Good evening, and thank you for joining us. Together with Stefano Beretta, our CFO, we'll go through 2024 nine-month results and guidance for the last quarter of this year. As usual, the Q&A session will follow at the end of the presentation. Results achieved in the third quarter are in line with guidance shared in August, confirming trends rose in our reference markets. Exposure to artificial intelligence has been the main driver of our revenue, partially offset by weakness in the automotive and industrial market. Looking at different segments, we have automotive and industrial remains sluggish, suffering of the ongoing inventory correction started in last quarter 2023, which is expected to last at least until the first half of next year. More consumer-exposed markets and related inventories are completing the process of normalization, as expected in 2024.
Moving to artificial intelligence, as said, it has been the biggest driver of this year and will remain the same in the short to mid-term in the same space overall. It was powered by edge AI and inference applications that boosted data centers. Looking at artificial intelligence devices, you see chips being split into chiplets, new architectures, and some other factors feeding the device complexity. It means that we expect an overall growth in the test intensity, driven by the need of a customer to prove and test each component of chiplet during the assembly process. Also, the average selling prices are higher on these complex devices, so customers are increasingly willing to perform more tests in order to avoid the cost of failure. From a technical standpoint, these are challenges that we did not have to face a couple of years ago or even last year.
Now we have this new wave of complexity driving test content and test time. In addition to test time, another key contributor to test demand is the effort to bolster quality, assurance, and reliability. That said, it is difficult to forecast the testing market growth given our limited visibility. It is especially difficult to predict changes in test time and test content for AI devices. That being said, if you look at the architecture of next-generation SoC devices, the complexity is only increasing, so there is no reason for us to believe that the test content is going to do anything but continue to increase. Now, let me turn to Stefano Beretta, who will give you more colors on our numbers.
Thank you, Stefano, and good afternoon to everyone. As you will have seen in our press release, revenues in the third quarter were EUR 145.8 million, at the lower end of our guidance, registering an increase of approximately 31.2% compared to the same quarter of prior year, with a sequential increase of 4.7% compared to the second quarter of 2024. It's worth to remind you that on a sequential quarterly base, the lower increase registered in the Q3 compared to the previous quarter is related to an important campaign that came in the second quarter of 2024. Gross profit increased at 8.8% compared to the same period of 2023, up to EUR 59.4 million, representing a 40.7% at the midpoint of our guidance. As per gross profit, also the EBITDA closed at the midpoint of the guidance.
It decreased approximately 3.4% compared to the third quarter of 2023, down to EUR 34.6 million, representing a 23.7% margin. If we move to the year-to-date figures, total revenues were EUR 386.9 million, with a year-on-year increase close to 26%. Gross profit was EUR 160.1 million, 2.2% up compared to the same period of 2023, and representing a margin of 41.4%. The EBITDA closed at EUR 95.3 million, down 5.2% compared to the same period of 2023, and representing a margin of 24.6%. For the completeness of information and comparison purposes, it's worth to remind that all the above figures include a new perimeter consolidation, inclusive of the contribution of Harbor Electronics, nine months in 2024 compared to little less than two months in 2023, and DIS Tech Business, four months in 2024 compared to nil in 2023.
If we move to the next page, so now you can see a summary comparison between the financials at the end of the nine-month period, 2023 and 2024. Just to comment that further, revenue year-on-year increase of 25.9% was half-driven by the gradual recovery of the consumer segment and by volumes related to artificial intelligence, together with the other half by the contribution of Harbor and DIS, all of them partially mitigated by the weakness in automotive and industrial, as already reminded by Stefano Felici. If we want to give just a note on the FX impact, revenue expressed at constant currency, so using the same average FX rates on the nine months of 2023, would have been approximately EUR 3.2 million higher than reported revenue, meaning only 1% unfavorable impact.
On a gross profit level, the decrease in the margin from 51% to 41.4% was the result of the expected dilution from the mentioned acquisition, together with a relevant increase of depreciation following the high investment in fixed asset made in the last 12 months, and also due to the growing complexity of our products and the related and temporary lower performance registered in the second and third quarters of the year. Consistently, the EBITDA reflected the same trend, more or less showing a decrease in the margin from 32.7% to 24.6%, including a negative FX impact for approximately EUR 1 million.
We would highlight also that the EBITDA was impacted by a significant amount of R&D cost, approximately EUR 45 million, 3 million more compared to the amount recorded in the same period of 2023, and representing approximately 12% of our revenues, confirming once again the mindset and the commitment of our leadership to continuously develop new products and solutions. Finally, just a few comments on the net financial position that is up to almost EUR 280 million in the nine-month period, mainly due to EUR 64 million generated through the operating activities, together with EUR 385 million generated through the capital increase following the Teradyne acquisition of 10% stake, both of them partially offset by EUR 55 million absorbed by the investment of the period, EUR 80 million absorbed by the acquisition of DIS, and approximately EUR 23 million absorbed by the share buyback program, this amount at the end of September.
So now, finally, we project our guidance for 2024, where we see an overall stabilization in our reference market for the last part of the year, with a slight sequential increase in revenue, together with a progressive recovery of the profitability. That said, the last quarter of the year is expected to show the following, so revenues to be about EUR 150 million ±3%, gross margin in the range of 42% ±2%, EBITDA margin in the range of 25% ±2%. This is all I have, so thanks everyone for your attention. Now we can move to the Q&A session.
Thank you to the management team. Our first question today comes from Mr. Gianmarco Bonaccina. Please, the floor to you.
Yes, good afternoon, everyone. A couple of questions, please. The first one is on the order push-out. You have indicated that given the complexity in producing the new AI chips, there is some inefficiency in the supply chain, and so there is some push-out. Can you maybe quantify the amount that had been pushed out, and if that means that in Q1, which normally maybe is down sequentially, we can see a sequential growth? The second one is on the profitability. You mentioned also here that the development of the new products is having at the moment a negative impact on the profitability. Can you also quantify how much is the profitability in terms of margin that you can recover in 2025 and thereafter? Thank you.
Thank you, so first of all, we want to highlight that the push-out is not due to the complexity of the products, but it is simply due to our request of our customers. So the cutoff date of delivery is always very important for us at the end of every quarter, but it happens that some customers require you to push one week later rather than one week before. In this case, in the Q4, given also the Christmas Eve period, the last week of December, we have been asked to push out approximately EUR 15 million revenue to the Q1 2025. So just to be clear, this is not a cancellation. This is not a matter of cyclicality. This is not a matter of complexity of product. It's just a delivery date push-out for a couple of weeks.
About the second point, the profitability related to the new products, their profits, their contribution to our margin is not different from the other contribution for the products already on the market. The reduced profitability we incurred in the second quarter and in the third quarter is strictly due to the constraints we have in our manufacturing processes to manufacture and develop new products. So this is not a matter of selling price or higher cost of production, not recognized by the customer, but this is simply due to manufacturing processes that we have now resolved. So for the Q4, we do not expect in our estimates such extra cost for production, and we also expect in 2025 when we will sell these products to be in line with the other profits of all the other products.
Thank you.
Thank you, Mr. Bonaccina, for your questions. Our next question comes from Mr. Giovanni Selvetti. Please give the floor to you.
Hello, everyone. A couple of questions on my side. The first one, if you can please give us an idea of the contribution of DIS in the quarter and the one expected in Q4. The second one is about CapEx. In the press release, you mentioned roughly EUR 55 million related to upgrading of production line. I was wondering what you exactly mean by that, and if you can give us an idea of how much of these EUR 55 million roughly is related to the development of MLOs. Also, two more general questions. Let's just say you talk about stabilization of the consumer market, which is interesting, and as your main competitor said, in the fourth quarter, they still see some softening in PC and mobile. So I was wondering if you conversely can see an improvement there.
The last one, it's more of a strategic question, let's just say. Again, your competitors seem to hinge on the fact that going forward, selling protocols for the testing of HBM and thus memory will be key for maybe winning more business in the testing of GPUs as well. So my question is, do you think that entering the HBM market will be mandatory at some point if you can win a lot of business in the foundry logic when it comes to AI, I mean? Thank you.
Thank you, Giovanni. So back to the first question about DIS contribution. So for the Q4, considering that the business where DIS is involved, especially with a big customer in Taiwan, this is a bit cyclical. So the Q4 is always flat compared to the other quarters of the year, especially the first two quarters. So we do expect a little less contribution in that quarter from DIS compared to the Q3 and the second quarter. So for the Q4, we expect a contribution in terms of approximately EUR 25 million from DIS compared to the EUR 27 million they raised in Q3. And the contribution in terms of EBITDA is almost close to 0%-1% because DIS is now investing also in new headcount.
We are planning to hire new additional headcount, especially in Taiwan and China, to increase our capacity of design and development, considering the capacity we expect to be needed in the first quarter, in the first half of 2025. We talk about additional 50-60 people just in the Q4. That's an increase in headcount and labor cost, especially in Taiwan and China, and the contribution will be almost nil for the Q4. Back to the second point about the CapEx. Out of the EUR 55 million already cashed out in the nine-month period, more than half has been invested in Taiwan, especially in the second fab to increase our capacity for MLOs. We do expect to continue also in the Q4 such trend of investment, even though initially we expected to spend approximately EUR 70 million only in Taiwan for this increase in 2024.
We have a bit of delay in the investment expenditure, so we plan to end this investment program in the first quarter of 2025. So we talk about approximately EUR 20 million that has been postponed to the Q1. But this is due to the contingency in the country because it's not easy to find, again, headcount, it's not easy to find a provider of services, and also materials for machinery, equipment, and improvement of the building. So it's just a matter of delay of the investment for a quarter. Overall, it will have a total value of approximately EUR 70 million when the fab will be completed. For the other points, I'll leave to Stefano Felici to give some comment.
Yes. Starting from the consumer market, actually, we don't see the consumer market softening. Of course, it's not increasing very rapidly, but at the same time, it's not softening. So for this year, it was kind of flat, but we see for next year a low single-digit growth. And because also consider that now we start to see requests, and it's more common to see also this type of end customer to talk about edge AI. So since now, the biggest driver for the AI business was, of course, on the cloud AI, and this is where you hear a big announcement and big investments. But many of these consumers and customers now are moving in the edge AI because it's going to be much more efficient to have an AI that can communicate, can be done locally, and then communicate on the cloud.
So for this reason, again, we don't see this softening. Again, we see a single-digit growth for next year, but it's really also depending on how quick the AI application will be deployed in the market and used by the end customers. About HBM, so this is probably you know already that the HBM business is kind of parallel to the AI business because for each AI computing chip, there are several HBM chips around. Okay? So the two businesses are parallel. So as far as if you see the AI growing, same is doing the HBM because it's parallel. So what is interesting in this market for us that is not a conventional memory, but is much more difficult and much more complex as far as testing. And here is where we can leverage some of our technologies that we use in SoC. And this is becoming very interesting.
So our target is not to come as an additional player. There are already many players in the memory market with different technologies, with ASP pretty low for the RAM and NAND. But for HBM, it can be a completely different story. And that's why we were focusing now to penetrate that market.
Okay. Will this require, let's just say, new additional capacity? Because if I'm not wrong, let's just say for each AI chip for one GPU, you need roughly four HBMs. So I guess it's more volumes and so probably an additional production capacity needed. Or do you think you can, let's just say, start to produce something for the testing of HBM with the capacity that you have in place now?
We require some additional capacity, but our structure is already basically based on a MEMS factory. We have already the infrastructure to do many more wafers. So it's just really, as far as, let's say, equipment and the infrastructure, our structure can absorb much more, can increase much more the capacity. So it's just basically a matter of labor and can be done very quickly. So we already invested this year some of the CapEx were in that direction to get some of the components that are required for this type of protocols. So it was already put in place this year in the CapEx, in the investment that we have done this year. So we think that we are going to be ready.
So now when we are going to be qualified and we will start to receive orders, we are going to be ready to support the capacity needed.
Okay. Thank you.
Thank you, Mr. Selvetti, for your questions. Our next question today comes from Mr. Florian Treger. Please, the floor to you.
Thank you. I hope you can hear me.
Yes, we can indeed.
Perfect. Thanks for taking my questions. I got two. One is a financial one. The other is more strategic. So on the financial one, could you say what your margins would be if you would get your capacity utilization back to levels of 22? And the second question, more strategic, is on HBM and final test, can we expect any revenue contribution here in 2025? Thank you.
Thank you, Florian. So for the capacity question, if I well understood, we currently run our business with a capacity utilization between 75%-80% in our manufacturing sites, depending on the various departments, but the range is between 75%-80%. So currently, we can stretch a bit our utilization. We are underutilized. If arriving to 100%, currently, we can stretch our revenue for the organic business for more than the revenues we had in 2022. So when I talk about organic, I mean only SoC business. So without considering Harbor and DIS, that have a different capacity and different utilization. So just to give you some more reference, the contribution we had in the first nine months of the year only from the SoC business was in the region of 29% EBITDA, considering all the limitations we had in the manufacturing yield.
But given that we have almost resolved everything in the Q4, we expect to be back with this level of revenue with more than 34.25% contribution, so EBITDA. So it's a very high level for the organic growth, considering also the current level of revenue that is not, of course, the desirable level we have now because we want to be back absolutely to fiscal year 2022. Having said that, so our structure of business has not changed during the year. So we still have the same number of headcount for this business and the same level of investments and R&D employed in our business. So the utilization is just a matter of volumes out of the market.
Then the second question was about expected revenue for HBM and final test, correct?
Yeah. Yeah, that's correct.
For HBM, right now, again, I think we cannot say we don't have enough visibility to tell you what can be the expectation because there are qualifications in progress. And once this will be over in the positive way, of course, at that point, we are going to have some visibility. And again, I repeat, the good thing that we think is good for us is that these chips are much more difficult to be tested compared to regular DRAM. And this really is an opportunity for us to use some technologies that typically are not used for memory cards, but are used for SoC cards, which are much more difficult. So we are positive about our qualification plans, but we will have some more visibility starting from Q2 next year. And for final test, we have already, actually, business, but it's going through DIS.
Here, all the final test strategy will be consolidated in the DIS strategy. This is already something that is ongoing right now.
Thank you.
Thank you. Thank you, Mr. Seger, for your questions. Our next question today comes from Mr. Domenico Ghilotti. Please the floor to you. Mr. Ghilotti, kindly unmute your line.
Okay. Can you hear me?
Yes, we can indeed. Thank you.
Okay. Sorry. So a couple of questions. The first is related to the trend for data center AI. So if you can give us a sense and on the other side, on the out and out. So if you can give us a sense of what is the trend that you have seen for both segments into Q4 and what can be, say, the expectation for 2025. And the second question is related to, well, a question related to duties. We are all concerned about potential tariffs coming from the new U.S. administration. So how can you deal with this risk? Are you structured to, say, to shift or to have some flexibility in your production capacity?
Okay. About the expectation for data centers and AI, again, here, for the visibility we have, we can see low single-digit CAGR in the next four years. So this said can be unpredictable how fast once the big company like hyperscalers, other company will deploy, and these applications, the application in the public, in the public domain might be even better. But for now, if we go by the predictions of the market, we think it would be low single digits.
Sorry, just to interrupt you. So are you referring to the, say, more consumer-driven AI? So you were mentioning before the edge AI or even on data centers or more cloud-based AI?
Actually, we see still a big push on the cloud, but it is still with some capacity constraint from supply chain. But anyway, so this is still very hot, I mean. But overall, let's say if we combine the two, will be again the low single digit, the total growth. But AI might be faster, but this will be, we'll have to see to know.
Okay.
Second question.
If I may add, the total growth is high single digit. What Stefano mentioned is related only to AI in data center, so overall, for the logic segment, it is expected to be 2025, high single digit. For the second point of your questions about the duties, we are not currently affected by any particular increase in duties, especially for the U.S. market, and also, you should also consider that our main market currently is in Taiwan, so it's in Asia, Taiwan, and Korea also. It's not only the U.S.. Most of the big U.S. players currently got the testing in Taiwan as well, so they are U.S. companies, but the testing is done in Taiwan, so our products are very often addressed outside the U.S., but in general, we are not at all affected by this kind of restriction for the moment.
There is no clarity, no sense, or no rumors on the increase of duties.
Okay. So even when you are billing U.S. clients, typically the business is executed outside of the U.S.?
I would not say typically, but it's a good part of the business is addressed in Asia.
Okay. Thanks.
Thank you. Thank you, Mr. Ghilotti , for your questions. Our next question is actually a follow-up from Mr. Bonaccina. Please, the floor to you.
Yeah. Just a quick follow-up. Reading the PR, there is a reference to basically expanding the product offering by type and customer. I thought that in the logic segment, you have basically all the top semiconductor companies as customers. So what you are referring to, you are referring to some small new customer or because you are expanding into memory? So this is the reason why you are indicating that you will get new customers. Thank you.
In terms of customers, you are correct. So we already work with all the main customers of the industry. What we referred to in that case was to expand our business with them into new segments. So for example, basically final testing. For example, in Asia, we were not in for the final testing. Now we are, and we can expand for already existing customer a new portion of the business. But there are not currently customers not served by Technoprobe. I mean, in terms of the major customers. And any new customer would have been negligible in terms of additional contribution.
Okay. Clear. Thank you.
Thank you, Mr. Bonaccina. Our next question comes from Mr. Luca Bacoccoli. Please, the floor to you.
Hello. Good afternoon, everyone. Can you hear me?
Yes, we can indeed.
Okay. Good. So my question is on building block profitability for next year, if I got it right. The fourth quarter EBITDA margin on the, let's say, core business is going to be back to 34%-35%. And I assume that this level of volume growth, this threshold could be also reached next year. So I was wondering how to look at the DIS and Harbor Electronics EBITDA contribution in 2025 just to have the overall picture. Thank you.
Thank you. So when I mentioned the increase in the profitability in the region of 34% or 35%, is the profitability of the organic business, so the SoC business standalone for the Q4, that is more or less in line with what we have in the past before having the troubles in the manufacturing phases. So this is not to be projected to the whole company because the whole company, as mentioned many, many times, is diluted through the contribution of Harbor and DIS. So when we talk about more than 30% of EBITDA, this is strictly related to the SoC business. So the contribution of DIS and Harbor is currently pretty low compared to the numbers we are accustomed to. What we are planning now to increase, I mentioned many, many times, we are still integrating the two businesses.
What we plan in the next few months in the very short-term period, we are already elaborating a reorganization strategy that will largely affect DIS and Harbor business. That will be, let me say, very combined with each other. It will be very strictly related to each other. And this reorganization will happen more likely in the Q1 2025. And we expect to see impacts and benefits since the beginning of this reorganization. So already since Q1 or even Q2 2025. This will improve our efficiency, internal efficiency. Also, the way to work, all the processes will be impacted. All the line of change, the supply chain will be impacted, relationships with customers and supply chain as well. So we have an ambitious plan that will be achieved in the really next few months.
Okay. Thank you, Stefano. Very clear. And related to this, should we expect some external costs related to these restructuring activities on DIS and Harbor Electronics?
Actually, we do not expect a big reorganization cost. So the reorganization will have, of course, will bring some, let me say, extraordinary expenses, but will not be dramatic. So it will be under control. We have our estimates, and it will be more and more than compensated by the benefits.
Okay. Thank you. And my last question regarding the CapEx plan for next year. So you said that EUR 20 million CapEx expected for 2024 have been shifted to next year. Overall, how we should see the CapEx plan for 2025, please?
We are still working on the CapEx plan. So the CapEx should have been approved already still by the board of directors. So we cannot disclose that much in 2025 apart from this shift of the already approved investment of 2024. We do not expect to have the same level of investment as we have done in 2024. So we believe to be more or less on the same level of prior years. So in the region of €40-€50 million CapEx in 2025. But this is still to be confirmed. I will give you more and more information in the next quarter once all the plan will be approved by the board.
Okay. Thank you very much.
Thank you, Mr. Bacoccoli, for your questions. Next question comes from Mr. Fabio Pavan. Please, the floor to you.
Yes. Thank you. And thank you for taking my two questions. The first one is follow-up on profitability. Stefano, thank you for the details on the reorganization plans for DIS and Harbor. So I was wondering if also thanks to this, we may expect EBITDA margin to go back to above 30% in 2025? A second question is for the CEO. Do you think that AI contribution could exceed 20%-25% of new products next year? Thank you.
Thank you for your question. So the profitability expected after the reorganization is still early to disclose because it also depends on the month when that will be completed. So one month or two months later could affect the profitability for a couple of points. So probably it's not fair right now to show exactly how much we expect. But let me say overall, if everything is completed and running on a 12-month basis, on a TTM base, we could gain at least 3%-5% EBITDA only from the reorganization. That doesn't mean that it will be reached in 2025 because, again, it depends on when it will be improved. But on a normal run base, it will be in that region.
About.
3 to 500 basis points increase in EBITDA and.
3 to 5.
Yeah.
3%-5% EBITDA recovery.
Okay. Thank you.
Only from the reorganization.
Thank you, Mr. Pavan.
Yeah. There was a second question about the weight, basically, of the AI in our business. We can say that for this year, 2024, it's already reaching the 25% threshold. And for 2025, considering that we still see strong demand on the cloud AI and adding also the edge AI now, so we might reach around 30%.
Okay. Thank you.
Thank you very much, Mr. Pavan, for your questions. Next question comes from Mr. Alberto Berizzi. Please the floor to you.
Hi. Good afternoon, everybody. My first question is on the industrial automotive, if you can share additional flavor on, in particular, what you expect for next quarter and 2025. Then a second one on HBM, if you can give us an idea of the addressable market of HBM probe card in 2024, and the market expectation also on the next year. My last one on final testing, when do you expect more material contribution from new products, I mean, the new integrated products for final testing market on top of the revenues of DIS? Thank you.
So for auto industrial, the market will be impacted in the slowdown in the automotive segment. Also, what is in the first half, at least in the first half of 2025. And our exposure to this segment is about 20%. So we don't think that would be any growth for sure, but would be kind of flat, this market, at least for the first half of the year, 2025. For HBM, so right now, as I mentioned before, we think there is all the HBM eventually after there are, as you know, three major players doing this type of, let's say, two major players with one that is really, really very strong now, and the other two, they're following, basically. So we are engaging in qualifications with all of them.
But to understand, really, to be part of this business, we really need to go through the positive results of this qualification. That point, that would be we'll have really visibility on what can be our growth plan, and we can have more we can give you more visibility on that. There was a third question.
There was a third question. Sorry, probably we missed one.
Yes, on the new final testing products.
Okay. So for the final test products, the strategy here is really part of the reorganization that we are carrying out, Stefano Beretta just mentioned, and because DIS is already in that, is doing that business already, and so part of the organization will be basically to transition this type of business into the DIS business, which is already active in the final test business, so for the numbers for the DIS, it is expected also there to have also in the final test market to start to grow next year, and we are going to have also in this case some better visibility starting from next quarter, but the expectation for DIS, including the final test market, including the integration of the new products for the final test market, starting from Q1 next year.
Thank you very much.
Thank you, Mr. Jegra. Next question is a follow-up from Mr. Giovanni Selvetti. Please, the floor to you.
Hello?
Yeah. Sorry. Maybe the last one on my side is about the cash because as far as I can see, if you don't do buybacks, more or less, you can cover the CapEx with the cash from operating activity. So I was wondering, should we think that this cash will be invested only organically, or should we think about other M&As in the pipeline? If so, what is that you think you're missing and what could be of interest? Thank you.
Our strategy on, of course, having cash is we are in a good position, a strong position financially. And our position has always been to invest in technology. So this is really the point. And at this moment, we don't have any big target. But of course, we are considering many different options to improve some of the technologies, some of the processes, some of the components that we're doing. So in the course, having cash will be we will have the possibility to take the opportunities as soon as we will see. So.
As you can see from the numbers, all the cash from the operating activities is currently engaged into our own business. So we really believe that any penny invested in our company will give more profit than in any other way. So everything is invested in CapEx to increase our capacity, increase our technology.
Okay. Okay. Thank you.
Thank you, Mr. Selvetti. Next question came in via chat. It's from Mr. Spagiani. Good evening. Just for a pure sake of knowledge, how much of your revenues on an annual basis have a delivery in Taiwan? Thank you.
On annual basis, I think we have more than 50% delivery between, let's say, in Taiwan, yes, you are in the region of 50%, regardless of the geography of the customer.
Thank you very much for your answer. There are no questions queued at the moment. We will wait just a few moments just to give everyone the chance to raise their hands. Thank you. As there are no further questions, I will now leave the floor to the management team for any final comments. Thank you.
Thank you, everyone, for joining us today. And stay tuned for the next quarterly call. Thank you.
Thank you. This presentation will now come to a close.