Good afternoon, everybody. Thank you for being here and also to be connected to our first Capital Market Day. I take only a few minutes in order to highlight our agenda. Sorry, Stefano, can you give me this one? Perfect. Okay. To highlight the agenda and the panelists which are with us today, our CEO, which highlights the vision and the strategy of the company, and to Stefano, we'll follow Marco Beretta, our Chief Commercial Officer, who will give you the market perspective and our feelings about the main segments of Technoprobe. Then he will hand over to Joe Parks, our Chief Technology Officer, who will give you more details about technology and testing, and then at the end, financial outlook with our CFO. Thank you again for being with us, and I hand over to Stefano Felici.
Thank you, everyone, to be here today. It's a pleasure to meet all of you, and today we'll try to give you some additional information about our vision, what we are doing, and to let you understand that you are investing in the right company. Thank you for all investors who invest in us and believe in us. Let me go with our vision. First of all, I'd like to say that our results are very positive, and more than any number for us, what is the most important parameter is the market share, because the market, as you know, can go up and down, especially in these days with everyday different announcements. What really we need to care about is the market share.
For now, in 2024, we could achieve this incredible result of 60% of market share for the, of course, logic and advancements technologies. Another thing that I'd like to highlight here is we have very big supporters and sponsorship by the main semiconductor companies, like in this case, Intel and TSMC. In TSMC, last December, I was invited by TSMC because Technoprobe won an award as Best Supplier, Excellent Performance Award for 2024. It was presented to me, and I'd like to say that Technoprobe was the only company about testing that could receive this award, special award. Just to give you the importance of this, TSMC has more than 15,000 suppliers, and only the best 30 could get this type of award. Technoprobe was the only testing company.
It is very important for us and showing the commitment and the result, the very good result we achieved with this company. Another thing that is really new, and I just got the permission to share with you, was the 2025 EPIC Supplier Award that we just got from Intel. Intel Supplier Day will be next June, but we already got the permission to share with you. Last Friday, I got the permission to announce this to all of you. We also, for this year, have been awarded with this special award. Also, in this case, for this year, we are the only probe card company that can receive it. This is just showing you that we have very big supporters and we are doing the right thing with the main customers.
Our main, let's say, vision here and target how to achieve this type of results is the innovation. Every year, we develop new technologies, and we have over 600 patents, international patents, and we are applying for new patents every year. This is really the core of our strategy, innovation. What is good now is not going to be good for two years from now. We know for sure. That is why we always develop Probe Cards and solutions for the following years. Right now, we are already working for what is going to be needed in the next four or five years. As far as global footprint, we are everywhere. We are a global company. We are in Europe, United States, Asia. We are very close to all our customers because it is very important, the uptime for all these main customers of the Probe Cards, of the testing.
Our strategy here is to be very close to them in order to minimize this type of uptime, I mean, the downtime of the probe cards, and allowing them to save money. That is why we are everywhere, close to all the customers. This slide is, I'd like to highlight here, you can see the revenue and comparing with the SEM, the serviceable available market. I'd like to, first of all, to say that you can see here that Technoprobe growth was faster than the market. The market in the last five years grew by 11% CAGR, but Technoprobe was growing by 20%. This is, of course, a very easy explanation because of market share. We gain market share year- over- year. This is again because of technology.
We develop technologies, and we also went through strategic acquisition like MicroFabrica, and then Harbor, this tech. Finally, we have some partnership also with the major tester companies, and I will go through this in the next slides. Basically, where do we play? We play in the front end, what we call front end, in the wafer testing level, and back end on the final test level. For front end, what we mean is wafer, of course. This is a standard way to test chips at the wafer level. More and more, we see more a need for single-layer die test and advanced packaging test because it's really our customer really needs to make sure to have non-good die. This is not possible for many reasons that also Joe will explain later.
It's not possible to do this just at the wafer level, but you really need to go to on single-layer die testing. On the final testing level, of course, we have socket and system level testing with device interface board, and this is more on the TIS business side. Talking about wafer level testing, this is just to give you a simplified overview of the applications we are testing. You can see the non-memory on our side, and Technoprobe is really, we are expert. We are focusing so far, we focus on the non-memory testing. HPC, RF, you have sensor power, and ASIC and GPU CPU. On the other side, you see memory with DRAM, HBM, and other memory. Till now, our business, our core business was related to the non-memory, okay, and not so much on the memory. Why this?
The reason is simple. The non-memory required us to develop new technology, was much more complex, and it is where really the MEMS technology could make the difference on the non-memory testing. That is why also you can see it is more attractive as far as margins and so on. Memory, till now, was possible, let's say, to test the memory chips with pretty standard probe cards, so let's say not requiring the level of complexity that non-memory is requiring. The good thing here is, especially with starting from HBM, we see that this is happening, a transition, and also for memory chip is going to be required more complex probe cards. They are becoming more complex, the chips. Just to give you, making an analogy, in the past, also for logic, the main technology was cantilever. I am talking about before the vertical probe cards. It was 100% cantilever.
Memory now is 100% micro cantilever. It's not even a vertical pin. Logic was exactly the same, was 100% cantilever. Because of the complexity of the devices, the semiconductor companies decided to or needed to put testing pads everywhere in the array of the chip and increasing the complexity. It was impossible for cantilever technology to follow the roadmap. That's why came the vertical technology and then finally MEMS. For memory, we are still in the phase that micro cantilever could work, but running out of steam in our vision and requiring very soon to use only vertical solutions. That's why it's very interesting. As you know, we already said in other different calls, we are engaging with the main players for memory with the idea to change really, to transition from an existing technology to a vertical MEMS technology. This is the idea.
We see that there is a lot of good traction there. Really, the interest is there because the vertical technologies will allow to test much more complex chips, also for memory. I guess that probably most of you know already the structure of Probe Card, but I'd like here to give you some piece of information. You can see here, basically, that this is a very simple schematic about the main part of Probe Card, starting from what we call the probe head, where we have the probes inside guide plates, and then we have an interconnection, the main PCB, and some stiffness mechanics supporting the whole structure. Typically, the core technology of any Probe Card was just related to the probe head, where basically we insert the MEMS pins.
This has been and is our main business, still is our main business, how to develop new MEMS probes to satisfy new requests. Right now, also this interconnection in the past years became a bottleneck for us. Why that? You can imagine to have down here the wafer under test, with a very complex geometry. Let's say testing pad and structure with very close to each other and many hundred thousand, can be up to 100,000 pads to contact or bump, same. You can imagine here a vertical probe going from the wafer up, just vertical. This probe should put in contact a pad on the wafer and a pad that will be here in the interconnection, lower part of the interconnection. Here is a problem.
I mean, I don't want to be too much technical, but it's very understandable that here you have a wafer that is done with wafer technologies. It's very, very with very fine geometries. It's very, of course, you know it's very, very complex, and you have geometries that only wafer technology can reach. Here, how you can do this? Because this is you have a probe in vertical going here. It means that the lower side of this interconnection will have the same type of geometry of the wafer. This really became the critical part because till a few years ago, this interconnection was done by suppliers using technologies similar to PCB technologies. It's not wafer technologies. How you can keep up with the PCB technology, keep up with wafer technologies? The two technologies are not talking to each other, are not compatible.
That is why, and this was the big investment we have done in the last five years with GWAY. We created from scratch two new wafer fabs. One is 18-inch and one is 12-inch, just in order to make this interconnection. This interconnection is not based anymore on PCB technologies, but it is based on wafer technologies. We are doing this interconnection using wafers. We can say we are the only company here, probe card company, that can do that, I can tell you. That is why I am very confident also that we figure out how to follow their roadmap in the future because the critical part is not only the probe head anymore, but also this interconnection. It became very, very critical. Of course, you have PCB.
PCB remains more a commodity type, even though it can be very complex, but I would say that several suppliers can achieve the right complexity. You have mechanics. Of course, after this, Joe maybe will have more explanation, and if you have a question, we can go over the question. I just want to give you, and sorry if I've been too much technical, but some to understand why we are doing things and why we are investing so heavily also in this case for this interconnection because it was the only way really to follow the roadmap for the future. Just to give you a summary, here is, you see again different parts of the Probe Card. Where you see our logo is basically our capability to do this internally. We vertically integrate the process to do this internally.
The reason is again because it was too difficult to find technology outside and was impossible basically. We brought all these technologies inside and developed this technology inside, giving us also a great advantage not only for technology, but also cost. We are still continuing with this, let's say, strategy for the future. The only thing that you see with partners following, let's say, using partners is for the PCB, as said before, because it's more a commodity. The assembly of all these parts together is becoming also very complex. That is why the rationale of Harbor that now is not used anymore, but the rationale of Harbor was to bring in-house some technology to assemble all these parts with special technology to assemble all these parts together.
To build, let's say, just the PCB, we achieved an agreement with Advantest, and you'll see the next slide. About Teradyne and Advantest. Both of them decided to invest in Technoprobe. We are very glad of this because we like to have industrial partners like them. Those are the only two tester companies, basically, the major ones. Probably they share 99% of the market among them. Both of them have some strength and weakness. They do pretty much, they can do logic and memory testers, both of them. Of course, for several reasons, for example, Advantest probably is a little bit stronger in AI or Teradyne in different areas. Bottom line is that both companies will be the only companies doing testers, and Probe Cards are becoming a critical hardware for them to develop new testers.
In the past, these companies were not talking to Probe Card suppliers to develop testers. They could develop testers and then give the package to the Probe Card suppliers and say, "Okay, this is the package. You can build the Probe Cards for the customers." It is not working in this way anymore because without this company, without knowing what the Probe Card can do or what the Probe Card can achieve, they cannot develop the next tester, the next generation. That is why it is very important for us, but also for them to be a partner. Make sure that we align our roadmap with their roadmap. The next generation for them should have at the end a working solution for the Probe Card.
For us also, it's a key point to understand and to drive this process in order to be there when the new tester will be ready to have the right solution. As both companies again, they will still be in the market for many years and they will share the market. For us, it was really important to collaborate with both of them. What we see for the future, of course, you can see the three main challenges we see, the evolution of chip technology, the increased complexity, and of course, client satisfaction. How to address this is we have just one recipe for us. Of course, it's innovation, R&D. M&A, again, is used for us as accelerator of technological development. As I said before, we like to develop technologies in-house.
We are highly vertically integrated, but M&A can accelerate the process in case we need some specific technology. That has remained our target for M&A. We have strategic partnership like Advantest, Teradyne, just to mention. Vertical integration, support on site. Marco maybe will give you more color on that. Commercial agreement. These are our answers. What's next? For the front end, the wafer testing level, our expectation is to continue to be leader in the market. What we see the challenge is coming is in high-speed, high-voltage RF, silicon photonic. For sure, as mentioned also before, we see some good traction also in the memory because of the technological transition. For backend, what we expect is to strengthen our position in the final test, basically.
Now let me leave the floor to Marco, our CCO, for more comments and information about the market. Thank you, everyone.
Okay, thank you, Stefano, and good afternoon, everyone. I'm also very glad to have you here. I will start from what Stefano just mentioned about complexity, about new challenges presented by the advent of advanced packaging and no good dies. Joe will add a lot of colors to this later on. Whenever I see or I hear a word challenge, then I'm very happy because this means opportunity. We think we have big opportunities for us to continue with the growth and to expand Technoprobe further. If you look at this chart and we look at what happened in the last few years and what we expect to happen next, we see basically the Probe Card market to continue in its growth.
We grow at about the same pace in the next few years as it was in the last few years. That is the only analogy because what was the driving force in the past was completely different from what we see today. In the past, around the COVID pandemic, there was a big need for consumer laptop, mobile phones, and so on. This was the driver for the past growth. However, in the future, we see AI and advanced packaging that will drive the growth and will drive the increased demand in Probe Cards. We see this as a big opportunity again, and I will try to get across a message here that we are set to succeed in this and capitalize on this opportunity. Let me go a little bit in details. Stefano already mentioned about the split between the two main categories, memory and logic in Probe Cards.
We see the logic to be the biggest market, and we expect this portion to grow for the last in the next few years, reaching a remarkable number in terms of a value market for us. Obviously, it was already mentioned that memory is also interesting. I will comment on this later on because HBM especially is becoming very interesting for a company such as Technoprobe. We think we will have Joe will add colors to this again, but I think we have some very good idea and approach on how to be present in that market as well. Another level down, we can split the market by technology. Stefano was already referring to this when he was talking about MEMS. MEMS definitely is the fastest growing segment on the logic Probe Card. This is the focus of Technoprobe.
It's always been the focus for the reason that we already mentioned. We believe this is going to grow also. Eventually, it will displace any other technology. Cantilever and vertical will be definitely displaced by logic MEMS. Again, we want to continue to be leader there and continue with the supremacy in our product. Now, looking at the market share, the market share was already presented. However, if we look at what we could do and achieve in 2024, and we base the market share on the overall Probe Card market, it was about 34% the logic, but we actually serve our server volume market is logic MEMS Probe Card. On that, we reached the remarkable number, 60% that was mentioned before. A few comments here. 60% means it's a nice number, but there is another 40 that we can try to get. It's my job.
I hope we'll be able to do so. Obviously, to reach 60% market share was a combination of technology for sure, supremacy in technology, being better than a competitor, being faster in developing, but also a very strong and long-lasting customer engagement that we have been developing for the last few years. That also allows us to get to know what the customer needs, try to achieve what they need in the fastest investment way possible. This is a combination of all these aspects to reach this remarkable number. Talking about final test, also final test is an interesting opportunity for us. Before the IS acquisition, Technoprobe really never addressed that market. After the acquisition and the merging was completed as of last year, we started to be pretty active there as well. The market can be split in two.
We focus on the device interface board, not on the burn-in. We also see this market that is going to grow in the next few years. The opportunity is there also for final test for us. Market share-wise, on the overall market, the market share was 6% for final test. If you look at the real market we serve, which is the device interface board, it was 8%. Obviously, this is a more fragmented market. There are many more players. 8% means there is another 92% we can try to get. I look at this in a positive way. Why am I positive? First of all, by having the IS as part of Technoprobe, we can now increase the number of opportunities for final test. Before, the IS as a Teradyne company could only deliver boards for the Teradyne platform.
Today, we can also, we're already delivering to Advantest platform. The number of opportunities increases immediately. Even there, we want to have the same approach we have been using for the last few years on Probe Cards. Deliver the best-in-class technology that can be a real game changer for us in the way we address this market. Joe will comment on this on Fusion Link. That is definitely something we believe can be a game changer for us in final test. Last but not least, we were scratching our heads in order to represent graphically some different numbers we are getting about HBM opportunity from various reports. There is no unique number over there. It can be anything. This $10 billion, $15 billion, $20 billion market in a few years. For sure, it's a market that is about to grow.
There is a big IPO, obviously, on HBM. What we think is important and interesting for us, it was already mentioned before. There is something that is changing in this market as well that makes Technoprobe potentially be a D-player because there will be the need for a transition from micro cantilever to vertical. This is our specialty. We believe we have a big opportunity, and we definitely want to capitalize on that. As we speak, we also have active customer engagement with all the players on HBM. My key messages, I hope to get a very clear message across. I think we definitely are committed to grow. We have all the way to grow, and we will base our growth on three pillars. We will continue to focus on MEMS logic Probe Cards. The segment that we saw before is growing, and we have a supremacy.
Our market share is a nice market share, not enough to me, so we can grow there, and we will grow. On final test, as said, we have opportunities now to expand on Advantest platform for device interface board, and we have Fusion Link as a potential game changer for us. In HBM, obviously, an interesting market, a memory that can resemble vertical and the logic. So we definitely want to play a key role, and we will play a key role. I'll let you, Joe, go over your presentation and give more colors to what I mentioned here. Thank you.
Thank you, guys. My name is Joe Parks. I'm the most recent addition up here. I've been at Technoprobe for almost seven months now, but I feel like I've known Technoprobe for a long time. My first engagement with the company was in 2016.
Ever since then, I've really admired the technology of the company, and I'm honored and privileged to be part of the leadership team here. Following up with what my colleagues have said, we'll talk a little bit about Front-end test and Back-end test and how we are doing technology development along both of those lines. Starting with the driving advancements, continuing logic semiconductor test advancements, obviously, this is our key core business today. Like Stefano mentioned, it is extraordinarily important that we stay ahead of the competition and we continue to have customer engagements to satisfy all the customers' requirements as they are advancing through this. Let's spend just a minute talking about advanced packaging. There's been a couple of conversations about advanced packaging so far today. Just trying to drive home what is motivating advanced packaging in the industry.
There are really two key vectors that are driving this. The first one is we continue to see some slowdown in the rate of transistor density scaling. Simultaneously, as we try to put more and more transistors into a complex, we are continuing to bump up against full reticle size limitations of Fab photolithography. When you take a look at both of these combined, plus the yield impacts associated with it, obviously, your larger die is going to have a lower die per wafer yielded output. There is a huge push to move to an advanced packaging or a wafer-level assembly type of application. Technoprobe, I think, is very, very well positioned to monopolize and to grow into that space. Here is an example from Intel that has been around for a number of years now, about one of their HPC devices that really drives home the die disaggregation.
For a device like this where you're putting many transistors into the part, into the complex, this gives an opportunity for having improved wafer-level yields, deploying smaller chiplets, reconstituting yielding die from multiple wafers back into a full complex. It also allows you to do some things that AI companies are taking a lot of advantage of right now, which is mixing functionality capabilities. Putting CPUs, GPUs, and HBMs all within the same complex and in the same stack gives a much better power performance profile as well as yield. I think the next thing that we're going to start to see is more in the line of co-package optics or silicon photonics coming into the play.
The disaggregation or advanced packaging really plays into that space as well because you can take technology that ordinarily would not fit within standard silicon processing, and you can drop it in actually relatively easily into a co-os type application. The last thing here is it allows for mixed technology node application and IP reuse. I think we're going to see more and more of that as you're pushing your most complex applications to the newest technology nodes, and you're keeping some of your other potentially input-output drivers and GPU-type devices maybe one step back in the technology node profile. Of course, I'm talking about the GPU high-performance compute type applications here, but this sort of technology is also getting applied generally across mobile client server space as well.
We firmly believe that this is going to be the wave of the future, and we need to be in a situation where we're prepared to handle it. I'll describe some about how we are. First, let me just walk through, I would say math, but it's not. It's cartoons about how the technology and how wafer-level assembly really requires advancement in probe card and probing technology. I'll walk through a couple of build foils with you here. The easy example, of course, is you have one die that goes into a substrate. This is what we've lived with for many, many years, one or two die into a substrate.
In the scenario where we have listed here, not good die, but in reality, what it means is it's a die that we said was good at front-end test, but ended up being bad after it gets put into the substrate, runs through burn-in, runs through final test. In the situation where you have a couple of percentage die loss associated with this in a single chip package, it's not a terrible situation to be in. Any misscreening of sort does directly impact the final yield, but it's going into a relatively low-cost, generally organic substrate package, which in and of itself is not that expensive. The overall return on investment or opportunity loss is not great. You end up still with relatively high yield. There is a lot of discussion about balancing the front-end coverage, increasing front-end coverage versus the relatively low substrate cost.
To this end, generally, you want to have good screening. You don't have to have perfect screening of your device. However, when you go into advanced packaging, chiplet-type architectures, the calculus changes very, very quickly. This is an example where I have, say, four die going into a package. Here I have selected one of those die being misscreened at sort. Sort says it's good. However, once it gets into the final application, there's something that happens. It's an infant mortality problem or a VMIN or a margin failure that causes that die to be bad in the final application. In this case, it's actually a relatively large impact because for each misscreened die, the chances are the other three die were healthy and good. We're not just throwing away the relatively inorganic package, but we're also throwing away three perfectly good die.
The cost becomes much, much higher. As I mentioned a couple of times, a relatively inexpensive package, that package actually starts to become not so inexpensive in the applications in the AI space and so forth. Here, front-end test coverage actually becomes extremely critical. If you can get fractions of a percentage better screening at front-end test, it pays huge dividends when it comes to final unit to die yields of the assembled product. Here's another way of describing it. If you take a look at a single die with a yield loss associated with it, we have the single die represented as the square. The little itty bitty red square in the bottom represents the amount of cost that results from having a misscreened die.
When we get into the more advanced packaging or the disaggregated space, your manufacturing cost actually scales, maybe not linearly with the number of die, but it definitely scales with the number of die, especially when you start to add very expensive HBM into the co-os stack. This manufacturing cost goes way up. The cost of yield actually goes up by the power of the number of die. The manufacturing cost goes up linearly, roughly, and the cost of yield loss actually goes up to the power of the number of die. Just representing how when you start to put more and more die into the same complex, the necessity of screening those die becomes very, very important. What does that mean?
That means we need to be able to drive higher voltages, higher power, manage thermals to a greater extent, provide better opportunity for increased content, so either toggle content or high-speed I/O content. The last thing is spending more generally test time screening the devices. For all of that, probe cards are required. For every additional second or every additional die, probe cards are required for that. Again, the role of test here is we're really emphasizing the need for additional front-end test coverage. Known good die is absolutely required. Known good die means different things to different people. Known good die, in my perspective, is when I call it good at front-end test, it does not fall out even at speed at final test. Really having known good die. Again, how do you do that?
You do that by maximizing voltage, temperature, content, and time. All of those have either technical or capacity requirements for probe cards. Again, the solution here is accentuating more test at the wafer level or singulated die level. That, of course, means more probe cards. Moving on beyond that example, how are we as Technoprobe well positioned in order to really be an enabler for advanced packaging? There are several ways that we're doing that, and I'll highlight a couple of them here. Up in the upper left, you can see that we talk about fine pitch and ultra-large pin count. As we move into an advanced package architecture, the bumps or the pads that are used in order to do that die-to-die interconnect become smaller. The smaller pads, the higher your bandwidth, the lower energy cost per transition going between die.
There's going to be this acceleration or this drive to move to finer and finer pitches, less and less conductivity between the die in terms of metallurgy, capacitance, and so forth. Additionally, going to larger pin count. If you remember the graph a few pages ago, a lot of these dies, GPUs in particular, are really pushing the limit of reticle size. To that end, more power is going into the part. We have a smaller pitch, greater number of probes. It's really important for us to be able to effectively probe at tight pitch with very, very high pin count. This is something that plays well into our development space. We have roadmaps across multiple families of probes where we are pushing capabilities into lower and lower pitch while still maintaining contact resistance, force, lifetimes, and so forth. The next one is high speed.
As we look into the future, high-speed I/O becomes more and more important. HBM technologies are driving, PCIe, DDRs, all of these things are driving to higher and higher speeds. In order for us to test that effectively, it really requires us to go to a shorter needle and to have RF or high-speed I/O specific needles in order to manage the input and output of the device, including silicon photonics. I'll talk a little bit more about that in a minute. The next is having high-density interconnects. Now, Stefano talked some about this. This is really figuring out how you can get from that tight pitch at the needle level up to a looser pitch at the probe card level.
A lot of this has to do with the MLO, GWAY technology, where we are advancing things like stacked vias, finer and finer pitches, greater number of layers, and so forth. The final part over here is high power and thermal. Everybody is interested in pushing the power envelope, getting more and more current into the device. Of course, when you get more current into the device, you have to have a way of extracting the heat out of that device. We are well positioned in both of these areas. Our needles, we continue to push higher and higher current carrying capability across a series of pad metallurgies and doing that in a very reliable way.
We have technologies that allow us to do heat dissipation both at the probe card or MLO level as well as at the die level by extracting heat through the thermal head on the die. All of this combines actually to make a very, very good means of giving us a good enablement as more and more of the industry moves towards advanced packaging. As I talk a little bit more next about high speed, high voltage, radio frequency, or high speed I/O as well as silicon photonics. This graph is very similar to many that I'm sure you've seen before. The world continues to get hungrier and hungrier for wide-band applications. More and more applications are pushing into higher speeds. Again, HBMs, PCIe, more and more parts are showing up on the device.
The interconnectivity into server farms and so forth becomes more and more important. We will talk a little bit about all three of those. As we look at the increased number and types of interconnected things, this is really pushing into things like RF applications. We do have RF technology that we are partnering with several companies on that gives us pure, true, good RF interactions as well as high-speed I/O technologies for more digital-type parts. Silicon photonics, I believe silicon photonics is going to be something that we see a lot more of in the next 5years- 10 years. As chiplet to chiplet interconnects continue to drive, I think we are going to see more and more limitations from the plug-in photonic devices. People are going to be more interested in driving that fiber straight to the stacked unit.
We need to be able to test those picks as well as stacked units ourselves. With respect to silicon photonics, our technologies, like I have already said, enable fine pitch probing. That gives us good alignment with the advanced packaging roadmap. Radio frequency probes, this is something I think is very important. You cannot do silicon photonics in a vacuum. You have to do silicon photonics in conjunction with high-speed probes. You have to be able to feed the part an electrical signal at the same time that you feed the part an optical signal in order to do the test. We are well positioned there. The next one is we really want to continue to use standard wafer prober, so vertical probe cards, standard ATEs. We act as the personality between the ATE and the silicon.
We also want to make sure that we maintain compliancy with the customer's preferred test method. Whatever their ATE is, whatever their probe solution is, we really feel like silicon photonics being part of the personality of the device, just like the probe card is part of the personality of the device, is going to be very important. Lastly, pushing into higher density and higher speed I/Os is using some of our phantom technology in order to help meet customers' needs for high-speed measurements and high-speed loopbacks during front-end test. Both of these gentlemen talked about HBM and DRAM. If you take a look at the graph on the left, you can see that HBM, we believe, market continues to grow.
What's very important about this graph is how you can see some very quick gen-over-gen migration up into HBM gen 3e, gen 4, gen 5. Why that's important is this is going to drive a lot of high bandwidth applications. When you get into the 9.6 Gbs per second type transfer rates, even though that rate might not necessarily be used during test, the higher bandwidth requirements become more and more important. People are really starting to look at how we can do better tests at these high rates. The way to do that is to migrate to a short vertical MEMS probe in a full wafer contactor type solution. This is where we are driving a tremendous amount of our focus, really looking at next generation products and driving the MEMS into that space.
Finally, as we talk about final test, we've already talked a little bit about Fusion Link. This was the big technology that came with the Harbor acquisition. This is an area where we think that there's a lot that can be done in the probe card space. This is effectively using the same disaggregation, if you will, but at the test hardware level. We can take the right material, the right chiplet, if you will, and do bonding of substrates together to where we can have the very best main motherboard coupled with the very best device substrate, coupled with the very best probe substrate. Those can all be different materials, and we can do a performance and cost optimization with this technology.
To summarize where we are today, we have a very strong position in advanced packaging probing. We have a strong position in Fusion Link. In the future, we've got an additional push into radio frequency and silicon photonics and in HBM. These are new markets that we think that we can grow into. At the end, we end up with an expanded leadership across a very large portfolio of both front-end and back-end test. Thank you very much. I'll hand over to Stefano.
Thank you, Joe. Thank you, Marco and Stefano. Thank you very much. Let's get to the point. In the previous sessions, we have addressed two of the three main pillars where Technoprobe is based, technology and market trends, and where Technoprobe will base the growth for the next two or three years. We talked about co-ops. We talked about the new packaging architecture, which are expected to grow both in terms of volumes and also in terms of complexity.
We talked about market trends segment by segment with particular attention to the impact of the AI that is expected to be pervasive in every end market where Technoprobe operates, including, of course, the high volume market like consumer goods. Technoprobe has always been linked to these two factors, so technology and volumes and markets, because they remain what we can really control of our business. When we started preparing this deck, the third point was not considered so decisive like it is in the current weeks. All of you and all of us experienced in the last couple of weeks, in the last month, the vulnerability of the markets, the vulnerability of the capacity of every company to plan their own business.
That is why this is a key point right now. Unfortunately, this is a point we cannot really control. Once again, that is why our strategy, our business strategy, and consistently our communication strategy remain about concreteness and prudence because we want to focus exactly on what we can control. Where we are now, 2024, on the left side of the page, you see the actual figure divided into three main blocks in terms of market share, market percentage of our total revenue. From the bottom, you can see consumer goods, AI and data center, industrial and automotive, really simplified. What we can expect for 2025 is on the right side of the page. Organic growth, mid-single digit. This is the best KPI we can use to estimate our revenue and our growth.
The segment growth, we can see clearly from this page, AI and data center will lead the growth for 2025 with a double-digit range of growth that will bring the total revenue, the percentage on the total revenue of Technoprobe in order of 38%. The other segments are expected to be stable or at the middle single-digit growth. By the way, if we want to consider the annualization of DIS for 2025, meaning five months of annualization, we have simplified the model and put all of them into the consumer goods. That is why it seems a bit stable at about 49%-48% of total revenues. Looking further, this is a mid-term scenario.
We never discussed about this in the past, but we truly believe this is now the right moment to address where we are going in the mid-term scenario, meaning three years- four years. On the right side, you see total revenue in the region of EUR 850 million-EUR 900 million, all of them driven by the growth of all the reference markets. We expect in the mid-term, all of them that are currently suffering today will bring a contribution to the volume of our business. Again, our business is definitely linked to volumes of the end markets other than technology. What we can tell about this, again, is that this target model, according to our estimate, is achievable considering the current structure we have right now, meaning that we do not need further extraordinary investments in infrastructure, in CapEx especially, but only ordinary investment to maintain and sustain our business.
This is even more evident in this page when we can see the strategy growth of our profitability. We ended 2024 with a 25% EBITDA. We expect, as already announced, the Q1 in the region of the 30s. We expect, again, our target model to reach the level of 38%-40% in the mid-term. You can clearly see the operating leverage and product mix blocks that can lead the growth, especially the operating leverage, highlighting again the scalability of our business, assuming all the investment we have done and we have achieved in the last couple of years, two, three years. The product mix will bring benefits. Again, this is more a mid-term benefit because to benefit from that, we need to complete the vertical integration of the recently acquired business together with the development of the new products we have in pipeline. We talked about HBM, for example.
While in the short term, we will see additional costs for integration and additional costs for the innovation and ramp-up costs always linked to what we have already showed in the prior presentation. Finally, investment, this is an important moment also to highlight what we have done in the last three years. Significant investments representing more than 15%, up to 18% of the revenues, a significant portion of the cash flow. Everything has been self-financed. Most of them, as you can see in the cake on the left side, address it to the increase of the CapEx in Italy and in Taiwan to the two fabs we have built for the production of the MLOs and the MLC. Our target model, and this is already starting in 2025, will be much lower compared to the share on the revenue.
We expect to invest in the region of 8%-10%. This is not to give up about the investment, of course, but this is to sustain the capacity, the automation, and the renewal of our equipment. That will be another key area. This is all I have. I hope you really enjoy the duty-free lunch, the villa, and also the presentation. Thank you very much. We can now move to the Q&A session.
Yes, hi, Fabio Polan, Mediobanca. Thank you for this presentation. It has been super helpful. I would have many questions. Let me just start with two. First of all, my take from this presentation is if the answer, the question was, what's the real driver of growth for Technoprobe putting together innovation, technology, new products, and high customer satisfaction level? Probably the conclusion is that you need to have everything together. You need to have clearly good relations with the customers, but also you need to remain on the forefront for innovation. Probably this is why you feel confident in increasing your market share regardless of what the market and the industry could do in the short term.
The second question is, in order to sustain this strategy, and given that you are sitting in a very decent amount of cash pile, which area would be better supported in terms of external growth? Thank you, if any.
Yes, maybe I'll start from your first question. Yeah, I can tell you I agree with you. There is a bunch of factors that are contributing to our success, as you said, innovation, but not only that, also customer relations and so on. Probably I'd like to say what's behind this. I think probably if I have to summarize why, without saying exactly what, I think it's a matter of mentality. This is what is making the difference in all the points you just mentioned. The mentality is that we always take the challenge, the next challenge. There are also many Technoprobe employees here, and they know that nobody can say impossible to any customers. Otherwise, they get fired after that. No, just kidding, but this is almost true.
If we get any request that can be really difficult, really, really difficult or almost impossible, it's an opportunity. It's considered really an opportunity. We're happy, actually, to get it because we know that if it's very difficult, we can make the difference. This is, I would say, valid for technology, but not only for technology. Anything we do is, I think, this is the right mentality. That's why at the end, we can be successful. This is not your question, but this mentality, I think I have this mentality also regarding the market situation, which is not so stable in the recent day. Maybe I'm anticipating a question later, but also in this case, I really see this as an opportunity for us, actually, and not worry. Of course, we cannot, let's say, drive the market.
Of course, the market is something bigger than us, but I can make sure to be ready for the next big wave. This is, I'm sure. That is why also the stability of the market is not worrying us because I think we have the right mentality to move forward. As far as the second question is.
Yes, about the external growth. This has been always an option, not the option, but one of the options we have in our strategy. We have demonstrated in the past years, in the recent years, that sometimes we prefer to invest in ourselves, and sometimes we needed to catch the opportunity to grow externally. We are now in a very sweet spot. We have cash enough to take this decision if it happens. The market now is on a roller coaster. There are opportunities that probably six months ago should have been not considered because of the high prices. Right now, we are looking for some opportunity again, but as always, the important thing is to be ready to catch the opportunity because this operation, this M&A happens very quickly.
For example, what happens with Harbor Electronics, with DIS, the discussion was open one day and closed one month later. In just one month, you have to close the deal. This is what we expect again for the next couple of years. The market will be very volatile and very, very quick to move. We are in a good spot. There is a one page where we show exactly how we are vertically integrated right now. For our production, we do not expect in the short term the need of new acquisition, external acquisition. Again, if there will be an opportunity to catch some technology and not some company, because what we look at is technology. If there is the opportunity, we are ready to catch it.
Yes. Second question. Sorry. Okay.
Yes, good afternoon. Gianmarco Bonacina from Banca Akros. First question is about your 2025 outlook for organic growth. Just doing the math with, let's say, zero growth for auto industrial, small growth for consumer, and double-digit growth for data center, I get more high single digits. Just what's your definition for mid? It can go up to 8%-9%, or is it 5%?
I'm not sure if you have summarized also the growth of DIS, probably yes. The five months missing for the annualization.
I'm just looking at the percentages, so the year-over-year growth. Okay, but maybe we can have a follow-up.
In terms of segments, this is exactly what we show in the pages. We expect that when we say double-digit for AI, we expect even more than 20% for the growth. Low single digit, of course, in the region of 2%-3%. Middle, of course, 4%-5%. No more than that.
Okay.
Overall, if you summarize firstly the annualization of five months for DIS, you will arrive to organic 5%, middle single digit.
Okay. The second question is about your target model. If you can give us an idea of the, let's say, different products, because basically today, clearly, your main product is wafer testing for logic. So what could be, let's say, a range for the testing for HBM and maybe for the other products that you are planning to, or if you want to say in percentage, just to have an idea. Thank you.
Starting from the HBM portion, this is very dynamic right now, also because we do not have technology ready for tomorrow. We are still in a qualification phase with all our customers, potential customers, of course. We do not expect, as said already many times, any revenue for HBM in 2025. We expect the qualification at least. We need time to ramp up the production, to industrialize the product and to get the first orders from customers. This is a key point because customers are not only customers. We consider them definitely partners. They need to decide, let me say, to change technology, to move further from cantilever, micro cantilever to vertical lens. This step is not easy to be done when the market is in a difficulty, where there is no ramp, where there are no volumes, especially for the edge AI.
We need also to, let me say, wait or to convince them to move to this technology. The target, when we show the mid-term target between three to four years, that really depends on when the vertical lens for HBM will become a dominant technology. Can be 2027, we do not know. Can be 2028, we do not know. 2026, we do not know. The volumes, again, are the driver. There is a bit of uncertainty. There is, of course, a portion of this in our model. It is something we prefer not to disclose because we probably will be taken as misleading. For sure, the bigger portion of that growth will be driven again by our core business, Probe Cards, and the DIS business for the PCB solutions and final testing.
Good afternoon, Domenico Ghilotti from Equita. I have a question on the final testing market because you mentioned that you are enlarging the addressable market, also thanks to Advantest agreement. I am trying to understand where are you really in getting new revenues from Advantest, in the commercial relation and in unlocking this opportunity. Maybe just a follow-up on the HBM, that is another, say, key component, key driver for the medium term. If you can give us any color on the qualification process ongoing.
Okay, maybe I'll answer to the first one. When I said, I mean, once the acquisition was complete and as part of Technoprobe, we started immediately with opportunities for the Advantest platform, and we are already having business there. There are opportunities we are capitalizing on. There are businesses we are already doing. We want to continue to expand there because it's a nice opportunity for us. It's already a business we are having.
On the HBM?
Yeah, on the HBM, maybe I add some comments to what Stefano already said. The point is that here we're not trying with a similar technology like MicroSpring to gain some market share. That would be much more predictable. Let's say we try to go with the market with the same technology used by other competitors. Okay, at that point, we could say, okay, we can predict we can gain 5% market share here or 10% after, but this is not the case. Here we are in a situation where there is a turning point, a transition point, that it's a kind of on-off situation. When really, and it's difficult to predict exactly the day or the month when this will happen, but it's on-off.
That means that when the customer will decide to make this transition, very quickly, they will from zero go to 100% because it will be the transition. We have been in this situation in the past, also for Logiq, where it's going to be after the qualification is done, it's going to be very challenging then to be ready for the ramp because it will be very, very steep. That's why, again, it's difficult in the target model. We think this will happen, and we cannot give exactly, predict exactly the month or the year. We see this coming, and we are, that's why we are preparing not only the technology, but we are also preparing the capacity because we know that when that will happen, there will be huge, huge demand immediately. I don't know if this is answering to your question.
Next question.
Hello everyone, Giovanni from [Bermeguia]. Thanks for taking my question. I have two. The first one is maybe you spoke a lot about how the interaction with Teradyne or Advantest has changed through the years that previously was more like, okay, give me the Probe Card, now it's more about interacting and so on. Is there a chance that one day these guys will want to have this internally and say, okay, I'm going to buy the whole thing instead of keep asking Technoprobe what are the features, right? That's my first question. The second question is that you spoke about that Technoprobe is always interested in new technology. What can they bring in?
Now, without disclosing maybe too much, is there an area where you see, oh, I'd like to invest there or without saying names, but is there a technology that you may like to invest in or acquire or whatever?
Okay, so about the first question, yeah, it would be interesting to ask this question directly to them, but probably they might say yes. I mean, it's a matter of fact that both of them invested in Technoprobe. So it's a kind of sign that the Probe Cards are becoming a critical hardware also for them. Okay. On our perspective, of course, we are in a very good position to play with both of them and be partnered with both of them. Anyway, I cannot exclude that that can be one of their plans, of course. I don't know.
About future, yeah, we are working on many new technologies. Again, we have a vision that go from here to for the next five to ten years. Without, let's say, disclosing too much secrets, of course, three secrets, but we can say that one of the challenges for the future will be the power, power delivery and thermal dissipation of AI chips and CPU, GPUs. They're going to be very difficult to deliver power and make sure that we have the right thermal dissipation during test. As said by Joe, you want to stress these parts as much as you can because, again, otherwise, you don't really have the 100% coverage during the test, and then you have all kinds of yield loss and cost, a lot of cost. How to stress your device or your SOC, how you stress it.
You need to deliver a lot of power, very high voltage, high power, high temp. It is very, very stressful. I mean, it should be the most you can do. At that point, the temperature will raise so much that everything will start to fail, not only the chip under test, but also the probe card and also the tester, because you cannot deliver so much power in so small area. I think that will be for sure a challenge for the future. I do not know, Joe, if you want to add anything else.
Yeah, no, I think you hit it. The power dissipation, as you're getting to pushing 2,000 amps at whatever, 1.4 volts, something like that, it becomes a major challenge, both from extracting the heat out of the probe card itself as well as extracting heat out of the die. Some of the technologies that we have will fully enable us to help customers extract heat in both of those areas. To give you an example of what the opportunity is here, if you have the ability to, say, keep thermal runaway from happening to the extent that you can increase the voltage during stress by 40 millivolts or 50 millivolts, right? Not a lot of voltage. If you can keep the die from doing a thermal runaway, you basically gain an extra top bend for your part.
This becomes really important for companies that are trying to squeeze out an additional top bend. If you can manage thermals just even to enable an extra 40 millivolts, it becomes a really big deal, right? That, I think, is a really good opportunity for growth for us, right? Because I think we bring the technology that could very, very well enable that.
Thank you very much. Don't answer the mic.[audio distortion]
Good afternoon, everybody. Alberto Gegra from Equita. My first question is on AI data center, just to better understand which are the real drivers of your growth in this segment, if it's more a matter of volumes, complexity, or launch of, for example, new GPUs, just to, let's say, derive your trends from the indication and the data we see on a quarterly basis from, for example, TSMC, NVIDIA, and so on. On Edge AI, when do you expect the first impact from the first AI PC, AI smartphone, and if something has changed compared to a few months ago? One last comment, maybe on tariffs.
I know it's still a volatile environment, but I'm interested in both the direct impact of a product shipped to the U.S. and potentially also the indirect one, considering that more or less half of your business is related to the consumer segment. Thank you.
Okay, for a question, I think today, most of our application we serve is mostly on data centers. AI is still in the industrial phase. There, I think we try to list the few factors that are basically making Probe Card and innovation as an enabler. We were talking about thermal management, advanced packaging, and so forth. All these new requirements are basically something that if we are able to deliver performance to them, are making, I mean, the difference. It is for sure complexity, test intensity, and we nicely see also some volume. It is a combination of what you were asking. Definitely Probe Card complexity, new challenges, new requirements to deliver performance to, and volumes. This is across the board. GPUs, obviously, ASICs, so everything that has to do with AI is in our portfolio and we are delivering. The second one is Edge AI.
I cannot predict that. When I was showing my chart and we predict the demand for Probe Card to grow in the next few years, we expect Edge AI to become important or relevant in the second half, so I would say in a few years from now. Still too early. We see some sign, but it's not yet there. For us today, AI means the industrial side of AI, data centers mainly. We definitely see Edge AI to emerge in the next future. You want to take a comment on tariffs or whatever?
I wasn't expecting a question of tariffs. Joking aside, this is something we were wondering every morning when we wake up. We passed through no tariffs to 100% tariffs. Now I changed my mind, take a pause, six months, no for the semiconductors, yes for the iPhones. That is a very complicated scenario. It's really a puzzle. Even if we try to make a rationale on that and build some metrics of the scenarios, what we can do in case of the metrics is probably 25 scenarios different. Depending also if there will be a tariff sooner or later, how much it will be because we don't know. Which is the direct impact? We currently ship in the U.S. approximately 15%-20% of our goods.
Even if the total of the customer in the U.S. currently prefers to be shipped in other countries, not only in the U.S., so that those shipments are theoretically free of duty. We do not know the indirect effect if those goods will be then re-imported into the U.S., if they will be subject to duties in that case. Will the customer be able to support the duty? Will the suppliers be charged too? We go down to the commercial negotiation. This is another x2 for the scenarios we have just built. What we can do is just react when something happens. We have plans, of course. We already talked with some customers. We have analyzed internally other solutions for shipments or to reduce tariffs, benefiting from some special treatment allowed by the U.S. customs.
For the moment, we are still on the window. We are still waiting to see what will happen. In one word, it's unpredictable right now.
Maybe I'd like to add, again, something about this. Because in addition to what Stefano said, I'd like again to reinforce the fact that we are leaders in the market. The customer cannot find an alternative. In any case, they come to us. It is not unthinkable also to change our pricing structure. I mean, in any way, we have technologies. We are leaders in technology. Everybody, all the major customers, will continue to use our technologies because we are leaders. It would be a much different position for someone that is not a leader that can be replaceable. This is the first point.
What really can be probably the bigger worry is not the direct impact of the tariff on our revenue, on our sales, but it can be the general downturn that the tariffs can cause for the market itself, the consumer market, for example. This is what really can hurt any company at this point, not only us. I mean, if there is downturn, of course, all this business will go down. For Cloud AI, this is my opinion, by the way, it might not follow also in the worst case because right now it's difficult to predict if we'll see a downturn, really downturn or not, also in the consumer market. Cloud AI might be more resilient on this because the major player might continue to invest in any case to be ready.
Because if one of them loses the train, then it will be very difficult to catch up. I do not think also in the worst case, Cloud AI will see an impact in the demand. Okay. In any case, for us, also in a difficult time, as I mentioned before, I think it will be an opportunity to continue to invest and go faster than other players. Also in the worst case, that right now, again, nobody can tell if that will happen or not.
Hi, Rory Pike from Nevat. I just had a couple of questions on HBM. I think if you look at, obviously, the different process flows for all the different vendors, but there are generally two sets of test insertions, wafer sorts, and kind of performance tests. I'm just curious if you expect both of those to go to vertical or one of those to go to vertical, only wafer sort, only performance. If there's a difference, which would be most likely? Second question would be, do customers get to use their existing process flow, or do they need to develop a new process flow with your vertical solution? Finally, what generation are you currently being qualified for, 3E or 4? Thank you.
Yeah, so there's absolutely no change to the process flow for HBM to adopt our technology. The vertical MEMS spring interacts with a whole variety of metallurgies, pad metallurgies, bump metallurgies. I do not envision a scenario where we would have to have any sort of compromise or negotiation with the HBM supplier. As for the technologies that we're certifying, obviously, the qualification is going to be on an HBM3 solution. That's what's available now. We're really targeting what happens in the next couple of generations when you go to 3E or 4, right? Because that's when we think that the data rates start to get to the point that this becomes a very important thing.
To your first question, the technologies that we have with the signal integrity benefits that you get from the short needles, you should be able to adopt that to any test insertion, either a hard defect screen or a performance socket.
Thank you. Good afternoon, everyone. A few questions on the target model. You said that you want to increase your market share from 60% to something above that level because you think that you have the capabilities to increase your market share. What is the assumption embedded in the three, four years target? The second one is on the profitability. You said you are targeting a 38%-40% EBITDA margin and looking at the chart, most of these improving margins are coming from the operating leverage, which is, of course, connected to the top line. I was wondering what are the headwinds or the challenges that you are seeing in order to deliver that target sales.
The other question is on the effects, if you can remind us your sensitivity to the U.S. dollar changes because not only the tariffs are moving very quickly, but also the U.S. dollar. Finally, on the advanced packaging, you mentioned that you have a feature strength to be a leader also in this space. Are the technology or the technologies that you are using already registered in this business? Thank you.
Sorry. The first question about the market share in the model. That was a combination, of course, in the model because it includes both the Probe Cards, also HBM, and also final testing and all the new product mix we're going to develop in the next three, two, four years. For the increase of our core business right now, maybe I leave this to Marco. Yeah.
Yeah. I mean, again, we definitely are committed to grow. I mean, the target model or in the short term, I think the recipe is always the same. We came to this point to 60% market share because of multiple factors. Definitely customer engagement. If you stay close to the customer, you know what they need. If you know what they need and you're willing to satisfy this need, take the challenges, Stefano is saying, obviously, I mean, you're rewarded. This is also allowing us to, in a way, to align the roadmap to develop what they really need. One of the traits for this company has always been to develop the next solution and innovation in the fastest way possible in the market. We've always been very fast in developing the technologies the customer needs, assuming we know what the customer wants.
This is what we do every day. With a strong relationship, we will definitely have all the ingredients to cook the recipe and continue to grow the market share on our core business.
Thank you. About the operating leverage, if you remember the page we showed before, on the left side, there were the initial cost integration cost. On the right side, there is the operating leverage and product mix positive impact. Operating leverage will be evident starting already in the second part of 2025. We said already we invested a lot in our internal capacity and in automation. We made also a very recent restructuring and reorganization in the United States, especially at the production level. All of them will bring positive effects already in 2025 because of labor cost reduction and automation increase. That operating margin, sorry, operating leverage will be more evident as sales grow. This is the key. Always the volume is the key for our business. That will be visible starting in the short term, which will be the challenge you ask.
The challenge will be, as always, the R&D. If all our efforts will be finalized in the defined time we have in our plan, then we will be able to present to our customers the solution and the prototypes we want to hit the market. Luckily so far, maybe not luckily, but thanks to all the R&D department, all the efforts are now addressed in the right pipelines, in the right timeline. The challenge will be always the same to conclude and to finalize these solutions. That will be, of course, expensive in terms of R&D investment that, again, represent more or less 13%-14% of our revenues, only about labor and material cost, without excluding the investment in CapEx that are addressed to research. In that case, we are currently seeing already positive impact, especially also in Taiwan.
Taiwan, for E-wave, the two fabs were included in our P&L in the R&D line until 2024 because they are still in development. Starting in 2025, they are now, I do not want to say fully operating, but operating, especially in the second quarter of 2025. All that cost will be moved to above gross margin. You will see a lighter R&D cost and higher cost of goods and labor cost in the first portion of the P&L. Again, the R&D will be the key for our strategy. You asked about effect sensitivity. Of course, the U.S. dollar is still one of the main drivers for our revenues. Most of our revenues are exposed to that currency.
We have seen in the recent weeks a strengthening of the euro against the U.S. dollar, about 5-6, or even 7% difference compared to the initial part of the year. There is, of course, an impact. We have measured that. We have a sensitivity. We can estimate for every 1% of change in favor of the euro, we will see a 0.75% impact on our revenue. It is not directly one to one, of course, because there are other costs in U.S. dollar. There are other currencies impacting our revenues, the Taiwan dollar, the euro, and of course, the Korean won. This is a rough estimate of the impact for the currency volatility that, by the way, we expect to be normalized in the next few weeks.
I do not want to make provisions because it is something right now unpredictable like tariffs, but we do not expect that to be consistent in the next few months. The last point was about advanced packaging.
Yeah. On the technology front, the concept of advanced packaging is not new, right? There are multiple semiconductor manufacturers that are already producing chips with advanced packaging. We are playing in that space already today, 100% qualified. Any type of metallurgy, aluminum pad, copper pad, micro bump, C4 bump, hybrid bond pad, any of these sorts of things are already technologies that we have. The development that we need to do is to make sure that we meet specific customer needs. There may be specific needs around bump layout, what angle the probe is interacting, some things to optimize in terms of current carrying capabilities and so forth.
As far as the core technology for anything that's advanced packaging, I think that we're already in that space and we're already fully qualified in that space.
Okay. Thank you all for being here. Unfortunately, we had no more time to answer questions. We came virtually, but I'll take care of them. I will answer using the mail they sent to us for attending the meeting. Now, for once you want to attend the site visit, you can go back at the reception, and in a few minutes, we will be ready to start. Okay? Thanks to you again.
Thank you.