Technoprobe Spa U/Adr

Good evening, everyone, and thank you for joining us. I'm here today with Stefano Beretta, our CFO. Together, we'll walk you through Technoprop's full year 2025 results and provide our guidance for the first quarter of this year. As usual, we will leave time for Q&A at the end of the presentation. Let me start with the key takeaway. 2025 confirmed Technoprobe's leadership in testing solutions for artificial intelligence logic chips. Based on our initial assumptions regarding organic growth and market mix, we delivered on our revenue targets and expectations in terms of AI exposures. This performance reflects the investments we have made over the past few years, together with TechnoPro's strong and well-established expertise in MEMS technology, which continues to underpin our leadership in advanced testing. This result was mainly driven by a strong increase in data center GPU volumes, which is expected to remain the key growth driver also this year. At this stage, demand appears both solid and well supported. It is also worth highlighting that AI chips, as well as the most advanced semiconductor devices, require extremely high performance, low power consumption, and very fast data processing capabilities. These requirements are enabled by advanced packaging architectures, which integrate multiple components in the single electronic device using highly sophisticated technologies. As a result, these architectures require advanced testing solutions based on MEMS technologies. And this supports our expectation of a progressive and steady increase in demand for advanced logic testing solutions. Moving to the consumer segment, which for us mainly includes PCs and mobile devices, 2025 was essentially a year of stabilization. Looking ahead, we expect moderate growth in 2026, although some uncertainty around the replacement cycle remains. In the automotive and industrial segment, we expect a gradual recovery despite the long tail of the stocking process that is likely to continue through this year. Overall, our outlook for the current year remains positive. Our focus remains on strengthening technoprops leadership in the AI space further consolidating our distinctive expertise in advanced testing across both front-end and back-end, and continuing to work on ambitious projects to expand into the high bandwidth memory market segment. Now, I turn over to Stefano Beretta, who will give you more color on 2025 results.

Thank you, Stefano, and good evening. Thank you for joining us. As you may have read in our press release, revenue in the fourth quarter were 161.8 million, slightly above our guidance, registering an increase of 3.5% compared to the same quarter of prior year, with a remarkable sequential increase of 15% compared to Q3 2025. Gross profit increased 13.4% compared to the same period of 2024, for an amount of 71.8 million, representing a 44.4% margin. And consistently, the EBITDA increased 34.2% compared to the fourth quarter of 2024, up to 55.3 million, representing a 34.2% margin, slightly below the mid-range guided during our prior poll. Moving to the year-to-date figures, Total revenue were just over 628 million, with an year-on-year increase of 15.7%. Gross profit was 279.1 million, 24.9% higher compared to the same period of 2024, and representing a margin of 44.4%. Finally, the EBITDA closed at 201.4 million, up 47.5% compared to the same period of 2024 and representing a margin of 32.1%. For the completeness of information and comparison purposes, it's worth to remind that all the above figures include a new perimeter of consolidation, inclusive of the contribution of DEIS tech acquired on May 27 in 2024, then present for the seven months only in the same period of 2024. Moving to this page, you can see the annual trend by quarter of our main KPIs. So excluding the cyclicality we observed in the third quarter, the steady growth in all our KPIs was consistent with our prudent and realistic view of the market and confirms the solidity of our plan, despite the uncertainties that characterize 2025, particularly in the consumer and automotive segments. as well as the headwinds of the dollar weakness that we have been experiencing through 2025. The organic revenue in 2025 versus 2024 was, as expected, a high single digit in the range of 8% approximately. On this page, you can see a summary comparison between the financials at the year end for 2024 and 2025. Just to comment that further, revenue year-on-year increase of 15.7% was mostly driven by the change of perimeter, so DIS inclusion, partially offset by Santa Clara shutdown for approximately $49 million, in addition to an organic growth of approximately $36 million, as I said, 8% approximately, largely sustained by artificial intelligence volumes, as well as a softer recovery on the consumer segment, and all of them partially mitigated by the weakness in automotive and industrial. The mentioned growth components have been partially upset by the FX headwind you can see here represented for about 28 million. In fact, the revenue expressed at the cost and currency, so using the same FX average rate in 2024, would have been approximately 28 million higher than reported revenue. meaning an unfavorable impact between 4% and 5%, with a 12-month average FX rate at 1.13 compared to 1.082 for the same period of 2024. On a gross profit level, the increase in the margin of 330 basis points from 41.1% to 44.4%, As already mentioned in our prior press releases, confirm once again the recovery efficiency in our production processes, together with the operating leverage that our business model allows us when certain volumes are achieved. All of these impacts have more than compensated by the increase of depreciation following the investments in fixed asset made during the latest months to expand our capacity and increase the automation. Gross profit at cost and currency would have been approximately 20 million higher by using the 24 average rate, so bringing the margin well above 45%. Consistently, the EBITDA reflected the same trend, showing an increase in the margin of 700 basis points from 25.1% to 32.1%, benefiting also from savings attributable to the US organization, and even including the negative mentioned Forex impact for approximately 18 million, which would have brought the margin above 33% at cost and currency. Finally, net financial position is up for approximately 28 million compared to the end of December 24. with the variance mainly attributable to 212 million generated through operating activities, 26 million generated by the sales of Santa Clara building, partially upset by 99 million absorbed by the investments of the period, 60 million absorbed by the purchase of treasury share according to the buyback program, 20 million absorbed by the acquisition of the remaining minority stake in EUA, that is now owned 100%. And 7 million absorbed by the acquisition of a minority stake in InnoStar Service. And finally, 22 million theoretically absorbed by the unrealized FX impact on foreign currency bank accounts. Moving to this page, it's worth highlighting the shift In strategy compared to what was announced during our latest capital market day back in April 2025, when we expected investments to normalize around 8-10% of annual revenue. However, the acceleration that characterized certain markets in the second half of 2025, along with Technoprop's market share, contributed to the decision to significantly increase our investments, aiming to double production capacity by the second half of 2027. And this decision has already shown its effects in the final quarter of 2025, when the investments were almost equal to those made in the first nine months of the year. I mentioned before for a total of approximately 99 million euros in 2025. I hand over to Stefano Felici to comment further this page.

Thank you, Stefano. As we look ahead to 2026, the geopolitical environment will continue to shape the global economic landscape. Ongoing tensions in the Middle East, combined with evolving trade restrictions, are creating a context of uncertainty that is influencing both demand and supply chains. In semiconductors, the industry is responding with a clear shift toward geographic diversification. The concentration of leading-edge air capacity in Taiwan is accelerating investments in the United States, Japan, and Europe. At the same time, governments support across all major regions is reinforcing the development of local ecosystems. This is driving a more regionalized industry with greater complexity and different approach to capital allocation. Importantly, we do not expect these dynamics to weaken overall semiconductor demand. What we do expect is a structural revolution where capacity is built how supply chains are organized, and how products are segmented will continue to change, increasing the overall complexity of the industry. From a testing standpoint, the underlying drivers remain very strong. Artificial intelligence, high-performance computing, and advanced memory are fully sustained demand. As devices become more complex, the level of testing required per chip continues to increase, effectively expanding our addressable market. At the same time, the transition to chiplet architectures and advanced packaging, particularly HVM, is shifting the industry toward more stringent quality requirements at wafer level and beyond. This is increasing the importance of the advanced probing, final test and burn-in, especially for high-performance devices. As a result, we are seeing a strong demand for high-end probing solutions, HVM-related applications, and advanced interface technologies. Growth is being driven not only by volumes, but also by mix, performance requirements, and additional testing steps. While traditional consumer and industrial segments remain uneven, the AI and data center ecosystem continues to provide a solid and structural growth foundation for our business. Against this backdrop, 2026 will be a year of execution for technical We are focused on strengthening our support to customers and positioning the company for the next phase of growth. From a product perspective, we remain fully committed to the high-end logic segment, while accelerating our expansion in HPM, where we expect to qualify with two additional leading customers and investing in new opportunities such as a CD company. Overall, we are confident in our position and in our ability to execute leveraging strong structural strength and our technology leadership to drive sustainable long-term growth. Now I turn over again to Stefano Beretta.

Thank you, Stefano. We are very excited to announce this important event we wish to highlight in this session. That is the progress that Technoprobe is making to expand into new markets where it currently has a marginal presence in terms of market share and customer base. As summarized on this page, Technoprobe has defined the goal of expanding into the Chinese market over the next couple of years with its proprietary TPAC technology by transferring the patents and machinery required for production to its subsidiary The key to the success of this project will be the know-how of technoprops engineers and operators, combined with the skills and knowledge of the domestic market from local investors, who will contribute approximately an amount of 35 million US dollars in a capital increase in exchange of a consideration of approximately 30% of the share capital of the EIS. The operation are currently expected to gradually ramp at the end of 2026, beginning of 2027, and will concern exclusively the Chinese customers on the Chinese territory. Moving to this page, it's worth reminding again the differences between the information we provided to you or to the market during the capital market day and the latest press release and we want to confirm once again the trajectory for the organic mid double digit growth in our reference market on a yearly basis together with the consistent strengthening of the profitability So bringing forward to 2027, the medium-term objective originally planned for 2028. You can see that on the chart on your right. While on the chart on your left, on the left side of the page, you can see again the capex trend that initially we were planning to normalize from 8% to 10% of the revenues, but now considering the change in the strategy so the expansion plan would wish to invest more than 200 million euro in the next 18 months so the shift is well above this amount on revenues finally that said first quarter of the year is expected to show record revenue for technoprobe although it will still be affected at least for the quarter by the expected depreciation of the US dollar against the Euro, together with a robust growth in gross margin and EBITDA. So revenue to be about 182 million plus minus 3%, gross margin in the range of 46.5% plus minus 200 basis points, a beta margin in the range of 35.4%, plus minus 200 basis points. Thanks everyone for your attention. Now we can move to the Q&A session.

Thank you to the speakers today. We now have an opportunity for question. As a reminder, if you would like to do so, please use the raise hand function on your screen, or for those dialing in, it's star nine on your keypad. Once your name is announced, please unmute your line, state your company name before asking your question. Thank you. The first question comes from Andrea Todeschini. Please, Andrea, the floor to you.

Good evening, everyone. Can you hear me correctly? All right. I'll assume so. So I just wanted... Perfect. Thank you. So I just wanted to ask a quick update on HBM contribution looking at 2026. So I remember that you guys were talking about probably an impact at the top line level between 10 and 15 million. I would like to know if it's still like that. There's been any evolution or... And if you can tell us a little bit more about those qualifications that you were thinking they're going to come in 2026, can we expect something coming in the first half, both of them? So if you could give any color on this. And then moving on, talking about Silicon Photonics opportunity, if you could just expand a little bit more as of how is the situation right now, or maybe any type of projections that you guys have regarding the possible addressable market for it. Thank you.

Thank you Andrea for your question. Let me take the one related to HBM. So currently there are no particular news about qualification on the other two potential customers. So we are still with the one already qualified. Revenue again is confirmed in the range of 10 million. Let me say, let me repeat that it is a placeholder for 2026. We don't have any orders so far related to HBM. This is something it can be additional. It's included in our estimate, but again, just like a placeholder.

Yes, I will take the question about the silicon photonics. So silicon photonics, I think right now is still too early to say what kind of opportunity will be. And the reason why I'm saying this is because it's still not clear for our customers how many test insertions they are going to do. So how much testing and how they are going to do the testing is still unclear for our customers, so for the semiconductor companies. So what they're doing now is basically is more on the engineering development and the lab type of testing. And for this type of test, anyway, we, the silicon photonic test, we already working on technologies, pro-car technology that can enable this type of test and eventually also to go in mass production. But again, it's not yet our call. So we are ready, I can tell you, but at this moment, there is not a clear strategy how big would be the test, let's say the test volumes done in this segment for our customers.

I understand. Thank you very much. I just have a quick follow-up. When talking about silicon photonics, I know that it's very unclear still, but is the technology in question going to be able to do EIC testing and photonic testing using the same probe card? And will this probe card potentially use the same type of machinery or will require a new equipment from an equipment manufacturer? Thank you.

A very good question. Of course, here at Silicon Photonics introducing one extra element that typically is not there. So which is the light, of course. But you still need the pros because you wouldn't have enough power without the reading electrical connection, electrical conductor between the tester and the chip. So basically you need both. You need the probes and you need the fiber optics to really be very small fiber optics to get to the device under test. So the first approach that this was taken a few years ago at the beginning was really a lab system where you have a small probe cards and some handler or manipulator to get some fiber optics on a side, okay. Our solution is to integrate the two things together in order, exactly in the same way that basically you're suggesting already. We have one ProCard that is very similar to a standard ProCard, basically it's a standard ProCard with this additional feature. And eventually this type of ProCard will dock on the standard tester with some, of course, some other instruments inside the tester for the silicon photonic testing. But the overall thing will look like exactly like a test standard test cell. And this is really the point how to make this testing is in high volumes, for high volumes, not at the lab level, but The volumes should go in a standard test flow. I don't know if this is answering your question.

Yes, absolutely. Thank you very much.

Thank you, Andrea. The next question comes from Alberto Giegra. Please, Alberto, the floor to you.

Good afternoon. I hope you can hear me. So I have a few questions. The first on the full year guidance. You confirmed, if I'm not wrong, the mid-teens growth for 2026. My question is, which kind of seasonality should we assume? You typically have a pickup in the second quarter. I'm wondering if you still see this in 2026. Then if you can maybe guide us through the main drivers of profitability in this year between operating leverage, maybe an increase in OPEX and a possible mix effect. Then a follow up on Silicon Photonics. My question is if you expect to have more material opportunity with more, let's say, advanced architecture like packaged optics, or you believe it would be sizable even in less complex architecture like pluggable one? Thank you.

Thank you, Alberto. So let me think about this. increase the growth we expect in 2026. We confirm the mid-double digit, so between 13% and 16% overall, without raising a specific guidance because you know that our visibility is very short, but we see the trend and we can confirm the trend as is. We can also expect a certain cyclicality, so Q1 and Q2 are expected to be more interesting compared to the second part of the year. So we're in the third quarter, we do expect a sequential decrease, so a kind of cyclicality. So the same we experienced also in 2025. And then at the end of the year, the latest quarter, the latest month, an increase compared to Q3. So kind of Q we have already experienced during 2025. With the first half, much more involved, much more robust compared to the second part of the year. In terms of profitability for the mid-term, so until the year end, we cannot confirm any particular margin right now. There are so many projects and also so many operational activities carry on. We do expect, of course, that over the level of revenue we are currently reaching, so over 180 million per quarter, the margin, of course, has a significant positive effect. And we also hope to have some benefit along the year also from the US dollar impacting our financials. I have included also in our estimate of effects in the latest page of the presentation, so you can see the guidance according to the numbers calculated with those. But about profitability, of course, the more revenue we have, the increase in the profitability will be more than proportional.

Yeah, let me maybe add a few more color about the silicon photonic. So basically, yeah, that is, it can be a huge opportunity if, again, and can be any type of a chip that is related to photonics. This for sure will happen. Means, I mean, the customers, I mean, these chips will for sure be used or for interconnecting the GPUs, CPUs with the other systems because of the, you need speed. So high bandwidth. And this is, we are reaching the limit of the standard silicon. So silicon photonics will happen. So there will be many, many, many, million, billion of chips regarding silicon photonics. The point here is again, Coming back to the point I made before, is that right now, the customers might decide just to do some very limited sampling. So the difference between, let's say, if we consider standard chips, all the chips are tested 100% with PropCast. With silicon photonics, the photonic part now is not sure if they will want to do 100% testing or maybe just 1% sampling or 0.1% sampling, for example. So that's why it can be a huge opportunity if they decide to test 100% all the photonics chips. But if they decide just to do a very limited testing because of the cost, for example, or because the yield of this photonic part is so high that it is not worth to test, for example. So there are many, many factors here that can influence the decision. But right now, again, there is not a very good visibility to say that what would be the real direction taken by customers.

OK, thanks.

Thank you, Alberto. The next question comes from Henry Blakelock. Please, Henry, the floor to you.

Hi, there. Thanks for taking my question. I've got a few. The first is just related to the China market deal and just a few questions in relation to that. Is it focusing on specific end markets, specific use cases? How are you looking at it from a protecting your IP and technology perspective? And then also, at the moment, what does China represent as a percentage of revenue or absolute figure of revenue? And how much do you expect that to grow as the partnership starts to ramp up?

Thank you, Harry. In terms of IP, this is a real key point for Technoprobe. So that's why one of the main reasons of this presence in China, the increase of our presence in China is to protect the IP for the TPEG technology. That is not, it doesn't represent the latest available technology for Technoprobe. So the Microfabrica technology, let me say. It relates to a technology developed many years ago, but still used in many, many products, especially for the Chinese market. And currently is already used on the market by, let me say, other players. We want to reinforce our presence to introduce in the market Let me say the real TPEC technology, appropriatory technology of Technoprobe. So this is the first aim. And we know that the best way to protect your IP is to be physically present with your products and with your patents. And this is the reason why the patents of this technology will be transferred with an asset transfer from Technoprobe Italy to DIS China. So BIS China as a Chinese entity will be the owner of this patents and will be able in that market to get the right protection. So this is what we hope and what we believe will happen. In terms of markets, currently Technoprobe is very limited in that area. So at least as a customer base, our revenue in China do not exceed, let me say, 15 million per year, something like that. But we are present already with BIS as a design center with more than 200 employees. We are present with another company, Technoprop China, representing the repair center for the area. So we have already a privileged point of view on the market. We will be able in our expectation to better represent Technoproba also from the manufacturing standpoint. Back to your question about the market, about the revenue, we don't have now a large customer base. So we do not serve the major customer in that area. because we know the market is very tough and it's very often served on a local by local base so that's why we will start from the local partner that is involved in these in this project to approach and develop new business with these new completely new customers they are the main customers present in the area it's still too early now to understand we have our assumption of course but it is too early to introduce this revenue in our figures. And again, we do confirm in our estimates for 2026, there are no revenue coming from this project that will affect the numbers. So that's why we confirm it again, the same growth announced a few months ago.

Got it, that is very comprehensive and super clear. My next question was just around whether you could comment on the progress you're making with the engagements with custom ASICs customers and whether you're still seeing the same kind of shift towards vertical MEMS technology from vertical probe guards.

Yes, this is still what we're seeing, custom ASICs. are following, as far as the complexity, following the GPUs, of course. And in terms of power demand and computing power, also, they're following, again, the big brothers. And this means for us that also those type of devices requires right now, the high end technology, and we constantly see this, this push. And of course, we are gaining continue, I will say that we are continuing to gain market share on for all those customers.

Got it. That's super helpful. One last question, if I may. Just around kind of like the timing of or the phasing of reaching your 2027 numbers, your 2027 top line guidance. And it implies quite a big step up in 27. And I know you've mentioned on the call already that kind of your visibility is generally quite limited. And obviously you speak to customers about kind of their expectations, but you don't have super clear visibility beyond kind of a few quarters. So I was just wondering what kind of gives you confidence in reaching those numbers in 2027 and yeah, what are the main kind of drivers you're seeing? I guess HBM, one of them sounds like Silicon Photonics maybe not contributing that much next year, but yeah, it'd be helpful to know what's driving that.

Yes, correct about silicon photonics. About the rest, in general, we are very confident to be at that level already in 2027. If you consider the CAGR starting from 2025 to 2027, we say it's still 13%, 14%. So it's a grown number that is pretty reachable under our assumptions, especially considering how the market is moving, especially from the AI segment and considering the volumes we are already experiencing in our portfolio in the next couple of months. So we do expect to be able to reach that amount that we all will also be reaching thanks to the capacity increase So that amount of revenue in 2027 cannot be reached only with internal efficiency growth, but need to be consistently expanded through investments in CapEx, in equipment, in FABs, new FABs, hiring, etc. So all the project is consistent with this growth. And we are very, very confident to be at that point already one year in advance compared to 2028.

Got it. Thank you, guys.

And without considering Chinese market. So Chinese market is not embedded in these numbers.

Yeah, makes sense. Thank you.

Thank you, Harry. We now have a follow-up question from Alberto Gegra. Please, Alberto, the floor to you.

A couple of follow-ups from my side. The first on the ASIC business. Maybe if you can, generally speaking, give us a sense of growth in the AI segment that you expect in 2026 and what is more or less the split between GPU business and ASIC business for 2026. The second on GPU, we know that your US competitor is quite vocal on the progress that he's been making in the GPU market. So what is your expectation on your market share with the leading GPU players for 2026? Okay.

It's a good question. Maybe I like, first of all, to say how we classify the AI business here, because what has happened in the last quarter, I would say probably in the last year, is that everything is classified as AI. I mean, it can be awesome because of marketing in some case, but the reality is that any type of chip is in some way connected to some, another AI chip. And then, so it's, it's somehow considered AI. So that's why now for the first time, this time we are classifying AI or just for GPUs, basically. and ASICs. So we are talking about cloud AI, substantially, and the core chips of the AI. So this is now how we classify our AI. And that is the most advanced segment in the market. So for 2025, we presented a market share. I mean, I want to say on the, our revenue, I mean, the was 38%. Uh, due to the AI and again, classified in this way, I just explained for 2026, we expect, uh, the range of 40, 45%. So we, so we are growing the revenue and the, is growing also the percentage of our revenue related to the AI classified in this way. So I would say that as far as market share, we still have the lion market share in all the main customers, AI customers. So we are developing the next generation of technologies for the next generation of chip. So already projected for the next two to three years from now. So we think to be in very good position, yeah.

Sorry, but just a follow up. So of that 40% to 45% of sales, Can you give us more or less how much GPUs and how much ASIC?

Well, this is tricky to split this detail, especially for the full year. So this is something we don't have on our hands for the year end. But basically, the most part of this split is related to GPUs rather than ASICs. But then, you know, ASICs is becoming more and more important in the latest months, in the latest year, especially from the hyperscaler user. And when the hyperscaler will become more and more important in terms of customer base, their percentage on our cake will increase accordingly. So for the moment, GPUs represent still the most part of our revenues.

Okay, clear. Thank you.

Thank you, Alberto. As there are no further questions queued, I will now hand back to the speakers for any final comments before bringing this presentation to a close. Please go ahead.

Thank you, everyone, for attending today and stay tuned for the next important call. Thank you. Thank you.