Good morning, good afternoon, ladies and gentlemen, and welcome to the Besi Quarterly Conference Call and audio webcast to discuss the company's 2023 fourth quarter and full-year results. You can log in to the audio webcast via Besi website at www.besi.com. Joining us today are Mr. Richard Blickman, Chief Executive Officer, and Mr. Leon Verweijen, Senior Vice President, Finance. At this time, all participants are in a listen-only mode. Later, we will conduct a question-and-answer session, and instructions will follow at that time. As a reminder, ladies and gentlemen, this conference is being recorded and cannot be reproduced in whole or in part without written permission from the company. I would like to now turn the call over to Mr. Richard Blickman. Thank you.
Thank you. Thank you all for joining us today. We will begin by making a few comments in connection with the press release issued earlier today, and then take your questions. I would like to remind you that some of the comments made during this call and some of the answers in response to your questions may contain forward-looking statements. Such statements may involve uncertainties and risks, as described in the earnings release and other reports filed with the AFM. For today's call, we'd like to review the key highlights for the fourth quarter and year ended December 31st, 2023, and update you on the market, our strategy, and the outlook. First, some overall thoughts on our performance. Q4 2023 operating results were significantly better than both Q3 and Q4 2022, as our favorable market positioning offset continued weakness in demand for mainstream assembly equipment.
For the quarter, revenue of EUR 159.6 million was up 29.4% and 15.9% versus Q3 2023 and Q4 2022, respectively. The increase was due to higher shipments for hybrid bonding, photonics, and other AI-related 2.5D applications continuing trends we saw last quarter. Of note, we shipped our first inline flip chip system for 2.5D HBM logic applications to address the needs of this growing market. Orders of EUR 166.4 million were up 30.7% versus the third quarter last year, of which a portion is anticipated to be shipped in Q2 and Q3 of this year. Operating profit also improved versus prior guidance, as gross margins increased to 65.1% due to a favorable advanced packaging product mix and net forex benefits, as well as cost control efforts, which kept overhead levels relatively constant versus the fourth quarter of 2022.
As such, net margins rose to 34.4% versus the 28.4% in the third quarter of last year and 29.2% in the fourth quarter of 2022. Overall, we are encouraged by our performance this year, as Besi's leadership position in advanced packaging lessened the adverse effects of an industry downturn as severe as the 2017 till 2019 period. For the year, revenue, orders, and net income: EUR 578.9 million and EUR 548.3 million, and EUR 177.1 million declined by 19.9%, 17.4%, and 26.4%, respectively, versus 2022. Revenue and order weakness reflected significantly reduced demand for mainstream computing applications by both IDMs and Asian subcontractors, and to a lesser extent, reduced demand for automotive applications following strong growth over the past two years.
Such weakness was partially offset by increased demand in the second half of the year for silicon photonics, hybrid bonding, and 2.5D logic memory applications, as customers began to build out their AI and high-performance computing capacities. In particular, hybrid bonding orders and year-end backlog approximately doubled versus year-end 2022. Of note, approximately half of Q4 hybrid bonding orders were represented by our most advanced 100 nm Generation 1+ accuracy hybrid bonding systems. We achieved peer-leading operating and net margins of 36.9% and 30.6% in 2023 due to the alignment of Besi's operating model to difficult market realities. In fact, gross margins increased to 64.9% versus 61.3% in 2022 due to successful product introduction supported by a keen focus on cost control efforts, effective supply chain management, and net forex benefits.
From an end-user perspective, Besi's 2023 revenue decrease was primarily focused on mainstream computing applications, as well as ongoing weakness in Chinese demand for mobile handsets. As a result, computing decreased by 24% of our end-user mix versus 30% in 2022, while mobile and automotive each rose two points to 30% and 18%, respectively. Besi's revenue and profitability has increased significantly since the last industry downturn. As seen in this next chart, revenue, orders, and operating income in trough year 2023 grew by 62.5%, 57.2%, and 132.2%, respectively, versus 2019, with operating margins up by 11.1 points. Besi ended the year with a solid liquidity base consisting of cash, cash equivalents, and deposits, aggregating to EUR 413.5 million. Of note, we completed the EUR 300 million share repurchase program in October 2023 and launched a new EUR 60 million program due for completion in October 2024.
As such, share repurchases increased by 45.4% to EUR 213.4 million last year, or 2.6 million shares. In addition, we proposed to pay a cash dividend of EUR 2.15 per share for approval by Besi's 2024 AGM, which represents a payout ratio of 94%. Including such a dividend, we will have returned approximately EUR 1.9 billion to shareholders since 2011, or approximately 30% of cumulative revenue during this period. Next, I'd like to speak a little bit about the current market environment and our strategy. We believe we are in the early phase of a next assembly upturn after a 40% market decline from 2021. Industry analysts anticipate a rebound 2024, 2026 driven primarily by recovery in mainstream assembly equipment demand and Chinese markets. Additional capacity needed for next-generation AI, logic, and memory applications, and new wafer fab facilities coming online requiring advanced packaging capacity.
TechInsights estimates market growth of 78% between 2023 and 2026, reaching a new peak of $7.3 billion. However, the slope of the recovery in 2024 is uncertain given restrained demand for mainstream applications and weakness, in particular, in automotive markets currently. We made significant progress from a strategic perspective this year. Besi maintained attractive levels of revenue and profitability relative to peers due to our significant R&D investment in next-generation advanced packaging systems and rapid alignment of production and overhead levels to difficult market conditions. In addition, we completed an in-depth strategic review to better position Besi for growth over the next decade and industry upturn. We also expanded our operational footprint in Malaysia, in Singapore, and also in Vietnam in response to customer reallocation of certain production outside of China and in anticipation of the growth of hybrid bonding and other advanced packaging technologies.
Significant progress was also achieved on our ESG agenda, a s we made advances in the sustainable design of our platforms, positioned ourselves to meet and exceed challenging targets set for 2024, and launched many new initiatives to further reduce Besi's environmental footprint. In addition, we formed a technology advisory board with leading industry experts to advance our core technology competitive position and growth prospects. Our leading position in advanced packaging was on display this year. We introduced new products as planned both for 2.5D and 3D assembly, including our TCB Next inline flip chip system and the next-generation 100 nm accuracy hybrid bonder. Our installed base grew to over 40 hybrid bonding systems at year-end, with adoption increasing from three to nine customers encompassing North America, Europe, Taiwan, and Korea IDMs, foundries, subcontractors, and research institutes for both logic and memory applications.
Now, a few words about the guidance. For the first quarter this year, we expect revenue to decrease between 5% and 15% versus the fourth quarter, and for gross margins to range between 64% and 66% due to a favorable advanced packaging product mix. Baseline operating expenses are forecasted to increase by 0%-5% versus Q4 last year, with total operating expenses increasing by approximately 50% due to a $15 million increase in share-based incentive compensation expense. That ends my prepared remarks. I would like to open the call for questions. Operator.
Thank you. If you would like to ask a question or make a contribution on today's call, please press star one on your telephone keypad. If you change your mind and want to withdraw your question, please press star two. Please ensure your lines are unmuted locally a s you'll be prompted when to ask your question. Our first question comes from the lines of Charles Shi from Needham & Company. Please go ahead.
Thanks. Hey, Richard. I think you've said for quite a while that Intel is going to put hybrid bonding into the Forests and the Lakes, I mean, the code names for their data center and the client CPUs. Yesterday, they did announce that the first Clearwater Forest, which is the first Forest to have hybrid bonding. My question is: do you still expect that, I mean, hybrid bonding will come to the Lakes? Presumably, the Forests, those data center products, they're probably high value but probably a little bit low volume, but the client CPU side probably will drive a lot more meaningful volume for you. Do you see any ordering activities in preparation for that ramp, hybrid bonding in the Lakes? Thanks.
Well, we see, thanks, Charles. We see ongoing development process qualifications for many applications. And as said in previous calls, the industry is evaluating, more than only Intel, where hybrid bonding is advantageous compared to, for instance, a TCB interconnect solution. So over time, we will understand ever better which applications are being chosen to use the hybrid bonding. What is important with the message of Intel yesterday is that this is a very clear signal confirming the establishing of a main volume capacity onshore in the U.S., fully focused on using hybrid bonding. So a benchmark in that sense, which is critical for further adoption of this technology.
Got it. So may I ask a little bit more on your largest end market, I mean, at least for now, the mobile side? The mobile revenue seems to be down a little bit last year. And you also mentioned it seems like in last quarter, you did not see pulling orders from your high-end smartphone customers. I wonder, what's your view on 2024, first on the overall mobile revenue and the high-end smartphone revenue for you this year? And is the lack of the pulling orders from that high-end smartphone customers last quarter a reflection of a potentially lower demand this year, or maybe just a reflection of maybe the supply chain has really returned to normal so that customer doesn't really need to do any of the pull forward this time? Thank you.
It's the latter. So from supply chain constraints, especially in 2022-2021, it has come back more to the normal pattern. Although your statement is not 100% correct, we did receive orders for applications in high-end smartphone in the fourth quarter. But it's more according to the model before COVID, where we will understand in the next couple of weeks how much and for which applications new features will be added, with then orders delivered in the later part of the second quarter, early part of the third quarter. So that is very much tied to anticipated market demand of a next generation. On the other hand, overall, what we read in the general press, the expectation is not that, let's say, huge in volume as we had in 2022, also driven by COVID. So on the one hand, it is back to sort of a normal pattern.
On the other hand, when you read, again, the general information, whether it be a very strong high-end smartphone year, is still to be seen.
Thanks, Richard. Maybe lastly, I want to ask you about those Inline flip chip tools you shared into the 2.5D application. So I was kind of surprised that this is the first tool. And I don't fully understand what that inline here means. And what is the significance of this tool? And what kind of incumbent tool you are replacing? And can you give a little bit more color?
This tool builds simply 2.5D modules, you could say, in the 2.5D family like cores. And it's a complete line of placing devices with a flip chip technology in particular, and also the final mounting onto a substrate. And we have received, yeah, very important first automated lines for these modules, 2.5D. It's in Korea. We have already delivered similar concepts to Chinese high-end smartphone manufacturers, by the way. But this is mostly for computing devices. So in that 2.5D family, you have, on the one hand, the cores, and then you have several other devices with a similar architecture.
Maybe a quick follow-up. Speaking of cores, does this mean this tool can not only do the OS part but also the entire DOW plus OS? Thanks.
Yes. Exactly.
Thanks. So that will be all for me. Thank you.
Thanks, Charles.
The next question comes from the line of Madeleine Jenkins from UBS. Please go ahead.
Hi, everyone. Thanks for taking my question. My first is just on a comment you made in your release around the hybrid bonding orders in 2023 being double the comparable levels of last year. I just wondered if you'd give us some color on this. On my numbers, I would imply a big hybrid bonding order in Q4, but I just wanted to check if that was right and also if this went to multiple customers or just one. Thank you.
Well, first of all, as we also mentioned, in 2023, we went from three customers to nine. That's already a major step. It is from multiple customer orders. The interesting thing is, of course, it is a mix between the first generation, which is 150 nm-200 nm accuracy, to now the 100 nm accurate system, which we call Generation 1+. If you remember, we shipped the first system end of Q2, and that has been qualified, and that resulted in additional orders in Q4. So that is also our statement, that the mix of Generation 1 and 1+ is about 50/50 in orders for Q4. From a number of customers, we received orders in Q4 from four customers, four distinct customers, being IDMs and foundries.
Okay. Thank you. And my next one is just on the photonics demand you saw last quarter. I was just wondering if this has been kind of sustainable, have you seen more orders, and what you sort of expect going forward into 2024? Thank you.
Well, that's all of a sudden. We mentioned that also in the call end of October. We have been active in photonics for a long time, goes back 12 years. And that has suddenly increased in demand quite significantly, multiple customers, five customers, we mentioned at end of October. Order intake is continuing also in the first quarter. We also categorized that under advanced packaging. So advanced packaging is not only hybrid bonding. And so to answer your question, that is not a one-time event that could lead to, let's say, continued demand, certainly in the first half of this year.
Okay. Thank you.
The next question comes from the line of Robert Sanders from Deutsche Bank. Please go ahead.
Yeah. Hi. Good afternoon. My first question is just on HBM. I was just wondering if you still maintained that hybrid bonding adoption would be likely on the second generation of HBM4, or do you think there are customers that are considering introducing hybrid bonding for the first generation of HBM4? And the second question would just be around Intel Foveros. It looks like they're using TCB for substrate and hybrid bonding for die-to-die. On the TCB side, do you see yourself potentially becoming a preferred supplier at Intel, or is that a little bit of an ambitious hope? Thank you.
We're always ambitious. Your first question, whether it's Generation 1 or 2, is not yet clear. But if you ask me, I would guess more Generation 2, although there's an enormous amount of testing going on both at the center of excellence in Singapore, the combined centers of excellence of AMAT and WASI, and then with two major memory customers. We mentioned earlier Micron. We've also shipped the first system into a Korean customer. So that's also good news. And that's all in qualification mode. Your second question, the key answer to that or what we hear is that as long as the bumped pitch is above 25 microns, that customer continues with its suite of TCB tools currently installed.
Once it goes below that, and that's why we have shipped also a system to that customer, ordered by that customer, that's all for development, let's say, below 20-micron bump pitch. There's a similar opportunity and demand, and we will ship a system at the end of this quarter to another customer also for tighter bump pitch dimensions. So it's not a question whether we will have an opportunity. We have opportunities. It's a question of whether the design of those devices move to dimensions which require a more accurate TCB Next solution. And that's.
Got it. Thanks a lot.
By the way, also, Rob, a die-to-wafer solution. That's not die-to-substrate. So one should bear in mind that that is for die-to-wafer applications.
Got it. Thanks, Richard.
The next question comes from the line of Didier Scemama from Bank of America. Please go ahead.
Good afternoon, Richard. Thank you so much for letting me on. I just wanted to clarify a couple of things. On the press release, you mentioned that your install base of hybrid bonding systems was just over 40. I could be wrong, but I thought you said your install base was around 28, or at least you shipped 28 in 2022. Does that mean you shipped about maybe 13 or so hybrid bonding systems in 2023, or do you not include maybe the one you're selling for R&D? Just trying to clarify that for ourselves, I think.
Those are revenue-recognized systems, the 40, 41. Then we have development systems, so test machines which may end up in orders or may be exchanged for high-volume production versions. So, to clarify the question, how many bonders are up and running? That is the answer. So we made significant progress last year.
Right. But I think at the end of 2022, you were hoping to be flat on hybrid bonding system shipments in 2023. Did you achieve that if you take into account also the sort of systems sent for R&D, etc.?
Well, it's probably flat, maybe even some less because don't forget we shipped in 2022 quite a number of machines, but which then were producing or let's say they were showing yields of only, yeah, let's say, below 50%, even at the beginning, 20%. And once that yield moved up, capacity simply was added through system performance. Last year, that continued significantly. So these machines are pretty reliable as they are, and they're simply following the demand. So when there's more demand, more bonders will be added. And that's also the guidance for Q1 for multiple customers. And as we said earlier, the backlog end of 2022 was half of the backlog which we have at 2023. And that's also confirming further adoption and orders.
Can you say how many systems are in the backlog end of 2023?
No. We haven't mentioned that, and also we said that already earlier because it doesn't help. I think it's more important that on a regular basis, we inform those numbers of machines which are revenue-recognized and which you can see in our revenue.
Okay. Understood. And then for this year, I mean, I think market expectations are for hybrid bonding system shipments of, I don't know, around mid-40, something like that. Is that something you're comfortable with when you look at the adoption of hybrid bonding at your main foundry customer but also North American customer? And obviously, now, the Korean memory makers and U.S. memory makers also trying to work on high-bandwidth memory. Is that something you are comfortable you would like to underwrite at this stage, or is it a bit too early in the year?
It's too early in the year. We don't guide further than one quarter simply because there are many factors influencing this further proliferation. The first question also alluded to that. What's very fascinating, customers are very and they also explained that to us in a phase where they decide, will they use hybrid bonding or still a flip-chip type of process like TCB or even flip chip? And that can lead to many more systems or less systems. But the adoption at nine customers is very important. And that means that the drive to adopt hybrid bonding has increased quarter by quarter. And to how many bonders that will lead is hard to tell. So we will update quarter by quarter how things are going. But with the orders in Q4, things went definitely in the positive direction.
Okay. Thank you. Maybe my last question. ASML, on their last earnings call, they effectively said that demand from high-bandwidth memory is driving demand for EUV systems. In the medium term, they could be upside to their 2030 capacity of 90 EUV systems to address not only the doubling in die size of high-bandwidth memory DRAM chips but also the layer count. I think in the past, you also mentioned that HBM was probably a much larger opportunity for WASI than perhaps the logic accelerator market. So has your view changed, ED? Are you even more bullish on HBM than you were before? Related to that, do you see an avenue to penetrate the main sort of AI accelerator player that you don't have yet? Thank you.
Those are very good questions. Number one, we maintain the view that the volume in HBM is potentially significantly larger than for logic. Maybe a ratio 1:4, 1:6 is what is often used. It can even be more than that if you look at the designs, 12 stacks, 16 stacks. So that view has not changed. On the other question, yeah, how should I answer that? Well, I can't share more detail in that sense to that question. These are all forward-looking or, let's say, very much in the past, what I can say. We did update our model in a sense to bean count again, low case, a medium, and a high case. It's definitely intact. If you look at also, again, the announcement yesterday of Intel, which we see as a major confirmation that definitely the world is going in that direction.
Yeah, that's how we view the longer-term future.
Thank you. And maybe one last quick one from me, and I'll pass on to my colleagues. On the high-end smartphone adoption, are we still sort of around 2026 for the sort of main player? And then let's say the Android camp, high-end Android moving in 2027 via ASE, is that your way of thinking at this stage?
That is our thinking a lthough comments from, for instance, Qualcomm, they are more aggressive on that adoption for chiplets. As we have said also in the previous call, we are now already in 2024. So they have to hurry up to include that in a 2025 model. So in our view, it's more likely 2026. But it's certainly being developed, and it's developed on our equipment. So yeah, what the timing will be precisely. As you know well, I'm always a bit conservative in that sense. Customers are maybe somewhat more, yeah, let's say, aggressive in how that adoption will happen. Time will tell. But for us, the key, again, is are we involved in that development? And the answer is yes.
Wonderful. Thank you so much, Richard.
Thanks, Didier.
The next question comes from the line of Ruben Devos from Kepler Cheuvreux. Please go ahead.
Yes. Good afternoon. Thanks for letting me on. Just a clarification question on memory in HBM. I think the leader in HBM, they've been recently talking about 16-layer DRAM stack with the help of hybrid bonding technology, which they were quite explicit on. It could be helpful if you would help us understand where you are today in terms of engagements with them. Did I understand correctly that they are not included yet in the nine customers? That's the first question.
No. They are not yet included in the customers, but they are equally testing, evaluating to use the hybrid bonding technology for various HBM generations. It also has to do with, and you may have read that, about the thickness JEDEC standards. If they are not raised, they are forced to go to hybrid bonding. But some are arguing that that thickness will be raised or that limitation, and that will allow TCB to be used for a next generation. So there are many technical or technology considerations. But your question, are we involved in that? And the answer is definitely yes.
Okay. Great. And then with respect to the subcontractors, I think in the prior call, it was mentioned a major subcontractor had placed an order, and now it seems you're suggesting multiple have ordered. I believe, unlike before, the OSATs are sort of expected to be the late adopters, but it seems they're relatively early at this stage. I was curious to hear your thoughts, how significant these orders are to you, and how that reads for a potential acceleration of adoption.
No, no. Then I have not been clear enough. We still have only one subcontractor. The others are foundries and institutes and, of course, the big IDMs. But on the subcontractor front, there's, of course, a lot of interest. There's a lot of, yeah, ongoing exchange in test information with the big ones. And I can name them. We have, of course, ASE, as we mentioned in the order. Amkor is also actively looking at it, driven by an end customer. Those two are, yeah, most advanced in the OSAT, let's say, environment. So I don't know where we have given this information, but it's only one so far.
Okay. Yeah. I guess it was just interpretation. Just the press release was saying plural, subcontractors, when referring to the nine customers. But okay. That's clear. I would say final question just on the cross margins. They've gone, again, up well ahead of your own guidance and partly helped by a better mix, including more advanced packaging sales. I was curious, could you give some indications how the 2.5D assembly, 3D hybrid bonding tools, and the rest of your equipment basically relate to each other in terms of cross-margin contribution and maybe relative to group average, which was now at 65%? Thank you very much.
Well, we have always sort of guided that the more new technology, the higher the margins. It's, as a note, very important to understand that these margins, even in the early phase of this hybrid bonding, but the same is for 2.5D, that means that the gross margin potential is still higher than what we accomplish today. We have also guided for somewhat higher margin in Q1 simply because of the order mix. And as soon as the overall, the conventional business comes back, you will have a slightly different mix. But this between the low 60s% and, let's say, the higher 60s% is where the margins are currently. It also depends on Forex. We have had a favorable Forex relationship in Q4, second half last year, also with a very low Malaysian ringgit, and that helps in the cost. So on the one hand, strong product position.
On the other hand, favorable cost. I should also mention here that one of the silver linings of this whole COVID constraint in supply chain is that we have many more suppliers developed in that time, and that helps to negotiate also better pricing at this point in time. Also a favorable effect, yeah, in this mix for unique solutions in this 2.5D with the multimodule attach platform. So it's not just one product. It's a broad advanced packaging product range. And then you have more of the conventional products. And that mix will determine going forward how those margins will develop. But as we guide right now for the first quarter, it looks positive.
All right. That's very helpful, Richard. Thank you.
The next question comes from the line of Rolf Bulk from New Street Research. Please go ahead.
Yes. Thank you for taking my question. You mentioned that around half of your hybrid bonding orders in the fourth quarter were for the leading-edge 100 nm accuracy tools. Can I ask, are all of these orders still for standalone bonded tools, or are you seeing an increased interest and perhaps orders for the integrated tools, what you refer to as a cluster tool solution?
That's a perfect question. As we have shared in previous calls, the two major customers right now, one in Taiwan and one in the U.S., they have a different philosophy. In Taiwan, and probably because they have a foundry concept, they are still expanding standalone bonders and now also the 100+. In the U.S., it's very much focused on integrated lines. That is a conceptual choice. For us, it doesn't matter that much because the bonder is the bonder. The interface is, of course, different. Software is slightly different, of course. We've had now quite some time to design the interface between the bonder and an automated line. For the future, we expect that high-volume manufacturing will most likely use these integrated lines because they have a big advantage in terms of cleanliness. We need zero particles for a reliable hybrid bonding process.
And that automation leads simply to higher yields. But on the other hand, it is less flexible as a standalone setup. But your question in Q4, it was a mix. Some are for, and we said that also, that we expected these orders in the U.S. to happen in Q4, and that has happened. And some more standalone bonders for Taiwan and also for others. So yeah, that's the current situation.
Thank you, Richard. It's very clear. Maybe a quick follow-up. Could you remind us what kind of ASPs we should model for the 100-nanometer accuracy tool versus previous generations?
The first generation, Generation 1, the 150 nm-200 nm, is between EUR 1.5 million and EUR 2.5 million. The Generation 1+, so the 100 nm, is between EUR 2 million and EUR 3 million.
When you move to 50 nm?
That is still early days. We are now reaching 80 nm in the laboratory consistently, and we have to move down to 50 nm. Yeah, that will be more expensive. That will probably be somewhere between EUR 3 million and EUR 4 million. But that's too early to tell.
Thank you. Very clear.
The next question comes from the line of Timm Schulze-Melander from Redburn Atlantic. Please go ahead.
Yeah. Great. Thanks for taking my questions. A clarification, and then a lot of stuff's already been asked, so I just had a couple of housekeeping questions. Just on the bonder side, the non-hybrid bonders, but the TCB side, you began talking about applications potentially or indicating chip on wafer-on substrate. So would it be reasonable for investors to conclude that you are going to be exposed to growth in AI accelerators over the course of 2024? And then I had a couple of quick follow-ups.
Yes. That assumption is a reasonable one. But again, as I tried to explain earlier, our system is developed in particular for dimensions below, so the bond pad pitch below 20 microns. And that is where our system has a great benefit over the other systems.
Okay. So that's just a function of time and node shift and therefore transistor density?
Yes. Exactly.
Got it. Okay. Okay. Very clear. Maybe two housekeeping questions. You referenced the FX benefit a couple of times. Could you maybe just help us give us some kind of quantification for the 4Q gross margin impact?
Well, let me first explain. We are in a dollar market. We have cost in euros. We have cost in Swiss francs. We have cost in Malaysian ringgit, Singapore dollar, and Chinese renminbi. So it's a mixed I wanted to say bag, but it's a mixed pattern. And we have this overall reporting in euros. So what was favorable in Q4 is still a strong dollar, a weak euro. There was a weak Malaysian ringgit, which also helped. The Swiss franc remained strong, so that was yeah, the cost in Swiss francs is negatively influenced. But in total, you're talking about maybe one gross margin point. So it's not that spectacular, but still, it is a point. But the best to watch is always the dollar-euro relationship.
As soon as the U.S. dollar is weak and the euro would be strong, that has a negative impact overall because our business is in U.S. dollars for 80%.
Got it. Very clear. And then maybe just one other thing on the operating expense line. Clearly, 4Q orders, very solid, very strong. The pace of improvement in the hybrid bonding offering has been very strong. You're working very hard on version 2.0. Just on the stock-based compensation uplift, I know it's a seasonal pattern, but maybe you could just give us a little bit of color. Is that almost all the increase because of a higher share price, or is also the share count increasing because certain KPIs were met or exceeded? Maybe just some color around that stock-based comp number would be helpful.
No. Because of the increase of the share price, that has a negative effect on the number of shares in the compensation, in the variable compensation plan. So it's simply because of the high share price, so compared to last year. And the number of shares have gone down. But anyway, that is, in a nutshell, explaining the impact.
Got it. Very helpful. Thank you.
The next question comes from the line of Karl Lidsle from Sifter Capital. Please go ahead.
Sorry. I can't hear you. There's an echo. Again? I'm sorry. I can't hear you. Operator, can you?
Karl, your line is breaking up. We'll proceed to the next question. Next question comes from the line of Johannes Ries from Apus Capital GmbH. Please go ahead.
Yes. Good afternoon, Richard. Four questions very fast. First, you gave some market forecast for this upturn with 78%. Is it right to assume that you are because in the maybe high-growth path with this new technologies have some clear leads that you could maybe cross these expectations over this upcycle, or is it too optimistic to think in this direction?
Well, if you look at the last cycle, and as we mentioned, the year 2023, the overall market for backend equipment and our peers, competitors, the impact was more closer to 40%, and ours is 20%. So that is because of the focus on the advanced packaging. That could have a similar pattern in the next upturn. But the question is, will that happen? The market forecast is, as we know, a market forecast. But anyway, so we should do better if that happens based on the current market position and proven strength over the last 12 months.
Great. Maybe partly a follow-on. Given what all happens to AI, what we heard tonight from NVIDIA, this boom we see in HBMs, is it right to see that maybe some market opportunity of hybrid bonding is larger than we maybe have seen one or two years ago?
Well, as I explained to an earlier question, we think, based on today's data points, that the model which was developed is more or less intact. It can have some shifts in time, but you may remember that nice graph with a low case, a medium case, a high case, depending on whether it only remains logic and memory and then kicking in for high-end smartphones. So that's as it is right now. But also, I answered two earlier questions. It's very hard to find definite applications. You can argue below a certain bond pad pitch, you must use hybrid bonding. Will that accelerate is always a question. But the bottom line is so far, so good. If we look back to the past three years, hybrid bonding has developed beyond expectation.
The fact that there are now nine customers using this technology, partly also for mainstream products already for some time, that looks very promising, Johannes.
Great. Maybe to another also interesting, promising, yeah, business line or maybe activity you have for photonics, can you give us more feeling how big this market could be in some years? Maybe is it something between TCB and hybrid bonding for you, only to have a feeling? It's an interesting business, and if AI is growing, you also need faster communication. Therefore, maybe you have some ideas about the size of the market, which gives us an orientation for the next years.
We look at it from, yeah, let's say, a simple model. It could reach volumes like flip chip. That's the best analogy. The machines are more accurate, so they are slower, so you need more machines. But that is a market potentially, which is around EUR 100 million per year, and in a peak year, going to EUR 200 million, EUR 300 million. So a bit similar in size to other advanced packaging platforms.
Great. Yeah. Yeah. Great. Totally different question at the end. You formed this technology advisory board. Explain a little bit why, because you have so far did the right decisions, but otherwise, you have really put very well-known, great guys from customers and maybe partners in. Explain a little bit the logic behind why you're doing this and what you are discussing here.
The answer is very simple. The risk in choosing the wrong direction is increasing over time because the technology is ever more complex. The R&D spend is increasing year by year. This is not uncommon. There are many technology companies who have this concept of an advisory board. So what you want to achieve is challenges, also support in views. We have developed an agenda of meetings six times per year, three times face-to-face, one meeting in Japan, which is coming up in three weeks' time, next meeting face-to-face in the U.S. in early July, around semicon, and then end of September, the third meeting here in Europe, also meeting Besi supervisory board, and in between, virtual meetings. Yeah, the objective is to find out and challenge, are we understanding the directions in the right way, or is there another way to look at that?
It's testing our own intelligence ever more.
Great. Great idea. Thanks a lot.
Okay. Thanks, Johannes.
Bye.
We've come to the end of the call. We have to stop here because we, yeah, have set this for an hour. If you have any further questions, don't hesitate to contact us directly. You know where to find us, and we'll be happy to have an ongoing dialogue.
Thank you for joining today's call. You may disconnect your lines.
Thank you very much. Bye-bye.