Q&A session with the whole panel on stage, so if you have any questions during the presentations, please make a note and save them for the Q&A at the end. Let me also remind you of the forward-looking statements. The presentations and the Q&A today may contain information relating to ASM's future business and results in addition to historical information. Please also note that today's presentations are being webcasted. The recording of the webcast and the presentation slides will be available after the event. So with that, let's get started. And now I'd like to welcome Benjamin Loh to the stage. Benjamin?
First of all, good afternoon, and, you know, welcome to ASM's 2023 Investor Day. It's actually great to see, you know, everybody here in person. Two years ago, we had to do a small scale one in Amsterdam, partially because of COVID. So it's great that we can do it in live in person. The other thing is, I want to thank everybody for taking the time to join us this afternoon. Very much appreciated. As Victor has introduced the agenda, so we have put together a couple of topics, and we will be going through them this afternoon.
They basically range from, you know, market, technology trends, industry trends, sustainability, and we'll try to show you and explain to you how our Growth through Innovation strategy fit with all of this. I will start with an overview of what will be the presentations that will take place after me throughout the day. In other words, I'm going to frame the day with the presentation that I will start with. And my colleagues will then elaborate on specific topics, you know, in their respective area of responsibility, and ending with Paul Verhagen, our CFO, basically giving the projections for 2027.
Before jumping into the first presentation, I would like to just share two minutes of your time, you know, to take a look at the latest corporate video that we have made. Hopefully, you enjoyed it.
Since the sixties, we've been making the equipment that builds the chips found inside your telephones, gaming consoles, laptops, and more, way more. Our technology has been the driving force behind almost everything you've grown to love and rely on every day. The technology that bridges gaps, powers connections, and drives us forward. We stand at the crossroads of life's most pivotal moments, enabling life-saving medical marvels and strengthening human bonds through 5G and cloud computing. In other words, we're the people to help make this happen. And this....And next. But it's not just any people, it's these people, her, him, them, and everyone in between. Because we're more than just a tech company, we foster a community of brilliant minds who propel progress, shape the evolution of what's possible for AI, biotech and electric vehicles, and empower global advancements and global sustainability. We're not just enabling the world of today, we're moving the world into tomorrow. At ASM, we're ahead of what's next.
I hope you enjoyed that as much as we do. We are really excited about the whole industry and what we can offer. I'm going to jump straight into my presentation: Growth through Innovation. As I said, you know, I am going to provide basically the framework so that the rest of my colleagues are going to take you through for the rest of the day. We're gonna start with, you know, a couple of key takeaway messages, and I will also repeat them at the end, because I think that's the most important. So if you fall asleep, just remember these four or five, six messages. You know, semiconductor continues to grow.
We are very bullish about the growth in the semiconductor industry, despite what you call a downturn that we are facing this year. Secular trends remain, and in fact, when you look at AI, electric vehicles or electrification of vehicles, it's even accelerating. We believe that our Growth through Innovation strategy is working. We are delivering results. If you look at the last three years, we have, you know, grown revenue with a CAGR of 35%. And if you stretch a little bit longer horizon, you'll see that we significantly outperform our WFE growth. I have a chart that speaks to that later on. We are witnessing a lot of technology inflections. Technology is changing fast and coming at us fast, but we are very well positioned.
We are well positioned for the forthcoming transition to Gate-All-Around, advance our memory with our ALD and with our epitaxy, silicon epitaxy products, where, you know, in, in terms of ALD, we continue to maintain our market-leading share. And in silicon epitaxy, we are continuing to build and grow our share from there. We have also, since last year, added, you know, a new product, silicon carbide epitaxy, in a very fast-growing market, and that's going to help boost our revenue as well, going forward. And this is driven by electrification of vehicles and other power applications.
Sustainability is something that is one of the six strategic objectives that we already shared, 2021, and, you know, I think we have made a lot of progress there, including the most recent, which is our Net2035 targets have been verified by Science Based Targets Initiative. And last but not least, we have also, in this morning's press release, kind of upgraded our 2025 targets to EUR 3.0 billion to EUR 3.6 billion. And we are giving a new target for our new midterm, which is 2027, of EUR 4 billion to EUR 5 billion, growing at a CAGR of about 11% to 16%, and we reiterate our operating margin of 26% to 31%. So before I go any further, just a quick introduction of my colleagues.
So early last year, we decided to create what is called an executive committee within the company as the company was growing, which consists basically of the three management board members, myself, Paul and Hichem. And we have also included three senior executives. One of, you know, Kent Rossman, who is also here with us, who runs our operations and will be also presenting to you what we are doing as far as supply chain operations and so on. And we have recently added Edyta Jakubek, who is now our Chief People Officer. So maybe just to recap, what are the achievements that we have made since 2021? Revenue, I have spoken to that. And, you know, the subsequent slide will show the outperformance compared to WFE.
We have invested significantly in growth and innovation. I think that's something that is a lifeline. And you know, because of all the opportunities in front of us, we continue to invest heavily in R&D, growing our headcount by 54%. From an operating CAGR point of view, I think we are also doing well, and we have more than triple free cash flow during the last three years. Expansion of our facilities to make sure that we have enough capacity. I will show later on that we have completed our expansion in Singapore, and we have just broken ground in Korea. That's good enough to take us all the way until 2027, based on our plans. Last year was a little bit something new.
First time in 18 years that we did acquisitions, and I will touch on that later on as well. But we have now, with the acquisition of LPE, entered the very rapidly growing silicon carbide market. Leadership in terms of logic, foundry and, you know, wins for, let's say, process tool of record. We are coming to the end of, you know, the first 2-nanometer Gate-All-Around development, and we know where we stand. We continue to be in a very good position, maintaining our market share, and have secured many, many wins. So we are waiting for, you know, customers to go into pilot, and then subsequently high-volume manufacturing. One of the 6 strategic objectives that we have is also to increase our share in memory....
Over the last two years, I think we have, you know, done that. First of all, through penetration of high-k metal gate, and subsequently with ALD gap fill in 3D- NAND. I think that's helping us to increase our market share. And at the end of last year, memory became about 19% of our total equipment revenue. Last but not least, the net zero target again, but at the same time, I think, you know, we have been trying to do as much as we can, you know, in terms of renewable electricity, and last year we were at 76%, and we'll continue to increase. So this is a chart that shows that over the last seven years or so, if you index WFE, you'll see that we are growing at twice the rate of WFE.
We do believe that going forward, we will continue to significantly outgrow WFE growth over the next couple of years. Maybe a little bit on what we see in the market. Trends in general, I think digital transformation will continue to drive semiconductor usage. And the general consensus is that by the end of the decade, you know, semiconductor sales is going to be, you know, almost twice of what it is today, hitting $1 billion, and continuing to grow beyond that, past the decade. We think that, you know, all the secular trends, whether it's AI, whether it's, you know, edge computing, cloud, you know, they are all still very valid and still growing. So we are very bullish on this. What does this mean then for WFE spending?
We think that WFE spending, after coming down this year, should start to recover, starting from next year. Now, of course, you know, this is just the projections that we are getting from, third-party, you know, data research companies, so still very much left to be seen. But there are a couple of factors that are driving, you know, WFE. Of course, you have the changing of technology, further scaling, transition to Gate-All-Around. You also have, multi-year investments that have been announced. Some of them actually don't come into a high-volume manufacturing until even 2027. Sovereign efforts is also helping to build more fabs, and the good thing is, a lot of them are on the leading edge.
Last but not least, I think over the last couple of years, we have witnessed fairly significant investments are coming from China. If we look at where our projections are, which my colleague, Han, will share with you, our 2025, you know, projections are based on 2025 having a WFE of $100 billion, and for 2027, it's based on $120 billion. AI will become, in our opinion, a very significant driver, and, you know, we are very excited about this. It's going to result in more data centers, more processing power, more high-speed stuff that is going to be required. And you know, there is from Gartner that by 2027, at least 30% of all logic devices will have AI-specific functions.
I think what is more important is what does it really mean for us? The good thing is, you know, all the AI stuff, especially the recent generative AI boom, they are all using advanced semiconductors. So GPUs, NPUs, you know, that are made at the leading-edge nodes. At the same time, they require a lot of memory, high speed, high-bandwidth memory as well, and that's also good for us in terms of, you know, adoption of high-k, you know, more metal layers, and also adoption of Epi silicon epitaxy. Our strategy hasn't changed. So two years ago, we shared, probably for the first time, our Growth through I nnovation strategy, the six pillars, and actually, it hasn't changed.
The only change is that we have added selective growth now, silicon carbide epitaxy, and that has been the only change due to the acquisition of LPE. Other than that, we are just busy executing on the strategy and trying to grow the company. Of course, the strategy is underpinned by certain a couple of key enablers, best people, you know, flawless operation, you know, leading-edge innovation, et cetera, and we continue to build on that. All this with the purpose of, you know, improving people's lives, you know, by advancing technologies and unlocking potential. Last year was the first time in 18 years that we made acquisitions in this company, and we did 2 acquisitions: Reno Sub-Systems in March, LPE in October.
You know, going forward, M&A will be an opportunistic growth opportunity for us, but we are also very focused. We will only look at acquisitions that are in the semiconductor deposition space, either equipment or a technology that boosts or enhance our equipment. That's what we will continue doing. So Reno Sub-Systems was a technology acquisition, and of course, LPE, you know, was an equipment acquisition that allows us also entry into the silicon carbide space. And with LPE's acquisition, we have now really entered into a very rapid, rapidly growing market that is still evolving. And, you know, we are, I would say, doing significantly well since the closure of the acquisition in October last year.
We have shared with everybody major customer wins in the U.S., in Europe, and we continue to have a lot of traction with the other major customers. So we are very bullish on this part of the market. And, you know, the main driver here is, again, electrification of vehicles, renewables, power generation, et cetera. If you look at the portfolio of products that we have, so with the acquisition of LPE, we now have added silicon carbide epitaxy. The rest kind of remains the same, you know, thermal ALD, plasma ALD, silicon epitaxy, PECVD, and vertical furnace... We are also growing our spares and service business very well. If you look at the last three years, we have kind of grown it 17% CAGR, and we continue to target healthy growth over the next couple of years.
Of course, we have, you know, kind of changed our model also a little bit over the last couple of years with outcome-based services, which gives us more traction, more long-term engagements with our customers, and that's been helping us. And we have a slide here that basically says, you know, a lot of our spares and service business is increasingly being driven by outcome-based services. We came from a time when we were very transactional. And over the last three to four years, we have become much more focused on developing longer-term relationships with the customers, helping them to get better on-wafer performance, you know, reducing costs, better productivity, while in the meantime, having a longer relationship with them.
So that has been very helpful for us, and we do believe that over the next couple of years, we will continue to be able to build this capacity, and my colleague, Kent, will talk about this. At the beginning of this year, we completed the fit-out of our Singapore facility that became ready at the end of 2020. So with that, we are basically fully fitted out in Singapore, and we have quite some capacity, and mainly this is preparing for future growth. We do see the growth coming in the next couple of years, and we want to be ahead of it rather than behind. At the same time, we have broken ground on a new building in Korea.
So even that, even though that is mainly focused on expanding our R&D lab space, capacity, it also adds a manufacturing capacity. So with that—with those two added together, we are very confident that we have enough capacity to meet our 2027 targets. People is really critical for the company, and I would say people really lies at the heart of our success, and we have been trying to do a lot. If you look at what has happened over the last three years, we have more or less doubled our headcount within the company. We had had to hire a lot of people in very challenging circumstances, but I think we have been quite successful. Now, hiring the people is one thing.
We have also been trying to build on the culture of the company, making sure the employees are engaged, that they understand how they contribute to the strategy. We are also working on things like gender diversity inclusion, and if you look at, you know, for example, our gender diversity, we are slowly creeping up. That's something that is also important for us. And of course, in terms of, diversity, geographic, or let's say national, diversity, we have, 60+, you know, nationalities now working in the company. Sustainability is something that we have already, you know, decided two years ago when we did our Investor Day. It's going to be one of the six strategic objectives for us, and I think over the last two years, we have done a lot.
Our sustainability strategy is focused on the five pillars: innovation, people, planet, responsible supply chain, and governance. I think in all areas we have improved significantly, and we are still continuing to do a lot. My colleague, Hichem, will also share with you, for example, in terms of, you know, embedding sustainability into our product design, what are we doing about that? So good to share with you that we are doing a lot there. Besides the net- zero 2035, you know, targets being verified by SBTi, we have also been fortunate that we were actually one of the founders of the SEMI Semiconductor Climate Consortium. In fact, my colleague now chairs the consortium on behalf of SEMI. We were one of the first, together with 60+ other companies.
The reason for doing that is, you know, if we want to get to net- zero, we cannot do this by ourselves. It is an entire industry effort, so we need to work with the suppliers, we need to work with our customers, and hence, the birth of the Semiconductor Climate Consortium. Again, I just want to end with the key takeaways. We believe semiconductors remain robust. You know, despite the downturn this year, I think the growth will come back very quickly and will continue to grow over the next couple of years. We are actually excited and bullish about our strategy. We think that we have delivered over the last three years with growth in revenue and have significantly outperformed WFE.
In terms of the technology changes that are coming in front of us, we think that we are very well positioned, whether it's the transition from FinFET to Gate-All-Around, whether it's advanced DRAM, advanced 3D- NAND. I think we are all, you know, prepared and, very well engaged, with our customers for all the forthcoming opportunities. We continue to, maintain, and try to increase our share in ALD and at the same time trying to increase our share in silicon epitaxy. Those are the two of the si- key targets that, we have as our strategic objective. We have added silicon carbide, a fast-growing area that we are very excited about. The net- zero targets, I won't, I won't repeat that again. And last but not least, we upgrade our 2025 targets.
Revenue-wise, we are going to target EUR 3 billion to EUR 3.6 billion, and for our new mid-term 2027 targets, we are projecting EUR 4 billion to EUR 5 billion, which is a CAGR of 11% to 16%, and we reiterate our operating margin of 26% to 31%. With that, I'm going to, you know, hand over to my colleague, Han, that will talk to you about the market outlook, market share, etc. Thank you.
... Thank you, Benjamin. And also, first of all, I would like to thank you tremendously for all showing up here, spending the time to listen to how we're doing and what our plans are. And also in my case, I will start with key takeaways, and also here, please try to remember this, and you can fall asleep in between. I'll remind them at the end. We see continued digitization, growth in AI. I will speak somewhat in detail on AI and what it could mean to us. Vehicle adoption trends are all expecting to continue to drive growth in semiconductors strongly in the coming years. In my presentation, I will update on the market size of single-wafer ALD.
Two years ago, we gave an update on the size of that market to reach by 2025, between $3.1 billion and $3.7 billion in size. We are now updating, giving a first update for 2027, $4.2 billion to $5 billion size by 2027, and we're basically confirming the number for 2025. Our views on the size for 2025 have not changed since 2021, when we gave that forecast. We're updating it for twenty-seven, with the new numbers, $4.2 billion to $5 billion. Silicon epitaxy market. We are also updating the forecast that we gave two years ago for 2025.
We're forecasting now a range from $1.9 billion to $2.3 billion by 2025, and by 2027, we forecast that size to go up somewhere between $2.3 billion to $2.9 billion, and we'll explain you the key drivers behind that market. The transition from FinFET to Gate-All-Around is ongoing or is being prepared, I should say, eh? We are expecting pilot investments next year and full steam HVM in 2025. But of course, for years, we have been working with customers on qualification of certain applications that we have at try to have in that transition.
We will share with you that we are expecting a $400 million SAM increase, a served available market increase, as a result of that transition for between single-wafer ALD and Epi combined, and I'll speak to that more in detail later on. Then the market outlook for silicon carbide epitaxy is very strong, so I'll share with you a forecast, a third-party forecast on silicon carbide devices. We will not be ready yet to really share with you a silicon carbide equipment forecast because the market is so much in development, very healthy development. But if we would share a forecast with you right now, it could be different three months from now.
My colleague, Hichem, will speak in detail on the technology side of what we're doing in silicon carbide epitaxy. We are basically saying is that we believe we're gaining share in this market in this year. Developments are going well, and it is a market that we are actually very pleased with, that we stepped into, also because of the synergies between silicon epitaxy and silicon carbide epitaxy. So the market outlook. Let's first look at the mega trends behind what's driving the semiconductor market. Smartphones have become, basically become a de facto requirement for life.
Even young kids, older people, everyone has a smartphone, and already for, you know, a decade, for a decade, decade and a half, continuous new capabilities are brought into these smartphones, which means, IC development, chip development, and it also means, more equipment that we provide. Security is also requiring more capability of of chips. Autonomous vehicles, electric vehicles, not only silicon devices, but also silicon carbide devices are needed here. And then, of course, AI, which is on the top right corner of this list, a tremendous boost behind the semiconductor market that we expect in the coming many years.
Specifically on AI, you see, in the wheel on the left-hand side, a human face that I will speak to more on the next slide. But first of all, all of these trends that I shared are basically driving the semiconductor market to roughly double by 2030 compared to where we are today. And yeah, there is basically a general understanding that that development, that growth, will continue beyond 2030.
And one of the reasons is that, artificial intelligence, basically to mimic or, or to copy or to help the, the human, the human brain in many of the things that, a human being is, is used to doing on his own, 30, 40 years ago, tremendous development in this area. So this particular picture signifies that silicon, yeah, silicon basically is, in small pieces printed on this face because all of this capability is driven by semiconductors, by silicon. So all of these different, capabilities, like machine learning or generative AI, is basically driven by, by silicon, which means more, more silicon fabs, which means more silicon, epi silicon, equipment that, that, we can provide.
Also, if you look at the recent developments in the last six months, there is so much new capability being thought of and being developed. Certainly in a few years, there will be products delivered that we are participating in, that are basically not invented yet today. If you look at, Benjamin showed this chart already. If you look at the development in logic chips, yeah, Gartner predicts that by 2027, more than 30% of semiconductor chips will have an AI function. So it means, what it means for ASM, increased capacity. Many of these advanced learning chips, for instance, require advanced computation capability, a lot of memory capability, and that goes into the heart of what we do in ALD and Epi.
So many of the, let's say, ALD, leading-edge ALD and epitaxy applications that we provide are needed to build those advanced chips used for AI. If you look at the wafer fab equipment forecast, wafer fab equipment forecast is down this year, but is generally expected to rebound in 2024 and continue to grow thereafter. Many third-party forecast companies are sharing that view. We're sharing here two external forecasts. They differ in absolute value quite a bit. 2026, 2027, the difference in views between these companies is, yeah, could be because recently, TechInsights has updated its forecast for the future years. Gartner is expected to do that within the next week or 10 days.
We have based our long-term plans on a WFE number for 2025 of $100 billion, which is, and $120 billion for 2027, which is roughly the average of these two forecasts. So we, we use everything that we share in a forecast, that number for 2025 and 2027 as the WFE base. What you see also is a Gate-All-Around transition from FinFET to Gate-All-Around, is will start to play an increasing role in the, in the actual investments in 2024 and beyond. So we'll, we'll, we'll expect pilot investments next year, but towards 2025, 2026, and 2027, it is our expectation that roughly 40% of investments are, are driven by Gate-All-Around.
So we share in this circle on the middle right a number of $50 billion. 40% of that $120 billion WFE number is 48. So roughly $50 billion, a very sizable number, is related to Gate-All-Around related investments, and that's smack in the heart of where we are developing our new ALD and Epi products. So growth opportunities. I'd like to speak now about growth opportunities. Here I show on one slide a number of different applications where my colleague, Hichem M'Saad, will speak more specifically on some of them. ALD high-k gate, an area that we have been active in for since the beginning, in the mid-2000s, and that is progressing from FinFET to Gate-All-Around right now. Hichem M'Saad will speak about that.
Gap-fill is an important application that drove quite a bit of 3D- NAND business for us last year. ALD metal is also an area where we're very active, specifically on ALD metal. Hichem will speak more about that later on. Silicon carbide power, so in the top, on the bottom left, an example of a device where a silicon carbide layer is shown as part of making a power device. Epi, general activity, and selective ALD, also a key area that we are active in, and we will be sharing more about that later on. If you now look at the single wafer ALD market, the size of it, this is what we showed two years ago in 2021.
In 2020, the market was $1.5 billion in size, and we forecasted that market then to grow to $3.1 billion to $3.7 billion in 2025, with a larger than 55% market share. Driving this market are the different applications that are shown here on the right in logic foundry, in memory, metals, high-k gate, VT tuning, and also metals in memory and selective ALD in both logic foundry and memory. Now, let's first look at how the market has behaved since 2021. So last year, the market developed basically as forecasted, $2.6 billion was the size of the market.
We have a 55% market share for 2025, and we still have that market share. We had that market share last year, and we confirm that market share for 2025. We also confirm the size of the market for 2025. We have been in this market for a long time, since the beginning, so we model the size of that market in detail, working with our early development groups, with our product lines, with our sales, with customers. We have a good view of what applications are being worked on, what applications are being considered, and then assess whether certain applications will be adopted, whether we will win share, whether the competition will win. So basically, that is all behind these forecasts.
So also for 2027, we have updated the market size now, somewhere between $4.2 billion to $5 billion, market size, and we are reconfirming also for 2027, a share of larger than 55%.... The CAGR of that market is then from 2022 to 2027, somewhere between 10% to 14%, so still overall, a very healthy development. If we do the same thing for silicon epitaxy market outlook, we shared market was 0.8 in 2020. We forecasted that market to be somewhere between $1.5 billion to $1.8 billion in 2025, driven by leading-edge devices, where epitaxy is used, and non-leading edge. Non-leading edge, specifically non-leading-edge foundry, so basically devices of 28-nanometer, 40- nanometer, we call that non-leading-edge foundry, and wafer power analog.
Those two segments are driving that non-leading edge market, and we have the leading-edge section. So let's first look now at what happened in 2022. In 2022, that market, that Epi market, grew very strongly. It ended up being EUR 2 billion in size, strongly driven by the non-leading-edge market. There was growth in the leading-edge part of the market, so there was growth in investments, but by and large, the growth of that market was driven by non-leading edge. Power wafer analog, there was healthy growth there, but the strongest momentum for us in that segment was in the wafer market. Very strong development of our Intrepid ES system that we announced a few years ago.
The non-leading-edge foundry market, there were stronger than expected investments there, also driven partly by China, in which we have a weak position. So as a result, our market share went up a couple of percent. It was around 14%, in the year prior, in 2021. It went up a little bit, 16%, but that was our share in 2022. So now for 2025, we're providing an update of the size of that market to grow to $1.9 billion to $2.3 billion, so larger than the forecast that we gave two years ago. And we're reconfirming the share target of 30% that we gave two years ago. So we're still aiming for a target of 30% share in the epi market, driven by leading edge.
So there is structural growth in that leading-edge section, which you can see on the chart on the bottom, greenish color. In that size, you see that's bigger than in 2022, where we expect to grow share. And also in the non-leading-edge section, in the wafer power analog, we could expect to continue to do well, and we also do not think that the strength in 2022 of that non-leading-edge foundry will sustain. So we also think that that size of that portion of the market will not be as high as in 2022. So because there is... We expect there to be a change in the market mix towards leading edge, that we will grow share and maintain that size target of 30%.
Then for 2027, we give you, for the first time, a forecast for that market to grow to somewhere between $2.3 billion and $2.9 billion by continued growth in, in the leading-edge epitaxy, market, driven before an important part, by the transition from, FinFET to Gate-All-Around. The wafer and power market, also in silicon, we expect to do, continue to do well in, and also in non-leading-edge foundry. That's an area where we don't expect that strength to remain. We will continue to have there a weak position, but the share growth for us, for us, will mostly be driven by the leading-edge, portion, is our expectation.
So the CAGR for silicon epitaxy market from 2022 to 2027, because of its strength it had last year, and the relative strength also still of that, the non-leading-edge segment, we expect the CAGR to be 3% to 8%, but the leading-edge segment of that to be 10% to 15%. It says 12% here, but we mean a 10% to 15% range, CAGR for the leading-edge segment of the Epi market. If we look at the transition from FinFET to Gate-All-Around, and we look at what it means for us in terms of served available market, between single wafer ALD and Epi, we forecast that the transition will be a $400 million increase of our served available market, just by that transition from FinFET to Gate-All-Around.
Gate-All-Around is, will have more ALD applications. We have a good view of that. We have assessed that that is the size, increase that we see in the served available market. And in that transition, we expect that we will maintain our leading share in single wafer ALD, that we will grow our share in Epi, like I shared on the previous slide, that 30% share target, and we'll get to a $4,400 million SAM increase, which is a very healthy development for us, and that's what it means for us, that the industry is going to, Gate-All-Around. Then on silicon carbide epitaxy. Silicon carbide epitaxy, we're sharing here a forecast of silicon carbide devices.
This is an external, third-party forecast from the company Yole, showing that, between, let's say, automobiles, electric vehicles, inverters, for that market, that market is strongly growing. There are other drivers for that market that are still relatively small, but expected to continue to drive it later on, beyond this time horizon. And this is driving an enormous amount of development in silicon carbide epitaxy equipment. So the key drivers for this investment, again, are EV adoption, which is accelerating, power devices for electric vehicles. And we experience, at the moment, a strong pull for our silicon carbide epitaxy equipment that is in need of, let's say, better, overall better, more productive, providing silicon carbide epitaxy layer with fewer defects. And Hichem will speak more in detail about that in his presentation.
To summarize, the key takeaways: we see a continued digitization, growth in AI and electric vehicles, adoption trends that will continue to drive the semiconductor market. That means more fabs and more equipment. Market for single wafer ALD, we're confirming the forecast for 2025 that we gave two years ago. The forecast for 2027 is shared $4.2 billion to 5 billion, with a 55% share target. Silicon Epi market is expected to grow to $1.9 billion to $2.3 billion by 2025, a higher target than two years ago. The new target for 2027 is $2.3 billion to $2.9 billion, Epi market for 2027.
The transition from FinFET to Gate-All-Around will deliver a $400 million SAM increase for us, and then the market for silicon epitaxy is strong, supported by a growing market for electric vehicles. Strength for us, we expect a strength in the silicon carbide epitaxy equipment. With that, I conclude my talk and hand it over to Hichem to continue the presentations.
Good afternoon, everyone. It's my pleasure to talk to you about the latest development we have in our R&D, and also products. Let's first start with the key takeaways. First, our products and R&D portfolio are very well positioned to capture growth opportunity in both logic and memory. The second takeaway is that we see trends toward new materials and 3D scaling actually accelerating with further adoption of ALD. Here we see three inflection point. The first inflection point happening is that molybdenum ALD is replacing tungsten CVD and copper PVD in metallization. The second takeaway is that for ALD, selective ALD adoption is happening right now, and it's gonna be in production in the first generation Gate-All-Around. Selective ALD has benefit of performance improvement and process simplification.
The third takeaway is that with this very high aspect ratio device structure, there is a need of many new plasma ALD technologies. The third point is that we are in Epi, silicon Epi, we are poised to gain market share because of our closed-loop, unique closed-loop wafer temperature control, which enables unmatched process control and capability. In, we see also very good momentum in our furnace and PECVD products. In our furnace, we are seeing... we introduced the SONORA, and we see good momentum, and also in PECVD with some niche application. As was shown earlier, for silicon carbide Epi, we see significant market adoption at leading customer because of our unique silicon Epi reactor, which leads to best-in-class silicon carbide epitaxy. Last, we see opportunity happening in heterogeneous integration in area of ASM strength.
So let's look into first about this acceleration that's happening in both logic and memory. First, looking into the logic roadmap. Here in the logic roadmap, you see that the FinFET is nearing its end of life. FinFET started in 2012, now it's in fifth generation with 3-nanometer. And in 2025, we're gonna see Gate-All-Around in HVM production. A Gate-All-Around is gonna be in pilot production in 2024, and is going to be in HVM production in 2025. Then by the end of the decade, we're gonna see CFET, which is stacking of different Gate-All-Around on top of each other, to be the next technology node. Then, if you go into DRAM, the scaling in DRAM is really based on the 6F-squared architecture. This architecture suffers from low capacitance and high leakage current.
So the industry is looking, how can we going to scale DRAM? And the first thing, they're looking into 4F-squared, which is smaller cell size. You go from 6F-squared to 4F-squared, so you have some shrinking of the cell to improve performance. But at the end of the day, DRAM is gonna go to 3D. When? Most likely by the end of the decade, and earlier, 2030s. Then, if you look into flash, flash is the first technology that went into 3D, and the scaling happens by adding more and more layers. Now, right now, in production, there are 200+ layers, and by the end of the decade, we're gonna have 1,000 layers of flash memory. So what does this 3D scaling means for opportunity in both logic for opportunities for ALD and Epi?
So let's look into the Gate-All-Around structure for logic. In the Gate-All-Around structure, you have the Gate-All-Around device along with RC challenges we have in the middle end of line, and better usage of space in the back end of line to add functionality. These three things provide huge opportunities for ALD for more ALD layers. So let's focus right now on the Gate-All-Around device. Here, the gate stack is the heart of the Gate-All-Around device, and as you can see, there are many different colors. All of those colors are ALD layers. So it's very important to deposit these ALD layers to, and to have them wrap around the nanosheet. So this ALD layer have to have very good conformality. And some of the layers here is like, are the dipoles and work function metals.
The dipoles and work function metals, that's what controls the VT in your d-device, the threshold voltage. From one generation to the other, you have more and more VTs. So the knobs that were used to get the right VT in FinFET is actually no longer applicable for Gate-All-Around. What does it means? It means for Gate-All-Around, you need additional dipoles and additional work function metal. Indeed, for each dipole layer, you actually need an additional pattern, ALD patterning layer. At ASM, we've been developing this, dipole and work function metal layers, and we have released a few of them to our customer in logic devices. The second thing that's happening in the Gate-All-Around is really around Epi. Here, for the first time in the industry, the channel space area is actually defined by Epi. It's no longer defined by etch.
For the past 50 years, that was the case for the past 50 years. Now, Epi is defining the channel width. And with that, we have what we call a, a silicon, silicon germanium nanostructure, which they have to have very good uniformity. The other paradigm that's also happening in, in Epi is the deposition of 3D Epi source drain and very low resistivity. Next, the things that happen in the Gate-All-Around are also happening in the metallization, and here we see a new era in metal, in, in metal, in the semiconductor industry, which is the adoption of molybdenum as a replacement for tungsten CVD and copper PVD. Why this is happening? This is happening because CVD tungsten and PVD copper suffer from higher resistance because they have to use a barrier layer.
So, for example, tungsten has to use titanium nitride barrier layer because tungsten has fluorine, and fluorine diffuses, because fluorine comes from the precursor. And copper, you need to have a tantalum nitride barrier, you know, titanium nitride, tantalum, tantalum nitride and copper. The reason you use tantalum nitride barrier, because to stop the copper diffusion. Copper wants to go everywhere, diffuses like crazy, so we need to stop it, and we need to put tantalum nitride. So when you add this barrier layer, the overall resistance of the metal is very high. Here comes around molybdenum. Molybdenum doesn't need any barrier layer, so when you replace copper and tungsten with molybdenum, you have lower resistance, and this is what's happening right now in the industry.
The other, the other thing that we're saying is, seeing right now is selective deposition becoming more and more, more and more used for Gate-All-Around. And selective deposition, which is ALD selective deposition, is actually used to improve device performance and integrity. And here, I'm gonna talk to you about two things: dielectric on dielectric, that means you deposit dielectric film on dielectric. So to explain to you selective deposition, what does selective deposition means? That means you deposit something on something else, okay? You don't, you don't deposit everything on the wafer. So I'm gonna show you how we deposit dielectric on dielectric, and also depositing molybdenum on dielectric.
The third thing that's happening for Gate-All-Around is the addition of functionality in the back end of line, and there you have two things: for additional memory, and the additional memory use MIM caps, and the MIM caps needs ALD high-k oxide, titanium nitride electrodes. It's all great ALD, and also for power management, you need lots of ALD with oxide channels for the thin film transistor. Then, if you go to DRAM, the opportunity there is in DRAM for Epi and ALD, especially ALD, is tremendous. Here, for DRAM, we have the DRAM, as I mentioned earlier. It has a high leakage, so to improve the leakage of the cell, you go to low-k spacer. It has low capacitance. To increase the capacitance, you go to ALD high-k.
And also, you do, you have high-quality oxides, and you add Epi layers for storage node contact and bitline contact. Therefore, DRAM also. DRAM, as everybody knows, high bandwidth DRAM is going into AI. But that AI, for that high bandwidth DRAM, needs high capacity and high performance CMOS, and high performance CMOS needs ALD. So it needs high ALD, high-k, high-k silicate or high-k oxide. It needs many dipoles, ALD dipoles with lanthanum oxide, aluminum oxide, and also work function metals. It also needs some Epi, especially a SiGeB, for strain enhancement in the channel. So we are, we think we are very well positioned with our leading innovation to address this technology node. Why do we think this? We think about this because for these different reasons.
For reason number one is that, you know, ALD is in our DNA, and, for us, ALD started 30 years ago with our acquisition of Microchemistry Oy in Finland, which is the birthplace of, of ALD. Since then, we have, opened R&D centers in both the U.S. and also in Asia.... The second thing is that we have very long-term R&D in ALD, and we're working with some partners. One of them is University of Houston, which is focusing on precursor for ALD. The other partner we have is IMEC, where we work on process development and process integration and device integration. So the combination of this long-term R&D and our own R&D, makes us focus on developing new products and new precursors, and also new plasma source in the industry.
So from like three to four layers 20 years ago, right now, we're working on 80+ ALD layers in our lab for the next generation of logic and memory devices. The other thing that's happening, our strength in ALD is also reflected by our strength in our IP portfolio, and this is the third. So this is data from LexisNexis, which is generated this year, that shows the strength of our IP. And the strength is really based on the number of citations that you have and the scope that your IP protects. So let's look into metal ALD, which is the first inflection point that's happening, and this is really very exciting thing that's going on in ALD.
If you look into the first here, if you look into the device structure on the left side, I'm sure this is like cartoon showing you tungsten in the middle end of line. And at the back end of line, you have single damascene and dual damascene with copper. As you can see, tungsten has the green, that's the titanium nitride, and copper has the black, underneath it, that's the titanium nitride underlayer. So as I mentioned earlier, this underlayer increases the resistivity of the whole structure. What's happening in the Gate-All-Around node? And this is really simplification of what I'm showing. In the Gate-All-Around node, we see that some of the tungsten contacts, the tungsten contacts are actually moving to molybdenum.
In the back end of line, the copper dual damascene is actually we see that the via is gonna change to molybdenum. Some of the work that we have done, which shows in the graph on the right with our imec partner, shows that when you replace copper with molybdenum in the via, your resistivity of the via goes down by 20%. This is really for CD of 10, CD of 10, which corresponds to 2- nanometer Gate-All-Around technology node. Replacing copper in the via by moly, you improve your via resistance and the device performance. Why? Because copper in the via is very, very small, and the tantalum nitride takes more of the area. The overall resistance of the copper module is higher than molybdenum. These are things that are happening.
And what I'm showing you right now, this is the start, and in future, you're gonna see more and more moly in those orange areas. The second thing, which is very exciting right now, happening in selective ALD, which we actually we've been working on for many years, but right now, guys, it's happening, and it's happening, and it's gonna be in Gate-All-Around structures. Selective ALD overcomes many interconnects scaling challenges. Here, I'm gonna give you two, two cases. The first case is metal on dielectric, MOD. Here, like I mentioned earlier, in a via, you have copper, and the barrier layer, you have tantalum nitride. As you shrink, you scale, you make it smaller, you see that the barrier is taking more of the area. That means the whole resistivity there becomes higher.
If you can take the bottom of the—If you take out the bottom barrier, you only deposit the tantalum nitride on the sidewall, then you actually reduce your resistance and make it better. And I'm gonna show you a case where we do this. We do metal on dielectric, and on the side, which is not shown here, that's the dielectric. The second case is really showing DOD or dielectric on dielectric. And here, one of the issue that's happening in the industry is actually is shown on the left side. On the top, you have a via, VX, that's a via, misaligned on top of the metal. That means you have... You know, you want to have your via on the metal, but there's a misalignment. As devices shrink, it becomes very difficult to have exactly good alignment.
And that misalignment means that the contact between the via and line becomes smaller, resistance goes up, and you have higher resistance and worse integrity. And, you know, the misalignment is really called, is defined by EPE, which is the Edge Placement Error. So what does DOD do? DOD solve this misalignment. By using DOD, we're being able to have fully self-aligned via. How do we do this? We do this by depositing an aluminum oxide layer on top of the low-K. So when we put the via on top of it, it goes directly to the line. And this picture here shows aluminum oxide on top of the low-k, while there is no aluminum oxide on the copper lines down there. So how can we do this? How can we do this, this self, this, this kind of, selective deposition?
Let's look into the process flow for the DOD, a simplified process flow. Here, from the left, you start with your structure. You have copper, and you have your low-K. And first, you know, when you have copper, the wafer has copper, copper oxidized, so you have oxide on the copper, and also copper diffuses everywhere. So you even have copper residue on the low-K. So you take that wafer, you put it in your reactor, and you do surface prep to remove the copper, and then you take the wafer to another reactor, and you deposit an ALD film, ALD inhibition film, whereby you deposit that ALD only on the copper line. And then you take the wafer to another reactor, and you deposit your aluminum oxide, and of course, the aluminum oxide gets deposited on the low-K.
After that, you etch away your inhibitor in a different reactor, and you end up with aluminum oxide, and then you continue your metallization for the self-aligned via that I just showed you in the previous slide. Now, if you can go look into the MOD, or metal on dielectric, this is, you know, where we only deposit the tantalum nitride on the side wall and not deposit the tantalum nitride at the bottom of your structure. First, your incoming wafer has either tungsten or molybdenum. Metal oxidized, so the red thing there, that's an oxide. So you take the incoming wafer, and you do surface clean in one reactor. After that, you deposit your inhibitor. So the same inhibitor, the blue here, is actually the same one we use for MOD is using for DOD.
So we have developed this new inhibitor that deposit only on metal, and it's an ALD inhibitor, and it's something that's really specific to our technology. And after that, after you deposit the inhibitor, you take the wafer out. The inhibitor is very stable in air. You take the wafer out, and then you deposit the tantalum nitride on the side wall, and then after that, you remove your inhibitor. This way, you only have tantalum nitride on the side. You don't have it on the bottom, and this way, you actually improve your resistance. So it's very exciting to do these kind of things, but in order for you to do these kind of things, you have to do deposition, inhibitor deposition, ALD deposition, clean, etch, et cetera.
In order to do this, we have to develop a platform for it, and we have used our XP8 platform, and we extend it to six facets. And with this platform, six facets, we can do all these different things. And here on the bottom left, you can see different reactor. They are different because you have, clean, surface clean, you have inhibition, you have different ALD going on, and we are shipping this reactor as we speak to our, some of our customers. Now, let's talk about the other, the third inflection point that's happening in ALD, and that's the gap fill of the very high aspect ratio, the new plasma technology. I've shown you earlier that everything is going 3D.
You know, logic is going 3D, 3D- NAND is already went to 3D, and DRAM, little by little, is gonna go there. But we have this very high aspect ratio. So if you look into the very high aspect ratio, you see them for channel hole gap fill. You need to do channel hole gap fill. You have to do staircase gap fill. For example, for 3D- NAND, you have TSV liner, you have slit slit liner. And what we're talking about here, we're talking about 100 to 200 to 1 aspect ratio. What does it mean? It means that we need to deposit films in the bottom of this structure.
In order for us to have species go all the way down, you have to have a very high density, and the way to have very high density, plasma density, you have to increase your frequency. All the plasma processes usually happen at 13.56 MHz, but if you increase your frequency, as you can see on the, on the graph here, as you increase your frequency, the plasma density increases. Once you increase the plasma density, then the species can go down in the structure. But once also, when the, the plasma density, when the frequency goes up, you will see also the ion energy comes down. It's good for the ion energy to come down.
You have less damage to the film that you deposit, but you don't want to have it too much down, because you want to have some ion energy to densify the film, to make your film much better, higher quality. It's for that reason that we have developed new plasma ALD sources with different frequencies, 60 MHz, 100 MHz, 2 GHz, to address these many challenges in the industry. The next challenge is also for gap fill happen in the Gate-All-Around. And in the Gate-All-Around, you don't have like vertical structure. You actually have reentrant profile. You know, you saw the nanosheet. You know, nanosheet, you have to deposit your ALD film between the nanosheet. That means you need to open, so you have reentrant profile. It's very difficult to gap fill in the reentrant profile.
You need to have to go to a different plasma technology, and that's radical plasma technology, which is developed by ICP or inductively coupled plasma. And as you can see, see on the left side, a deposited film, that's very small layer thickness, that's ALD film, that we have deposited with this technology. And after you etch it, you wet etch it, you put it in HF, okay? What you can see two things. You see that the film thickness is the same in the bottom of the structure, deep into the structure and outside the structure. If you have normal ALD, you will see no thickness down in the cavity and much higher thickness outside. But this is not normal ALD. This is ICP ALD, and it has very good homogeneity and very good film property.
So this is the kind of technology we actually have to develop in ALD to be able to have this kind of liner fill and gap-fill films in Gate-All-Around structure. So let's shift gears, move from ALD to Epi, and in Epi, I want to talk to you about the opportunity we have in Gate-All-Around and in DRAM. So if you look into this slide, this slide shows the roadmap for both logic and DRAM. The first thing you can see is that, okay, opportunities for Epi in both logic and DRAM are increasing. The first thing you see is the Gate-All-Around structure. The 3-4 silicon, silicon germanium channel superlattice. That's what's happening in Epi. This is depositing silicon, silicon germanium channel. When you go to the FinFET, instead of having 3-4, you're gonna have 12-16.
But one thing I want to tell you, this 3-4 silicon, silicon germanium channel, if you in the FinFET, you might have only silicon channel or only silicon germanium channel, or both, or some customer of ours have no channels at all. So in Gate-All-Around, nobody cannot-- everybody needs to have this superlattice structure. So this is significant improvement in, significant increase in the Gate-All-Around Epi technology. The other thing we see in Epi, in the second generation, is the move to low temperature, LT, that's low temperature Epi. Because of thermal budget requirements, we have to go to low temperature. But when you go to low temperature, you have to have-- the film property degrades. So we have the precursor that we use, or that the gases that we use at high temperature are no longer applicable for low temperature.
So because of that, we have developed new precursor, which we have developed new precursor, in order for us to give us low resistivity Epi at low temperature. And then in the DRAM, you see there are quite a few in high-performance DRAM. It use Epi in the channel. It use Epi, as a c-SiGe, but also there are many new Epi application in DRAM, in the contact area, both for the storage node contact, which very, very difficult to make, and also in the bitline contact. So let's first look into these advanced Epi films in both logic and memory. Here, let's start with the first figure on the left-hand side. This is actually showing you CFET, silicon, silicon germanium superlattice. And earlier, we showed you the GAA superlattice structure. Here, we show you CFET.
With the CFET, you're gonna have even more germanium content. In a Gate-All-Around structure, germanium can be between 20% to 30%. CFET, it can be as high as 40%. For people in the Epi realm, they know that as you increase germanium, you have more defectivity. But this is a structure that shows that we can deposit very high concentration germanium without any crystallographic defects. The top graph on the left-hand side, top left, showing you low temperature Si Epi for NMOS, and also low temperature silicon germanium boron, using the new precursor that I just talked to you about. Epi is also gonna move into memory in the future, 3D DRAM. And for 3D DRAM, you need to have not only the 3-4 superlattice structure that I showed you for Gate-All-Around, you need to have hundreds of those silicon germanium Gate-All-Around structures.
The thicknesses of this is 8 micron, and this is some of the development we're generating right now. And if you take-- zoom in some of the structure there, you can see the extreme pristine non-uniformity between silicon and silicon germanium. So the black, that's silicon, the white, that's the germanium. Silicon germanium content. How can we achieve this kind of performance right now? The reason we can achieve this kind of Epi performance is because we have a differentiated technology. And our differentiated technology, we call it Turino- CL. CL stands for Closed Loop, and this is the first and unique technology in the industry, whereby we measure and control the wafer temperature directly. Everybody right now, when we measure the, the Epi, we say Epi wafer temperature, we're actually not measuring the wafer temperature. We're measuring the susceptor temperature.
So the wafer sits on the susceptor, and we have TC in the back or pyrometer from the back, and we measure the susceptor temperature, and from the susceptor temperature, we imply the wafer temperature. So we don't have direct measurement. We have changed that game, and we actually have direct measurement of wafer temperature. And we don't have only measure the wafer on one point, we measure it on 3 points. We have three pyrometers, one in the center, one in the edge, and one R/2 in the middle of the wafer. And this pyrometer measure the wafer temperature and can control it by changing the powers, the flame powers, in such a way to achieve 0.1 degrees Celsius delta on every part of the wafer. In CVD, you know, 2 degrees Celsius is okay.
In Epi, 2 degrees Celsius delta on the wafer is a crime. The wafer is not gonna be good. Your Epi that you're gonna have is very, very bad. So you need to have this kind of control, and it's only with this technology that you can achieve this kind of control. So when you have this technology, you can achieve what we have on the left side, which is a thickness of the silicon, silicon germanium, throughout the wafer. So this is line scan of throughout the wafer, 300 millimeter, showing the silicon thickness of the silicon in the nanosheet structure and silicon germanium. On the Y-axis, it goes from 95 to 105, so the non-uniformity is less than 1 angstrom. That's less than 1 monolayer. So we have 1 monolayer.
That's why we call it monolayer layer control, monolayer control on Epi in our technology. So this Turino- CL gives you this great control, but also it actually improves our process. For example, in the middle, we're showing the Turino- CL, which is our technology, versus the susceptor control, which was used in the industry until now, up to now. As you can see, you can actually reduce the time it takes to stabilize the temperature by 25%. Once you reduce the time, then you have higher throughput. When you have higher throughput, you have less energy usage. So actually, we've been able to reduce the energy usage on our Epi by about 35%. Next, let's talk about this niche application in PECVD and vertical furnace. First, let's talk about PECVD.... PECVD, we look into, you know, is there any inflection point?
We go and go after it. One of them has actually happened a couple of years ago. Right now, we have developed this technology, and we are PTOR in one key customer on this, using what we call PECVD carbon gap fill. The PECVD carbon gap fill replaces the spin-on carbon. Spin-on carbon is used up to now in the industry, and it's gonna continue to be used to pattern many structures. But, you know, once you go to smaller and smaller geometry and very high aspect ratio, spin-on carbon has problems. It's not very planar, so you have to make it much thicker. But since it's a spin-on carbon, it's not strong. Then you have to anneal it, and after that you have to CMP it, and you continue with your lithography.
As you can see, many steps. With our PECVD flowable carbon, you just one step. You just deposit, it's flowable, and it's very, and has very good planarization. So it's flowable and very good planarity, and as you can see, you can achieve very high, very good gap fill, up to 16-to-1 aspect ratio, which means it's actually, it can, we can extend this to three to four generation from now. And we've been shipping some of these tools to our customer. In vertical furnace, here in vertical furnace, we actually, in the past few years, what we have done, we have freshened up our lineup. In 200 millimeter, we have introduced the A400 DUO in 2019. This is in replacement to the A400 C and A400 XC.
Just last year, we replaced the A412 with the SONORA 300 millimeter. What we have seen, we've seen extremely good acceptance of both these platform in the industry. If you look into the A400 DUO, this year, the A400 DUO is actually the leading platform for all the more, more than more technology, silicon carbide, gallium nitride, power, and NEMS. It's the leading platform. For SONORA, we see significant acceptance in the industry, and actually, more and more customer are putting more application into this platform. Now, let's talk about LPE for silicon carbide. The question here: Why did we buy LPE? Well, the reason we bought LPE, because it has the best reactor technology. Why does it have the best reactor technology? Because it's very similar to our silicon reactor technology.
Now, to be very serious, Epi, you have two technology to do. You have cross flow reactor or you have perpendicular flow reactor. So the cross flow reactor, you have the wafer sitting there and the gases come in horizontally, parallel to the wafer, and you deposit. So the cross flow reactor, every silicon Epi reactor is a cross flow reactor. So when we looked into LPE and we found out that they have this cross flow reactor, we are very happy because the cross flow reactor gives you the best defectivity, uniform defectivity in the wafer. And if you have the best defectivity, lowest defectivity, that means you have the best Epi, that means it's good stuff. Then, the other technology is using actually perpendicular Epi.
Perpendicular flow reactor, that means you have the gases in terms of the wafer, so it's perpendicular to the wafer. So when the pieces come to the wafer, they don't take their lowest energy state, and you have some vacancies, and those vacancies give rise to crystallographic defect. And that's what happened in some of the journal, Journal of Applied Physics, that shows in 2009, showing that if you have defects, what you call the triangle defects, these are the crystallographic defects, you have high leakage current and your Epi is no good. If you don't have those defects, you have lower leakage, and you can, and the device is very good. And indeed, we actually have side-by-side comparison from our customer. This is customer data.
I cannot show you exactly the numbers, but what I can tell you, this is in orange, that's our cross-flow reactor, and the other one is a perpendicular flow reactor. As you can see, the triangle defects, which has the crystallographic defects, we are 30% lower than the other reactor. The downfall defects, these are the foreign material, this is the defects that fall on your wafer. We're actually 40% lower than the other technology. And of course, thickness uniformity is better and doping uniformity is better because you have a laminar flow. Okay? So because of all of this, you know, customer have seen this, this difference, and when they use our Epi, they have much better device performance, and for that reason, we're seeing extremely good acceptance of this Epi technology in the market.
Last part of my presentation is just to talk to you about some opportunities we have in advanced packaging. Advanced packaging is growing, it's fun, but also it is many new materials, and we like it. I personally like it very much because these films are very thick, and I'm working all the time on ALD. We're talking 10 angstrom, 15 angstrom, and then I find I have to deposit 40 micron. I'm very happy because, you know, we can sell more tools. Okay? So the opportunity we have in packaging start with the first one. This is through-silicon via. The through-silicon via needs oxide liner and very high aspect ratio. And what's best is you deposit ALD with some of the plasma technology that I showed you earlier.
This is, you can do this, and actually, we have some PTR position and in oxide, in this technology, in many of our customers. The second thing is really looking into the MIMCAPs and the deep trench capacitor that actually are being used right now in the interposer. And for these MIMCAPs— Of course, MIMCAPs is high-k, ALD, and so on. So for this, we see, we see many customer are, are using this MIMCAPs and deep trench capacitor and using our high-k material for this application. The third thing is actually in the space of chiplet isolation. Here, you have many chips, and you need to isolate them one from the other, so they don't cross talk, and the way to do it is you deposit very thick TEOS, 40-micron thick TEOS, to isolate the different chips.
So that's using the PECVD TEOS. The fourth application here we have is the backside power distribution network. This is the PowerV ia that Intel has come up with, and here there are many ALD applications in the liners and also the moly fill. Okay, you can fill these aspect ratio with molybdenum, and we have many new engagement in this area. The other thing, we see engagement in Epi, with our Epi in silicon photonics. With germanium, you can use germanium as a photodetector from that point of view. Last but not least, hybrid bonding is really growing a lot, and here, in order for you to do the higher hybrid bonding, you need to have dielectric layers.
To end my presentation, I hope I've shown you in this presentation that we, in our products, our technology roadmap is actually is well positioned to, to capture many new opportunity. I hope, I hope that I have shown you that there are three inflection points right now in the industry. Number one, metal is changing. No longer tungsten, molybdenum is it, and molybdenum is gonna be in production in 2024 and 2025. The second thing that's happening is selective ALD. This is an area of growth. This is happening, being adopted in Gate-All-Around, and gonna be in production in 2025. Last but not least, ALD is about new plasma sources. In order for us to deposit this very high aspect ratio, in order for us to deposit in very re-entrant profile in Gate-All-Around, you need to have a new plasma source.
And the number of plasma source and the complexity of plasma source in ALD is much bigger than what happened in CVD. Because the structure that we're talking about right now is just crazy. It didn't happen in the CVD world. So that's why we need to have new plasma technology to enable in order for us to be able to deposit these very conformal ALD film. I hope that I've shown you that with our new Turino- CL technology, that we are very well poised to gain market share in Epi, and we feel very good about it. We have good momentum in our offering in furnace, and also we have good momentum in our PECVD offering. Silicon carbide Epi is great. We love the technology. We love...
It gives a much better film, and customers know it and see it, and they are adopting it. Last but not least, in advanced packaging, we have many new applications and many new films to target this growing technology. Thank you very much for your time.
Thank you, Hichem. We will now have a break of about 20 minutes. After that, we will continue with the second part of the program, the presentations. For the webcast audience, please be back at 10 past three. For the guests here in the room, please join us for coffee or tea in the foyer. Thank you.
Okay, welcome back, everyone. We will now continue with the second part of the program. Let me quickly show you the program again. So we will continue with Kent, who will talk about global operations, followed by the second presentation of Hichem, about product sustainability, followed by Paul, with a financial update. After a wrap up by our CEO, Benjamin, we will move on to the Q&A. Kent? Please.
Thank you, Victor Bareño. So about four years ago, I followed Hichem M'Saad here to ASM, and now I get to follow him on stage. It's a real honor, although it's a tough act to follow, I'll say. I think Hichem M'Saad talked about our products and technology innovation going on in the industry, which is, which is unbelievable. I'm here to talk about operations, and we have to make whatever he thinks of and his team thinks of a reality in terms of supply chain and the ability to get out the door working. And what I will say is that innovation is also not lost on us. So we've got a lot of innovation that we're working on. We've got big plans, and I'm going to share a little bit of it with you today.
So the key takeaways that I want to try to leave you with here today is, number one, the company's been growing very, very quickly and, during really this COVID period since 2020, at a very, very high pace. And despite that, and despite very significant supply chain headwinds, we've been able to keep up with customer demands. Actions have been taken, so that going forward, we can really be ready in our supply chain for what we're calling the new reality, which is a very difficult climate that we've been in and we believe will continue to be in for some time. As Benjamin mentioned earlier, we've recently finished an expansion in Singapore, and we've got an upcoming expansion in 2025 in Korea, which we believe sets us for hitting our targets of revenue in 2027.
We have increased focus on manufacturing innovation, which I'll go through, and then ESG is becoming a critical requirement now as we look into the supply chain. It's no longer a nice-to-have. All right, in terms of the global operation, I wanted to start by showing you a global map of our sourcing. So this is where we buy our parts and assemblies from globally. You can see we're very heavily footprinted in Asia, which is also where we're heavily footprinted from a manufacturing standpoint, and we've got significant supply base in the US, and then some also in Europe. Main manufacturing, where we do most of our manufacturing, is in Singapore, followed by Korea, and then we've got a couple sites in Europe, one in Milan and one in Almere.
We use a flexible outsource model, which effectively means that we don't assemble and build everything in-house. We actually have suppliers that we call contract manufacturers that do some of that for us upfront. In fact, about 60% of our assembly happens outside of our own doors before it enters our factory. We've got 1,400 employees in operations globally, and we've increased our output since 2020 by 80%. Now, a lot of these pictures, I think, will resonate with you. I'm talking about supply chain headwinds now and what the new reality is, but I don't want to just rush through this and talk about supply chain effects. The human toll here is incredible, and it definitely overshadows anything I'm going to be talking about from a supply chain standpoint.
In fact, I think the first one on here has probably touched every one of us in a very personal way. But what I will tell you is, again, we see this as a new reality. We see it as a difficult challenge. I think despite these headwinds, we have—we've been able to deliver last year, delivering a 39% growth, far outpacing the rest of the industry in a supply chain constrained environment. And what I wanna talk about is what we're doing going forward. Now, any one of these is a significant event when it comes to supply chain disruption, but when you have all of these happening concurrently or in succession, it's really a major challenge that we've had to figure out how to overcome. So really, the past approach that we had to supply chain had to adapt.
All right? We've, we've done a lot to look at supply chain differently. We've kept up our pace, and like I said, I think we've outperformed the industry pretty significantly in terms of delivering. But I can tell you, it was very painful taking this punch in the mouth. Very, very painful for us. So what we're looking to do is innovate and look at our supply chain differently in such a way that we're ready. We get another COVID, we get another group of problems like this, or even worse, we wanna be on the front lines of knowing how to get in front of it, so we can avoid the punches in bunches that we expect.
So in the past, I would say we looked at ourselves as really sandwiched between our customers' needs directly and our first-tier suppliers, the ones that actually deliver directly to us. Everything below those guys, we more or less expected them to manage outside of a few key areas. So of course, the present, we've really exploded back into the supply chain, deep. All the way to when minerals come out of the Earth to make the parts that we use to make our tools, we've got the entire supply chain mapped, and we look at it very carefully. We look for any vulnerabilities, risks, ESG, potential concerns. We have to think about geopolitics these days, potential trade barriers and tariffs, and certainly any capacity bottlenecks.
So we're working very deep in the supply chain to get a handle on it well ahead of time, and I'll talk a little bit about what we're doing differently going forward with this expanded scope. All right, so one thing we're doing is thinking about when we bring a part or an assembly out to our supply chain for the first time, are we bringing one that's gonna work with the supply chain? So we've actually invested a lot. We've invested a lot to find ways to improve the manufacturability of the parts that we put into our new tools, improve the reliability of them as well, and to commonalize them so that our parts, we have less part numbers across all these products than we do today. That's a big focus of us. How we've done it? We've brought in experts.
We've brought in experts that have 20+ years' experience at suppliers doing the work and working with suppliers from this end. They can look very differently at a part when it's in its design phase. I mean, once a part's designed, a lot of the value is already captured. So we're working upstream with Hichem's teams to design things in such a way that they not only work for the product, but they work for the supply chain as well. So that's something quite different. We've got a team, what we call the centers of excellence of the supply chain. Now, if you look at, let's say, our scorecard of the supply chain, back in middle of last year, it was rough.
When I say red, it means it was all hands on deck, trying to figure out how to get parts in, how to still deliver the customer needs, and getting that 39% growth year-over-year. A lot of work, a lot of pain, a lot of getting punched in the mouth. You can see it was practically across the board. By the end of last year, we were in a little bit better position. You can see some of this turns yellow, and some of it even has a little bit of green for the first time in a few years, and we still have a couple reds. Currently, we look at our supply chain in a way, mostly green to what we want to do.
It doesn't require the expediting, the you know, micromanagement of the supplier, et cetera, to get these things in. But we still do have challenges, and I'll tell you, when it comes to specific commodities, and this is really all of the commodities of interest that I'm showing here, specialty materials such as carbide, graphite, quartz in some cases, still are quite constrained. So it's still a challenge, but we are working very, very hard. And what we're doing differently, let's say, than we have in the past, I think the first one, the first knob we've been using, of course, since the beginning of time: How do you get more capacity or supplier? What we're doing differently now is going down to the sub-suppliers, all the way down to the mining, and making sure that we're working on that.
We're innovating the part manufacturing techniques, so not only designing the part up front to work better with a supply chain technology that exists today or leveraging a new one that's coming about, but also to make the part manufacturing process itself innovated. So we have certain components already where suppliers globally that have done nothing but this for their whole lives will do it a certain way or certain ways. We've come up with new ways, new IP, to bring to them that enable them to make our parts more robustly and more cost-effectively for us, and the ability to ramp more quickly. The other thing we've done, of course, from a BCP standpoint in the past is dual source parts. These days, we're actually multi-sourcing parts. We don't want to have eggs in one basket, one geographic basket.
We don't even want to have them in two, so we've got a lot of multi-sourcing going on. We're improving our supplier efficiencies through working with them. We actually put people on the floor with them, looking at what they're doing and coming up with innovative ways to improve their efficiencies. Again, help them help us. And then we're signing long-term agreements and some other even more creative approaches to making sure that when the next big boom comes, and the next 39% year-over-year kind of thing comes, that we're ready with the supply that we need all the way through the entire supply chain. All right, we've been investing ahead of needs when it comes to manufacturing. This is a normalized plot of our manufacturing capacity, starting in 2020, and you can see through 2027.
Through 2025, we'll have increased our capacity by 3.7x in internal manufacturing capability within ASM. We believe that sets us up to hit the revenue targets that we have set for 2027. Now, how did we get here? In 2021, we moved from a Yishun plant in Singapore to a Woodlands plant and opened up a level there that increased our capacity very significantly to 1.7x of what it was. The acquisition of LPE last year added some incremental manufacturing capacity. We opened a third level of the Woodlands facility. We did that back in January of this year. That added a significant global capacity to us, bringing us up to 3.3x. And then, as Benjamin mentioned, Korea, we'll be bringing online in 2025.
That'll bring us to the 3.7x that I talked about. Now, if you look at the demand profile we were at in 2020, it was getting a little tight. In 2023, where we've got plenty of room. In 2027, we've got room compared to where we expect to get. But I will say next year, we're going to be doing a very concerted analysis of where our next footprint should be, and maybe even make a decision that could impact the 2027 investment. All right, manufacturing innovation. First thing we're going to do is fully digitize what we're doing in the manufacturing floor within ASM. That'll allow us to make sure we're efficient, understand the bottlenecks, really diagnose what's going on.
If we have downstream problems with customer performance, we can work them back and look for where to improve it. So this is one major improvement we're working on. Employing AI, so we've got the big data collectors. We've got hundreds of sensors that are constantly kicking out oodles of data that otherwise goes generally lost. We're putting AI brains on this data, and we're looking for the types of trends that could have an impact on the performance of our tools downstream. We're fingerprinting so we can learn a lot more going forward and adjust. Increasing factory automation. There's limited ability to do that on our own manufacturing floor, but as we go upstream in suppliers, we want to take the human error out as much as possible.
Full modular operation, so as our demand shifts from this product to that product, this customer to that customer, the more modular we are, and we are modular today, but increased modularity will also be flexible and be able to meet the demands in a fast manner. And then, where we're really heading, let's say in the medium term, is getting to an AI-assisted final test. The idea here is we've got this brain of data correlating things upstream of our factory and downstream into customers with what kind of test should actually be done as the tool goes out the door, and actually conducting the test. So that is, we hope to completely innovate in this area and have a way to send more robust tools more efficiently than anybody has before. All right.
And lastly, sustainable supply chain. As I mentioned, it's no longer a desired approach with our suppliers. It's a very much a requirement. Starting first with the carbon footprint, with our net- zero goal for 2035, we work very closely with suppliers and third parties to make sure we're assessing them right. We put a lot of effort on this, and I can say that our supply base is actually quite willing and been very cooperative in getting back to us and working on this with us. So I think that's been very good. We've got very high rate of carbon disclosures. In terms of surveys, we've got 100% renewable electricity being used in our Singapore factory.
If you look at the circular economy we put into our service business that Benjamin had talked about earlier, we believe we've managed to save 370 tons of carbon dioxide effectively by us bringing the part back from the customer that needs to be replaced, sending it back through our supply chain with something that will allow it to work like a new part again, and putting it back into the customer over and over and over again. This not only saves cost, but it saves the environment very significantly. I think in terms of responsibility, treating people right... super critical. I think there's no question about that.
So we are working, again, all the way down our supply chain to make sure that, you know, even down to the mineral mining and smelting of our metals, that we are doing the right thing and that people are doing the right thing. And this is an area that I would say that we've uncovered some issues and risks that we need to mitigate and remove. And the other thing we've been doing is trying to save materials, and we've been using reusable crates, for instance, on lanes that make sense, rather than crates that would be thrown away after you ship tools.
Then lastly, on resiliency, other than what I talked about earlier on supply chain, from a resiliency standpoint, we're constantly sharing best known methods with our customers, with our suppliers, to help all of us get a little better at protecting ourselves from cybersecurity threats, and also protecting our ability to operate, should we be threatened. Okay, one last thing on human rights. We put more due diligence on it all the time. There's been a lot of work that has went on to today. Going forward, if you look at what we're looking into the mining operations, we're gonna continue to increase the minerals we're looking at. Today, we're looking at the 3TG, very aggressively.
Our industry collaborations, I think we've been setting a very high standard in the industry of collaborating, and we're gonna continue to put a lot of emphasis there. And we'll continue to look at our supply chain and make sure that everybody's on board with what needs to be done there. So hopefully, these will be somewhat recognizable now. We're growing very quickly. We've been able to take the punch. We're looking to be better at not getting punched, despite expecting a lot of punches and bunches coming at us. That's what we've been doing in terms of dealing with the new reality. Our expansions have us set to hit our revenue targets by 2027, though we might have increased footprint in manufacturing by then.
ESG is now a requirement and something we're working very closely through our entire supply chain with. Thank you very much. Actually, now I want to hand it over to the good doctor, Hichem M'Saad, again.
Thank you. Thank you, Kent, and good afternoon again. It's my pleasure to talk about our product sustainability, which is becoming a very hot topic the past few years. So first, accelerating sustainability for us is really part of our six strategic initiative, which is focused on people, planet, innovation, governance, and supply chain. And the target for us, which is aggressive target of net- zero 2035 for all scopes, has been recently verified by SBTi just a few weeks ago. We have found out that the majority of our greenhouse gas emission is actually attributed to Scope 3, which means our products. So accordingly, we have a very strong focus right now on sustainability, on product sustainability, and we've been working with our customer and partners to make sure that sustainability is part of our product development.
Both at the product level, which is on the fab level, and also the sub-fab, which I'm gonna talk about later on. So reducing Scope 3 footprint is also highly dependent on our customer, which and also on the upstream value chain. In this presentation, I'm gonna give you some example of the sustainability focus. I'm gonna show you how we've been able to reduce plasma source, have more efficient plasma source, more efficient use of chemicals, and much better heating technology. So let's first talk about what's our usage of greenhouse gases. So looking into our emission, emission is divided in three scopes: Scope 1, Scope 2, and Scope 3. So Scope 1 is the direct emission from the company-owned facility, from what we own, and it's actually less than 0.1%.
If you look into Scope 2, which is the indirect emission from the generation of energy that we purchase from our suppliers, it's about 0.5%. Last but not least, Scope 3 is actually 99%, and Scope 3 is the indirect emission from the downstream and upstream of our value chain. And if you take Scope 3 and you zoom in, you'll find out that zoom in using, you know, like, emission of tons of carbon equivalent, you find out that the use of sold products is 59% of our Scope 3. So the product is the culprit. Second is the purchased energy with a 33%.
Transportation and distribution is at 2%, and capital goods is at 1%, and the rest, investment, business travel, fuel and energy, they constitute about 5%. So as you can see, the product is really the key in the emission, and our focus right now on sustainability is really focused at, on how can we make our product that we develop using much less energy and more sustainable? So that's our vision. So the vision is to develop differentiated products and processes while maximizing the energy saving through product innovation. Our aim is to develop these very high-quality film. At the same time, we wanna lower the energy. We wanna usage. We wanna lower the gas usage. We wanna have less emission and also reduce the cost per wafer for the customer. So it's no-brain.
You know, sustainability, it's good for the customer, good for the environment, and also it makes our product much more competitive. So how do we look into the emission in our product? What contributes to these greenhouse effects of our product? First, we need to define what the product is, and the product is basically process modules, or these are reactors, and they are connected with a platform. The platform, that's where the wafers gets moved, the wafer get transported. And in that one, we have four energy sources. One of the energy source is electricity to power our products. The second one is the process chemistry, and process chemistry we use gases, we use liquid, and also use some even solid from that point of view. And also we have cooling liquid.
All this reactor have plasma or they have high energy, they happen at very high temperature, so we wanna make sure that we can cool them down. Otherwise, you know, you have too much heat. With that, you know, we have a cooling liquid to, to cool the reactor, to cool the reactor. The fourth energy source is actually in the carrier gases, and the carrier gases is the nitrogen, the helium, the argon, the CDA, which is the compressed dry air, that are really necessary to power the tools. So the cooling liquid that we, that we have, that we use, is actually, is connected to a chiller, which is a heat conductor, heat exchanger, which is in the sub-fab. Let me talk a little bit about the sub-fab.
So every tool that we have is in the fab level, but the tool is connected to pumps, it's connected to scrubbers, it's connected to abatement units. So where do we put all these pumps, the scrubbers, and everything? We put them in the sub-fab. So you have one floor where the fab is, and underneath is where you have all the pumps and so on. That's why we call it sub-fab. And why do you use sub-fab? So that the connection between the tool and the pumping and exhaust and the abatement is very short. You don't wanna make it on the same level. So once you do your reaction chemistry, and you have your deposition or whatever you have, then what you have, you have exhaust, you have the byproducts.
In the byproducts, which are the unused gases and chemicals, they leave, they get pumped out. The pump that's in the sub-fab is powered with electricity and nitrogen, and the pump takes all these gases to the abatement. And in the abatement, they burn all these nasty gases, and after that, you go, it goes out without any greenhouse emission. So our scope is really looking into how can we reduce emission from our products and also from the sub-fab. So our scope really is because the sub-fab component makes our tool and related to our tool, so we wanna make sure that we concentrate on both our products and the sub-fab part of the equation. So this is really, this is shown here, when you have the product level and the sub-fab level.
So when you look into the sub-fab, and the sub-fab, you have both the process module and the platform, we see three levels where we are working on sustainability. First, the design foundation. When you design a new tool, a new reactor, you wanna make sure that it uses less energy. And one of the things you wanna do whenever you design something new, for example, you wanna make sure that the reactor volume is very small. The smaller, is you use less energy, you use less precursors. So that's, that's some of the foundation. And the component selection, you wanna make sure that the component that you select for your reactor or your platform doesn't use is, is efficient. For example, if you're looking into RF generator, you wanna make sure that RF generator has the highest efficiency.
If you wanna use a chiller, you wanna make sure also the chiller has the highest efficiency. And also, in the process optimization, when you develop your processes, you wanna make sure that, okay, you use less gases, you control your process, you can go to low temperature, it's even better. So these are the kind of things that needs to happen from the get-go. On the sub-level, sub-fab level, you have also the component selection, so the pumps, the abatement, the scrubber, and so on. You wanna make sure that these components use the least energy possible. And actually, we require from our supplier for, from our supplier, abatement supplier, that, you know, they have much lower energy usage. The other thing that happened on the sub-fab is recycling.
You know, in your product, you use nitrogen, for example, so you need to recycle nitrogen. We also use water recycling. In our lab in Phoenix, 90% of the water is actually recycled. In a place like Arizona, where water is a very scarce resources, recycling is very important, so we recycle the water. Of course, if you can use renewable energy for this component, it's you're better off. The third thing in the sub-fab is you wanna make sure that these components that are in sub-fab are really synchronized with your tool, operation of your tool. For example, if the tool is idle, then the pump needs to be idle. The abatement has to be idle, not powered on.
And also, for example, when you use Epi, Epi uses hydrogen, and some processes use high, high hydrogen, some other processes use very small amount of hydrogen. So the abatement unit, which really burns the hydrogen, needs to know when are you running a high hydrogen process. So, so the tool, your tool needs to be synchronized with abatement, and if the abatement sees that your, your process is gonna run very high nitrogen, then it, it goes into the high fire mode. And when you run lower hydrogen content, then you go to a lower fire mode. So this is the synchronization that needs to happen between the sub-fab level and the product level. So that's why, for us, in order for us to reduce energy usage and precursor usage, we need to really work with the sub-fab component. So what's our strategy for sustainability, product sustainability?
Our strategy is really based on three things: more efficient plasma sources. Many of our products use lots of plasma. I talked earlier about plasma. More efficient use of chemistry... and more efficient thermal technology. Our thermal technology, silicon carbide, vertical furnace, epi, you know, these are all thermal energy. So if you look into the more efficient plasma sources, the plasma sources uses RF power. They use also matching network. You match the RF power to the matching network. You have RF generator, which are removed in the fab fab, and you have the RF cables, which are long, like can be, you know, even 50, 50 ft long, and then you have losses. So number one thing for us to improve sustainability here is to reduce the RF energy usage per wafer. And how do we do this?
Actually, the first thing we have done is actually, using the technology from Reno Sub-Systems. Reno Sub-Systems is a company that we bought a year and a half ago, and actually has a unique technology, which electronic variable capacitor which is used for matching. Every other technology in the industry use VVC, which is a Vacuum Variable Capacitor. This electronic variable capacitor is great for our plasma technology. I showed you earlier, there are many plasmas, and the plasma power can be very high. By using this VVC, we actually can reduce the reflected power by 5, 5%. That's a huge when you have very high power level you're running. And also, we can tune the power much faster, 100%, 100 times much faster. What does it mean? It means your process is shorter, it means you use less energy.
So the acquisition of Reno really helping us to reduce, to improve sustainability for all of our plasma processes, be it, you know, ALD and also PECVD. The second one is more efficient use of chemicals, and using chemicals is very important right now in our industry. I can tell you, it's right now, for me, it's the most important thing. Why? The reason for it is that, as I've shown you earlier, everything is going 3D. When you have these 3D devices, and you need to do deposit, the surface area is huge. What does it mean? That means you're using more and more chemicals. Okay? How can we reduce this chemical usage? Okay, and this chemical, to be honest with you, are not the best chemicals you, you, you can think of, okay?
So we need to make sure that we have better usage of this chemical, and also, when we develop a new process, a new film, the selection of the precursor is actually we look into sustainability to select which precursor from the, that point of view. So the less nastier precursor is the one we select. And also, we need to make sure that, okay, in the reactor, we can purge this precursor much faster. So precursor dose per wafer, we need to reduce it. In order to do that, we actually, in our process, we need to do modulate the residence time. And residence time is how much those precursors stay in the reactor. By optimizing your process, we've been able to reduce precursor consumption of 50%. If you reduce the precursor consumption of 50%, you have less abatement of the precursor.
Abatement uses lots of energy, so you reduce the abatement energy usage. The third part of our sustainability strategy is more efficient thermal technology. Our processes, as I mentioned, are very use lots of energy. Silicon carbide Epi runs at 1600 degrees Celsius. Silicon silicon Epi can be from 700 degree Celsius to 1200 degrees Celsius, so this is lots of energy. So how do we choose the heating technology is very important. Do you use resistive heating? Do you use plasma heating? Do you use UV heating, microwave heating? So the choice of the heating technology is very important, and we choose the technology that has the highest efficiency. And also, you want to go to low temperature processing.
Everybody, you know, I tell my guys, "Can you lower the temperature of any process you do?" Why? Because you use less energy. But the problem is when you go to lower temperature, usually when you go to lower temperature, you have more impurity, and the film is not very good. Quality of the film degrades. So we have to develop new processes and new precursor to do that. So reducing the thermal energy per wafer is very important. So say, for example, the Turino- CL that I showed earlier, you know, the fact that you can get the temperature very, very quickly, you can achieve a much faster ramp that allows you to reduce the energy by 35%, I showed you earlier. And actually, with a combination of better cleaning, we can reduce the energy usage on our Epi reactor by 50%.
Now, let's give you an example, a real example about sustainability. In this case, we took the SONORA. The reason we took the SONORA, because this is a product that we released last year, 2022, and from the get-go, we put sustainability as part of our release criteria. There are two things we looked into. We want to have higher throughput of the platform, because when you have higher throughput, then you have less energy usage per wafer. The other thing I wanna have- we wanna have much lower precursor usage, and for that, you know, we want a much smaller mini environment. We wanna make sure that also the mini environment is leak-tight. We reduced the volume, we looked into our reactor, and we changed it, made some differences in our reactor to reduce the precursor usage.
So the combination of the higher throughput and lower precursor usage allowed us to reduce our energy usage by up to 40% in the, the one on the right, polysilicon deposition. This is 40% less usage than the previous platform, which is the A412. This is using the exactly same process. We can achieve 40% less energy reduction, and for oxidation, we can achieve 14%. So if you go to this next slide, it shows you the energy sources for some of these three processes. These three processes, you have oxidation, which is chemical oxidation, happen at 1,200 degrees Celsius. Silicon nitride, polysilicon, these are CVD processes going at lower temperature. So that you see in all three cases, that electricity is the highest energy usage for all the three processes.
If you look into the definition with the three top electricity consumption, you see that heating element is the most one, second vacuum pump, and third is the blowers. So the heating element, you see that the one, the oxidation, use 67% because use most heat, because oxidation happening at higher temperature. The vacuum pump, the pumping is only happening for nitride and poly. It's not happening in oxidation, because oxidation is a atmospheric process, so you don't need to pump anything. So we showed you what we have done on the product. Right now, in order for us to be successful on the product sustainability level, we need to make sure that we take care of other things. One of them is we want to make sure that we have a reduction roadmap.
And otherwise, the only way to do reduction roadmap is you to be able to measure every year your sustainability. So we have SEMI S2 baseline for key application, and every year we have SEMI S23 measurement on our processes to see the improvement from one year to the other. The second thing, you need to monitor what you're doing. And here, you know, we're putting some sensors in our tool, okay? To measure the gas flow, to measure the temperature between the tool and other tool, and also between the tool and the sub-fab component. The third thing, we need to digitize. We have to have a database.
We like this database because the learnings that we have right now on reducing, improving sustainability, we can actually use it for as a theoretical model for the next development of our processing. We need to be able to trace the chemical uses, okay? This chemical we measured somewhere in the sub-fab, which tool it's coming from, and what process coming from. We want to make sure that we synchronize, as I mentioned, between the sub-fab and the fab and the product. Here, we need to use smart communication. And actually, we're working with our some customer to see how the sub-fab component can work with your product. Are you going to use Wi-Fi? Are you going to use other technology to do that? It's really, there's lots of work from that point of view.
Last but not least, we need to make sure that we, we measure the PFAS, the PEP, and of our products, and make sure that we have screening from that point of view. So to to conclude my talk, I hope we still have a very we have a very aggressive roadmap for net- zero. I hope I showed you in my presentation that in the product point of view, that we've taken it very seriously. There is many things that we are, we are, we are improving, and most importantly, it's our strategy is reduce emission on three components: better plasma efficiency, better use of chemicals, and more efficient heating, more efficient heating technology. Thank you very much. Now, I'm gonna have our CFO, Paul Verhagen, talk about the numbers.
Thanks, Hichem. So, good afternoon, everybody. I'm happy to be able to present the numbers to you. You've got an update on the market opportunity, technology, operations, sustainability, and what I will try to do is to show to you what that means in numbers. Five key takeaways. The first one, we believe that our growth through innovation strategy is creating significant value for all stakeholders. The second key takeaway, as you've seen, we've increased our 2025 guidance upwards to EUR 3.3 billion to EUR 3.6 billion, with the target range of growth and operating margin maintained. We've given you a view beyond 2025 up to 2027, where we target a revenue of EUR 4 billion-EUR 5 billion, representing a CAGR of 11% to 16%.
Operating margin maintained as well, but with some more color, namely that we do expect an upward trend in the outer years towards 2027. Annual CapEx, EUR 100 million to EUR 180 million. We used to have EUR 60 million to EUR 100 million, and that's mainly to support growth, and I will come back on that one as well. Capital policy, capital allocation policy unchanged, so investment in growth remains a key priority, and excess cash will be returned to shareholders. And the fifth key takeaway is, plan to achieve net- zero target by 2035 for Scope 1, 2, and 3, and our plans have been verified by SBTi. Further few KPIs, just to show you that why we believe that our growth innovation strategy is actually creating value.
These numbers are for the three-year period, 2020 to 2022, so an annualized shareholder return of 34%. Total cash returns to shareholders, EUR 500 million. Accumulated free cash flow, EUR 660 million, 12% of revenue. That might sound a little low. Maybe it is a little bit low, but that's because of the working capital built up. The last year, of course, before 2020 was 2019, we had a revenue of EUR 1.1 billion, and we grew to EUR 2.4 billion in 2022. So that's an additional EUR 1.3 billion, for which, of course, there is working capital needed.
You see in the footnote, if you exclude this change in working capital, it's actually 18% of revenue, close to EUR 1 billion free cash flow, excluding ASMPT dividends, and excluding acquisitions. An average ROIC, also excluding ASMPT, of 26% over this period. Revenue CAGR, you've seen that number a few times, 35%. Gross margin, operating margin, nicely within the guided range. Our renewable electricity over this period has been increased by 66%, so quite significantly. First shareholder return, a very nice slide, obviously. This is normalized to start at 100 in January 2019.
So if you would have invested $1 million or EUR 1 million, you would have had a return and see that capital grow to around $12 million over a 4.5-year to 5-year period. Significantly higher than what you would have if you would have invested in the stocks or in the, in the AEX. As I said already, excess cash is returned to shareholders. We have a sustainable dividend policy. You see a gradual increase of our dividends. No increase in 2022. Why not? Because we did two acquisitions, and they were mainly funded by cash. As you know, we issued a few new shares for LPE, but the bulk was cash.
We wanted to conserve some cash as well, and that's why we decided to keep the dividend equal to 2021. And as you most likely have seen, we just completed our EUR 100 million share buyback that we announced in 2021. Strong revenue growth, especially if you compare it to the WFE markets. So we grew 35%, WFE market 23%, so significantly outgrowing the WFE markets. Key drivers for revenue, obviously, the end markets. That's the most important one and also the most difficult one, especially today, to assess what's gonna happen in these end markets, what's gonna happen in communication, what's gonna happen in computing, what's gonna happen in any all the end markets in which we play.
The growth and the composition of the WFE market is, of course, derived from that to a certain extent, and then more specific to, to ASM, of course, yeah, our leading share in ALD, our growing share in EPI, of course, is important. Our selective growth in vertical furnace and PECVD, our growth in the spares and services, Benjamin is showing that, but all these factors contribute, of course, to, to our revenue, growth. Sustained healthy margins. We've guided, yeah, 46% to 50%, with the most important, element there is our application mix. We've talked about that many, many times. Limited operating leverage in our gross margin because the bulk of what Kent does in his operations, at least in terms of cost, in terms of finance, is purchased material, purchased subsystem, purchased components, so it's variable.
We have, of course, some fixed costs there. You've seen the factory expansion, we have depreciation, but still there is some benefit from operating leverage, but not a lot. Of course, productivity and supply chain improvements, of course, will help. Continued value engineering that we do with customer projects can either maintain and/or improve our margin. And, of course, sales price management, another important one, especially in the last two years, where we've seen quite some inflationary pressure. And as for operating margin, of course, it's the development of gross margin, but also, of course, SG&A development and R&D, and I'll come back to that one. So first, SG&A. You have not seen a lot of operating leverage in the SG&A line in the last two years.
We had to significantly actually step up investments to, one, grow in a controlled manner, but two, also to prepare ourselves for the next wave of growth, and that's what we've done. And you see also where we invested it, it's actually across the board, most in sales and, yeah, basically product and business development, which you would expect, I guess, with the growth that we've seen. The second biggest bucket is HR and recruitment cost. You've seen that we more or less doubled headcount over this period. Was an important area where we had to invest in, and then, of course, strengthen and professionalize a number of the functions, as you see here on the slide.
Then on the level of investments, you see actually now for four quarters in a row, that the investments in SG&A are leveling off. We are at a level of EUR 70 million to EUR 75 million per quarter. You see EUR 147 million and EUR 144 million, which of course, is a big increase compared to where we were, let's say, in H1 2020. But at least now, four quarters in a row, it's leveling off. It might go up still a little bit, but a big chunk of the investments for now have been done. Of course, at a certain moment in time, there will be new investments needed, but before that, you should see some benefits from operating leverage with the revenue growth that we target. R&D is a different story. R&D is not leveling off.
Actually, we're stepping up investments in R&D. You've seen, Hichem's presentation. There's, there's an enormous amount of opportunities out there. We are stepping up. We will move towards the high end of the range, maybe even slightly higher, and then, over time, it will come down a little bit. But there is so much going on now, we need to try to capture these opportunities. So we are actually, try to stay within our range as guided, but in the, let's say, this year, next year, the year after, most likely at the high end of that, of that range? And you see here the gross R&D, which, in the first half of 2023, moved up to 13.8%.
As you most likely know, but you can see it from the left, part of my graph, the difference between gross and net is typically 2.5% to 3%. But the gross R&D is the real money we spent, is the cash money that we spent in R&D, so I believe that's the most important number. The other one, of course, goes into the P&L, but gross goes into your cash flow. Working capital increased somewhat, especially, of course, in the last two years. Again, talking about receiving punches and now trying to avoid punches, especially inventory, went up quite a lot. Still, I believe in a controlled manner.
It was a deliberate choice to do that, to be able to deliver to our customers and to actually enable the growth that you've seen in the last two years. So you've seen that coming up a little bit. We see also some pressure on DSO, given the current market circumstances. So customers, of course, are trying to extend a little bit, not a lot, but anyhow, there is discussions that go on also now and then. So you see the working capital moving up a little bit, but still nicely within the guided range, and we expect that to continue. Then two acquisitions in 2022. Sort of LPE, EUR 470 million, and as you know, an earn-out of up to EUR 100 million, which still needs to be paid.
Strategic rationale, I think, is clear, and I think Benjamin Loh talked about it. We want to access this high-growth silicon carbide market. It's a very interesting segment, also from a sustainability point of view, because it's an important enabler for electrification of transport and other markets. We believe because of the technology, it was the right acquisition, a very experienced team there, and with our footprint, our technology, a lot of synergies that we believe we can bring to the table for LPE. The other one is more technology acquisition, Reno Subsystems, price of EUR 42 million. As also Loh showed, new ALD, PEALD applications are driving more complex process solutions, and this fitted perfectly in that for us. So it was really an important technology solution.
Also a company with strong IP, and also a company that also contributes in terms of sustainability because of, let's say, yeah, accelerated matching and improved cycle time. So it's less energy consumption and as a result, also improved sustainability. This is just a summary of our progress against our initial 2020 to 2025 targets. So nine dots, seven green, one yellow, one red. So the SG&A is slightly higher than what we anticipated. Our target still remains a high single digit. If we get there in 2025, I don't know. We'll see that, but it remains the target. But beyond 2025, with the growth that is coming, for sure, we will get there, no, no doubt. And the other one, of course, is CapEx. We had underestimated the, especially the investments we're doing now in our R&D footprint.
I think manufacturing was clearly on our radar, when we did our 2021 Investor Day, but literally across the globe, everywhere, we are outgrowing our facilities, given the number of opportunities, the number of tools that we need. Clean rooms are getting too small, so we need to upgrade and invest, and that's exactly what we are doing now. On all the others, I think we are on track. You've seen the numbers, and looks all pretty, pretty green. The new targets for 2025, so two things changed. The revenue, you've seen that, EUR 3 billion to EUR 3.6 billion. Basically, it's the addition of LPE to our previous targets, which were EUR 2.8 billion to EUR 3.4 billion. The CapEx, of course, increased to EUR 100 million to EUR 180 million.
You know, for 2023, so for this year, we guided EUR 150 million to EUR 200 million. So this EUR 100 million to EUR 180 million is for beyond 2024 and beyond 2027, EUR 4 billion to EUR 5 billion , representing a revenue growth of 11% to 16%. That's lower than the 16% to 21% that we used to have, but also the compare, it's on the back of the 2022 numbers that you've seen. That was two years and three years in a row, yeah, 35% CAGR, so it's not an easy comp to have that CAGR, but still significantly outgrowing the WFE market, at least that's our expectation.
Gross margin maintained, SG&A high single digit, R&D high single digit to low teens, and initially, you should expect that to be especially in the upper part of that range. Operating margin maintains with an upward trend expected. CapEx, I talked about. Tax will come back, and working capital basically remains as is, so no change there. Now here you see the revenue figures, same drivers. I don't repeat them, but yeah, from EUR 2.4 billion to EUR 4 billion to EUR 5 billion, it's 11% to 16% CAGR. WFE markets, based on this assumption, would grow with a CAGR of 4%. So 11% to 16% is 3x to 4x faster. That's quite significant.
And, yeah, that's what we currently, based on our assessments and based on the numbers that we have, believe. What is important to understand is that our revenue cost is really based bottom up by application, by customer. So Han and team spent an enormous amount of work in there. It's we know pretty well what kind of layers there will be. We know pretty well what that means in terms of revenue. The key uncertainty, of course, that we have is how do end markets develop? So how many tools and how much capacity built up will there be? That's always, of course, an assumption that is a little bit difficult to assess, but I know this is currently our best estimate and projection that we have.
Margin maintained, and the simple reason is, as I mentioned already, there is limited operating leverage at the gross margin level. The most important one, of course, is application mix. As you know, competition is also looking more and more at ALD. We are also entering areas where competition used to be strong, so also that, of course, is not always easy. So they will defend, of course, their positions. So to expect a significant increase in gross margin, I don't think at this stage that would be very, very realistic. Of course, we continue to drive supply chain improvements. Kent has shown that to you. And what we've done, I think, reasonably well, is offset the inflationary pressures that we've seen in the last two years by increasing sales prices.
SG&A, high single digits, and there we should start to see real operating leverage, going forward once the revenue starts to grow. For instance, in the second half of this year, compared to what I just showed you the previous slide with the half-year SG&A, the percentage will go up. Why? Because we guided that revenue will come down 10% or more second half compared to first half. So you can expect, of course, that the SG&A as a percentage of revenue then will go up. R&D will grow with revenue, for all the good reasons. It's a lifeline. It's important for us to invest in R&D and to stay ahead of what's next, and that's what we try to do.
Operating margin, yeah, the key improvement and the reason why we actually add the additional color of, yeah, the upward trend in the outer years is simply because of the operating leverage that we will see in our SG&A line. Maybe to refresh your memory, in 2018, so five years ago, our SG&A was still 15% of revenue. So it has come down, maybe not as much as you would like it to be, but it has come down. It will also come down further in the years to come. But again, the investment that was done in that line was needed for all the reasons that I mentioned. Tax, there's a lot of things happening in tax land. Maybe the most important one is the global minimum tax.
Of course, it will kick in starting next year, 15%. That in itself can have an impact on our tax rate. I mean, if you are in a country with certain tax incentives, and if the overall tax rate that you pay is below 15%, then of course, you will see impact. If the overall is higher than 15%, then you will still benefit from the same incentives in the same manner as you have them today. So there will be most likely some upward pressure, but still, we believe it will be limited, so we still guide for high teens to low twenties. Allocation of profits between countries, of course, is important because there's different again different taxes between countries.
We follow an arm's length transfer pricing principle as good as we can. We benchmark that and underpin that regularly to make sure that we update it when necessary, so we do that. We don't have any artificial tax structures in place, so it's really based on substance, what we do, which I believe is important. And of course, we stay on top of all the tax developments that happen and try to, of course, manage that as good as we can and try to optimize our ATR as good as we can, but within the rules of the tax rules that we have. Working capital, the absolute amount of working capital, of course, will go up.
You saw in the free cash flow, also 2020, 2022, the free cash flow as a percentage of revenue, with and without was quite substantial. I hope that the difference going forward will not be as large as what you've seen in the past, because we had two effects. One, of course, was the growth, but the other one was also just the increase in working capital because of all the supply chain challenges that we had to deal with. CapEx, EUR 100 million to EUR 180 million. You've seen our announcement on the Phoenix land. You've seen our announcement in Korea, both R&D and manufacturing. We also have plans to expand in Europe further, not yet announced.
We're evaluating that now, so once we know more, we will let you know, of course. There's continuous investments in products, in our products, but also in metrology tools, but also in clean rooms that need to become larger and larger. And as I said already, for this year, 2023, CapEx guidance is EUR 150 million to EUR 200 million. Capital allocation strategy unchanged, so growth, growth, and growth, allocating money to R&D, to CapEx, and opportunistically to M&A will continue. We want to keep a strong balance sheet, EUR 600 million net cash position around that level. Sustainable dividends and excess cash returned to shareholders. And then on net- zero, two or three more slides. I think what is important for us is that, one, we take sustainability very serious.
It's now, I think, ingrained in the whole company. It's part of our product development strategy, as you've seen, product development process, product improvement process. It's part of supply chain. We actually invested quite a lot. We added also quite a few people in this area, which is important. I mean, the semi industry has a footprint. The semi industry needs to attack that footprint, but I do believe that we are part of the problem, but we are a bigger part of the solution. To basically meet all the, yeah, the computing power that this society needs, you need innovation, you need more powerful chips, you need scaling, you need new materials, and that's exactly where we are very well positioned to deliver on that. So that's our aim to do that.
Our target is verified by SBTi in July 2023, which was a big deliverable, and you need to deliver a lot of material to them. You're challenged, you get questions, so they go through it in great, great detail. It's not an easy exercise, but we managed to get it verified. In all our key sites, we have 100% renewable electricity, and by 2024, which is next year, it's around the corner, we want to be 100% based on the renewable electricity. As you've seen in Hichem's presentation, Scope 3 is by far our biggest challenge, and that's something we cannot do by ourselves. We need our customers, we need our upstream, our supply chain.
That's why we've also taken the initiative with some other players to actually create and found the Semiconductor Climate Consortium, because that brings the whole supply chain together. Various work streams will start working there to actually drive sustainability and, of course, ultimately to achieve our net- zero target, but we cannot do it by ourselves. We will do whatever we can to work with our customers, to work with our suppliers, but we're also dependent on them, so that you should recognize as well. Near-term actions, most, I think, have been covered. It's a lot of work. It's something we like to do. We put the people in place to actually manage this, to get it done, so it should be doable.
We are now completing energy audits for all our sites to see what can we do further to improve energy efficiency. With the expansion that we have, of course, we put new buildings in place with the highest sustainability standards, so that in itself will already help and will contribute, but still, Scope 1 and Scope 2 is less than 1%. So the real challenge is in Scope 3, and that's where we need, again, customers, suppliers, and we do, of course, our part as well to get that improved. And with that, my key takeaways. We believe that our strategy is working and that we are creating value.
We have increased our revenue targets for 2025, introduced new targets for 2027, capital allocation policy unchanged, and our plan to achieve net- zero by 2035 verified by SBTi. So thank you. For sure, So...
Oops. So I was just taking a look, and I think it was about 142 slides, so I hope we did not overload you, but the intention, of course, is to share with you as much information as possible. And, you know, as I said at the beginning, I hope, you know, me and my colleagues have been able to, you know, explain to you what is our view, you know, of the, market, the outlook, the technology trends, the industry trends, and how our strategy, growth through innovation, fits into it. You know, if I were to summarize, you know, what we are really trying to convey to you, is that we are very bullish as far as the semiconductor industry is concerned.
Despite short-term headwinds, we think that the industry is poised for a lot of growth. Having said that, I also believe that we, as a company, we are very well positioned in terms of the technologies, that are coming. You know, my colleague, Hichem, has shown you, and it's very exciting, stuff like metals, stuff like, selective, ALD, you know, all the great stuff that actually, you know, Hichem and team have been working on for already a couple of years that is now ripe for us to harvest. So we are very excited about that. You know, our biggest play, even though it's a leading edge, logic, foundry, and memory, you probably got a glimpse from a part of his, presentation that we are also trying to explore some smaller areas.
Maybe they are not as big as, you know, advanced logic foundry type of size, but it is meaningful, and it is also leading edge. Stuff like advanced packaging, heterogeneous integration. We are not doing the full value chain, we do what we can. And of course, you know, we have started to become opportunistic in terms of our growth strategy, and we will continue to look at M&A that fits well, just like what we have done, you know, last year.
LPE as an acquisition, was, you know, great for us because it allow us to, you know, develop and explore the synergies that we think, are there, but at the same time, to look at a market that will add, you know, a chunk of meaningful, you know, revenue to our, portfolio as well. And last but not least, we are going to continue to focus on the other factors. Paul ended with sustainability. That's a, you know, a, a big thing with the company, and, it's well embedded in the company. We're going to look at how we grow our culture, people, you know, gender diversity, et cetera. So we are very excited. I think I speak on behalf of all my colleagues. We are really excited about what's, what's coming.
With that, I hope we have been able to impart and convey all the information that we have planned. I will close the presentation and then shift over to Victor for the Q&A. Thank you very much.
Thank you. Thank you, Benjamin. So we will now start the Q&A in a moment with all the, all of our presenters on stage. We will be taking questions from the audience here in the room, and if you want to ask a question, please raise your hand, and we will be passing around some mics here in the room. Before asking your question, please state your name and company name, and please limit your questions to not more than one at a time, please, so that as many as possible have the chance to ask a question. We will now prepare the stage for the Q&A, so please bear with us for a moment.
Want to sit here?
Yeah, that's what I was told.
Ah, I didn't know.
Okay.
It's interrogation time.
Seems we are ready for the first question. Didier?
There's your mic.
Didier Scemama from Bank of America. Thanks so much for your presentation. Very useful and, high conviction that you will outperform WFE over the next few years, at least on my end. Maybe starting with the ALD, market opportunity. I'm just curious, you gave us your 2022 ALD TAM. Can you, can you give us an estimate for 2023? I mean, probably grew double digit, I would imagine. So that would bring us quite close to $2.9 billion and put your 2025 TAM in, sort of within the vicinity of that, if that makes sense. So just give us a sense of 2023, and, and I've got a, a follow-up. Thank you.
I'll take that, and then, Han, maybe you can add on. I think 2023 is still in progress, so we are not going to give any kind of number on that. But I think Han has explained the look. You know, if you look at where we were in 2022, the market size was EUR 2.6 billion. And by 2025, you know, we do expect that that's going to be between EUR 3.1 billion to 3.7 billion. Let's just say a midpoint of EUR 3.4 billion. So you're going to see significant growth over the next couple of years. And a big part of the growth is going to come from, of course, the transition to, Gate-All-A round.
I think, you know, if you maybe do the math, so you can find a way that we get there, but at this moment, we are not giving any guidance for 2024 yet. Han?
Yeah, I think the question was for 2023.
Yeah.
But same answer. We're of course tracking in detail how the market this year and the first half next year is developing, but it's too early to give really a color. We're confident about the 3.4 for 2025, but for the shorter term, we will not give any guidance.
Okay. The other question is on sort of the current markets. Obviously, China has been a big market driver for ASM and for the WFE this year. I think it's about 20% of your revenue. I was surprised to see on your slides that the mature node investments in China actually are not a major contributor to your revenue. So I'm curious if you could give us a sense of where your strength in China is coming from, presumably LPE, but also leading-edge logic and foundry. Memory, presumably, has mostly disappeared, I would imagine. So just give us a sense of where that strength in China is coming from.
So, maybe I'll take that. I think when you look at China, of course, you know, they have various, you know, pockets, where, you know, different segments of the industry. So one, if you maybe start with, really leading edge, you know, I think that part of the business, you know, today is more limited just because of export control. So there isn't really going to be a lot of investments as far as 14 or 14 and below. Now, if you look at, maybe what has actually been strong for us, you know, in the past and also I would say today, is the power analog and wafer segment has been strong for us, and that's actually where we have, you know, quite some revenue coming in from China.
If you look at then the silicon carbide area, that's just, you know, starting, but we do have a good presence there. Maybe not, you know, in terms of absolute numbers are that big. Now, what actually is a big part of the market in China is what we call the non-leading-edge logic and foundry. So the stuff that is like, you know, 28, you know, 32, 40, 45, you know, 65- nanometer type of... That is an area where, you know, traditionally we don't play much there. You know, and, you know, if you want to go in now and you are not one of what is called the baseline, whether it's in deposition or whether it's in epitaxy, it's very hard because a lot of it is just copy exact.
So that's an area where it is less beneficial for us, and that actually is reflected in Han's presentation why, you know, in 2022, we did not benefit as much in our silicon epitaxy market share.
Yeah. That's exactly right, and no, nothing to add.
Yeah, hi, Sandeep Deshpande, JP Morgan. You know, you had a very good presentation on the new opportunities you have with your products. You know, when you spoke to the market last time at your investor day, you talked about Gate-All-Around, and that is actually happening now. Which of these new opportunities that you've talked about will begin to contribute, and in terms of wins that you already have in these new opportunities for your 27 revenue opportunity? And I have one quick follow-up on that.
Okay. I think maybe, Hichem, you can speak about that, the ALD and the Epi opportunities.
Okay. I think when I look into the Gate-All-Around, I mean, there are many new opportunities in both, ALD and Epi. I think in ALD, the opportunities are mainly in the gate stack area, with many new dipole and also work function metals. When you go to Gate-All-Around, you need many of, many more of these dipole and work function metals. And for each one of them, you know, for each dipole, you have a new patterning ALD layer. So we see strong growth there. And for Epi, we also. When you go to Epi in the Gate-All-Around, what happens is that, you know, everybody needs to have the nanosheet structure. Before, like we have in FinFET, not many customer have a channel films.
Some customers do, but many of the customers don't have any channel films in FinFET. But when you go to Gate-All-Around, you have to have those channel films, the nano, nanosheet structure. So that's how, you know, you gain market share, and that's how, you know, your market increases from that point of view.
My follow-up then would be, you know, with ALD, you were, you know, a pioneer in that market. When you look at Epi and now some of, you mentioned, Paul mentioned earlier, that you're now coming across some competition in some of the new areas that you're going into. So how do you see yourself positioned? Is this now that you're going to be directly competing with some of your, particularly U.S. peers, on these markets, and how are you going to fare in this competition? And then particularly, do you see them in your ALD market?
Yeah. So I think, of course, you know, with all the new ALD applications coming up, you know, it's kind of open, so everybody has a fair chance. And, you know, of course, the competitors are trying to get a piece of it for themselves. But I think we, you know, coming to the end of almost to the end of a 2- nanometer Gate-All-Around development, I think we see that we have maintained our position there. So we are confident that, you know, we are getting our fair share based on our leading market share. That's one. And going forward, of course, we continue to work with the customers, you know, for the next node, et cetera.
One of the things that we looked at, actually, is that if you look at ALD, that's the situation, but if you look at, for example, epitaxy, it's actually on the contrary. When they go into a leading-edge Gate-All-Around, it will be, you know, much more beneficial to us because we have been so far just a small player as far as FinFET. You know, we book in a 7-nanometer and have continued, but that was about it. And, you know, the, what do you call, a non-leading-edge logic and foundry, as I've just explained, we don't really have a position there. We have a position in power, wafer, analog. Now, moving to Gate-All-Around, that's where I think it's going to help us because we will have many more Epi layers.
We would have many more, much more penetration, and that is what, you know, we are looking at to get us to, for example, that, that at least 30% market share by the time we exit 2025. Maybe just one addition to your question, Sandeep, is you asking when do we see all of this becoming revenue for us? Actually, it already starts next year with the pilot lines. Sorry, can we maybe move to the middle, Janardan, I think?
... Hi, thanks. It's Janardan Menon from Jefferies. Two questions which are perhaps sort of related into one, so hopefully, Victor, you're okay with that. When I look at your guidance, and I break it into two segments, 2023 to 2025, and 2025 to 2027, and I'm using sort of consensus, which is for 2023, based on your own guidance of more than 10% drop in the second half and your Q3 revenue guidance, et cetera. I calculate that your growth will be around, give or take, 13% CAGR between 2023 and 2025. And then you are forecasting a sort of acceleration to about 17%, 2025 to 2027, based on the 3.3 and the 4.5.
I'm just wondering what you think is gonna drive that acceleration, because, as you've said, you know, you are confident that the 2-nanometer Gate-All-Around is going into high-volume production in 2025, and that has been the big driver for you in ALD and epitaxy. So what is gonna drive the further acceleration among various things from 2025 to 2027? And sort of maybe related to that is, you know, compared to your previous capital markets day, you've talked about some new areas in atomic layer deposition, which is like the molybdenum interconnect metal layers, the advanced packaging, the selective ALD, et cetera. So these seem to be new and seem to be coming in around the 2025 time frame, but you didn't raise your ALD market expectation.
So, you know, why is it that you've got new, many new areas, as you just said in the previous question as well, but, but the market sizing is not changing on ALD?
Sure, Janardan, maybe I will answer the second part first. And I think one of the reasons why we don't share so openly all the different applications is, you know, because of the competitive nature. And, you know, we are a very, let's say, focused, dedicated deposition company. So, it's very easy to, you know, kind of figure out what we are doing, you know, from our competitors. And actually, some of the things that, you know, Hichem has mentioned, whether it's metals, whether it's selective ALD and so on, it is now finally happening. So before it happens, you know, there wasn't really, I would say, a need or, let's say, a necessity to, you know, try to oversell such stuff. But this will happen.
I mean, Hichem has already mentioned, this is happening in the first 2-nanometer Gate-All-Around. Pilot lines, you know, will come into place in 2024, and if everything goes on track, high volume in 2025, we are going to see this happen. So this is now and the reason why you are getting this information now. You know, when you look at the, the, let's say, the growth rates, as you have correctly mentioned, I think, you know, we should also consider that we are in the year 2023, where we are seeing headwinds. If we did not see that, the, the curve would probably, you know, be much more smoother. The growth rate would, probably, be better. But having said that, I think, we stick to what we have guided.
I think, you know, by the time we get to 2025, we hope or we believe there will be a recovery, and there will be significant investments in leading edge logic foundry memory as well. And, you know, the steep drive towards 2027, I maybe want to bring you back to, you know, one of the slides on my colleague Han's presentation, where, you know, it's basically showing that from 2024 onwards, you're gonna see a $50 billion, you know, let's say, increase due to Gate-All-Around. That is a big part of it. Our biggest driver is still, you know, in leading edge, you know, logic foundry, and memory. Yeah.
Hi, Stéphane Houri from ODDO BHF. Quick question maybe on silicon carbide, because the upgrade of your guidance for 2025 is mainly based on the silicon carbide upgrade, LPE. And when you make the calculation, EUR 200 million of revenue makes, you know, 24% CAGR over the next two years. And I've heard you many times saying that 25% CAGR was a bit conservative. So is there a delay in the deployment of your product, or is it just a conservative view of the next two years? Thank you.
Well, we did not really look at whether it's 24, 30, but I think if you look, see from Han's presentation, the yield data for devices is basically showing it's going at a... Or the latest information is going at greater than 30%. And, you know, I think, you know, if you look at the industry today, it's growing very rapidly, and it's also, I would say, evolving. And, one of the reasons why we have not provided, you know, for example, the market sizing and market share, is we do believe that whatever we give it to you today, it will change or become outdated, maybe three to four months down the road, because the market is moving so fast. Now, on the whole, I think we are doing very well.
Having taken over or, let's say, closed the transaction in October last year, I think we have shared that, you know, with the continuing presence that we have in China, which we are still doing well, we have also secured, you know, new wins for major customers in the U.S. and also in Europe. And maybe just to add that these are not eval tools. These are multiple tools that goes into high-volume manufacturing. And, you know, we are in a lot of discussions, you know, demonstrations with, let's say, the other major players, and we are confident that this, some of them will lead to also positive results for us.
Hi, thank you very much. François-Xavier Bouvignies from UBS. I have quick one on pricing. I mean, we talk a lot about your layers and your opportunities to increase your footprint there, but what do you assume in terms of pricing as... I guess, as well, you have, you know, an improvement roadmap in terms of tools, throughput, you know, quality, and so forth. So what are you baking? And the second question is on selective ALD here. It seems to be a very significant opportunity if, you know, you showed the platform, you know. What kind of ASP are we looking at for this kind of product, and how is the competitive landscape specifically for selective ALD, given the change?
I think maybe on pricing, Paul can give you some idea, and then on ALD, you know, Hichem can give you some insight. But we are probably not going to tell you how much the platform is for competitive reasons, but you know, maybe just more idea on selective ALD. Paul?
Yeah, on pricing, the key word is the cost of cost of ownership. So when we talk to customers, it's all about cost of ownership, the cost of ownership they target, the throughput time of the wafers that we see. They have certain targets, we need to deliver on these targets. So from a margin point of view, again, it is very similar with certain applications higher than others, very similar to what we see today. There's no big changes. There's also, as we say, always competition, yeah. For every layer we need to compete.
It's just head-on competition, many times, and sometimes, again, we're leading by a mile, and sometimes it's more close, so that will be a factor. But cost of ownership is there, the key point. On average pricing of the tools, it will, again, depend very much on the mix of the tools. I mean, the tool that Hichem showed with the 6 passes, the 6 chamber, that's of course a more expensive tool than some of the other ones. But it will depend over time, how much of these are we going to deliver, and of course, we made certain assumptions. So average price might go up a little bit, but it does not really have a direct impact on margin, because again, the magic word is cost of ownership.
Hichem?
Yes, I think to talk about our differentiation in selective ALD, I think that what we have done is actually when you do selective ALD, you need to use inhibition, and the inhibitor is a unique process and precursor that we have developed ourselves. This inhibitor is ALD based, which is different from what's used in the rest of the industry. It's very stable when you expose it to air. It's easily removable, and it's applicable to all of the selective ALD markets. So in selective ALD, today I showed you DoD, MoD, but also there is MoM, there is DoM, there is all of the things, and the technology that we have is actually applicable to all the different selective ALD deposition.
So we think that we stand a very good chance in the selective market from the technology point of view.
Any? Alright Didier? Yeah.
Yeah. Sorry. Two quick questions. One on, on Gate-All-A round, and just wondered if you could clarify one thing from the presentation. It seems like your strength in, I can't pronounce it, molybdenum, ALD, for basically the vertical channel is essential, as historically it's been based on copper. Can you confirm that that's just broadly the sort of industry recipe across the three major leading analog foundry customers? And if so, does that mean you've got them in the bag already, you know, locked in effectively? And I've got a follow-up on China. Thank you.
Hichem?
Okay.
Molybdenum.
Yeah, so talking molybdenum, I mean, molybdenum ALD is used, not in the Gate-All-Around device, but actually in the middle end of line and back end of line. We have worked with a few customer on metal ALD, on molybdenum ALD, in a Gate-All-Around application, and we have some good collaboration and some... And we have developed, and we are qualified in some of the customer, our customers' process flows.
All right. And on, and on China, of course, the current environment is somewhat uncertain, you know, with, concerns around, for instance, a full immersion ban in China. So that's why I asked the question earlier, and it feels like from your answer, your market position in the core immersion node, so call it 28 to 65, is actually reasonably weak. Does that mean that in the sort of worst-case scenario, a full immersion ban in China, the impact on your business would be reasonably modest?
I think, you know, that would be, you know, speculative, because it really depends on what is being restricted. You know, today, I think there are some, as far as we know, there are some restrictions on, for example, you know, immersion, but I do not believe that everything is restricted. So I think we still need to see, you know, what exactly is the restriction, and, also at the same time, for example, you know, for whether it's 28, whether it's 40, or whether it's 65, what are they using it for? I think that, that's still, you know, left to be seen. As of today, you know, maybe just to share with everybody... We do see a strong business coming in from China, this year.
So we, at the end of 2022, China was about 16% of our equipment revenue. This year will definitely be higher. And a part of it is, you know, for the customers that have not been able to, let's say, invest in what is called leading edge, which is 14 and below, they have changed to, you know, investing in maybe a, you know, a SOI or, you know, a 28- nanometer. So, those we still have some business there, and that's what is actually helping us to drive our business. In addition, I think the, the, especially on the power side, in China, it is still the investments are still continuing, and it's still, quite healthy. Yeah.
Mark, ING. Can you provide more, a bit more detail on market shares? If I recall correct from the previous capital market, you have a very good visibility a few years out because of the application wins that you expect to make. Then you also said that you expect the logic foundry market share to remain at ALD, to remain very high, similar level, and then win share quite clearly in the memory segment. Can maybe give a bit more detail on that? And also on the epitaxy side, you make the jump from 15% to 30%, and at the time, I think you said it's mainly driven by shrinkage. I assume that you'll give even more shrinkage and more towards Gate-All-Around, towards 27.
So why then you do not make more market share in 2027 than what you're gonna make in 2025?
Han, you want to-
Yeah.
Do that?
Yeah.
Yeah.
Yeah, so the market share, we gave some color on the ALD market share, then higher than 55%. It is still true, what, what you, what you quoted in... Generally, in logic foundry, it's higher, and in memory it was lower, and we're improving, and that is exactly what we see in the last year, two years. And we continue to invest in memory, so nothing is changing there. So, so basically, we maintain our goal of larger than 55%. That's what we publicly share. We think it is a healthy target. It has become a competitive market. So for many applications, we have to compete. So that is the color I would like to give on ALD.
On epitaxy, also there, to go from where we were two years ago, 14% to 15% to 30% is a big step. It's competitive. But we confirm it today, and as I said, the transition to Gate-All-Around we think will help it. So for the moment, it is too early to change anything on the 30% target for 2027. That's just too far out, and we wanna just feel more confident and be a few more years later down the road to actually have made that progress by 2025.
Yeah, Tammy.
Hi, it's Dominic O'Brien from Balyasny. I just had a couple of questions, please. One, I think in the previous guidance, you said in 2025, there'd be a $1.2 billion opportunity from no progression in Gate-All-Around. Could you give us an update for that number in 2027? And then when I think of the 2027 guidances, understandably, quite a large sort of range on the revenues for that, can you just help us understand what gets us to the lower end of that range and what gets us to the higher end of that range? Is it just simply the adoption of Gate-All-round in 2027, or are there other factors? Thank you.
Han, maybe you wanna take the EUR 1.2 billion,
Yeah.
Yeah.
So, the $1.2 billion we shared two years ago was meant to give an indication on what the increase would be for ALD and Epi in going from in logic and the effect of FinFET to Gate-All-Around. We have not updated it for 2027, but instead, we have given you today the impact of FinFET to Gate-All-Around through the $400 million increased SAM for ALD and Epi. That is a very clear, this is the SAM for us today. At FinFET, this is the SAM at Gate-All-Around. That is the $400 million SAM increase that we see.
And that, we think, is a clearer update than trying to update the original EUR 1.2 billion target.
Maybe the second part, I will try to, you know, take a stab at it, and then Paul can also join in. So the way that we have structured it, as Paul has mentioned, is we do a really bottom-up, you know, kind of trying to find out, okay, what are the applications that we will be selling our tools for? We also look at, for example, where the customers will be investing, how much they will be investing, and that's why we come out with the range. Now, we also try to look at, okay, is there any correlation between this and, you know, WFE? And, you know, as Han mentioned, the midpoint that we have taken is $100 billion, let's say, you know, WFE in 2027. Paul, anything you want to add?
No, it's, of course, in market development, for sure it's customer mix. Per customer, you see, of course, different penetration levels. It's 27, so not all layers for 27 are known already, which one we will win and not win, so we made certain assumptions there. So we can, of course, be better or slightly worse than what we currently estimate. So there is still quite a level of uncertainty, and in addition, which also doesn't help, is today's uncertainty.
... It doesn't help if you have to guide for the coming years, if you have, let's say, an environment that we have today with the level of uncertainty, with inflation, with high interest rates, with geopolitics, et cetera. So yeah, we just try to illustrate that, yes, through a, let's say, somewhat larger range. I don't think it would be realistic to give you a range of whatever, a hundred million difference. That's not... Unfortunately, we cannot do that. I wish we could, but not possible.
Corinne, Corinne, can you-
Maybe in the front. Yeah, I see Tammy's hand , like, the 15 times going up.
Thank you for taking my question. So firstly, it's regarding heterogeneous integration. So it's definitely a very popular topic, even one of the backend player do have a cluster tool with a front-end competitor of you. So based on my knowledge, you still do have this kind of, you know, holdership with ASM Pacific, but you're not, not necessarily working together. In the future, let's say, for example, for hybrid bonding, when plasma technology knowledge is more important, would you be considering working closer with ASM Pacific to create this sort of front-end, back-end collaboration again? And also, my second question is actually on your Epi market share. So if I'm not mistaken, your current market share in Epi is about 16% in 2022.
Is the 30% going to come straight in 2025, or are we going to see a progression in 2023 and 2024? Are you gradually gaining market share? Thank you.
You know, maybe I'll try to answer a little bit on the ASMPT part. And I think it will be good for maybe Hichem to explain to you what exactly we are doing on the advanced packaging. So of course, you know, we still are shareholders in ASMPT. We still own 25% of that. But what we do today in advanced packaging or heterogeneous integration has no relation to them. There's no collaboration, there's no joint marketing or activities. And what we are really focusing on is really what you call the front-end part of, let's say, advanced packaging. And Hichem can maybe share with you some, as he has mentioned, some of the things that we are doing, which also requires, for example, ALD.
So for heterogeneous integration, I mean, some of the things that I mentioned is in the area, for example, of TSV, okay? In that area there, you know, you need some ALD liner from that point of view. And so we're actually partnering with customer to develop that technology. And we see also lots of investment happening in the backside power distribution network with the PowerV ia, whereby you need to have many ALD liner films, but also metal cap film. And, you know, the metal cap film can be with tungsten. We're working with our customer to do actually metal cap film with molybdenum, okay? Which is actually be better because it has lower resistance than tungsten. So we're involved in that.
Also on the interposer level, interposer is really moving into the incorporation of MIM caps in the interposer and the deep trench capacitor. And for that, you know, the MIM caps, it's mainly ALD layers, you know, high-k ALD layer capacitor. You want to increase your capacitor, you go to ALD. So we see lots of partnership with our customer in that area, too. And in silicon photonics, which is really part of the heterogeneous integration, I mean, we actually the tool of record for all customers on in Epi with for the germanium detector and silicon detector. So we feel very confident of our position, and we're working very tightly with some customers to understand this market and to get something out of this market.
But we are very selective on what we do, you know? But if we see an area that we have some strength, then we go and jump in.
Maybe, Tammy, on your question on the Epi market share, I think, you know, the growth in our Epi market share is going to be quite dependent on Gate-All-Around. And that's why, you know, we need Gate-All-Around to come in. And of course, starting next year, we pilot, hopefully, you know, really ramping up high volume in 2025. You know, we have shared in the past, we are already in the leading foundry. We have won a second customer. We're not gonna say anything about the third customer, but we have a very good traction, let's put it this way. So that's all going to help us, you know, when it moves in the Gate-All-Around. Yeah.
Hi, Simon Coles from Barclays. Sorry, over here.
In the back, left.
Yeah. You've had the LPE business for nearly a year now. I understand that silicon carbide is a fast-growing market, but I was just wondering if you could expand a little bit on the strategy for the business, because I'm guessing now is a consolidation phase, understanding where the opportunities are. But given your expertise and their expertise, is there scope to expand the addressable market to other compound semiconductors in the future as well? Just how to think about LPE from here and then into that 2027 guidance would be great. Thank you.
Sure. I think on the silicon carbide epitaxy side, we are doing well with some of the activities and wins that I have described. I think the market is, at this moment, looking you know at buying or investing in what you call equipment that can do both 6- inch and 8- inch. And the reason for that is 8-inch substrates is not really available, and they would just want to make sure that they want to buy equipment that can do both so that they don't have to change. This investment, I think, will continue heavily driven by electric vehicles.
Now, one thing that I think we should, you know, also, make clear, and, that, that came from, you know, Hichem's presentation, we actually have already been selling, you know, 8-inch vertical furnace for silicon carbide and for gallium nitride applications, you know, for, for a long time. In fact, I would say when you look at 8-inch vertical furnace, we are probably, you know, the market leader for that, that product. So we have been doing that. So if you look at that, we are now trying to look at, can we also cross-sell, for example? Same silicon carbide customer. Some may not have our vertical furnace, some may not have our silicon carbide epitaxy, and we are trying to do that together.
... Oh, hey, it's Timm Schulze-Melander at Redburn Atlantic. Two questions. The first, in FY 2024, you talked about these Gate-All-Around pilot lines. On average, how many wafers a month is a Gate-All-Around pilot line?
Sorry -
Wafer starts per month for a pilot line Gate-All-Around.
For fiscal year 2024?
Twenty-four, yeah.
I don't think we are at liberty to disclose that. But generally, pilot production is small. Small. Few K, maybe maximum 10 K. Yeah.
Thank you. That's, that's very helpful. And then, you mentioned expanding your European capacity, and you talked about CapEx. Are you thinking about building a sort of European campus center of excellence, or is this just purely assembly and manufacturing capacity? Thank you.
Well, I think maybe Hichem would be in a better position to share the idea that we have.
Okay. I mean, for Europe, you know, what we see in Europe as the strength is really in R&D. So. But in Europe, we also located in many different area. If you look into U.S., we have a central R&D system. In Asia, we have a central R&D system, but we don't have it in Europe. And Europe is our hometown pretty much, okay? And we like also to put more and more emphasis on R&D in Europe, especially in the area of power electronics, which Europe is a leader here. So, that's really the thinking that we have. It's really strengthening our R&D in Europe, which we already have some investment in Finland with University of Helsinki and also with IMEC.
Gianmarco Bonacina from Equita. Your margin, your EBIT margin guidance, is pretty stable through 2027. So do you expect it to improve gradually from 26%, which is probably this year, up to the end of the range, or should be, always in the middle of the range because you are ramping up the, R&D expenses? And then second question, on silicon carbide, you have a single wafer technology. You have in the pipeline to expand, to multi-wafer, or you will stick with this because of, better, let's say, productivity in terms of defects? Thank you.
I will just quickly answer the second part, and then Paul can, you know, give you insights on the margin. So the answer is no. We will continue with single wafer, because it's proven that, you know, in terms of our performance, that is superior, so we will stay with that. Paul can explain to you the, the margin side.
Yes, so, so for the margin, it will not be a linear line up. As we always say, the most important factor determining the margin will be the mix, and the mix can be different from quarter to quarter, from year to year. It can be better or worse, et cetera. So it's definitely not one line up. One factor that will contribute and that should see a line up, once revenue starts to grow again, is the operating leverage on the SG&A line. So everything else equal, the margin will go up. But of course, if the gross margin would come down because of mix, then of course, also the operating margin will come down. But still better than it would have been today, given the operating margin, the operating leverage that we will expect on the SG&A line once again, revenue starts to grow again.
I don't know.
Yeah. Hi, yeah. It's Rob Sanders at Deutsche Bank. I just wanted to take a step back and look at the deposition market. It's about $27 billion. You only address maybe 10% to 15% of that deposition market. The largest part of the other is PVD, CVD from tools like Endura, from Applied Materials. So what is your ambition to penetrate the Endura market today? You mentioned this molybdenum opportunity. Is the goal to basically eat into the larger CVD and PVD markets, or is it just more of a niche strategy? Thanks.
So of course, the main part of our play in deposition is, you know, ALD and epitaxy, silicon epitaxy. And, you know, we will continue to focus on that. Now, what you have just mentioned about, you know, whether we would like to go after the market that is now covered by our peer, will depend. You know, the case of metal or ALD metal is one that is very clear because it is transitioning from CVD and PVD into ALD, and that plays to our strength. When it comes to this kind of situation, of course, we are going to go after it, but we are not going to, you know, try to go into, for example, a CVD, you know, tungsten or PVD, you know, copper, because that's not our main core area. So that's what we would do. Yeah.
Thanks. two, two quick ones. Thanks so much for the spotlight on your advanced packaging capabilities. So first question is, how big is that market for you in percentage of revenue? You can choose whether you want to disclose 2022 or first half of 2023. Second, on the TSV opportunity, should we conclude that you've got exposure to CoWoS within that? Thank you.
... So maybe I'll just take the first part, and then Hichem can answer the second part, whether we have exposure to CoWoS. I don't think we are going to you know, disclose any revenue specific for segment. And you know, for us, advanced packaging is still a growing market. Today it is not really that big, but maybe if you fast-forward five to seven years down the road, it could have a lot of potential, and that's why we are trying to do it. As far as whether TSV has exposure to CoWoS, I will let Hichem answer that question.
Yeah, I think we're working with our customer on CoWoS opportunity right now, in TSV packaging, but it's really under development phase right now more than anything else.
Maarten Verbeek, the IDEA!. Firstly, a bit puzzled on your, on your CapEx. Initially, you mentioned EUR 60 million to EUR 100 million for the next four years. First year was already ahead of that number. For this year, between EUR 150 million to EUR 200 million and then for the coming years, between EUR 100 million to EUR 180 million. More or less, initially, we would end up at EUR 320 million, now we are at EUR 550 million. Your sales guidance hasn't changed, if we taken into account the LPE acquisition. So what new signs, insights did you have, that you have changed your CapEx plan so drastically going forward, compared to what you have stated two years ago?
Paul?
Yeah, you have more insight in the number of square meters that we require. And, I think that was not on our radar that, with the growth and the number of the opportunities that we work on and the number of tools that we need in our labs to work actually on these opportunities, that we would literally, almost at the same time, instead of sequential, outgrow the facilities that we have, the clean rooms that we have. And that was just not sufficiently clear on our radar. Now, that we have better on our radar now. We are working now on these applications. We know better, I think, what is coming, and with that, we've updated our guidance.
But of course, if we'd have known this in 2021, we would have immediately shared that, but it was just not foreseen that our R&D facilities would need that space that we actually need today.
We have time for one final question.
Yes, yes. Ruben Devos from Kepler Cheuvreux. Just one short question on ALD, basically, where you talked about the addressable market being 10% to 14% growth rate. Basically, you're looking for a 55% market share, both in 2025 and 2027. So, just thinking about split between volume and pricing there, what would that be? And then the second question is really on the geopolitical risks and your export restrictions. Obviously, with... You've had them already in advanced equipment, specific types of technology. You've seen some in materials as well, in chemical compounds. I think you've talked about the fact that ALD obviously could drive on new materials. So basically, the question is, is there some tail risk that basically that could obstruct some of your ALD growth, basically, if-
Yeah
... if some of these materials would be further restricted?
I can take the second part of the question, but can you maybe just repeat the first part of your question? You are asking about volume versus pricing split.
Yeah, exactly. So 10% to 14% growth rate, right? That's sort of the idea. So what is the, the units, basically, growth rate you're looking at and the sort of pricing within the-
Okay. Okay.
Yeah.
Unfortunately, we don't disclose that. We don't give a volume. We also don't disclose our pricing. Now, on the so-called, I think you mentioned geopolitics or, or let's say, you know, export control. So number one, we comply, you know, maybe, maybe even more than 100%. You know, we are very cautious, so we comply fully to all the export control regulations. And, based on what we know today, including those that have been announced by the Dutch and the Japanese government, we don't see any additional impact to what we have already shared in the past. Now, of course, there are some rumors, some guess that, look, the U.S. government is going to announce new regulations or update their regulations.
Anything that we say will be speculative. We have to wait and see what's gonna come out. Today, you know, I think the restriction is really on, what do you call, equipment, critical equipment. Equipment that can use, can be used for 14-nanometer and below logic, in a 128-layer flash, 17-nanometer or 18-nanometer DRAM. That is the restriction today that we know of. How this changes, again, over the next couple of quarters or years, again, that will be speculative, and we just have to wait and see what happens.
If I can add,
Sure.
If I can add to your question. We really don't see geopolitical issues as far as getting the precursors from that we use. Many of the precursors or the gases or the chemical that we use right now, we, for us, we don't foresee any issue on geopolitical where we procure them from. Most importantly, you know, the precursor that we develop actually synthesizes. So many of the things that are happening are actually we develop those precursors ourselves in our R&D laboratory. So there is absolutely no geopolitical effect from that point of view.
I think we have to probably come to an end. I want to thank everybody once again. Thank you for your time today. I hope again, this has been very useful for you. And of course, it's always a pleasure to be able to see everybody in person. I think we have a short, I don't know what you call it, but it's outside. And hopefully, we can still continue, you know, just to chat a little bit. Thank you very much. Thank you.