Two years ago, when we met here for the last Capital Markets Day, the world has probably been a little bit different. Let our management explain today how we manage this situation, how we look forward for the next few years, what we plan in terms of technology, in terms of financials. Good morning, ladies and gentlemen, either here in the audience, in the webcast, or on the phone. Great that you took the time to follow the presentation today. You have seen probably already the agenda. We will have a first part with our CEO, Stephan Haferl, and our CTO, André Grede. After the break, we will have the business colleagues from IXS, PCT, presenting on progress in their businesses. The last part is for CFO to translate everything that has been told. How does it translate into financials? Before we start, two things.
First of all, I give you some time to read the disclaimer. Please be aware that we make forward-looking statements. The second point is, please switch off your mobile phones or put them on mute. Five speakers, as I said already today. The first speaker to which I hand over now is our CEO, Stephan Haferl. Stephan, please.
Thank you. Thank you, Ueli. Comet is a story of constant evolution. Over 75 years of adapting to and driving some of the biggest technological and, with this, also societal shifts in modern history. Comet's history began in 1948 with a focus on solving challenges in X-ray technology, more specifically medical X-ray, a commitment to innovation and serving society that has shaped us ever since. In the 1960s, we expanded from X-ray vacuum tube expertise into vacuum capacitors, entering high-frequency markets like broadcasting and telecom. The 1970s marked our move into industrial X-ray components, establishing Comet as a key player in non-destructive testing. In the following decades, our technology supported major breakthroughs, from the first CT scan to space missions, while we refocused on advanced industrial applications.
The acquisition of YXLON in the 2000s, today's division IXS, we transformed ourselves from a component supplier into a full X-ray systems provider. Now, in the era of AI and automation, additions like Canadian software company ORS, the development of the Synertia and our CA20 system platforms are enabling smarter, software-driven manufacturing and inspection for next-generation semiconductor and electronics devices. That is really been Comet's story. More than 75 years of adapting to change, staying ahead of and driving every major technological wave. From the first X-ray tubes in the 1940s to today's AI-powered inspection and plasma control systems, we have continuously evolved with the goal to innovate for sustainable progress of our societies, of humanity. Ladies and gentlemen, good morning. Welcome to our Capital Markets Day 2025. We sincerely appreciate that you have taken the time to join us today and to listen in on our journey ahead.
Over the next sessions, Ueli has already introduced them, we would like to share with you why we are confident that we can continue and further build on the successful past I have just introduced. Before we look into the future of Comet, let's take a moment to see where we stand today. Comet's current ability to innovate for sustainable progress is powered by our truly global footprint with strong roots in Switzerland. Our headquarters near Bern remains the heart of the company. It is where our technology leadership began and where our core R&D still takes place. Over the years, Comet has grown into a global player. Today, we operate from 12 locations across three continents, with more than 1,800 employees worldwide. We have manufacturing, application engineering, and/or R&D sites not only in Switzerland, but also in Germany, the U.S., China, Malaysia, and South Korea.
Close to our customers and to the innovation hubs that drive our markets forward. This global setup gives us both reach and resilience. It allows the ability to serve our customers where they are, adapt quickly to the local emerging trends, and ensure a stable, efficient supply chain. In 2024, our net sales reached around CHF 445 million, with 62% of sales generated in Asia, about 20% in North America, and 16% in Europe. That reflects the strength of our position in the world's fastest-growing high-tech markets, particularly in semiconductors and advanced electronics. With a strong global footprint and deep roots in Swiss precision, Comet is well positioned to serve the world's most demanding industries, with a clear focus on the semiconductor ecosystem, from chip design and fabrication to inspection and quality assurance.
Our business is thereby built on two core technologies: high-frequency power on the one side for plasma processes and X-ray technology for non-destructive inspection and metrology. Together, these two technologies are among the enablers of the semiconductor world, making modern electronics faster, smaller, less power-hungry, and more reliable. Let's start with our largest business, plasma control technologies, or short PCT. PCT is number one globally in vacuum capacitors and matchboxes, critical components for precise plasma control in chip manufacturing. Our products enable our customers to achieve stable, repeatable, and highest plasma processing performance, the foundation of every advanced semiconductor process. In 2024, PCT contributed 56% of Comet's total sales and continues to grow alongside the most advanced nodes in the industry. Next is our X-ray business, which is divided into the two divisions: X-ray systems, or in short, IXS, and X-ray modules, in short, IXM.
IXM is the global leader in high-end X-ray sources, providing the precision and reliability needed for inspection in industries such as semiconductors, electronics, automotives, and aerospace. With IXS, our X-ray systems business, we rank among the top three in industrial inspection systems globally, supporting customers from R&D to full-scale production. Together, these two X-ray segments contributed roughly 44% of group sales in 2024. Across all our technologies, we play a vital role in the industries we serve, especially the semiconductor industry. We're enabling our customers to push the limits of what's possible. At the same time, we are ensuring quality and efficiency to pursuit of highest possible yield in shortest time in every step of the process. Now, technology leadership is obviously very important, but it is also so that value creation is very important.
By enabling, among other things, the future of computing and data storage, we're not only advancing technology, we're building a resilient, growing business. The numbers tell the story. Over the past 25 years, our sales and EBITDA have risen to new levels through every cycle. As we continue our journey and transformation, our focus will turn ever more on semiconductors, in line with our aspirations of technological leadership and in pursuit of higher profitability. As such, very clear, semiconductors are at the core of our strategy and a tremendous growth opportunity. How do we plan to continue the long-term success story of Comet? As we just outlined, our technologies are integral to some of the most advanced industries in the world. The future is being shaped by semiconductors. We are not simply following the industry, but actively driving it forward.
This is why our focus on this sector is increasing, as we anticipate exceptional long-term growth opportunities. Where do we play in the semiconductor ecosystem? Comet serves multiple parts of the semiconductor ecosystem. At the top, you see the major players in wafer equipment manufacturing, such as Lam Research, Applied Materials, or Tokyo Electron. They all depend on highest precision plasma control equipment to enable the manufacturing processes that build the chips that power modern life. This is where our PCT business actually excels. We are one of the preferred partners for RF power delivery. Technology for these equipment manufacturers ensures stable, repeatable plasma processes that are essential for every wafer produced. As chips move through the production chain into packaging, testing, and quality control, our X-ray technologies come into play.
Our X-ray modules from IXM and systems from IXS are used by leading foundries, IDMs, and OSATs to inspect complex packages and ensure defect-free processes and products. We work closely with all of these companies, helping them secure quality and reliability down to nanoscale precision. Whether it's enabling advanced plasma processes or ensuring the highest inspection standards, Comet is right at the heart of semiconductor manufacturing. To summarize, Comet is deeply, deeply embedded in the semiconductor value chain, from front-end wafer processing to back-end inspection and testing. Our position gives us a front-row view of the industry and a very clear sense of where the next growth wave is emerging. That is very crucial because the semiconductor market isn't just growing, it's transforming. AI, electrification, connectivity, and automation are driving unprecedented demand for advanced chips and the inspection solutions that make them reliable.
Let's take a closer look at the manufacturing flow and where Comet delivers its progress-enabling solutions. From left to right, every chip obviously starts with a design and raw materials through a highly repetitive, cyclical series of highly complex and repeated steps, such as lithography, deposition, etching, polishing, ion implantation. Layers of functional materials are added in a structured manner, patterned, and refined to form billions of tiny transistors. Now, Comet plays a critical role at multiple stages in the front-end wafer fabrication, through the hundreds to in excess of 1,000 steps in the cyclical front-end manufacturing process. To the right of the slide and at the back end, assembly and testing process prior to holding a packaged and functioning chip in your hands. At the heart of these front-end processes is plasma.
Our RF power delivery system, matching networks and vacuum capacitors, creates and controls the plasma with extreme precision. This ensures the accuracy, the repeatability, and yield that advanced chip production demands. In other words, without stable plasma control, you simply can't build today's nor the next generation of chips. In the back end of the manufacturing process, where chips are diced, bonded, and packaged, our X-ray technologies take over. Our high-end X-ray and CT inspection systems are used to analyze connections, bumps, and detect defects in advanced semiconductor packages, a step that's becoming more critical as chip architectures grow increasingly complex. With our AI software-driven inspection solutions, we help customers guarantee reliability before chips even reach the market. Across both front-end and back-end manufacturing, Comet is essential, enabling precision, quality, and efficiency at nearly every stage of the process.
That's what makes our technology not just important, but indispensable to the semiconductor industry's progress. We've seen how Comet technologies are essential across multiple steps of chip manufacturing. Now, let's step back and look at the bigger picture, the market dynamics shaping this industry and where the next opportunities for growth lie. The semiconductor market is one of the most dynamic and strategically important industries in the world, the very bedrock of the world's electrification and our society's digitalization. It's driven by constant innovation, new applications, and structural demand that extends far beyond consumer electronics. We're seeing growing complexity in chip design, architecture, and manufacturing processes. Each new generation of chips pushes the limits of engineering physics, requiring greater precision and more sophisticated equipment. At the same time, miniaturization continues.
We're now moving from traditional 2D and 2.5D chips to 3D architectures, stacking multiple layers to boost performance and also power efficiency. Very important, increasingly important. Finally, the industry is evolving from individual devices to fully integrated solutions, combining compute, memory, and sensing into compact, powerful systems. Each of these transitions is accelerating demand for advanced plasma control and inspection technologies, which are the core focus areas for us. What is fueling this furious drive? The mega trends creating this momentum are very clearly digitalization, electrification, and increasingly AI, which together are shaping the modern world. The first wave, driven by Industry 4.0, cloud computing, and the Internet of Things, connected machines, factories, and data streams. The second wave, which we're in today, is powered by AI, autonomous driving, and immersive technologies like augmented or virtual reality. These applications demand ever more powerful and energy-efficient chips.
Ahead of us lies the third wave with 6G connectivity, quantum computing, and Industry 5.0, where humans and machines will work together in ever, even more, ever-intelligent automated systems. Each wave drives near-exponential growth in chip demand and, with it, the need for precision manufacturing and increasingly inspection and metrology. With this growth in demand, a market size capacity expansion goes hand in hand. Now, between 2024 and 2026, more than 70 new semiconductor fabs began or are expected to begin construction worldwide. Asia continues to dominate, but we're also seeing renewed investments in North America and Europe as countries seek to strengthen their local supply chains. For Comet, this global wave of fab expansion means the long-term demand for our RF Power systems and X-ray inspection technologies.
In other words, as the semiconductor industry scales up for the next era of growth, Comet is already inside the technologies and processes that make that progress possible. With over 70 new fabs being built around the world and more to come until 2030, it is clear that semiconductor capacity will be expanding rapidly and so does demand for our products. What is just as important as the scale is what kind of chips those fabs are making and what technologies are driving demand. Because as the applications become more advanced, the chips behind them require ever better products and partially entirely new technologies. How is the market and application landscape we are focusing on shaped? As we look at the semiconductor landscape towards 2030, we're talking about a market of roughly $1 trillion. What really matters isn't just the size, it's how that value is distributed.
Most chips by volume will continue to serve consumer products, consumer markets such as smartphones or mobile devices, IoT devices, or automotive systems. These volumes are critical for sustaining fab utilization and driving scale efficiencies. Yet the real value creation is shifting. Nearly half of all revenue is expected to come from AI and high-performance computing applications. These segments demand the most advanced technologies and highest utilization of leading-edge fabs. Now, the AI chip market is expected to grow approximately 4x-5x from 2024 to 2023, driven by large-scale infrastructure and also increasingly edge applications. While there are only estimates available of what share of total industry growth AI chips will represent, the rapid expansions make them pivotal, a pivotal driver of value creation. The large promises and opportunities of AI are driving semiconductor demand and reshaping the industry for years to come.
AI applications and requirements on the associated semiconductor devices are the most significant driver of front-end and back-end innovation. From generative AI models and data centers to edge AI and smart devices, the performance demands are skyrocketing and requiring faster, smaller, and very, very important, more power-efficient chips. That demand translates directly into needs for new manufacturing technologies. We are enabling the next generation of chipmaking with our leading-edge plasma control. Our new Synertia plasma control platform is designed to meet the precision and performance requirements of AI chip production. In short, AI is pushing the limits of what's possible today, and Comet's technologies are helping to make it possible. What needs to be achieved? Now, when we talk about pushing the limits and innovation in semiconductors, it really comes down to one key objective: efficiency.
The industry is working to deliver more compute power, more data store per square millimeter of silicon, per watt of energy consumed, per data package transferred and stored. That efficiency comes from two technology trajectories. First, continued and efficient physical scaling of individual devices makes it smaller, shrinking transistors to ever smaller sizes and moving, for instance, to gate-all-around architectures. Chip makers can pack more logic on every wafer, thereby increasing density and performance. This transition, which begins with AI and high-performance computing applications, will later cascade into mobile and consumer markets. In other words, as I explained before, today's value chips will become volume chips over time. Second, from so-called heterogeneous integration. As physical scaling reaches its physical and also commercial limits, chip makers are combining and stacking multiple chips into so-called advanced packages.
This boosts performance without relying solely on transistor shrinkage, as well as it boosts power efficiency. That is where Comet's X-ray inspection solutions are becoming essential, ensuring reliability and precision in these complex 3D architectures and packages. Whether it is controlling plasma in the front-end processes or inspecting stacked dies in the back-end, Comet's technologies are vital to enabling this new level of efficiency. Now, let me elaborate a little further. Let's first look at transistor scaling. Despite increasing headwinds to sustain Moore's scaling law, the pace of advancement from process node to the next remains at present unbroken, requiring accelerating speeds in technology development and also deployment while still yielding decisive performance improvements.
As an example, TSMC's plans for moving from the N2 to the A14 technology node in the next few years, with each new node improving performance decisively while also, and that is very important, reducing power consumption. For Comet, this rapid pace means understanding the industry, the industry's roadmaps, and speed in bringing solutions to the market is of the essence. We're focusing our R&D and investment exactly where our customers are pushing the limits. This means addressing rising process complexity, enhancing control, and staying close to all the leading equipment makers. Now, the reward for this is clear: growth opportunities, growth. As the semiconductor industry moves through these technology transitions, every new node, every new architecture, and every new application creates additional demand for advanced manufacturing tools, the very tools that incorporate our plasma control equipment.
The wafer fab equipment spend is expected to grow from just over $100 billion today in 2024, 2025 to nearly $170 bililion by 2030, driven primarily by foundry and logic applications. That said, NAND wafer fab equipment is also set to expand in the coming years with a projected CAGR of 15% between 2025 and 2030, growing even faster than the overall WFE market. AI continues to be the biggest accelerator driving the adoption of new process technologies like gate-all-around architectures and later CFET. At the same time, the industry is being shaped by structural forces from AI demand and advanced node investments to geopolitical developments and the global race for technological sovereignty. These are long-term growth drivers that create sustained opportunities for Comet. Now, what is driving this WFE growth? Demand for new technologies and new capacities.
Let's briefly turn to one of the big forces driving demand: the explosive growth in memory fueled by AI, data centers, and increasingly also edge applications. What you see here is the outlook for both DRAM and NAND through 2030. The trend is clear. Massive capacity expansion is anticipated, led for now by data centers. Starting on the left with DRAM, it is projected that the world will require almost three times more memory capacity by 2030, 86 exabytes in 2030 compared to 32 EB in 2024. While mobile and PCs continue to grow modestly for now, the real engine is, as mentioned before, data centers, which more than tripled over that period. That is because AI workloads are very memory intensive. On the right, NAND shows a similar story, but at an even bigger scale.
The market is expected to grow from 868 EB to more than 2.2 ZB by 2030. I do not even know how that number looks like, to be honest, but I promise you it is big. Again, data centers are the dominant driver, growing at 27% annually and accounting for almost half of all demand by the end of the decade. This is a structural shift. Memory is no longer just a PC or smartphone component. It has become core infrastructure for AI and hyperscale computing. Every new model, every new cloud service, every new GPU cluster increases the need for faster, more efficient, and more reliable memory. This circles back to the previous slide, with memory demand also contributing substantially to WFE growth in the next few years.
To sum up, as AI accelerates the pace of semiconductor growth and innovation, Comet is right there at the inflection points, enabling faster, more efficient, and more reliable chip production through its leading product portfolio of RF power equipment for plasma control. We've just looked at how AI is accelerating front-end innovation, but the story doesn't stop here. AI is not only transforming how chips are made, it is also reshaping how they are assembled, integrated, packaged, and tested. In fact, AI is now the most significant driver for both front-end and back-end innovation. AI applications are pushing semiconductor technology to its limits. They demand more computing power, as mentioned, higher bandwidth, and lower latency while consuming less power.
That means innovation must happen across the entire manufacturing process, from the front-end wafer fabrication, where performance is sort of defined, to the back-end packaging and inspection, where reliability is ensured, but increasingly also power efficiency and thermal management of such chips. That is exactly where we play a key role by enabling both ends of the process, with our leading-edge plasma on the front side and on the packaging side with X-ray technology. What is an advanced package, or what is advanced packaging? Now, as traditional spatial scaling will eventually come to an end, packaging will increasingly lead the way. By co-packaging individual semiconductor devices into one advanced package, increases in speed and cuts in power losses can be achieved by, for instance, drastically reducing transmission lines.
Thus, the industry is turning to heterogeneous integration, of which advanced packaging is a critical enabler to unlock another powerful lever for performance, power efficiency, and system innovation. In 2.5D and 3D packaging, multiple chips, CPUs, GPUs, memory, and specialized accelerators are very closely positioned side by side or stacked and interconnected to act as one high-performance system. This approach not only boosts compute performance, but also improves power efficiency and reliability, which is essential for AI and HPC applications. Today, advanced packaging is one of the most important innovation areas in semiconductors, and Comet's X-ray non-destructive inspection solutions are becoming indispensable in quality control, ensuring that these complex 3D architectures, 3D structures are perfectly bonded, aligned, and free from defects. How is this market evolving?
Now, the growth trajectory for advanced packaging is still sort of in its infancy, but advanced packaging is currently one of the fastest growing and most strategically critical areas of the semiconductor value chain. As mentioned, as device complexity increases, the ability to integrate multiple chips efficiently, as close as possible to each other, either side by side or on top of each other, becomes essential. The challenges of placing these chips closely side by side and/or on top of each other thereby drive demand for precise inspection and metrology solutions. The advanced packaging market overall is expected to grow at a solid 9.5% CAGR through 2030. The segment most relevant for Comet's back-end metrology and inspection is growing even faster at 11.3% per year, reaching about $850 million by the end of the decade.
It is also important to note that this market is highly concentrated and geographically focused. Roughly 90% of revenues are generated thereby in Asia, and the top six suppliers control more than 80% of the market. This is why we continue to invest in people, especially in sales and service roles in the region, to support customers who rely on us for testing and qualification. In this industry, speed matters. Support needs to be on site within hours, not days or weeks. For us, this represents a strategic opportunity as our customers push the limits of chip packaging and stacking and interconnect density. They rely increasingly on advanced inspection capabilities to ensure yield and performance. Now, with the multiple growth opportunities in our divisions, how does our overall addressed market look like?
Now, given my outline of the past minutes, unsurprisingly, we're in the right place at the right time, especially with our focus on the semiconductor value chain. Let's take a closer look at the size of the opportunity in front of us. What you see here is sort of a key takeaway for our investment case. Comet's addressable market is growing significantly faster than the overall market. On the left, you see the total addressable market growing from approximately $4.2 billion today to $7.1 billion by 2028, an increase of about 69%. The more important chart is on the right. Our serviceable addressable market, that means the part that we can realistically win, given our current product portfolio and product platforms, that market more than doubles over the same period from $1.7 billion to $3.5 billion , representing 105% growth.
The reason is simple. We are positioned in the fastest growing segments of the market, and we continue to extend our reach through forward integration and new revenue streams, particularly in semiconductor metrology and process control. What does it take to capture these opportunities? Now, let me elaborate in what follows on some of the most important building blocks and priorities, building blocks which enable us, building blocks that are required for us to capture these growth opportunities through focused and disciplined execution. First and foremost, our journey forward is defined by our further increasing focus on semiconductors. In today's rapidly evolving tech landscape, the boundaries between front-end and back-end process are blurring, fueled by the increasing demand for higher performance, higher efficiency, and heterogeneous integration. This convergence has opened new opportunities for innovation and growth. One key aspect of this convergence is the emergence of advanced packaging.
This trend is reshaping the semiconductor industry, and we see it as a significant opportunity for us to leverage our expertise and capabilities, especially on the X-ray side of our business. We are uniquely positioned to capture value from this convergence through our deep knowledge and experience in plasma control, X-ray, and process technologies. Now, these core competencies have allowed us to stay ahead of the curve and develop cutting-edge solutions that meet the evolving needs of the market. Furthermore, the transformation we are undergoing is being accelerated by geopolitics, which are reshaping the global semiconductor landscape. As the industry adapts to these changes, we see new avenues for growth and collaboration that will further propel our success. How do we move from where we are today to a scalable, high-margin growth company? Now, our strategic approach is straightforward: drive faster growth, operate more efficiently, and strengthen our culture.
These are our strategic pillars. We are dedicated to leveraging our expertise and market-leading positions in plasma control and X-ray technologies for the semiconductor market to drive innovation and deliver exceptional value to our customers. As we look to the future, we see a world of many opportunities to expand our reach and impact. By focusing on accelerating growth, we aim to capitalize on emerging trends and market dynamics to drive revenue growth and profitability. Through strategic investments in current and new technologies, R&D, our footprint, partnerships, and market expansion, we are positioning ourselves as a market and technology leader poised for success in the years to come. In parallel, we are committed to enhancing efficiency across our operations, optimizing our processes, and leveraging digital technologies to streamline our internal workflows and drive operational excellence.
By embracing a culture of continuous improvement and innovation, we are empowering our teams to work smarter, faster, more collaboratively, driving efficiency gains and cost savings that will fuel our growth and profitability. Perhaps most importantly, we are focused on strengthening our culture, a culture of excellence, integrity, and collaboration that defines who we are and how we operate, how we do things. By fostering a culture of trust, respect, and empowerment, we are building a strong foundation for success, attracting top talent, and driving employee engagement and satisfaction. How does this transformation then translate into concrete actions for each division? Each division has a clear set of priorities and initiatives, but they are all built on the same company-wide foundation: a unified and common strategy and market focus, a common operating model, shared infrastructures, footprints, functions, process technologies, and vision.
This ensures that while each business drives its own product roadmap, we move forward as one organization, one Comet, fully aligned in how we grow, how we execute, and how we create value. Our foundation is clear: a people-driven culture built on curiosity, integrity, passion, and excellence. Our core focus is semiconductors, driving quality, innovation, and technology leadership. In this industry, success requires true customer obsession, knowing their roadmaps, adapting ours, and delivering solutions that essentially and eventually win. To achieve this, we cut complexity, move faster, and stay focused on what really matters across all divisions. These priorities and several more included in the slide unite us.
To shed light on PCT strategy, and Joeri will talk a little bit or a lot more about that later, the division aims to transition from a component supplier to a full RF power system provider, offering comprehensive solutions for the semiconductor and electronics industry. By focusing on plasma process control for advanced semiconductors, PCT is further establishing itself as a key player in this technology. Scaling Asia-based manufacturing and R&D to serve global OEMs faster and more cost-effectively enhances the division's and obviously Comet's competitiveness. In the IXS division, and Isabella will talk quite a lot about that later, we further specialize in high-end X-ray and CT inspection systems for advanced packaging and semiconductors. These technologies drive efficiency and reliability from early prototyping to manufacturing, offering a range of solutions and integration services close to the customers and markets.
We cater to clients' needs from the lab to the fab with precision and expertise. In IXM, investments in dedicated semiconductor X-ray sources, software analytics, and AI empower customers to make informed decisions and excel in a changing industry landscape. Strategic restructuring in traditional markets enhances profitability and agility, positioning us for success. Our focus on semiconductor technologies, graphic expansion in Asia, and digital transformation drive operational excellence and growth opportunities. Common to all divisions is that high-performing product portfolios addressing the key points industry are key to our growth. Now, how do we go about this? Now, the growth elements of our One Comet strategy are leverage, expand, and broaden. Leverage is anchored in what we do best: high-frequency power, X-ray system design, and thus the continued advancement of our Synertia, CA20, and Dragonfly platforms.
These core competencies have been instrumental in shaping our success and will continue to be the foundation upon which we build our future growth, where we spend also the majority of our R&D investment leveraging our strong core competencies. Beyond the focus on our core competencies, we are actively expanding into complementary domains very close to the core, in line with rapidly evolving demands of our focus industry, exploring adjacent and complementary technologies and services that will allow us to offer superior process-centric solutions to our customers in the semiconductor ecosystem. Furthermore, we are strategically entering targeted markets to strengthen our position and broaden our relevance in the semiconductor value chain. By identifying and capitalizing on new growth opportunities, we are positioning ourselves for long-term success and sustainability in a rapidly evolving industry landscape.
Our expansion into adjacent technologies and services will not only enhance our offering but also solidify our presence and broaden our relevance as a key player in the semiconductor ecosystem. Having excellent products is not enough. Permanent, physical, and scalable presence in the market regions of our customers is of equal importance. Comet's gateway to Asia, where our customers sit and its customer, so our customer base, is for Comet Penang. Penang was once primarily an electronics assembly location and has evolved since the 1970s into a fully integrated semiconductor ecosystem, from front-end design to advanced packaging, testing, and equipment manufacturing. Today, more than 50 global semiconductor leaders operate here. In short, Malaysia is becoming the new Silicon Valley, the Silicon Valley of the East, with Comet positioned at the heart of it. In Penang, we're investing where it actually matters most: in scalable, cost-competitive, and future-ready production capacity.
Penang is a strategic milestone for us, our flagship expansion project and a key pillar of Comet's global footprint and strategy. Construction of our own building is progressing fully on schedule and within budget. Once complete, it will double our current production capacity, giving us the flexibility and scale we need to support long-term growth. The new site in Penang is more than just an expansion. It is a strategic hub that strengthens our presence in Asia and enhances our ability to serve customers regionally. We are executing this in well-defined steps. After securing land in 2023, we're now moving to our full operations in the second half of 2026, so actually next year. During this transition, we'll gradually shift appropriate production activities, shared services, and back office functions to the new facility. These moves ensure business continuity while scaling efficiently.
The building adds over 15,000 sq m of production space and follows a very clear roadmap to becoming our Asian hub the latest by 2027. This expansion brings multiple strategic benefits: long-term efficiency and scalability, a balanced and resilient global network, cost and scale advantages, and importantly, proximity to our regional customers. In short, Penang is how we are also securing future growth, besides investing into future product and service offerings. While our investment in Penang strengthens our operational backbone and global capacity, innovation, and people, the third strategic pillar remains also at the heart of our long-term success. At Comet, we believe that sustaining growth means more than scaling production and product platforms. It means developing technology and talent together. Our approach is simple but powerful. Technology improves our daily lives, and education improves technology. That is the cycle that keeps innovation moving forward.
We're deeply engaged in programs that strengthen that cycle, from our apprenticeships and training programs to our collaborations with leading universities worldwide. We're also proud to be part of the European Chips Diversity Alliance, co-funded by the European Union, which promotes inclusion and skills development across the semiconductor industry. Our efforts are being recognized, for example, with the Semi-Europe 20 Under 30, apologies, 20 Under 30 award in 2024 and again also in 2025, celebrating the next generation of innovators at and within Comet. All these initiatives reinforce one message: our commitment to innovation today ensures we can sustain growth and leadership tomorrow. Just as we invest in people and innovation to build long-term resilience, we also take responsibility for how we grow. With that, we make sure that progress today does not come at the expense of tomorrow.
For us, sustainability is not a side project but part of how we innovate. Under the same guiding principle, innovating today, sustaining tomorrow, we're advancing both technology and environmental performance across our operations, silently, calmly. You can see this in our progress on carbon disclosure and transparency. In 2024, Comet achieved a CDP score improvement from C to B, moving ahead of the industry, actually regional and global averages. This reflects stronger data quality, clearer targets, and tangible action plans for emission reduction. One of the key enablers is our expansion of renewable energy use across our sites, including the new facility in Penang. We are integrating solar electricity generation and green electricity sourcing, and we're on track to exceed 80% clean electricity by the end of 2025. These initiatives are not only about responsibility; they're about resilience.
By reducing our environmental footprint, we also reduce cost exposure, increase energy independence, and strengthen Comet's long-term competitiveness. This brings me to the end of my speech and the end of the introduction of the Comet Group. To emphasize the importance of technology, I'll hand over at this point to André Grede, Comet's CTO. André, please join me on the stage.
Thank you, Stephan.
There you go.
Merci. Good morning, everyone. Ladies and gentlemen, let me take the chance to welcome you as well here to our Capital Market Days 2025. I'm going to speak a little bit in the next 30 minutes-35 minutes about how we are positioned or how we are well positioned in the technology game in semiconductors. When the VP Investor Relationship asks you to speak about technology to the investors, it's always a little bit difficult, right?
I will try to explain it as simple as possible. On the other side, I want to be correct, but most important for us is, of course, we do not want to give away our IP. Please understand that we cannot share everything, especially not the newest things we are working on, but we want to give you some confidence in what we are doing and that we are doing the right thing. As I said, we are well positioned in the game. Let's start again with the ecosystem that Stephan already described, where you on top have more and more companies developing semiconductors, designing semiconductors, processors, certain application systems. On the other hand, you have fewer and fewer companies who are able to afford the environment that is needed to produce these semiconductors.
Actually, there are only some big ones for logic left, some special ones for things like high bandwidth memory and so on. With advancing technologies and challenges and performance in this industry, there are few and few companies who can really do the last technology node. They, of course, rely on their suppliers. If I am a little bit overconfident, I would say technology is actually not driven by, or at least not only driven by these companies like Samsung or Intel or TSMC. I mean, they build now two-nanometer chips. Who enables them to do this? All the companies who make the equipment, who develop the equipment, because you can think of the coolest, best, most performant structure if nobody gives you the materials to do it.
If nobody gives you a process and the tools to put the material into hardware that is doing what you once sought out or simulated, it will never become reality. That is why the supplier of systems like us or subsystems, like for example, PCT, really have such an important role. Although a lot of things for you may be like sometimes new, surprising, you see a new kind of memory coming, new kind of AI chips coming, new technologies you hear about, all these things are rarely, nearly never a revolution. There are roadmaps over many, many years. All these things are evolutions, step by step, learning based on a lot of experience. I mean, we have the same examples in our company. When we started with Synertia, we knew it would be a long way, a long challenge.
We also knew that it would not be an easy way into this because the generators were new for us. You will later hear from Joeri how he with his team now finally succeeded and really placed this new technology in the market. We are finally there. Honestly spoken, if you look back, how many companies were there five, ten years ago supplying generators into the semiconductor market? I really mean the top tech nodes at the top three-tier OEMs in this market. It is still the same number of companies plus Comet. Nobody else has really succeeded to get qualified at the top-tier customers on future high-volume tools. This is really an achievement, as I said, which does not come overnight. It depends on a lot of work and, of course, on a lot of experience.
That is the same on the CA20 side, by the way. We are not just talking about it anymore. We have not just people testing it. You will see that customers start to adapt this new technology to their processes. Isabella will speak about this. The dependency in this very complex, demanding environment is bigger than ever before, all the way up from the companies designing semiconductors all the way down to companies like us who supply the equipment needed to make the next or achieve the next technology node. I want to, as a next step, look a little bit at what is really driving these new next technology nodes. Of course, for me, like end customer or for me financial, from a business perspective, it is topics like AI, life science, autonomous driving, all these things.
If you look at these things from a more, like, a little bit more with the technology glasses, you will see it is mainly about what Stephan said, performance. Performance can be more computation power, but performance can also be, I want to load my or charge my electrical vehicle faster. Scaling up currents, scaling up voltages. At the end, it is not just about functionality per square meter or square millimeter anymore. It is now per volume. It is always about functionality, the cost. Because even if you think that a complex AI chip today is quite expensive, if you see it with relation to the computation power 20 years back to a simple CPU, which cost at this time, I don't know, $20, $30 if you bought it, you will see that this relationship is always improving as well.
You get more power per cost. The main drivers from a technology point of view are actually three, if you simplify it, of course, a little bit. The first one is scaling. Scaling is not just making things smaller. You know, scaling has especially developed in semi to the third dimension. We are scaling not planar structures anymore. We speak about scaling volumes where we make devices or systems. It is about advanced architectures. For example, you have not a single processor anymore. You have processors for certain domains, for certain tasks to increase, to scale performance. You have advanced architectures now, like the 3D scaling, heterogeneous scaling of systems. You also have new, for example, memory-centric architectures. On the other side, you have the technology.
Technology can be actually a lot, but a very famous example, for example, is the development from planar FETs over FinFETs to gate all around FETs. We are now speaking about complementary, about CFET. There are other technologies where you, for example, try to use the magnetic properties of material to store information, MRAM, instead of using the electrical properties. All these things, these drivers must be somehow enabled. Somehow at the end, you must do it. As I said, you can have an idea for the most brilliant concept. The question is how you do it at the end. It is about materials. For example, you know, there is a lot of scientific activities now around 2D materials like graphene. It is about materials like Moly that was just introduced for achieving better connections, especially for small structures.
On the right side, the other two points are manufacturing equipment and processes. At the end of the day, you need to take this idea, you need to take the material, and you need to make it a chip. Then it is about testing and inspection because what you can see and measure, you can't understand, and what you can't understand, you can't fix. This is actually what is enabling all these drivers in the technology chain. You see our playing field, Comet's playing field, is on the process equipment side, where we on one side help to make equipment so you can produce semiconductors. It is now with our X-ray inspection tools in semi, also about the inspection part, seeing what happens, fixing it, making it better. Let's start at the famous front end, which is, of course, not new to you.
I try to put the chain here again. As you know, the production chain, of course, is highly simplified. To summarize again what we are doing in the front end with Joeri and his team at Plasma Control Technologies, we are driving the process with energy, with power for plasma. If you consider plasma the tool to do the work, to deposit something, to modify material, remove it, you need an energy source to drive it. It is not just a simple energy source like a 220-volt wall plug. You need a special kind of energy. You need a special way to control it. That is what we are doing there with our whole chain now from the Synertia RF source generator.
Now this year released, we are the Synertia match using the same control system and a certain connectivity to the Synertia generator and the Vaccaps , which are still a very important element. If you look at the two most important process steps, which come in many variances, we speak about, with respect to plasma, we speak about deposition and we speak about etching. Deposition, we speak about applications like PECVD, which stands for Plasma Enhanced Chemical Vapor Deposition, ALD, Atomic Layer Deposition. There are also special processes, sputter processes, how you depth metals, all these things. They are run by plasma using RF as an energy source. There is not only RF, there is DC as well. DC is used for other things in plasma. It's a little bit more rare if we speak about advanced processes.
There are reasons why you cannot apply it, especially if you speak about processes for non-conducting materials like on wafers. There is an RF as an energy source I just spoke about, and RF can be anything from low frequencies, 400 kHz, all the way to several gigahertz as you use it in remote plasma sources. There are more and more other energy or power sources or ways to create energy and shape it, which I summarized under non-sinusoidal signals. This is things like bipolar pulsing, but also like tailored waveform. If you look on the right side, you see applications like etch, very famous dielectric etch, conductor batch, but also things like ALE, atomic layer etch.
You see the most important process steps, which you have probably heard about, like high aspect ratio structures where you want to drill very small holes through a very thick substrate, for example. There are things like Via-, spacer- , gate- etch, all these things you need to make a device. Last but not least, both sides come with more or less the same challenges for the generator. Repeatability is the Holy Grail in this industry. Our customers make hundreds of thousands of chambers. Every generator you deliver, which has such an important, or every match as well, has such an important influence on how the energy goes into the chamber and does the work. The repeatability starts really with the RF subsystem. This is a very important requirement, which is not easy to achieve.
Then it's about uniformity, about fast transience, especially in these new processes where things like gas mixture, but also the RF power you apply or other things you're using are changing very quickly. Efficiency is a big topic. I mean, if you see how much energy just a UV system is using to do an ANFF, and you have several of them, and then you have these RF generators, two and a half kilowatt, five kilowatt, and even more, efficiency becomes a very important topic. Of course, you add complexity with multi-frequency systems, with waveform shaping, with non-sinusoidal signals. These are all these things which you need to control the plasma, not just more precisely, but certain species in the plasma more independently. You can really make these new smaller structures.
I was asked very often in the last year about how this matching works or where it is going. There is this, will vacuum capacitors be replaced by solid-state matching? I can already tell you, no, they will not disappear. I want to explain a little bit why, because I read so much about it. As I said, we are asked very often, and matching RF power is one of the most important things in this supply chain of energy for the plasma. For a simple reason, you cannot directly create, for example, the voltage needed at the plasma with the transistors available today. You need to do something like what is called matching. I have chosen it as an example, as I said, because it is so important and we were asked too often.
I want to start with what is driving this part of the power delivery chain. Yes, there are more and more requirements which are about speed and transience. There is already a difference in speed, how fast you can match to a new plasma condition versus handling the transience of the plasma. What you can see here on the right side, without going to any detail, you see on the left side, there is coming some kind of shaped signal. We call this a multilevel port signal. You see the color in the red there. You see that there is a high reflected power. Then you can see how Synertia suddenly starts the match and the generator of the system matching in just one pulse train. All the reflected power is done.
What you see on the right side, these short spikes, this is handling the transience. The transience are still there because when the plasma, when you change the energy supplied or the power supplied to the plasma, is changing abruptly from one low point to another. You see how the system is adapting to it in sub-microseconds. This is one of the big challenges, no questions. There are more. There were challenges before. There were achievements before. There is performance we deliver today, which cannot be replaced by just being good in speed. It is all about the operation area. All these plasma tools are not just doing one process step. There are many process steps. There are different processes. There are different working conditions, different gas mixtures, different chemistries they are using, different pressure, different power they are supplying, and so on.
Finally, you end up with a lot of what we call load conditions, which you must be able to handle with your RF system. The thing is also, our customers develop these basic processes. They ship tools to their customers like TSMC, but the story does not end there. These systems are in the field for 5, 10, 15 years and longer. These end customers continuously redevelop or improve these processes with their growing knowledge over the years. Very quickly, it can happen that the load point, which is at a certain position in the shaft there, really moves dramatically because TSMC has a new idea for a new process. Our system, if it is a good system, should still work. You have to think about these things as well. The working area, the operation area, is a big challenge.
As I said before, everything is about efficiency, accuracy, repeatability. You must calibrate things. They must be all the same. This is what we are doing in our factory. This is what we have a special lab for that we have set up in the last years. Calibration is only removing the offset. The technology you're choosing will have a lot of influence on how your system is, for example, fluctuating or how it is drifting. All these things are connected. If something is not efficient, it means it burns energy. It becomes warm. If it gets warm, of course, dimension change, and this is where drift starts. All these things have to be considered as well. Last but not least, all these things we are doing in the front end, there are a lot of varieties.
You need to be scalable. Customers want scalability quite quickly. They are suddenly for more voltage or more current that the match must deliver, or they want a different frequency. Your technology must deliver the same performance, not at certain 0.56 MHz, but maybe at 60 MHz. We have a big library of network types. We are not just doing a match, as you see on these pictures, which delivers into a plasma. We have matches which split the power and feed several chambers. We have matches who split and can adjust the current so you can drive independent coils in an inductive coupled plasma. We have topologies like T matches, Pi matches, L matches. It is a big variety. It is a big product portfolio you need to be able to do with a technology.
If you summarize it, and maybe also to go a little bit to the core of this, yes. There are two extremes if we speak about RF matches. On the left side, you see the classical match based on vacuum capacitors, which is fantastic in efficiency and is extremely repeatable because for a simple fact, there is no better dielectric on this planet physically than vacuum. Everything else, if you do it differently, you will introduce losses, and very often losses come with reduced repeatability. It is not so good in speed. Just discovered a failure we did not find during our dry run. It is fantastic in the operation area, of course, but what it is not so good in is speed. Speed is the limitation for a vacuum capacitor match because it is a mechanical system.
You have the extreme right side where you have the same challenges, but you can only fulfill one very good. A classical semiconductor or solid-state match is extremely fast. You cannot do it faster. You can do it more or less infinite fast. Finally, it does not depend on the match anymore, but on the reaction of your plasma or the network. If it comes to repeatability, if it comes to efficiency, it is clearly the worst one because it has the biggest losses. It is also clearly the worst one if it is about operational area. If I say I want to go from 3 kV to 10 kV, it is difficult to do it, and it is then also much more costly. There are not just these two. There are a lot of things, a lot of flavors inside that we also look in.
For example, now with the new Synertia match, which came out this year, which is working really as an interconnected system with a generator, we can do new things. We can really leverage the repeatability, efficiency, the operational area which we have with our Vacc aps. The way we synchronize it, we can get tuning not for all applications, but some applications. We can get it as fast as a solid-state match. Again, it does not work for everything, but we can do it for a lot of processes. That is why we collaborate with our OEMs, especially on these topics, because that is then also the most cost-efficient solution. There are other concepts like what we call hybrid matching. I cannot tell you because there we have some really new ideas we are still working on.
There's a lot of IP involved, how and what we are doing it. There we try to bring these two best worlds of solid-state matching and classical Vaccap matching together at a reasonable price. Last but not least, there was place for one more box, and there is something what I hear about very often as well, and it seems it confuses people like matchless systems or direct drive. Honestly spoken, a lot of this what you hear is very often also some kind of advertisement. I can tell you the direct drive you hear about uses a lot of vacuum capacitors. Good for us. There are other technologies which I consider matchless. Matchless is all about these non-sinusoidal signals. That's things like bipolar pulsing. That's things like tailored waveform. What is important to understand, these systems are an add-on. They come together with RF.
They are used, for example, tailored waveform to better control the ion energy in an etch process. They are an add-on to the toolbox. They are not replacing 60 MHz, 40 MHz , 27 MHz , 30 MHz RF. This is important to understand. Of course, we know these systems. We are looking at these systems. We are working on things. Last but not least, maybe a last slide about the technology we are using in front end. You heard Stephan already speaking about our digital platform's capability. This is really a change we introduced with Synertia. Actually, it took quite a while until it reached the market, was accepted. Now with these advanced nodes, our customers really see the advantage of the capabilities in the data range about the synchronization we can do with equipment that we can do with a new platform.
I will actually leave it to Joeri to speak more about what we have achieved there and how this is now really hitting the market because you could say with all these data-driven analysis, two, three years ago, we were a little bit ahead of the market. Now is a time where it is really needed, and it is actually a unique selling point of the whole Synertia system. Short summary for the front end, where do we play and what do we do? As Stephan said, we are leveraging the technologies we have, which will not go away, which are used for 10, 15, 20 years after they were introduced into the semi-market. Do not forget, this is how the semi-market works. You add new technology. That does not mean that the old one is going away. It is running for quite a long time.
Of course, we are now expanding with Synertia, with our new ways to process data in real time, especially with a real-time control platform. We are processing and expanding also into new processes in the back end, which we will hear a little bit about because they are turning towards plasma as well. Very important for us, this new technology platforms we have developed give us a lot of capabilities. That is why I spoke about the match example. At the end of the day, we need to pick the right one. We need to pick the one which has a sizable market, which solves a high-value problem at our customers and which does it for the right cost. We are not a university. It is not about developing the fastest match, the generator with the highest output power or highest frequency or whatever.
At the end of the day, we must solve the customer problem, and we must do it for a certain cost. Let's jump quickly from the front end to the back end. I want to speak a little bit about the increasing importance of the back end because a lot of things you're seeing, as I said, these are all evolutions, not revolutions. There was a lot of scientific activities, but there were already also a lot of specialized solutions in the advanced packaging market before. You see a little bit the timeline, what was introduced when. For example, the embedded bridges which Intel is using today with their chiplets, this was actually already introduced like 10, 8, 9, 10 years ago.
There is a big change in this market, and that was on one side, on the commercial side, the driver, which is AI, which needs this kind of advanced packages to really deliver the performance. There is also another reason, which is that the scaling, the new technology nodes and the front end, they have become so expensive and so complicated at the same time that this opened the door for these technologies because they are also not cheap, right? Doing a complex setup with chiplets and HBMs and in-output communication into advanced package, this is not a cheap thing. Compared to what you need to spend now in the front end to make advances with respect to MUA, with respect to performance, this is now a viable option. I can tell you it will not go away again.
This is one of the reasons the homogeneous scaling down to smaller structures is not dead. There is a gap that is now filled and completed by advanced packaging. The second thing is that we are really going from devices to systems. Again, AI as an example, where you cannot reach the same integration level, even not with the most advanced PCB technology. You need to bring memory and CPU closer together. They are made in different technologies. You cannot make them on the same wafer. There is nothing else than an advanced package to do it to achieve the performance. AI being on the front end, the high-value game, the driver for the moment, this will clearly also be used for other things. Other things will benefit from this environment.
If we speak about advanced sensor systems, autark systems, maybe with their own energy storage on board, all these things will now be accelerated by this development. Last but not least, after from devices to systems, there's a third point, which is you can simply now combine also the perfect technology node with a reasonable or a balanced cost. You don't need, if you make something on a wafer, you need to use the same technology node for everything. That's not very cost-efficient. Maybe I make use a new processor with a 2 nm process, but all the in-output networks around digital or analog, of course, you can use a much older technology node, which is much more cost-efficient. Advanced packaging is also a way to optimize which technology for what cost to use in a device.
On the top chain, just to remember you, I have showed such a classical packaging process of the past. It is, of course, simplified. The message that should come across is so far in the back end, everything was more or less in 2D. Of course, there were already advanced packages which included processors and sensors, but normally they were placed side to side. They were connected via distribution layer. You see the normal steps. You had an overmolding, and then before you either do some bonding to the lead frame so you can connect it to a circuit board, or you build up these boards on the backside, which we call ball grid array, so you can solder it on the PCBA. These were mainly the steps done to do a package.
A big difference, especially with the testing, was you packaged a device. Maybe this was a $0.5 device. Maybe it was a $5 device. At the end, you tested it. If the device was bad, you swept $5. If you had a reasonable yield, and you normally had because the complexity was much lower, that was not such an issue. Now you have started building complex devices, and these complex devices are made of, or systems are made of high-value devices. The memory, extremely expensive, such an HBM, especially if you put eight in a package. All the chiplets, the CPUs, maybe different one, maybe some optic for communication. The value of these packages is extremely high.
Of course, you cannot simply, like I said before, "Oh, $5 device doesn't work, scrap it." You cannot do this with a $30,000 AI system or package, or you don't want to do it. The business case will not fly anymore if you work this way. There is a very simple strategy coming with this, which is called good die or good material. You need to check the devices before you put them together if they are good. When you have the different assembly steps, you want to check in between as well because in your value chains, if something goes wrong, you quickly want to take the chance to either repair it or you want to stop assembling and not put more good material in to finally scrap it because of one bad device in the system.
If you do this in 2.5D or 3D, there is an inspection gap because so far you could test everything from outside or you could look from the top or from the bottom, and you always saw what you had to see. This is not possible anymore. There are only two ways you can do it. You either take this thing if you have an issue and you cut it open and you look for what was going wrong, or, and Isabella will speak about how this works, you take a 3D X-ray or 2.5D X-ray system and look inside. You save a lot of time and find the problem much faster.
You increase not just your output, your quality, but you can also directly feed back the information through your production loop and increase your yield and make sure you do not lose another system or device. To do this, we hear a lot about 3D inspection, about the CA20, about us with inspection system in semi. By the way, it is an important step for us. Do not forget, so far we were delivering components, modules, subsystems to the front end, to the ones who are making the machines to produce semiconductors. In this case, we start the same way. We make the components and the modules, like the X-ray tube and the corresponding high-voltage generator. We make all the handling around it, but now we also make the system. We have advanced from delivering subsystems to delivering systems, at least in the back end.
I have selected, I would say, the four most important things in such a system. The message I want to bring across is all this is in our hand. Let me start with the light source, with the X-ray source, which is done by IXM, by the team of Michael Berger, who is here as well. If you have questions in the break, please take the chance and interview him directly. I want to shine some light on IXM and on this team and how important they are in this inspection chain. You will hear later from Isabella, there are many challenges you have to solve. At the end of the day, you want to see things very precisely with a high resolution, and you want to do it fast.
This all starts with the light source, with the X-ray tube, because it sets, you could say, the limits of what is possible. How much power do you get out, but also how much resolution can you at the end achieve? This is what I want to explain here, because you can play this game at home, right? If you have something, if you have like a big old light bulb and then in front you put an object and you have the shadow on your wall. Then you take a small LED where it seems the light comes out of one point, and you will see the same object, but the surrounding will not be so blurry anymore. The picture will look much sharper, much higher contrast. The reason and the challenge is the same for X-ray.
You want to have X-ray coming out of one single point. Your picture would look like this. That is not the case because we are accelerating an electron beam, and finally this thing hits a target, and it hits it at a certain surface. This defines what we call our focus spot, the spot we can focus the X-ray to. If you do it like this, you will see there is not just light coming out of one point, it is from the edges as well. This happens. This finally defines if I have two objects close to each other, can I still separate them? This is what defines your resolution. This comes with a lot more challenges the team of Michael has to solve. It is not just that you have such a small focus spot.
Because if you have a small focus spot, all the energy goes to one point. So sooner or later, you will melt away the target. The question is, how do you can still get power out of this? Because if you do not have enough power on the target, you do not get enough light. Again, no light, no possibility to see fine structures. You have to keep this focus spot and see, we speak about 400 nm resolutions. You have to keep it extremely stable. Of course, all these things are heating up. Now how do you make sure it is not wobbling around? Because that would mean your projection is wobbling around as well. If it does this slowly, if you can measure it, we can maybe use our AI-driven inspection software to correct for this, but you cannot correct for everything.
This is what needs to be understood here because you're used to your smartphones, which are getting smaller and smaller. To be honest, the whole optics compared to a professional camera system is like, it's nothing, right? You can do so many corrections by AI if something is looking bad. AI is estimating, guessing how it could look like and make the face or whatever looking nicer or complete. This is not possible here. If there's information missing and I use AI to generate this information and put it in, how can I make sure that I have not just made the picture nice, looking nice, instead of seeing the defect that was there? Again, the tube is so important because this is the beginning of the chain. Any resolution and speed you lose here, you cannot recover.
That is why there is such a competitive advantage for us to have this in our own hands and to have this strong collaboration between Michael's team, IXM, and Christian's and Isabella's team on the system side. Last technology I will speak about today, and I will not go into the automation, which is important as well, and the high-voltage generators, which are made by IXM, which now also have lifetime monitoring, which is extremely important for uptime and semi, right? Systems cannot go down. This is pure money running away. We leave this for today, but of course, I want to speak a little bit about what we are doing with respect to AI. I do not want to speak about how we use AI in the company to be more efficient, how to write and test code using AI. We are doing all these things as well.
I really want to speak about AI in products and a little bit about the technology behind our platform, Dragonfly, as used for our inspection system. You see the whole universe, let me call it like this, of AI there. What we are mainly using is the part of machine learning and deep learning. You are very used from the news to generative AI, generating additional information by AI. What we are more focused on for our products is learning, teaching machines by data. If we speak about deep learning, a special method of machine learning where you use neural networks. We are creating them on our own. This is not like public models or stuff you can use like Twin or LLM or these large language models. This is really done by our group in Montreal.
We also give our customers the possibility to train these models with their confidential data so that they can train the model to exactly discover and recognize their defects, that is the automatic defect recognition. Maybe I can go back there for a second so you see the video again. Another important thing we are doing driven by deep learning is the so-called segmentation. This is actually one of the most important features of this platform. You get a 3D volume made out of pixels after the scan, but now you must train the computer to understand, for example, this is an example of a battery. Where does the electrode material start? What in this gray picture, what X-ray finally is in this gray scale picture, is the electrode? What is the separator? What is the electrolyte?
So you need to teach the computer to do what you could do with your eye when you look at the picture and you say, "Yeah, this is where the head starts. This is where it stops. This is where the flower starts or the table." Recognizing these boundaries as envelopes, this is something the computer has to learn if you want to later use it for things like automatic defect inspection, but also, for example, like measuring how big something is. These are things we are doing there. We are really embedding AI into our products, and that gives us another big advantage, especially with our inspection system. Let me quickly summarize where our technology plays today and especially where it will also play tomorrow. As I already showed for the front end, yes, we leverage the Vaccap matches and the Vaccaps itself we have.
We have still a big demand also for new capacitors. Just to tell you, this development is going on as well. We are expanding, especially with Synertia, with our data-driven solutions further into the front and also into new processes there. For example, do not be confused, we are not building lithography systems in the future, but the photoresist you are making, the way you coat it, the whole process to make sure after the development you really have a stable EUV pattern. This is now going towards plasma processes as well. The same in the doping, right? If you have a gate all around FET, there is epitaxy doping you can do, but you can also use plasma for doping because you cannot just use the ion beam from the top. It is now a 3D structure. You also have to dope it on the side.
There are more and more processes, challenging processes where RF and plasma is used. Of course, as I explained before, we must make sure we have so many possibilities now with our new technology platforms, more than we could ever design in terms of products. We must choose the right things together with our customers. On the back end, we leverage our X-ray technologies, which we have developed over many years, not just for the system, but also all the technology we have in our IXM tube division. We are finally broadening into the semi market. Yes, we have achieved this again. Isabella will show you where we are with the CA20. This system is now used to inspect advanced 3D packages. There are also new challenges and chances, opportunities for PCT.
Because as you see on the picture there, building these redistribution layers, drilling these, we assume what material ever, glass or in future very likely. All this requires plasma processes as well. We have made a big step forward and will continue with our AI-driven software platform. There we have really a lot of possibilities. This Dragonfly platform is not just used to inspect X-ray pictures. It's actually quite famous in the scientific community. It can do a lot of advanced things like data fusion, and there are many chances for us to further develop the technology, but also the market we are serving. This said, I'm looking at the clock, and I think I can make the next announcement immediately, which is the coffee break. Thank you.
Cable fair, no time for the fight. Look like the end of this journey. Computer, who knows where I am?
Lost in this kind of a nightmare space. Rest up whenever you are. Give up and you're winning. Rest up whenever you are. Lead us to forgiving. Suspending in cryogenic sleep, rocked by the wave of a beautiful song. The lifetime that I left behind, waiting to come to the age of light. Rest up whenever you are. Give up and you're winning. Rest up whenever you are. Lead us to forgiving. Rest up whenever you are. Give up and you're winning. Rest up whenever you are. Lead us to forgiving. Last lifetime, I must have overdosed on your love because this time around I'm permanently tipsy with your name on the tongue of my tip. Can't even speak right. Can't even sleep, and I'm missing your voice and your choice and thoughts. I guess I ought to stop looking for intoxication.
No need to inflate this drum in my brain when I try to release this pain in my chest. I've got to rest. I need to chill. I need to chill. Then I see you smiling, all my hard work ain't worthwhile anymore. Just a festering sore. No amount of drinking milk is going to calm it. Just drinking rum to block out some of your, some of you, summertime. At least I try. At least a generation later, I'm still trying to ease my stone. Still trying to ease my stone. Ease my stone. Ease my stone. Ease my stone. Ease my stone. Ease you, some of you, sometime. At least I try. Last lifetime, I must have overdosed on your love because this time around I'm permanently tipsy with your name on the tongue of my tip. Can't even speak right.
Can't even sleep at night. Missing your voice and your choice and thoughts. I guess I ought to stop looking for intoxication. No need to inflate this drum in my brain when I try to release this pain in my chest. I've got to rest. I need to chill. Then I see you smile. I know my hard work ain't worthwhile anymore. Just a festering sore. Now I'm drinking milk. It's going to calm it. Just drinking rum to block out some of your love, some of you, sometime. At least I try. Last lifetime, I must have overdosed on your love because this time around I'm permanently tipsy. Can't even speak right. Can't even speak. Can't even sleep at night. Some of you, sometime. At least I try. At least I try. At least I try.
Last lifetime, I must have overdosed on your love because this time around I'm permanently tipsy with your name on the tongue of my tip. Can't even speak right. Can't even sleep at night. Missing your voice and your choice and thoughts. I guess I ought to stop looking for intoxication. No need to inflate this drum in my brain when I try to release this pain in my chest. I've got to rest. I need to chill. Then I see you smiling. Then I see you smiling, all my hard work ain't worthwhile anymore. Just a festering sore. No amount of drinking milk is going to calm it. Just drinking rum to block out some of your love, some of you, sometime. Or at least I try. At least a generation later, I'm still trying to ease my stone.
Welcome back to the next session. It gives me immense pleasure to actually introduce not a new face to Comet. She has been with Comet for many years, but a new Comet face to the Capital Market Day. We will continue the next session with the deep dive into IXS, Industrial X-ray Systems. The deep dive will be led by Isabella Drolz. She is Global Head of IXS and VP of the Commercial Stream. Later on during the Q&A or the Apéro Riche, she will be accompanied by Christian Driller, who is our VP of R&D and Operations and equally a Global Head of IXS. Without further ado, Isabella, please take over and give us a deep plunge into where we are with IXS.
Today is a very exciting day for me and also for IXS because I would like to present to you how smart X-ray inspection is enabling our semiconductor industry through precision to see better, through speed to be faster, and also through AI technology to see more. Before we go into that, how did this whole start? I would like to basically take off where Stephan left. Our daily life is influenced by AI, and we are basically living in that world. AI did not just come around the corner like that. It was evolving over time, and so also did our journey. Four years ago, when we decided to go into advanced packaging, nobody talked about advanced packaging. Today, 60% of the GPU's value is comprised of advanced packaging technology. This means tremendous change. What does this mean?
You've heard a lot about the technology from André and also from the market from Stephan. Let's take a deep dive into that. We can separate in two basic streams. We have the volume production and the volume market, and we have the high-end packaging market. If you look to the volume production, you see single-layer functions, and you see sizes that we are inspecting between 60 microns-100 microns, and on the single layers, around 10 microns. To put it a little bit in perspective, because sometimes we talk about micrometer levels, like what is that actually? If you take one hair, just one hair, I'll just pick one. I have a couple of them. If you take one hair, this is around 100 micrometers in diameter.
That means on the volume packaging market, you basically look at solder bumps that are as big as one hair's diameter or less, like 60%. On single-layer fronts, you're looking even at less power, like 10%. So 10% of a hair's diameter, which is when it comes to the volume packaging market, and it's a single-layer structure, it's not that challenging anymore. If you would have asked us four years ago, we would have probably given a different answer. We are now looking at the high-end market, at the high-end advanced packaging market. You heard that before from André. The 2.5D, 3D chips, they are becoming more complex. You see it up there, the upper layer, you see the complex structures. There are even the simple ones. There are chips now with 16 levels.
That means you need to have precise image acquisition, and you need to have an inspection envelope that helps you produce these packaging chips. You're not only looking at microbumps. You're looking at via, you are looking at TSVs, so through silicon vias. These are the bars that you can see, the small connections. You are looking at wafer alignments. You are also looking at C4 bumps. You're going even further down, even below 20 micrometers, so 20% of a hair's diameter, but sometimes even less. We are right now at 2% of a hair diameter. Something you cannot even see with your own eyes anymore. That's why you need AI in order to enable the semiconductor industry to enhance their inspection portfolio and be able to increase yield in the advanced packaging market. It's not only the yield that is challenging for our customers.
It's winning the time-to-market race. In the semiconductor industry, it's all about being first, having the first technology. It is crucial that you work together with the top leading partners in order to help them win the time-to-market race and enable them. Second, as I mentioned before, finding structural defects and critical defects in these complex structures is crucial. The inspection strategy in advanced packaging is also evolving. Last but not least, in the past, you had like one design. You had one product ramped up at a time. Now you have multiples a day to ramp up. You need to be very flexible in solution. That altogether brings a huge challenge to the advanced packaging industry. That is why they need insights in order to be able to create foresights to control the process. That being said, zero defects truly matter in the advanced packaging industry.
How do we contribute from an inspection point of view to this world? These are three levers that we are basically here hitting. That is clarity. I'm going to explain that in a second to see better. Efficiency to be faster and insights through our product brand, Dragonfly, our AI brand, to see more. Those three levers come all together. And André mentioned it briefly before. It's all in one hand. It starts at the imaging chain. If you don't see it, it's not there. That's why you need to have the perfect setup in your X-ray system. You need to have the perfect imaging chain fitting to the customer need and the customer challenge in order to be able to increase efficiency, to control the data flow, to control the inspection flows. Then you use AI on different levels.
You can have interactive inspection when you're looking for ground truths, when you're doing research. Then you use Dragonfly 3D Vault, and you can do a detailed image analysis, segmentation, and also measurement of your specimen that you're looking at. Second is when you are in a ramp-up phase, you want to have a guided inspection. That's what you do when you use deep learning models. You can bring in your own inspection workflows. You can basically, it's so easily set up. You can do it by yourself. I would challenge everyone that you can do it as well. You can see in real-life data what is happening. Last but not least comes the process flow. That means 24/7 process operation where you get feedback back from each process step.
That is fed back into your analysis tool so that you can see, is my process working or not? Do I have critical defects? Do I need to stop the process? Where does Comet X-ray come into play in the inspection envelope and where are we enabling the industry? These are three levels. First of all, it's root cause analysis. There is when you are looking at a specimen, you're looking at a part. In the past, it took you days, weeks, thousands of dollars to basically scrap a chip, dissect it, and basically look at every layer. Is it good or is it bad? You can use X-ray now, and this is where we brought it to that level that you can see that in hours. Second is failure analysis. As I mentioned before, product ramp-ups are now on a continuous basis in advanced packaging.
Sample inspection is crucial when it comes to these complex designs. This is where you need a fast input, and you need guided inspection so that you can use your deep learning models to help you become faster and to help you increase yield in a timely manner. Last but not least, the process control step in the inspection envelope where X-ray plays a tremendous role is when you are looking at every process step and you get the feedback back. I have talked a lot now. What have we really and truly achieved over the past years? In 2023, we launched our first manual inspection system for chip-on-substrate. In 2024, we brought out the automated system. In 2025, we brought the chip-on-wafer feature to that system and also for root cause, deep root cause analysis. We did not do that by ourselves.
We worked with the top players in the market to always have the pulse on the market and to develop a tool and a system that fits in the inspection envelope of the future and that really helps create our customers to create value. That led to something very important for us. We got invited to be part of the next development in the semiconductor industry. You can see that up there it says JOINT3. We are part of the JOINT3 consortium where we are developing together with the industry the next level of glass-level packaging. This is the next stage that is going to drive the industry. What does this mean? Inspection strategy, inspection envelope. As you can see, you would say like, well, you are at the end of the supply chain.
Why are you making such a big fuss out of that? This is because advanced packages, as I said before, are extremely complex. They need a lot of attention, are extremely expensive in manufacturing. Therefore, you are and customers want to talk to you on every level in the supply chain. That means we are talking to customers in advanced packaging on design level, on material supplier level, on chiplet foundries, on interposer foundry levels, and of course, OSATs and IDMs. Basically, we are covering now in the inspection strategy in our envelope the whole supply chain because it is so tremendously important to serve the advanced packaging industry and helping them to succeed winning the time-to-market race and increasing yield. Nice pictures are nice, but numbers are better. Therefore, what have we achieved over the past years?
In 2023, we had like five engagements, and we started out. We had good engagement with industrial suppliers and also with leading IDM. Right now, we are at over 60, and we are globally active. We are in 25 qualification stages, and we have five orders received compared to 2023 with zero. Is that enough? No. Advanced packaging is basically hitting off now and taking the next level. Here, we are going to continue working with strategic partnerships because for us, it's extremely important to have the pulse on the market and be globally active to be in co-creation with IDMs, with foundries, and also with OSATs so that we develop the right product for them in their right use case and also to have a global customer engagement on the complete value chain to be part of the semiconductor roadmap.
To summarize that, the semiconductor industry in advanced packaging has a huge challenge, and that is increasing yield by building advanced packages. With smart AI, we can enable this industry in the complete inspection envelope to support them to win the time-to-market race. We are also ready to scale. We are ready to scale when it comes to failure analysis. We are ready to scale when it comes to product ramp-ups, and we are ready to scale when it comes to process analysis. Last but not least, we are part of the future of the semiconductor roadmap, and we are engaging in partnerships globally from the U.S. to Europe to Asia, and we are part of JOINT3. Thank you very much. Now I will hand over to my colleague, Joeri Durinckx , President from PCT.
Thank you, Isabella.
Also a warm welcome from my side. It has been indeed two years since I presented here for the first time. When I presented two years ago, we were talking about PCT, PCT growth opportunities, transforming the organization, transforming the organization from really a contract manufacturer, which is where PCT started, into a high-performance organization that delivers RF system solutions to our customers. Today, what I'm going to talk about is that we're uniquely positioned for sustainable growth driven by technology leadership, deep customer trust that we've created over the last two years, and collaboration combined with strategic alignment that is aligned within the needs of the semiconductor value chain. We have a strong foundation in plasma technologies.
The PCT division combines decades of process control know-how with a culture of innovation and self-reinvention, which I'll also talk about, that enables really our customers to improve their own device performance, yield, and cost. Supported with a global footprint and a transformation which we started in 2023, we're at the moment strategically positioned combined with a high-performance team to meet the next generation technologies. In summary, PCT is really ideally positioned to capture the emerging opportunities, turning these market inflections into significant growth for the future. Without further ado, I'll continue the rest of the presentation. First of all, Stephan spoke about the end market of $1 trillion. If you look 2024 to 2030, the end market IC demand will grow to this $1 trillion market. Stephan also explained that this comes with unique challenges. It's driven by AI. It's driven by power consumption challenges.
There is a definite need in new technologies to reduce that power consumption and also that footprint of those chips. I don't know if anybody follows the news, but today, people don't talk in data centers about teraflops or something. They talk about gigawatts of data centers. It's the power consumption that goes in. Just to give an idea, one of those 1 GW factories or data centers that they built, they contain approximately 500,000 GPU chips just for one data center. That's without the memory that goes with it. That's without the power supplies that are needed to support it. It's a huge, huge investment, but it's also a huge driver for the next generation technologies. There is from the market an expectation to continue to downscale, but also come up with alternative technologies that reduce that power consumption and reduce that footprint.
André already touched base. Stephan already touched base. We're moving from the nano area to the angstrom area, which we talked about already last time. The technologies will need to continue to evolve from the gate all around, eventually 2030, 2035, or maybe later to the CFET technology. This is mainly for foundry and logic applications for the most advanced chips. As also explained, the memory will also need to continue to go through the transformation due to the increasing demand of the memory itself. There we will see a continuous evolution from 2D to 3D in these applications itself. What does it mean in regards of process equipment? In regards of process equipment, that means that the overall WFE will grow with a CAGR of approximately 8%.
What is really positive about it, and as explained by these new technologies, the number of applications, the number of steps will continue to increase. The complexity will continue to increase. As such, also the number of equipment and RF processes in those equipment will continue to increase. PCT is uniquely positioned because that means for us specifically that the overall CAGR for the RF subsystems is going to grow faster than the WFE itself, approximately with 11% in the same period of time. When we split then the plasma processing WFE, the total available market in 2024 was approximately $106 billion. If you subtract lithography, which is ASML, the remaining, which is approximately one-third, is plasma processes with a total available market of $36 billion for our end customers, which are the OEMs.
If we split that, we know that approximately $90 billion is going into Etch, which is the main driver of that market, and the remaining is deposition with $17 billion. If you take a little bit deeper look, you can also see that Etch is driven by three applications: conductor Etch, dielectric Etch, and then there is strip. Deposition, there is PVD, CVD, and ALD. Why is it important to mention? We are part of the most significant markets that are in those specific applications. That is really well positioned with the major OEMs. We spoke about becoming more resilient. I'll talk about it a little bit later, but this is also a piece of the puzzle. Understanding where you engage with which customers on which applications to enable future growth has been one of our focus areas.
Talking about transformation and customer obsession, when we spoke in 2023, we came from a contract manufacturing background. There was a definite need to transform the whole PCT organization to really become customer obsessed, to really become much more resilient, operationally efficient, creating a global team with a global approach because our customers went global. Having local regionalized approaches no longer worked. Of course, there was a need to broaden our portfolio with a much higher speed and expand our SAM, which Stephan also spoke about. We have been able to do so. In regards of operational efficiency, Penang is a piece of the puzzle.
Working on our processes, how we're globally set up, how we engage with our customers, how we do engineering, how we do NPI, how we execute in our FAPS by more automation are all remaining pieces of the puzzle which go into the operational efficiency. I will not go into too many details about that, but know that we're working on those. One global team getting close to our customers, really being engaged early on from the R&D, co-developing, co-creating with them on the new applications so you can be successful, embedded in is one of the things that we're working on. This is how we're set up with the account teams, how we engage with our product managers, how engineering collaborates, and we've been able to achieve that too. I will talk about broadening our product portfolio, and I will be talking about expanding our existing market.
When we look to expanding our existing market, we have, of course, the piece of the puzzle which talks about the overall products that we deliver, which are our vacuum capacitors, our matches, and our generators. Coming from a contract manufacturing background with limited key accounts, in the meantime, we've made such an amount of progress that we're no longer engaged with only one or two key accounts, but we really have five strategic key accounts globally where we have active engagements, close collaboration, co-creation with. This also results in that we're getting much more diversified in our end applications, whether it's foundry, logic, memory, that we are part of the value chain. This will enable us again to become much more resilient towards the future, a little bit less dependent on the semi-cycles because our whole portfolio is now rolled out into different applications and different end markets.
When you look to the broadening of the portfolio, there are, of course, activities in technology that we're doing. André explained already a little bit. When we look to vacuum capacitors, we are not just waiting on our customers to tell us what to do. We're continuing to drive the roadmap with higher power density, smaller form factor, innovating by different materials, also how we manufacture our specific applications. Match, and I will share it in a little bit more detail. We come from an old generation match technology. Now, in the meantime, we have this Synertia platform where we also enable that technology into. It allows us to do much more fast transitions with the matches that helps to solve our customers' problem. The generator is a SAM expansion. It's about expanding the portfolio.
There, the main focus areas are power accuracy and then fast tuning, pulse shaping. I'll show a little bit more. It really enables us to be the only one to really deliver a fully integrated system from the whole value chain. Looking back a little bit, in 2023 at CMD, we had really only this Synertia 13 MHz, 1.5 kW and 5 kW generator in the portfolio. It was a new platform. André explained that we had to make a lot of investments there. We still had the old match with Gen 1, Gen 2, Gen 2B technology in. In the meantime, because of the global organization being focused, customer obsessed, I'm proud to announce that now we no longer have only the 13 MHz , but we have in the 13 MHz expanded the range with two and a half, three and a half.
We've also been able to deliver to customers 27 MHz and 40 MHz and 60 MHz , 27 MHz and 60 MHz . 40 MHz is planned to release and go to customers towards the end of the year, beginning of next year. There is a vast expansion which happened in the overall portfolio, which broadens the SAM again and the number of opportunities which we're engaged with our customers. It creates way more opportunities, but it also requires a lot more improved features which we need to develop. That's where the co-creation with the customer becomes critical that we continue that. On the vacuum capacitors, we've made also a lot of progress. We've continued to develop with our customers. The big change compared to CMD 2023 is that we now have also deployed our generation three controls into our match. We launched that in March 2025 of this year.
In the meantime, we have already over 15 customer engagements because of the data that we were able to show, the improvements that it delivers, the tuning stability that it gives, the repeatability that it gives. That in itself created a lot of market traction, a lot of positive momentum. It also shows that with the development of the platform, we now have it. Rolling out new products at a much faster pace has shown that we've been able to create all these new products over the last 15-18 months. It also means that we have at the end now a fully integrated system where the generator and the match can communicate at equal speeds. We've been able to demonstrate to customers that this brings significant value if you are really a fully integrated subsystem supplier.
In regards to the Synertia market updates that we have, similar like Isabella mentioned in 2023, we had limited engagements under evaluation with our customers. We had limited qualifications, and we had no high-volume commitments or orders from our customers. Due to this close to the customer collaboration, the new global organizational setup, the transformation that we went through, we have more than 50 engagements with customers all over the globe. Engagement can be about multiple frequency, multiple applications. This does not include the match itself. This is purely generator. We, in the meantime, have also more than 30 qualifications ongoing where we are on the lab or on the tool that goes to the end customer for qualification. I think the most important is that we have also achieved that we have more than 15 high-volume orders confirmed, which will fuel the pipeline and growth for PCT for 2026 and beyond.
Because, as André explained, once you're qualified and you're in, getting then the additional volumes that come with it as the tools number increases is going to fuel that. We have more than 100 active engagements with more than 1,000 volume orders that we have in the pipeline. We are well positioned to enable future growth. This Synertia platform is finally yielding on the generator side, but also on the match side with the new projects that we're doing. What is also really important to realize is that we're not only focused on one area. We really have these engagements across the globe, across all these multiple new accounts that I spoke about. Being able to support that required a lot of work and a lot of effort in setting that up. We are really diversifying and making sure that that is the case.
Of course, the customer adoption is driven by all the advanced features that we bring, like the fast pulsing, the power accuracy, but also the pulse shaping and auto frequency tuning capabilities that we deliver. In summary, we're in the right market. The PCT growth will come from the customer intimacy that we've created over the last two years and that we will continue to work and continue to transform the organization with. Also, the speed that it brings in order to co-create with our customers. Everything needs to go faster. The cycle times become shorter. Continuing to reinvent ourselves so that we are in pace and in sync with our customers is key to succeed with that. Also, we keep focused on the resilience.
Resilience is not only how we optimize our overall operational efficiency and flexibility, but also how we become more resilient by being more diversified and more customer installed base. The focus areas are remaining focused on the customer obsession, expanding the product portfolio, and then creating that resilience. With continuous innovation, we will focus on continued customer success and overall success of PCT. With that, I hand over to our CFO, Mr. Christian Witt.
Thanks, Joeri. Hello and welcome, everyone, also from my side. After Stephan, André, Isabella, Joeri have presented where we are going as a company, what is the progress we have made in our key new product and business initiatives, I will take you through where this leads us in terms of sales and growth, as well as through what we will do on the other side to improve our efficiency.
First, let's have a look at what changed since 2023. Since 2023, a couple of things which have changed in the market and now in the environment have been discussed by the colleagues. It's the AI speed, flexibility, these things which you need in the business today much more than two years ago. The other point which clearly changed is FX regulatory and the general global environment. Altogether, that also requires us to have some additional business priorities. That means it's growth on the one hand, which is our key driver, but also looking much closer at efficiency and resilience, things we need to take and we are taking care about much more and focusing more on than we have done in the past.
Taking all that together, where will it take us in 2028 in what we think will be the next peak of the cycle? We have updated our midterm ambitions to reflect the change in exchange rate as well as the expected timing of the peak of the cycle. We expect net sales in CHF 670 million-CHF 770 million with an EBITDA margin of 22%-27% and the return on capital employed of 27%-32%. What are the key drivers behind the change or non-change, whatever you call it? We have considered the devaluation of the US dollar, which has not only happened in the past, but it is a systematic step which is expected to happen, also expected for the financial markets, the future rates to happen over the next years. It is something we should consider. That is as we are reporting in Swiss francs.
If you were to look at us as a US dollar company, it's a very clear message. In the next peak of the cycle, we are a $1 billion company. And that's basically the same message we've sent two years ago. In dollars, there's no big change. There's a slight change in when it will happen as the whole semi-cycle has shifted, as everybody knows. When it comes to turnover, it is the same thing which we have shown here two years ago. In terms of EBITDA, we'll go a little bit into what's happening. There is, of course, some impact no matter in which currency you report and show it. If a large part of the cost base is in francs and euros as well, and the revenue is mostly in dollars, that has an impact.
Altogether, that's our midterm ambition for the next peak of the cycle, CHF 670 million-CHF 770 million, or in other words, around $1 billion in turnover and net sales, 22%-27% of EBITDA margin. Let's get a little bit into how we are getting there and what's driving us to get there from today to a peak of cycle in possibly 2028. André, Joeri, Isabella, they've all clearly shown you have to plant the seeds early, and then it takes a number of years until you harvest. Where we've planted the seeds yesterday, we harvest today and tomorrow. Where we plant the seeds today, we harvest tomorrow and the day after tomorrow. We've always been ahead of the curve of growing sales and growing our R&D to really have the technological edge, the value add for our customer developed. We have done that.
We continue to do that. That's the core of our company. We'll be focused in high conviction projects and segments, highly adjacent, as Joeri has shown, highly adjacent also in the case of IXS, where we have a base and industrial base business which has allowed us, with our competency, to step into semi. That is our key, which we'll continue to do. As we are a heavy, resilient, strong balance sheet and setup, we'll do that also throughout the cycles. We are, we have been, and we will be ahead of the sales curve with our R&D. Let's look a little bit where are the specific factors which fuel our growth. That's a second focus we as a company now need, also where we enhance efficiency and strengthen resilience. Key aspects for growth.
First and foremost, it's the seeds we have planted in the past in PCT, especially in PCT, because that's where we have the longest cycle, where we continue to penetrate our existing customers with existing products, with new products, where we penetrate new customers, expand our markets, expand our reach in the markets. That's the biggest contributor to our growth. The second big block is that we have new products, new segments and services which we have been developing, which we have launched, which we are launching. That is on the one hand, inertia. That is on the other hand, pivoting the X-ray business towards semi. CA20, what Isabella presented for the systems as well as for the related modules, is a key part of that driver.
Last but not least, you've heard both of them, Isabella and André, talk about the software and how we do that with our Dragonfly. That's another platform which is key in order to fuel that growth in new products and services. Having a look at efficiency and resilience, the key point is this is the second focus for us in order to develop our company for tomorrow. What's the key drivers here? Number one, and Joeri touched on that as well, it is looking at the processes and becoming more efficient in what we do and how we do it. Second point is, how can we take advantage of our new best cost site in Penang? We've seen we are very well on track. We have a large facility we are building there to host, on the one hand, production.
That means shifting blue-collar work to Penang and growing blue-collar work in Penang, as well as significant office space because we want to shift more and more also white-collar work to Penang and use that as best cost site. Basically, that is a second key pillar when it comes to cost efficiency. First, how many resources do we need? Second, how can we shift some of those resources to Penang and grow there in Penang instead of high-cost countries? Another point, operating leverage, I think that's pretty clear. When you look at our capacities today, we run plus minus one shift, sometimes a bit more, but there is a clear case for the leverage in terms of our production capacity. We are set up to bring the volume in. We are doing the work with our customers. Whenever that comes, we'll see the operating leverage coming through.
Product line profitability, that's the multifold topic. Of course, we are looking into our different product lines. Are there product lines which need a turnaround? Are there product lines which we might or might not want to continue? Here we're talking small stuff, not big stuff. That's the usual type of cleansing a company should always do in its product portfolio. We'll do that. We'll do that diligently. All of that is needed in order to assure our growing profitability for tomorrow. Let's have a look at how that converts into numbers and what that means in numbers. When you look at net sales, it's very clear. The key part is volume and mix on our products. That is the key driver, as just explained.
You plant the seed today, you harvest today and the day after tomorrow and the day after tomorrow. Second part is the new business, and that's mainly, not only, but mainly CA20 and Synertia. And that is the contribution we will see from our new business in a year plus minus 28, which is not where we want to be in the future, but which is an intermediate step. Because, as I said, we plant the seed today, we harvest tomorrow and the day after tomorrow. Why we are really doing this? Synertia, as well as CA20, goes clearly beyond what we have in here for our 2028 ambition. And FX, as mentioned, as long as we talk in Swiss francs, we'll have this effect. In dollars, you wouldn't see this effect. Let's look at profitability. How numbers, how that converts into numbers over here.
Again, the volume and the price mix effect have the key impact. That is the volume from our current product line as well as the modernized portfolio, the new product lines which we are bringing in, which also come in at a higher gross margin. Second important point here is that efficiency will play a different role. In any case, we have an FX issue and we have opportunities on the efficiency side. Working on that will bring us between 5 percentage points and 7 percentage points in EBITDA between today and 2028. Focus points there. On the one hand, it is really the restructuring which we have started in the non-semi business from IXS in order to enhance profitability. It is also the efficiency measures I mentioned, and it is utilizing our best cost locations in Penang and possibly elsewhere. That is the key drivers behind our profitability improvements towards 2028.
Let's look at the divisions. This picture should be pretty familiar to everyone who has accompanied the company for a while. We have around two-thirds of our revenue from PCT, or we will have. We do not have it today, but we will have it as PCT is growing slightly at a higher rate. When we look at profitability, we have the same picture as we've seen in the past. IXM will be plus minus within our group range. PCT, being a highly differentiated pure semi player, will be clearly above that. IXS, and I'll come to that in a second, will be slightly below the group average. ROCE, same in general, just IXM due to its capital intensity is a little bit lower here, but that, I think, is also a familiar picture.
Let's take a little bit closer look at the divisions, what's driving the growth as well as the profitability development in the different divisions. It is, as you will see, slightly different drivers here. PCT, as mentioned, it's the penetration and it's the market growth, as well as it is the uptick in Synertia with our new product lines there. That's the growth side. On the profitability side, it's the operating leverage. That is for sure the strongest driver, but it's also efficiency measures. Here in PCT, as this is the main user of our new site in Penang, that shift and pivoting towards Penang has a significant impact as well.
Looking at IXM, IXM has renovated its product portfolio in the past years, so this will be kicking in in the next couple of years, improving sales and also improving the margins we can have with new state-of-the-art products in the markets. That is the growth side on the volume side. On the profitability side, it is volume on the one hand and efficiency measures. IXS, here the growth driver is the pivot to semi. It is not the industrial business. It is the pivoting to semi, as mentioned with CA20 and related products. When we look at the profitability, it is basically two key drivers. One is that pivoting to semi, which has a different gross margin profile than an industrial business.
On the other hand, we have started to right-size the organization in the industrial part, and that is the second important driver to get our profitability in IXS to where it needs to be. IXS, I'd like to shed some more light on how we look at IXS, because if you see IXS from the outside, it's one division and it makes a loss. That is not too transparent, and we look at it differently, and we'd like to share that view also with you. For us, okay, for us, IXS, we look at it as two business lines. One business line is our industrial business, and one business line is CA20 and the semi segment. When we look at 2024, we have invested an order of magnitude CHF 10 million-CHF 15 million in that year just to go into the semi market.
That is R&D, and that is market access, both important and both needs funds. So that is basically that part. If you take the other part of the equation, you have a 5 percentage points -10 percentage points EBITDA business. That is not where it should be for an industrial business, but that's where it was in 2024. When we look into 2028, then for semi and CA20, this is clearly the intermediate step to go to where we really want to go, what Isabella mentioned. In 2028, we'll be at least break even. That is pretty clear. And in our industrial business, we'll be at a benchmark level for the industry, which will be for sure more than 10% EBITDA for that part of the business. So this is the way how we look at it.
We look at it as two business lines, and we'll continue to share that view and a certain transparency on that with the financial community because we understand that it's important for you to know what we're doing here, what is the investment case, and what is the industrial business. Looking briefly at cash flow, we have an operating cash flow, as you know, which funds our operations, our standard investments, our growth, which also funds mid-size investments in, for example, Penang now. It's a facility of about CHF 80 million, including the equipment, which we are funding out of our operational cash flow of the years. It's also able to fund some bolt-on acquisitions, for example, to accelerate growth in certain areas. You can develop something if you want to accelerate and can buy something in terms of development of product that can make sense.
If it's smaller acquisitions, it's well to be financed out of the operating cash flow. The stability we have enables us to operate throughout the cycle and to conduct our business throughout the cycle. We'll continue on that path. Looking at capital allocation, key driver is we will continue to invest into growth. That's R&D and market access with an R&D ratio of about 10%-15%. That could also be the mentioned bolt-on acquisitions targeted M&A, for example, on the technology side. We'll maintain an adequate balance sheet for us as a technology company, so rather leverage-free and the financial flexibility to execute on our strategy. Shareholder returns, we will pay the dividends according to our policy. If there's other ways to return money to shareholders, and that would be the best way, we would, of course, not hesitate to do so.
Some housekeeping on other financial KPIs, and let me focus here just on two. CapEx between 3%-7%. We are low to mid CapEx intensity in our business. When it comes to Vaccaps and IXM, it's a little bit more in production. When it comes to IXS and to the matchboxes, it's rather assembly, rather low capital intensity. Maintenance CapEx is about 2%. About 3%-7% is what we estimate throughout the next couple of years. Networking capital, networking capital, we'll go to slightly below 20% as a group. When you look at today's where we are today versus where we will be tomorrow, the key driver there is inventory. We still have some high inventories from the last ramp, and we have, as Joeri mentioned, changed some of the ways how we do these things to build our resilience.
That includes that we know quite well how we get our inventory improved, our inventory levels, and we'll get mostly thereby to a networking capital ratio below 20%. Let me wrap it up. What's the key driver to get to growth and to profitability? It's the growth initiatives on the one hand, but on the other hand, it's also a second focus on improving profitability. That means looking at our cost position, looking at our cost efficiencies. It also includes looking into how to best use our best cost site in Penang, which we are finalizing in 2026, and the mentioned revisions of product portfolio and similar. Solid financials for growth, what does it mean?
At first, it means that we prioritize and focus our spending in terms of R&D or CapEx on the key pillars which we need to achieve our strategy and our future growth, and that we remain, other than that, a very healthy capital structure. How will that change and transform Comet? There are a couple of things without going into details, but it starts with a fully transparent culture. It goes on to infrastructure like IT systems. Look, we are a standard industrial company when it comes to execution. Why should we not have a standard system? That is much cheaper and much better. By doing that, we are able to really leverage our group size and, as Stephan mentioned in the very beginning, our fundamentals, which we have in common as a group. That is the key takeaways.
In the end, it's how do we enable Comet's transformation to growth and profitability? With that, I'm happy to hand back to Stephan for the wrap-up.
Thank you, Christian. Let me close Comet's Capital Markets Day presentations for today by trying to stitch and bring all together. We have looked at how Comet is executing today, delivering on our plans, expanding capacity, and investing in innovation, people, customer proximity, which is extremely important. The world around us is obviously, as you know, changing very fast. For us, that means our own transformation must continue with equal speed, determination, and also measured risk-taking. We are operating in an environment defined by shifting geopolitical and economic conditions, coupled with an industrial transformation that is reshaping the semiconductor landscape as well as our societies. Supply chains are being realigned. Technology roadmaps are accelerating, and regionalization is becoming a defining force.
These developments create both challenges and opportunities to which we need to respond proactively. That's why pushing ahead with our own transformation is not optional. It is essential. We are strengthening our organization, modernizing our processes, and investing where we can make the greatest impact in technology leadership, customer proximity, and operational excellence. Pursuing this transformation with determination will ensure that we maintain our leadership in RF power and X-ray and continue to be a relevant and trusted partner in the semiconductor industry. Our transformation is about preparing Comet for the next phase of profitable growth. That is reflected in our updated midterm guidance. Over the past year, we've continued to strengthen the foundation of our business by improving operational resilience, expanding capacity footprint, and sharpening our focus on the markets where we can actually lead.
As a result, we're now able to update our midterm guidance to reflect both the progress we've made and the realities of today's market environment. The midterm outlook remains weighed down by a challenging business environment marked by slower than expected adoption rates across key markets and continued currency headwinds. Persistent exchange rate volatility has pressured margins, as Christian explained before, while a more cautious spending climate is delaying customer investment decisions, tempering growth momentum in the coming quarters. Our updated guidance remains, however, ambitious, but it is grounded in solid opportunities and execution. It reflects a company that is transforming, scaling, and ready to capture growth as semiconductor demand rebounds and recovers. We see structural long-term growth in our core markets. At the same time, we're managing near-term volatility through disciplined cost control and operational agility.
With our transformation fully underway, we are confident in our ability to deliver sustainable growth, stronger margins, and improve cash generation over the midterm horizon. In short, our updated guidance is not just a forecast. It is a statement of intent. Comet is positioning itself for the next cycle, stronger and more focused than ever. Now, before we go into the next slide, before we go into Q&A, let me summarize why we believe Comet represents a compelling investment opportunity, not just for the near term, but especially for long-term sustainable value creation. First, Comet's foundation is built on mission-critical RF and X-ray technologies that sit at the heart of semiconductor innovation. As chips become smaller, faster, and more complex, the tools and technologies needed to manufacture and inspect become even more essential. We are uniquely positioned at the center of that evolution.
Now, second, the structural growth drivers behind our business are stronger than ever. We are aligned with local and global long-term global shifts, reshaping the industry onto which we have cast our focus. Our technologies are enablers of those transformations. Third, our strength lies in focus. We hold leading positions in high-barrier, highly specialized market niches backed by deep customer integration. These are areas where expertise and trust actually really matter because they define, decide the outcome. Fourth, we are developing a scalable and resilient operating model supported by a very global footprint. This combination of operational discipline and geographic reach gives us the agility to weather market volatility and the strength to fully benefit from growth cycles. Fifth, our capital allocation remains disciplined, supported by a strong balance sheet. We invest where it matters while maintaining financial strength and accountability.
Finally, we have a clear path to profitable growth and sustained value creation. This is the roadmap we are already executing. In summary, Comet's story is one of opportunity, innovation, discipline, and execution driven by people who believe in pushing technology forward and creating lasting value. We are proud of the foundation we have inherited, but also continued to build and even more excited about the road ahead. At this point, I would like to thank you for your attention and patience and continued trust in our journey. With this, we will go into the Q&A. Before we do that, we have to, I think, reconfigure seriously. No, no, you are not going to do that. We will reconfigure here. Thank you. The smallest from the easy. Joeri, good.
Thank you, Stephan, so while the colleagues are rearranging the stage.
You heard a lot about technology, strategy, commercial aspects, financials. Obviously, you will have some questions about that. We will start with the Q&A. I ask the colleagues to join me on the stage. Three housekeeping things. First of all, limit your questions to two, please. Second, there are two colleagues around with the microphones, so wait until you have a microphone that also colleagues and the webcast can hear you. Last, but not least, state your name and your company. That would be great. With that, I see that Michael is the first one to ask a question.
Yes. Does it work? Yes, thank you. Michael von Tobel, two questions. The first one is I was trying to piece together your market outlook for your service addressable market, which you expect to more than double versus your targets, which imply sort of 60% growth.
Even if you adjust for FX, the addressable market, it looks like somehow a piece of that market is going somewhere else. Trying to bridge that. The second question is a lot of growth drivers that you mentioned, but currently, things are not growing so much as we saw with your latest update. I was wondering what trigger really should we look for to see that growth accelerate over the coming 6 to 12 months. Thank you. Absolutely.
Thank you for that question. I understand the first part is you're puzzled by how fast our SAM grows versus our growth. Remember when I spoke about the SAM, I specifically said that is the part that we can address with our current and near-future product portfolio, product platforms.
As we start to and continue to expand that portfolio, you have to think of this in the following way. If you add a new feature, a new member of family to, for instance, PCT's Synertia platform, or a new feature set to the CA20, that opens up the SAM immediately, whereas you have to go through that painful process of qualifying that feature, of qualifying that new family member for Synertia. That comes at the latest stage. In other words, as Christian showed, we are always, when it comes to R&D, which then results into portfolio expansion, way ahead of the curve of sales. SAM grows typically in a company that increases its portfolio of offerings faster, sort of ahead of the curve versus your actual sales. The second part of your question, what are the proxies?
What do you have to look into? What we see now ever more clearly is that the CapEx, the real CapEx that has happened over the past couple of years, mostly with the hyperscalers, that is now really starting to pull in more and more activities and CapEx in the wafer fab equipment area that trickles down. Hopefully, that trickle will turn into a stream over the next three years. When will it happen? I mentioned in my last words, we will probably still see some sluggish behavior in the next few coming quarters, but it is just a fact of having to equip those many fabs that are in construction or near termination of construction, like I mentioned in one of my very, very first slides.
That will translate into real orders and real shipments converted into sales for both divisions, with the division IXS pulling IXM along.
Good. Thank you, Stephan. We have a next question from Nate.
Yes. Hi, [Nate Slawicz] from Octavian. Maybe my first question. Recently, you mentioned there is an increase in non-spending, but it is mainly on the upgrade side. Can you maybe just give us more clarity on that? What exactly is being upgraded? Are these existing machines that are maybe repurposed or just part of these machines that are changed? Just some clarity around that would be helpful. My second question would be China and competition there. What are you seeing there? I remember two years ago, there was conversation, yes, competition is coming, but 2027, 2028. I mean, this is now the peak of the cycle.
Are there some players there emerging? How is your market share shifting there?
Let me take that first, and then Joeri is best positioned to talk about upgrades because he is deeply involved in that. A year ago or maybe two years ago, I mentioned, oh, there are activities happening in China. It's kind of like a garden full of mushrooms, right? It's mushrooming. That is not surprising. Look at how China, as a country, is investing into the semiconductor ecosystem. You see that across the board. We see that from especially the wafer fab equipment makers and mushrooming of activities and applications that they can or at least say they can cover, that the investments that the nation, the state of China is making is probably bearing fruits, right? Is this something we should be concerned about?
For the time being, definitely not. We are not positioned in the area where their largest focus is in terms of really what subsystems or equipment they're investing into, but also the level of technology that they are trying to insource and nationalize. Do we see activities in our field of activities? Of course, on a low level, they're trying to do a lot of things, but it's not that we feel pressured or that there is anything coming close to a loss of market share. We have a lot of engagements with companies in China, always very clearly within the bounds of the regulations and limitations that exist. We will continue to do that, just like everyone else, also the Americans too. I hope that gives some color to that question. On the upgrades, Joeri.
We have taken with the team a significant deep dive specifically on the technology upgrades. We know how to distinguish between all the different versions. Technology upgrade can go from upgrading some of the subcomponents, which is a match, which we are part of the daily business with our customers. That is part number changes, which we then are part of the value chain, up to complete chambers, which is complete new technology, and then we are also part of the value chain. That is why you, by the way, saw a significant uptick in 2022 and also 2023 in our RMA business because we were actually part of the existing technology upgrades. We are part of it. We know where it plays. It depends on which technology upgrade the customer is talking about. We are trying to capture those opportunities also.
Thank you. Next is Michael, and after that, Felix.
Okay. Thank you very much, [Michaelina and Sakebi]. Two questions on IXS, actually. I'm trying to understand the Dragonfly story a little better because you mentioned it also in the last press release, first time. And I understand, or I'm trying to understand if there's a business model around specifically only Dragonfly. And if that is the case, what is it and what are the economics? And on the CA20, when I talk about it also to investors, the pushback I'm getting is, is it really ready to allow for 100% testing? Or is the throughput enough to be really in the fab with the CA20? Or is that not even a need that the industry has? So is sample testing still enough, for example? Thank you.
Isabella, you want to take that one? Yeah, sure.
First coming to Dragonfly, and how is it ingrained in the CA20 or in the semiconductor industry? It drives the inspection workflows. Basically, with our deep learning technology, we can create now easy inspection workflows so that you can measure voids, you can measure TSVs, you know exactly where you are. If you remember the picture that I've shown you, the slide screen, through the AI and Dragonfly, you basically can see in real time where you are. This is one of the heart next to also our imaging chain that defines the CA20 and the semiconductor approach. Yes, the Dragonfly software has its heritage, as André said before, in the science and research area. Here we have a standalone product, and you can use also Dragonfly 3D World alone. It's not only limited to X-ray. Here we are seeing also scalability potential.
The number two was inline inspection when it comes to X-ray inspection. Yes, there is a need for that due to the complexity of the advanced packages. When we look at the volume inspection, I would say, yes, you are right. Sample inspection is enough at some point, but due to the fact that the products are becoming more complex and you cannot see certain defects anymore without either 2.5 D or 3D X-ray, you need to have them inline.
You can do that.
Yes.
Good. The next question, Felix.
Yeah, Felix Lemons with Tech Capital. Two questions from my side. You anticipate by 2028, CHF 110 million-CHF 150 million additional sales from new business, so Synertia and CA20. How did you come up with that number?
Based on how is the market, or not the market, how are the applications evolving, right?
When we speak of, for instance, staying with IXS advanced packaging inspection, that is not a homogenous one application. It is multiple applications because there are many neuralgic points that need inspection to a certain point. Then that problem is solved, and it moves to the next one. Understanding from interaction with the customers where they are focusing and what problems they need to solve or where they really want to keep their eyes on and what the volume is, partially even together with specialists on the side of the customers, we have done our calculations, estimations, and then mapped out also with a competitive landscape, where are we strong, where are competitors strong, where are there me-too solutions that can do the trick, and where do you really have to have a CA20 type of technology.
We did our planning and came up with a number for CA20. The same thing Joeri and his team have done on the Synertia side, where things may be a little bit more clear as it is already a very much established market that is obviously also growing and evolving fast, but it starts from an established point, whereas the advanced packaging, especially the 3D inspection part, is something that is more emerging and also needs here or there some guesswork. All in all, it is a sound number. Believe me, Christian has been drilling down this number quite extensively, sometimes to the exhaustion of some of the colleagues here on the stage. We are very confident that these numbers are realizable to the peak of the next cycle.
Another question I had is, thank you for the answers.
I hope that answers it.
Yeah, I hope. I mean, the investors' expectations have been disappointed on the ramp-up of Synertia, on CA20, I would say. I hope you have better luck forecasting this time. Anyway, the other question is on, I mean, you mentioned recovery and uncertainty. Nevertheless, if you just look at the wafer equipment spend, it's estimated to reach a new peak this year, CHF 110 billion. Meanwhile, your sales in the semi-sector are off the peak in the very good years. What is your assessment from your perspective? Why is that? While we reach a new peak this year, your sales are below. What is the main reason other than FX?
Yeah, that's a very good question. The wafer fab equipment market is a very non-homogeneous market.
For instance, when you look at lithography as part of the wafer fab equipment market, in the years, especially of 2023 and 2024, ASML had the best years and were actually supporting the wafer fab equipment market, whereas on the plasma side, and Joeri showed the kind of the plasma part of the wafer fab equipment market was terribly down, basically mostly supported just by upgrades and service. It does not move, the areas do not move in lockstep at all, not even partially. Usually, the investment cycles, there are time shifts or phase shifts in between. What we see now, and we are being warned by our customers, is that they are saying, "Hey, now it is your turn. Are you ready?" Good.
Next question then, Serge.
Yes, [Serge Lomborghieri]. I will come back on the question of Felix, and I would like to ask this time, Joeri.
Three years ago, you mentioned our generator market is a CHF 1.7 billion market, and you go for 10% market share. Now we know we are delayed, but you have a much broader product range because you had one single product. Now you have numerous products. When I look into research, the market is expected to be clearly above CHF 2 billion by 2030. Let's assume CHF 2 billion around 2028, your target, if you can agree on that. In Swis Franc terms, unchanged, one to one, Swis Franc to dollar today, zero eight to one. The market is basically unchanged if you compare this. What kind of market share should we expect that you can achieve then by 2028? If you still go for 10%, then obviously it would be CHF 200 million.
If you go for 5%, it would be CHF 100 million, which would be aligned with the number from the CFO and gives no success to the X-ray product. Is this correct, this view?
Shall I take it?
Go.
Yes, you are correct. We are still aiming for the 5%-10%. We have been a little bit more cautious in our projections because of the slowness of the adoptions, hence why you see a slightly lower number.
This cautious was the CFO telling us. This probably goes to the CFO.
Let me shed some light on the number without dissecting it completely. What's in there? In there is CA20, that's one significant portion. Then there's Synertia, and Synertia consists of the generator and also consists of the Synertia matches where they are not just a replacement, but open up new markets.
These three are the key drivers between what are showed in new business. I think you are totally right with the question, the approach. When you put the numbers together, you just confirmed 5%-10% of the market is what we want. On the other hand, if you remember, you plant the seed today, then you harvest tomorrow and the day after tomorrow. In 2028, we will not have full penetration of what we can reach in penetration. Step by step, you complete the product portfolio, and with each completed step, you go into the market, you qualify for a tool, and then that tool is being sold by our customer to our customer's customer. Because of that, you always have a time shift in there. We do not have full penetration to our potential in the generators in the year 2028.
That's why I said it's the first step. It's an intermediate step on where we really want to go with CA20 as well as with Synertia.
I disagree. You mentioned you have 15 high volume orders with, I don't know the numbers, but was really impressing. I would say the visibility is much bigger than you tried to explain to us. Otherwise, I didn't cut it.
Yeah. Maybe there's also a little bit a misunderstanding how this market works in terms of winning market share. If we speak about a market of CHF 1.7 billion, this is all the tools delivered, right? A year with generators, matches, half power supplies on. You cannot like this win the market, even if you have the best, the cheapest product, whatever. This market uses, I don't know if you ever heard it, something called Copy Exact.
These systems are like extremely complicated. Existing systems, they are under this rule. No supplier, especially no user like TSMC, touches them if there's not a problem because it costs so much money to change something, to reconfigure something. What you can win, what you're going for, especially with new technologies, is the new kind of tools introduced every year. Joeri and his team have won some very important tools, which we will see in the future. They won the race being qualified on these tools. These tools will now come into the market, and they will scale step by step. As I explained also before, they will not immediately replace all the old tools. This is a little bit how things are set up.
As soon as these tools come into the market, there is a like step function where PCT wins market share, and then over the years, they scale and they are there for a long time. This is then the continuous market share you have, and every year you add additional market share by winning new tools. This is how the system works. Of course, you can see that this depends a lot on how these tools, which we have won now, scale in the next years with the new nodes.
I want to add to what André says. There is a distinguish between a new product, which becomes like mainstream normal usage, and sometimes you're lucky, you hit a jackpot, and one of those tools is a hyperscaler where suddenly the demand goes exponentially out of the door.
The problem is we're in this phase where we know we're on the tools, but you don't know whether they become a mainstream product or they become a hyperscaler. If we hit the jackpot with one of them, the number is way too conservative.
Okay, thank you. Let's go for the jackpot then.
Yeah. Serge, I see you're not totally satisfied, but I'm sure we can talk about that during the apero or ish. Come to me. We have a question here in front.
I have two questions. Can you profit as well from the investments in the military defense business? I heard once that maybe with your business, with the X-ray business, you could profit. The second question is how much do the U.S. tariffs hurt Comet, and how do you want to avoid it? Thank you.
Let me take the first one, and then you can talk about tariffs. By the way, has anyone followed up on the news? Has something happened in Washington today? No? Okay.
Time still yesterday.
Defense. We have some of the X-ray business, particularly on the side of IXM, and I invite you to talk to Michael Berger, who's hiding somewhere. Oh, he's standing back there. He can shed some details on it. Yes, also on the side of defense, be it for ammunition or for vehicles from fighter jets to tanks, there is non-destructive testing happening. We have customers who cater to that business. Essentially, Michael has seen over the past two years, definitely an increase in demand, an increase also in revenue in his business catering to that part of, let's say, the global economy.
It is not significant, but definitely something that for Michael's business, for IXM, is, I would say, not super important, but noticeable for the Comet group overall. It is a side business. It could increase. It has increased. Could increase. You know, I do not want to earn money on that stuff, to be very, very, very direct. It happens so that we do. Michael can give you a detailed download of where we are participating. If Rheinmetall has order books that are basically just exploding, then typically they require also non-destructive testing, provided among other things by X-ray systems that typically do not come from IXS. That is not something that IXS is or has been involved in over the past 10, 15 years. Maybe that answers it, but there he is, ready and prepared to give you a download on tariffs.
Let's look at tariffs.
How do tariffs affect us? Let's, for the time being, look at U.S. tariffs. It's very different for the different divisions. For PCT, there's hardly any effect. Why that? Because most of our customers import themselves. Many of our customers, also U.S. customers, don't even ask us to bring the stuff into the U.S. That is for us an administrative issue. We have to deal with it together with the customers so that in the end, these will be hardly effective anywhere. When it comes to IXM, we have a significant business with the U.S. A significant share of the company's business is there, and we produce in Switzerland. It's not a production you can easily double somewhere because it's high CapEx and high know-how. That is something where we are hurt with tariffs, but that is a one-digit amount.
There is a couple of ways how to temper the effect of the tariffs. There are ways like partially assembly over there. If it's really hard, there are some rather administrative ways how to do that, but that's the two structural ways how to approach it. It is not a major topic for us. The key is rather the uncertainty the whole tariff issue has brought to the industry. One more fact which is important, when you look at the setup of Comet, for what we do, we have one of the best setups and global footprints you can possibly have. We are a Swiss company, and we have production facilities in the U.S. if we need to, in Malaysia as a major hub for the future, in Germany and in China.
By that, no matter how the world turns, I think to have a better flexibility than us, it's difficult. I'm very, very happy with the setup we have. That gives us options.
Good. Thank you for your question. We have one from Sebastian, please.
Yeah. Sebastian Vogel, UBS. I got also two questions. The first one is regarding to the Synertia. You mentioned on this slide high volume orders. In that regard, two clarifications, if possible. A, what means a high value order? In that sense, what means high for you? The second thing in that regard is how quickly is such an order turning into sales? Does it have a translation time of like three months, six months, twelve months, someone beyond that number? The second question is going to your CEO.
In the beginning, you mentioned about the outlook that it could be a bit sluggish for the next couple of quarters before you see then a bigger of a stronger ramp up. Was that mainly related to semi or was it a broader perception for all the business in aggregate? If you could shed light on these two points, I would appreciate it.
High volume orders means in reality that we're qualified by the end customer, that we have officially been audited by the suppliers, that we are a high volume ready site. That means that we have worked with our customers on the forecast for next year, and we're locking in orders and POs for those volumes for next year. We're mainly working on forecasts for the moment.
In regards of delivery time, we are well set up for the moment with our suppliers and with our supply chain. We expect to be able to deliver the first units actually starting the first quarter of next year.
To the question about the sluggishness or the slow-tempered situation for the next couple of quarters, it is to a certain extent in the semi, but then definitely also very strongly in some of the non-semi markets that we're serving, be it automotive, with the exception of batteries where we expect next year that the glut that has been accumulated over the past two years will be resolved and lead over to, for us, a better investment climate.
Aerospace will continue to be on the positive side, but both automotive and aerospace become less important for the group, still rather important for IXM and to a certain extent also for IXS. My comment was mostly mainly on semiconductors as pretty much my entire overview was completely captured by semiconductors as this is really where we want to focus. Does that?
Okay.
Good. We have another question here in front.
Hi. Hello. Hi, Ali Wong from BMO. My first question is, you mentioned that currently you have for Synertia 15 high volume manufacturing orders and for CA20 five orders. What sort of level of orders do we need to get for you to hit your target for 2028?
It needs to be clearly higher, but we would not like to disclose any sales prices, related sales prices numbers or there.
I think as the colleagues pointed out earlier, what we do and what we have modeled very diligently is let's look at the different applications. Let's look at where we are qualified, where we're going to qualification, and let's look at how the buildup of these goes step by step. That is exactly the path we need to be on and we are on in order to reach out our 2028 targets.
Thank you. My other question is, I just wanted to drill down on the upgrades again. I was wondering, it seems like so far you're not benefiting too much from the NAND upgrades, whereas some of your competitors like MKS in the U.S. are. I was wondering if you could give a bit more detail on why is it that certain other components may be getting upgraded more? Is it a software thing?
What is it? Just because one of your largest customers, Lam Research, was quite vocal that they're seeing a lot of business from that. Thanks.
Talking about upgrades, Joeri shone some light on it. We profited back in 2022 and partially also in 2023, but upgrades, that's a large bag of options. An upgrade, and especially now also talking about MKS, look at the portfolio of instrumentation they have from vacuum gauges to also plasma processing equipment. Many, many things can be upgraded. You mentioned also software. It doesn't always mean that matchboxes or generators are part of the upgrade package. It really depends on what are they doing. Are they changing materials? Are they just changing the capacity? Are they upgrading the handling?
There are many, many things that can end up in an upgrade package, and it does not always mean that the plasma part is connected to the upgrade. Yeah, it is kind of difficult to quantify, but I hope it gives a little color to your question.
Thank you for the question. I see there are no urgent questions in the room, so probably Fabio, operator, there are questions from the webcast. No. If that is not the case, then there is a last chance to ask a question. Okay, I see two hands. Christian Brown first, and then Nate with a follow-up.
Thank you. Christian Brown, Finance and Wirtschaft. I have a question on resilience. It has been a bit of a problem in the past. Now the word has popped up quite a few times, which means it is being addressed.
However, I haven't fully understood how you want to improve it. Can you elaborate a bit on it? Resilience is also kind of an umbrella word for many things, right? Resilience can be. I mean the swings in the marching especially.
Yeah. Resilience comes in many different shades of gray. Right now, resilience in terms of being able to respond, for instance, on pressures pertaining to tariffs is to have a widespread geographic footprint, right? We can transfer things from A to B in order to lessen the impact of, for instance, tariffs. At the same time, a widespread geographical footprint is also a burden, especially a burden when markets are down because you have to carry the fixed costs.
Right now, resilience for us means that we can carry that geographical footprint, but especially also now with Penang coming online next year, build in that new geographical footprint. Resilience also on the profitability side by shifting more and more activities from high-cost countries to this best-cost manufacturing site, just as a part of resilience, the resilience package. Christian also mentioned a couple of things in terms of improving our resilience, and maybe you want to add to that point.
I think I'll expand a little bit on what you just said on Penang. If you lower your cost base, you automatically increase the stability of the EBITDA margin. That's one very key point. Second key point is when we look into what Stephan mentioned before, the ramp up and the ramp down.
There's, for example, inventory mechanisms where we need to be much faster than we've been a couple of years ago. We've been working on that. We are introducing that. That is a different type of measures to be more resilient. Number one is really the cost base. Number two is processes in the up and down. Number three is, and that's something which we have and we retain, you need to have the right financial structure and balance sheet structure that you as a company can still take your free decisions and you're not driven by banks. As our business can be lumpy sometimes because the market is, and the market is the market, it's the customer. That is something we have and we will continue to have.
Thank you. Stephan, Christian, Nate, follow-up? Just talk. Okay. No, no, no, with the mic. It's working. Perfect.
Sorry.
Maybe follow-up on the solid state and mattress technology. I mean, when you read Lam Research transcripts, clearly they're saying they're winning design wins in NAND. Now they're going to DRAM. The one advantage you suggested was speed, but clearly for them it seems to be a crucial technology going forward. Why is your perspective so different to what Lam Research is saying? Maybe that would be my first.
It is not. I maybe disowned it somehow how defensive in terms of speed. We are looking at speed. This is what Synertia can do extremely well. What I wanted to show you is that just from a part of physics, having a big advantage on one side like speed, maybe a disadvantage at the other. At the end, there is not one application. There are different requirements for different applications.
You need to build the right solution for the right problem. Some new applications, especially, they're very sensitive to speed. Maybe that's the most important point. There are others which are extremely sensitive to some kind of stabilities, the things you have to measure, repeatability, and so on. There is no magic to get it all in one package. We are looking at all of these solutions. I mean, we have technologies for all these solutions. You could scan our patterns and you will see we have developed a specific material combination where we can do super fast matching without vacuum caps. We have these technologies in-house. Then we sit down with a customer and we solve a customer-specific problem, also with Lam.
These tools I know which you are mentioning, where you speak about direct drive and the other thing they call tempo and snap. I can tell you, especially with a direct drive, Comet is involved. I cannot say more. As I said, some things are also advertisement, but we are involved with all these critical steps, solutions, new technologies. At the end of the day, we have to solve the most and high-value problems. That is what we are doing with our solutions. That is what I also wanted to show you, that there are other options we are looking in, which are maybe even better, which are not just fast, but fast and precise at the same moment.
I maybe want to add a little bit to what André says.
If you talk about specific customers and specific applications which you're referring to, those are also high power. If you look at the literature of solid state technology, then you will find easily that there is a disconnect and a mismatch in between. They are working on alternative technologies and it's not that what you're referring to.
Okay. Thanks. Maybe as a follow-up, quite some of the CapEx or let's say the big fabs, they're now in the US. Do you have some, I mean, how do you see your competitive advantage when it comes to partially shift in CapEx going there because you moved part of your production? Is that a risk? I mean, do you have to renegotiate with your clients or do against on the spec wins side or is that really out of question?
I don't know if everybody knows the PCT global footprint, but one of our biggest offices and key account structure together with application engineers, together with some form of engineering and NPI, limited low volume manufacturing is actually close to our customers in the U.S. We're actually working on a day-to-day basis with them. With the volumes of equipment going to the U.S. , we're actually ideally positioned in the heart of San Jose, actually next door to work and continue to work with them. We're part of the ecosystem since many years. That hasn't changed. We've moved high volume because their high volume manufacturing is also no longer in the U.S. to Penang, which is again close to the end customers manufacturing. We're actually, like Christian said, really well positioned with our global footprint, independent of where the markets go for the moment.
If we were to need to do any manufacturing or relocate it to the U.S. , for us, it is a very easy exercise. Very easy exercise and very fast exercise.
Good. Thank you very much for your interesting question. I'm sure there are a lot more questions you have over the next few days, weeks, months. You have our address from the IR department. Just call there or send an email. Thank you very much for being here or following us through the webcast. The people in the room obviously have now the opportunity to continue the dialogue with management during the apero riche that will be served in a few minutes. Thank you very much for joining us and have a good rest of the day.