Morning, good afternoon, ladies and gentlemen, and welcome to quarterly conference call and audio webcast to discuss the company's 2022 fourth quarter and full year results. You can log in to the audio webcast via Besi's website, 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 be followed 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 now like to turn the conference over to Mr. Richard Blickman. Please go ahead, sir.
Thank you. Thank you all for joining us today. We will begin by making a few comments in connection with the press release we issued earlier today, and then are happy to take 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 by management 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 of our fourth quarter and year-ended December 31st, 2022, and also update you on the market, our strategy, and the outlook. First, some overall thoughts on our performance. Besi's fourth quarter operating results were better than expected in an ongoing industry downturn.
Revenue of EUR 137.7 million decreased by 18.4% versus the third quarter last year, reflecting a number of headwinds, including weakness in high-performance computing and mainstream electronics applications and ongoing weakness in mobile end user markets. However, our backlog rose 12.2% to reach EUR 270 million due to a EUR 55.2 million or 44.1% sequential order increase. Such increase resulted primarily from higher bookings for high-end smartphone applications and hybrid bonding systems. Current order trends reflect customers' continued investment in high-end versus mainstream assembly applications, as well as ongoing weakness in demand by Chinese customers. Besi's profitability and efficiency remained at attractive levels despite the downturn, with gross margins reaching 62.3%, net income of EUR 40.2 million, and a net margin of 29.2%.
This year marked an important inflection point in our strategic development as we position Besi for sustainable growth over the next decade. Our business model generated revenue and profitability levels substantially higher than peers as we effectively responded to an assembly downturn following large capacity additions over the past 2 years. For the year, revenue and net income of in total EUR 722.9 million and EUR 240.6 million declined by 3.5% and 14.8% respectively versus 2021. Orders of EUR 663.7 million declined by 29.3%, principally due to decreased demand for high-end smartphone applications post-new product introductions in 2021. The decrease also reflected reduced bookings from Chinese subcontractors for mobile and mainstream computing applications linked to softening economic conditions.
Of note, revenue from Chinese customers declined by 33.5% and represented 25.9% of revenue in 2022 versus 37.6% in 2021. Revenue and order weakness in smartphone applications was partly offset by continued strength in Besi's computing/hybrid bonding and automotive end-user markets, as well as increased revenue from spares and service activities, which grew by 25.7%. We achieved peer-leading operating and net margins of 40.7% and 33.3% in a difficult environment as we successfully aligned production to changing market conditions. In fact, gross margins increased to 61.3% this year, due primarily to a 77% reduction of temporary headcount from peak first quarter levels, effective management of our supply chains, and price increases implemented to help offset inflationary cost pressures.
As you can see in this next chart, Besi's 2022 revenue decrease was almost exclusively due to lower mobile revenue. As a result, its percentage of the total revenue decreased from 43% in 2021 to 28% in 2022, primarily offset by an increase in computing end-user market applications from 19 to 30% and increases in automotive and spares and service to 16 and 17% respectively. Besi's revenue and profitability has increased significantly since the last industry downturn, as measured by a comparison of the years immediately following cyclical peak levels. As evident in this next chart, revenue, orders, and operating income in 2022 grew by 37.6%, 37.4%, and 70.3% respectively versus 2018, and operating margins expanded by 7.8 points.
Besi ended the year with a solid liquidity base consisting of cash and deposits of EUR 671.7 million or EUR 8.56 per basic share, and a net cash of EUR 346.5 million. We will propose a cash dividend of EUR 2.85 per share, which reflects a payout ratio of 93%. Including such dividends, Besi will have returned approximately EUR 1.6 billion to shareholders over the past 13 years, or approximately 25% of cumulative revenue. Shareholders were rewarded for their investment in Besi with an increase in dividends and share repurchases of EUR 236.8 million or 132% versus 2021.
Of note, we repurchased 2.7 million shares this year for EUR 146.8 million and have bought another 670,000 shares to date in Q1. Our objective is to further reduce Besi's share count to offset potential dilution from prior convertible bond issuance. Next, I'd like to speak a little bit about the current market environment and our strategy. As seen in this next chart, industry conditions have deteriorated significantly since the second quarter last year, highlighted by slowing memory, mainstream computing and data center markets, continued weakness in Chinese markets, and CapEx reductions announced by many of the largest semiconductor producers. The assembly market was the first to experience the full impact of the current downturn.
TechInsights estimates that our market decreased by about 10.6% in 2022, with an additional decline of 16.8% anticipated in 2023. Thereafter, a new upturn is expected, with the total market reaching $6.9 billion in 2025. We made a number of important R&D and business investments in 2022 to better position ourselves for anticipated growth over the next industry cycle. Development spending was increased by 48% to ramp hybrid bonding for commercial production, to introduce two new wafer-level assembly systems, and to upgrade our existing product portfolio. Operational resources were utilized to increase clean room production and service support capacity in Malaysia and Singapore as we prepared for anticipated hybrid bonding growth over the next 5 years.
To prepare for expected hybrid bonding market growth, we have increased development staff in Austria, Singapore, Taiwan, and the U.S. by 26% over the past 2 years and expanded capacity to 12 to 15 hybrid bonding systems per month. We retrofitted our Malaysian production facilities to better protect them against potential climate change-related events. Work has begun on a new tooling facility in Vietnam as many customers shift a portion of their production outside of China to other locations in the Southeast Asian region. Substantial progress was also achieved to help hybrid bonding become a market reality. Significant improvements in placement accuracy, throughput, yield, and lead times all contributed to its commercial viability. Full-scale production began in the second half of 2022. Besi has shipped 30 21, of which 10 are demonstration units.
Per revenue, producing unit, we shipped 4 in 2021 and 21 in 2022. In addition, we received orders for incremental 14 units from multiple customers subsequent to Q3 last year, of which 3 orders were received to date in the first quarter of 2023. Orders received in Q4 2022 are not anticipated to be shipped until Q2 of this year, of which several are to be incorporated into integrated production lines. Of note, the first integrated production line was shipped in the fourth quarter. Interest in hybrid bonding process technology is also gaining significant traction with the development community. We shipped an evaluation system to imec in Belgium, an industry re-research institute, and received an order in Q4 from a prominent American university for hybrid bonders as part of integrated lines.
In addition, chiplet interface standards are now being developed by the industry's largest players to help facilitate its adoption and utilization. Progress also continued on Besi's ESG agenda this year, where 2022 targets set in 2020 were met or significantly exceeded in key carbon emission categories, as well as in the areas of waste, hazardous materials, and renewable energy. A few words about our guidance. There's a high degree of uncertainty as to the outlook for 2023. The assembly market is in a classic downturn after two strong years of growth. We believe there are a variety of potential outcomes for Besi's prospects this year in the context of the current down cycle, including the outlook for smartphone and hybrid bonding demand, as well as the impact of the reopening of the Chinese economy.
The headwinds we face are many, including higher inflation and interest rates, decelerating economic growth, geopolitical tensions, and ongoing weakness in mainstream electronics, computing, and mobile handset end user markets. For Q1 2023, we forecast that revenue will decrease in a range between 0% and 10% versus the fourth quarter last year, as many orders received in Q4 are scheduled for delivery in Q2 and Q3 this year. In addition, we estimate that Besi's gross margin will range between 61% and 63%, and that baseline OPEX to decrease by 0%-5% versus the fourth quarter of last year. Total operating expenses are expected to increase by 15%-20% due to an incremental EUR 7 million of non-cash share-based compensation expense.
Further, we expect an effective tax rate of between 12% and 15% and CapEx of EUR 8 million-EUR 10 million for the year 2023. That ends my prepared remarks. I would like to open the call for questions. Operator?
Thank you, sir. A reminder to the participants, if you would like to ask a question, please signal by pressing star one on your telephone keypad. If you are using a speakerphone, please make sure your mute function is turned off to allow your signal to reach our equipment. Again, press star one to ask a question. We'll pause for just a moment to allow everyone the opportunity to signal for question. We will take the first question from Madeleine from UBS. Your line is open. Please go ahead.
Hi. Thanks for taking my question. My first is around hybrid bonding. According to certain press releases, we've heard that an American IDM will possibly have a slower ramp-up of tools this year than maybe previously expected. Could you just tell us what you're seeing from your end? If this is the case, will the like projected 50 tool capacity you have this year be satisfied by demand elsewhere? I have a quick follow-up, if I may. Thank you.
Well, as explained in the press release but also in the comments, hybrid bonding is certainly in an broader adoption trajectory, which started two years ago in Taiwan. Last year was critical to set up the first production capabilities in Taiwan. We have always shared with the investment community that according to our information, the U.S. should follow suit with maybe a year in between. Preparation is ongoing. As we also mentioned, we are setting up support group people in different locations. That could very well lead this year to a significant increase in the capabilities onshore in the U.S., likely with capacity coming on stream early next year. That's sort of the timing we hear. We are well prepared for t hat.
As you also mentioned, the 12 to 15 machines per month. If stressed, we can even build more. That is definitely sufficient to support the capacity required for the industry, simply bottom-up accounting, in the next 2, 3 years. It takes us about 6 to 9 months to increase our capacity further. This year will be very important. Further adoption, capacity expansion. You also see that in the orders we received in Q1. Let's say Q4 first, and then additional orders in the early part of this quarter. That basically fits that picture.
Okay. Great. Thank you. Another question I had was around your orders. You mentioned in the release that part of the increase you saw was due to high-end smartphones. Could you maybe provide some more details around the drivers within this? Is it, from specific content increase or more of a global market-driven? Thank you.
Well, it's content related. It is preparation for a next round. The question is always of how large will that round be? It's encouraging that these orders are placed in Q4. That also has to do with Supply chain issues which the whole world has experienced in the past two years. Yeah, it's simply preparing for a next high-end smartphone round.
We'll take the next question from Ruben from Kepler Cheuvreux. Your line is open. Please go ahead.
Yes, good afternoon. I just had the first one on hybrid bonding as well. I think in the press release you mentioned that you've made significant improvements on various KPIs. Could you maybe talk in more detail about the progress you've made? I believe you've mentioned in the past that sort of progress and placement accuracy and throughput would drive ASPs, and that if all goes very well, you could be looking for ASPs of EUR 6 million-EUR 8 million per system. Is that still your thinking today? Yeah, how do you see ASPs maybe evolve through the next 2 years compared to what it has been maybe for the 35 systems you've shipped so far? That's my first question.
Excellent. Well, let me first create some reference. Until the mid of 2021, we were in a development phase. As of Q3 2021, there was this famous first application and chiplet from AMD, which sort of ignited a mainstream adoption. That resulted then in orders which we started to install in early Q1 last year with some hiccup due to some production reasons. Basically, as of February, March, we really started to install machines to be used in full-time production. As we shared every quarter, this has been a typical road of many issues because these machines in the beginning are not fully tested in production environment.
We've come a long way, and we've shared in the last conference call, end of October, that the yield out of these machines has moved from as low as 20 to about 80% yield in end of October. That yield has further improved since, and that simply results in higher output per machine. One must understand, and we share that in the Capital Markets Day, simply comparing output of high-end flip chip machines to these hybrid bonders. With flip chip, we reach 7,000-8,000 units per hour. In this hybrid universe in the later part of 2022, we are now reaching somewhere between 1,500 and 2,000 UPH. About a quarter output.
We have also shared that there are continuously programs ongoing to make these machines more stable, higher yields, and that leads to further increase of output. That is needed to connecting chips to chips. As we enter into 2023, continued improvement on machines is ongoing every single day, we should reach that 2,000 UPH for certain die sizes as a standard, and that will simply cause further adoption of hybrid bonding. I'm explaining this a bit lengthy because this is with every new technology, you could ask the question, are you ahead of your expectation or behind? Well, today we can say, well, this is exactly what we have expected. Also the industry simply understands this.
That is why the adoption rate is gaining not only at specific customers but far more broadly. That looks all very positive, but we're still in an early phase, so a lot still has to come, and that will come. What we also should realize, we are now in generation one, and that is 150 nanometers placement accuracy. In production, maybe 200. We will ship the first 100 nanometers, so generation one plus or generation two is then capable of bringing this accuracy down to 100 nanometers. That is important because that is linked to a node size shift, around 3, 5 nanometers, and that will further create the possibilities of chiplet architecture design devices, and not only for high-end computing but also for high-end mobile.
Those developments are critical in the course of 2023, and then anticipating mainstream adoption in 2024. There's a big program. First of all, making the first generation more 100% production capable, reaching yields close to 100%. Then a next generation, going down to 100 nanometers, and that always starts with somewhat slower production capabilities because of the accuracy and especially vibration are critical issues in these applications. We will master that. A full program for 2023, heading up to a major expansion anticipated in 2024, 2025, and 2026.
Okay. Very clear. I think, Richard, you mentioned that there could even be adoption from high-end mobile. I think in the past, we shared that current market estimates do not really reflect adoption of hybrid bonding in this market. I mean, when you look at the slide in your presentation, that was always excluding high-end smartphone, right? Mainly on computing.
No, no.
If we have always, let's say, shared that the world is first and that is happening, high-end computing, data center modules. Graphics, so all logic-based, and then adding also memory to that, and then the high-end, let's say, smartphone arena, in 2024, 2025. That roadmap still stands. Roadmap and timing has not changed. Why is that important? Mostly, those roadmaps are very ambitious, and they continuously shift because of all kinds of technical reasons, and market reasons. It's fair to say that the adoption rate of hybrid bonding for every application is so far not disappointing. Let me share it in those words.
Okay. Very clear. Maybe a final question, a bit high level, just thinking about the Chinese market, for which demand was down materially last year. I understand you were not negatively impacted by trade restrictions, but what's your take on the production shifts that are taking place? I believe you also have some facilities there. What could be the impact, maybe both in the short term and in the long term? Thank you.
Well, it's common knowledge that every company is investigating and taking action to reduce exposure to China. For that reason, as we have done always, we follow customers, and we prepare for supporting customers in other Asian Pacific regions, which step by step will either be organized or set up for new products, but it may very well be that also existing capacities may shift. That's on top of the very high-end, where we have these restrictions now in place. To be very clear, we simply follow the direction of the U.S. in this case, huh. There's also no idea that this will change in any way and anytime soon.
Basic strategy is to be prepared for capacity expansions outside of China, for capacities moving out of China, supported by our facility in Malaysia, which we have expanded significantly last year with a major new facility added to our current facilities. Then We're in the middle of setting up our first operation in Vietnam. Those two should with the China, for China facility, which is ongoing, supporting the Chinese market, ready for how this market will develop in the years to come.
All right. Thank you very much.
We'll take the next question from Robert from Deutsche Bank. Your line is open. Please go ahead.
Yeah. Hi, good afternoon. I just had a question about the computing PC server business line. It seems like that is now the largest segment as per the last year at 30% of sales. That growth is pretty surprising to me, so can you just tell us a little bit more about what's happening there? Have you won more content at Intel, for example? Have you won some TCB business after ASMPT, or is there something else going on there? I have a follow-up. Thanks.
Well, the last one is still to come. As we explained in previous updates, we have or we are very close to shipping the first next generation TCB chip-to-wafer tool in April, which is fully ready for the next generation. So far in 2022, it also surprised us in a way. Let's say surprised in that sense, the drop in high-end smartphone was significant. The rise in the computing application simply adds the revenue recognition in hybrid bonding, which is for modules in data centers and high-end computing. Most part is ongoing strength in our market positions for a long time, and whether that's in the U.S. or in other regions.
Then on top of that, you have the hybrid bonding systems.
Got it.
That's also answering the question of some people, is this cannibalizing your current flip chip business? The answer is simply no.
Got it. On the order beat, you mentioned that your largest smartphone customer may have ordered more because of supply chain concerns. Does that concern you a little bit when you think about Q1 orders? I mean, are you kind of borrowing a little bit of the strength that you normally see in Q1 into Q4? Do you think that Q1 can still be as is traditionally the by far the least biggest quarter of the year when it comes to order intake?
Well, it's a bit of both. But you have to again segment the hybrid bonding orders. We had 11 in Q4, and decent ones, also ones to be connected to automated lines, and then 3 more in January. Then a nice round, first round for next generation high-end smartphones. How much more there is to come is very difficult to forecast. As we have shared over the years, the market timing decisions are very much February, March, sometimes even April. We have to be prepared. We simply have to be prepared to respond at the timing expected. We are. Anyway, it's exciting to see what may happen.
Got it. Thank you very much.
The next one is from Charles, Needham & Company. Your line is open. Please go ahead.
All right. Hey, thank you. This is Charles Shi from Needham. Hey, Richard, good afternoon, or maybe good evening. Maybe my first question, I wanna go back to your high-end smartphone demand. I think this may be the swing factor, because as you said, there's a variety of outcome for you in 2023. I noticed that for your Q4, you mentioned that year-on-year what's driving the bookings, the language you used in the press release seems to suggest that the booking in Q4 last year for a high-end smartphone customer seems to be relatively in line with fourth quarter 2021.
Looking into first quarter 2023, I mean, relative to first quarter 2022, I think that will likely tell us how your 2023 business is going to trend on a year-on-year basis. I know that the particular customer, they tend to place a greater amount of orders in Q1. We are at the end of February, now almost two-thirds into the first quarter this year. What can you tell us right now, relative to last year, the order coming from this high-end smartphone customer, is it gonna be better than last year, or in line or, well, I'm not trying to put words into your mouth? I don't think it's gonna get worse. How do you feel about that?
Are you feeling good right now?
Well, let me answer that with what we have said, also at the end of Q3, when we announced that we had some long lead item orders already in the third quarter, because of supply chain issues. If you look at the historic pattern of major new content in high-end smartphones, there was usually 3 years in between. The last time we had a major upcycle was 2021, a bit of an overflow in 2022. We have always been a bit careful in why would this 3 years be reduced to 2. On the other hand, we received these long lead item orders and we've also received very nice orders so far. There may be some impact because of significant market share gains of the US high-end smartphone manufacturer in China.
With an economy in China improving, maybe there's some anticipation of even stronger growth, to be expected. The only, yeah, let's say, the answer I can give you is that this is always uncertain in a way that we have to be prepared, but how much ultimately will come. I always remind you of the second quarter of 2018, which we will never forget, that we received a major round in February, March, April timeframe. In the third week of June, half of that was canceled. It's an unpredictable market. That's, that's it. Whether we like that or not, you simply have to be able to handle that. I can't give you more clarity, Charles.
Yeah. Thank you, Richard, for the color. I think the other, I mean, on a full year basis, right, 2022, your the weakness was attributed to two factors primarily, right? One is high-end smartphone. We just talked about that. The other one is China, and I don't think you've provided much specific view on your outlook in China this year. Can you give us a sense-
Yeah.
What is the order book you have you that tells you about the China strength this year? Especially we're out of the Chinese New Year. I think historically this is when new orders will come in from China. Over the last 1-3 weeks, how has your order book been trending out coming out of China? Thank you.
Happy to share. First of all, we were all, let's say, hoping for a next round after Chinese New Year because of a very slow market last year. With the growth ambitions and everything happening in China, one would expect that. Many of our peers and customers, sort of, also analysts in general markets, have expected a positive turn in the Chinese market. It's fair to say now, end of February, that that has not happened. Some are now expecting and voicing that that will definitely happen the second half of this year. The, yeah, Chinese economy is improving. If we simply bean count the utilization rate at the four major sub cons.
Those utilization rates today tell us, simply that they're not imminent to order a major next round. That fits more into as things develop, but also the worldwide economy, and if things don't deteriorate in general with China, U.S., then statistically with 6 quarters down, you should see a turn in tide in this summer and improving in the second half of this year. That's, that's, let's say, an answer of your question, from our view. If we are wrong, well, we can be wrong that it comes more early because some are saying, "Well, maybe it will come in April, May." Well, we're happy to see. If it takes longer, yeah, then you look into a model which the world does not yet share.
2024 should certainly be again, an up year. The question is, when does the tide turn? Well, Besi is prepared definitely for a turn. Any items in our machines and modules, also devices, critical devices, long lead items. You have to be prepared in this industry at short notice. The key is, Charles, we don't see at this moment a turn in this tide yet.
Thank you, Richard. Maybe just, lastly, I wanna say, your hybrid bonding shipment plus, new orders. You're approaching 45 total number of units and, almost there to reach the 50 tools. That's a magic number. I think that lots of people are skeptical about hybrid bonding, think it's ahead of its time, but I think your numbers are definitely showing the opposite. I wanna really say congratulations, and that's my last question. Thank you.
Thank you. Any further questions?
We have, sir. We have Didier from Bank of America. Your line is open. Please go ahead.
Yeah. Yeah, good afternoon. Thank you, Richard. I wanted to go back to the order intake for Q4. I'm gonna put words in your mouth. Sounds like EUR 60 million has been brought forward from Q1 to Q4. Should we just assume that-
Your Q1 bookings are gonna be around EUR 60 million lower than your Q4 level, or is there any reason why it could be flat given the seasonality?
Hybrid bonding is new. I can also answer it in this way. If we would not have the, let's say, success of hybrid bonding in the numbers, as explained to earlier questions, our numbers would reflect very much the overall downturn in the assembly market. Hybrid improves the picture, and that is continuing. That's a positive development on top of our conventional business. As we explained in the press release, and to repeat that again, we're in a typical downturn, a classical downturn, and some positive developments of new technologies, and one of them being hybrid, but there are also other new technologies. You can't simply say, "Well, 60 was ordered earlier because of supply chain issues, so Q1 should be 60 less." It's not that simple.
Anyway, I think the key is to repeat that again. So far we are going through this downturn in a very positive sense, with significant decline in the classical, and that is also, yeah, to some extent in the high-end smartphone applications last year. Chinese, as we discussed earlier, Chinese subcontractors, only very new things, but no capacity expansion whatsoever. We haven't discussed automotive. Automotive is still holding on. There are many strategic investments in expanding capacity for automotive. That also has helped us last year and will help this year. Yeah, that's roughly the picture.
Just wanted to go back to your point on hybrid bonding and high-end smartphones. You know, I think you said this is an important year to get to 99.9% yield on hybrid bonding. Any feel as to whether that happens in the first half or more in the second half? I just wanted you to repeat, if I heard correctly, high-end smartphone adoption for hybrid bonding in 2024, or if I misunderstood. Thank you.
Well, the first question, the 99.9%, we should reach this year, certainly 99%, whether it's 99.9, I don't know. These machines, as they are now running around the clock, are improving day by day. That is very positive development. What is key to understand, we shipped in Q4 the first, let's say, bonders, which were integrated to a line in Taiwan, and that is brand new. Although we have this first line in the center of excellence of AMAT in Singapore, we are now installing and starting up this line for production. In the course of this year, that also will take some time to become in the 99% category. Same will happen in the US.
A lot of focus on the integrated line concept and why this integrated line concept, because that improves the yield simply by less particles. That's very important for further adoption of hybrid bonding in all applications. To your second statement on high-end smartphone adoption, it's dependent upon very much the cost of ownership, which results from further improvement on the either standalone hybrid bonder systems or the ones which are integrated into integrated lines. That also will be critical this year and whether it will be ready for the major round anticipated in high-end smartphones in 2024. That's going to be the technology of choice, we will find out in the summer, in July, August, September timeframe.
Final question from me is on high-performance compute and, you know, the ChatGPT AI application we see. It feels like hybrid bonding is a perfect technology to tackle ChatGPT, you know, with the stacking of DRAM on top of server CPUs to run the inference or even to run the training of the algos. One of your customer is actually showing hybrid bonding or 3D IC packages with 3D DRAM stacked on top of the server CPU to tackle some of those applications. When do you think that might be a driver of the hybrid bonding adoption at your leading customer?
Well, that is also 2024, 2025. Remember last year we shared several times the progress on this stacking of memories. A US customer also high, not only interest, but engagement in Korea right now, which should lead to further, yeah, let's say adoption and then expansion into mainstream applications. There's a new one currently by a logic customer, stacking logic on logic, which is also a major development. These are all applications in hybrid bonding, using hybrid bonding technology. It will be, as I said several times in this call, 2023 will be a very important year to see where this hybrid technology can really become the mainstream, and then for very specific end products. One of them being, of course, high-end smartphones, that will change the landscape significantly with many more capacity required.
So many applications on the way.
We'll take the next question from Mark from ING. Your line is open. Please go ahead.
Yes. Thank you. My first question is actually also a follow-up on the hybrid bonding in mobile. Do I read it correct that if you get the performance up as planned that then that bottleneck is away? Or do you also need some step-ups from the fabless companies to really make that shift to hybrid bonding in mobile?
Well, that's a very good question again. They are preparing for that. Don't forget that, for any fabless company, they are fully dependent upon the capabilities of this fab. That's very transparent. Does that answer your question?
Yes. Yes, it does. Okay, thanks. The second question is actually on the ASP of hybrid bonding. I assume that in this current ramp it remains relatively stable. I also believe that when you really improve the performance, then the ASP should start to move up. Where are we on that on that process?
First of all, you can simply understand from our gross margin development in total that we are definitely in a good situation so far. Increase in ASP is always linked to increase in features. In these machines, there's a constant development with automatic tool exchange, with certain host computer communication software. Loaders are very critical in when customers want to build chiplets, you need to have different chips loaded into that machine. There's a high degree of development in preparing those different devices for connection in this chiplet architecture. Those features in the end will increase the ASP of these machines. You have underlying, and we also had a comment on that in this prepared notes.
Because of inflation and also other cost increases, we increase, of course, our prices of our machines because of that. You have an ASP inflation you can call it. On top of that, you have increases because of additional features, both hardware and software. We have guided in the ASP range in the Capital Markets Day between EUR one and a half million and EUR two and a half million for this first generation, the 150 nanometers. We have guided somewhat higher, EUR 3 million range with the 100 nanometer capability because that machine is more complicated, especially vibration is an issue with increased accuracy. That will also increase the ASPs.
So far that is, that is the development, but in a positive, in a positive outcome so far. I can also say differently, these ASPs are of course impacted today by the enormous support we have to provide in bringing these machines up to full production capability. Once these machines are running to expected and also committed specs, immediately you have less support, and that leads to higher margins.
Okay, thanks. Maybe a follow-up on the gross margin on the group level, very strong and already I think at the higher end of the previous long-term, medium-term ambition. I can imagine that today the mix is very favorable, but is it true that you made another step change there on the gross margin level?
What is of course very important to simply note, it's been a very difficult year, 2022. From a supplier, supply chain issues, from increased inflation over time, labor costs, simply increased labor costs. Offset by improved system performance, customers who are also faced with inflationary pressures, so difficult negotiations and improving pricing. Overall, it is definitely a higher gross margin than 2021. On an equal-for-equal basis, there was some impact on the gross margin in 2021 because of this flooding in Malaysia. Still, if you would take that out, margins have slightly improved, 2022 over 2021. As you said, the mix, it's always with new technologies, we have higher margins.
Simple question, why would you develop an application for a new technology if it doesn't bring you a higher margin? From that basic principle, that's where today we benefit.
Yes. Very clear. Thank you.
Thank you, Marc.
We have the next question from Nigel, from Morgan Stanley. Your line is open. Please go ahead.
Hi, good afternoon. Another question, hybrid bonding, obviously. Thank you for the color on order intake in both the fourth quarter and current quarter. I guess it showed a little bit of lumpiness as is to be expected. Also given your comments so far about the current generation versus the next generation and current yields, from the feedback from your customers, do you expect any additional platforms to enter production this year based on what I assume to be shipping the first generation tool still, or are these for development purposes broadly? That's my first question. Thanks.
No, no. The key is this year that they are definitely used in a broader end application. That's why the adoption is gaining traction. If it would only be development, you would not ship 35 machines.
No, that's clear. I mean, that's for the 41 customer that's already using the tools on.
Yeah, yeah.
Multiple platforms. You've mentioned the press with multiple IDMs.
Okay.
Sorry about that.
Let me start with that first customer. Supposedly, there are six versions so far. Then there are also the other customer in the US who is setting up now production capability. They have finally decided on three major families which will be, 100% using hybrid bonding interconnect technology. That is, let's say, explaining why this ramp could materialize in the second half of this year and certainly the first half of next year. Then there are also other customers, sharing roadmaps using hybrid bonding, and on top of that, the major subcontractors are now preparing to install a hybrid bonding capability. So they must have customers driving that capability, otherwise, why buy expensive tools?
There's a broad adoption, apart from what we mentioned, also in answer to earlier question, that we're now seeing the memory world, starting seriously to set up capabilities to connect, first of all, memories, so memory packs, which are then prepared to stack onto load. These are all, let's say, mainstream applications, a portfolio of many devices.
Excellent. That's clear. None of them is sort of waiting for the 2nd gen. I mean, you've mentioned those could be used for 3 and 5 nanometers front-end. Does seem that especially SRAM shrink doesn't really work then, so you'd expect major more application of hybrid bonding once we reach that node in the mainstream. Is that how we should think about this, or could you see customers use both 1st, 2nd gen, 3rd generation depending on their application?
No. The first, what you explained. That is how we read the rollout or the adoption of this hybrid bonding technology.
That's then true for. I mean, there's many questions on smartphones, but I guess also GPU, CPUs.
Yeah.
It's sort of tied into that 3/5 nanometer. Okay, that's great. Very clear and good to hear that this one major logic customer's finally also made a decision. Second question, I do wanna again ask about the order book for the current quarter. From the press release, I get that at least most of the hybrid bonding tools are due shipments in the second quarter. Obviously, historically, the smartphone camp has also asked for shipment in the second quarter. Is that, first, the right way to think about that? Second, maybe putting it a bit more bluntly, do you expect a positive book-to-bill for the first quarter?
The bill, that depends both on bookings and billings. Well, we don't guide for bookings, huh. As explained, some elements are hard to read how much more will be prepared for a next round, to say it in those terms. As always, exciting. In a market environment which as explained also by China, which is still a soft market. Automotive is continuing investments. Yeah. Well, Q2 and Q3 will certainly, as it looks today, be with also the orders received, higher revenue quarters. It's too uncertain to give guidance for more than the current quarter.
Understood. Maybe my last question, again, a follow-up on the order intake. Would it be fair to say, given supply chain conditions improving, that the backlog should normalize by the second quarter? Is that at least something we could assume?
Yeah. Well, but also don't forget in that backlog, you have hybrid bonder orders. They have manufacturing lead time still of about 15 to 20 weeks, depends a little bit. They're more than 1 quarter. That changes the profile, the historic profile, where we had typically quarterly run rates. Your point about the backlog, overall, yeah, it should be more normalized.
Understood. Thank you.
Thanks, Nigel.
Next question from Ries from Apus Capital. Your line is open. Please go ahead.
Yes. Hello. This is Johannes Ries. Hi, Richard. Maybe.
Hi
Three follow-on questions. First, on the topic again of hybrid bonding, it's clear it's a hot one. Maybe if I remember right, you talked also already about a third generation maybe which coming in the future of 50-nanometer bonding. Only to make it clear, this smartphone may be major adoption in 2024, 2025. Is it based on the generation 2 or generation 3, most likely?
High-end smartphone generation one and one plus, generation two-.
Mm-hmm
Not Generation 3. Generation 3 is so the 50 nano is in full development. Prototypes you can see in Radfeld, where we may have again a capital markets update in June. That will not be shipped to a customer before, let's say, early part of 2025.
Mm-hmm. Also will then use maybe also in the high-end smartphone market or other, or other, even other applications.
First the high-end computing application.
High-end-
That should be linked to 2 and 3 nanometer architecture. That's very advanced.
Okay. Thanks a lot. Maybe a market area where maybe I'm not so active, only to understand the development. You talked, said even hybrid bonding will be used maybe to stack memories or together. Is that maybe an application which also will use wafer-to-wafer technology, or is it just-?
Yes.
Okay. Mm-hmm.
Wafer-to-wafer, the limitation as we know.
Mm-hmm
Is 2 dies, not more than 2, and they have to be identical.
Yeah.
With the chip-to-wafer, you can stack more than two. In the past with TCB, we stacked eight. We still our machines do. That is a similar, let's say, technology trend for the hybrid applications.
I see. For you are also in the game, let's say, for the use cases in the memory space. Okay.
Yes, certainly.
Great. Maybe on your integrated solution, when it will be caught in the together with AMAT, go in the mass adoption, and what does it mean for you ASP, have you the same ASP in the integration solution? Is it different, and why?
Well, the integrated solution requires a mechanical handshake. You have to simply understand that that will be integrated into. We have the software part of that. The scheduling is the critical, let's say, piece in an automated line. We have started that development already over 1 and a half years ago. First system is installed at a key customer in Taiwan in Q4. 2 more in the US this year. As we also said, they will not be in full production until 2024. 2023 is the year of testing, qualifying devices, and aimed at the very high end.
How maybe you your account, maybe is this how you share the revenues with AMAT? Remind us now, get everybody on the partner.
It is very transparent. We deliver bonders to customers. AMAT delivers integration and also other tools for that automated line. That is separate. We have no deliveries to one another. We both deliver to those customers.
Okay. Since integrated to one system, I got it now. Maybe finally on another technology, only you said you have other technology which are growing. Maybe what is maybe the development of the TCB's technology in the next two, three, four years?
Well, we are about to ship in April, end of March, the first next generation TCB chip-to-wafer system to a major U.S. customer. That is definitely a key development for the next many years to come. What is special about this next generation? It's again, more accurate. It can achieve accuracies at around one micron. Other systems, competitive systems are not yet reaching that and it's also difficult to reach. It also has other capabilities which are unique and that will definitely define a next generation for all kinds of logic applications. Next to the hybrid bonding, we will then have a complete portfolio which can be used for let's say practically every high-end logic and memory chiplet architecture application.
For hybrid bonding and TCB will develop next, both as separate technologies which for different use cases and even TCB is an interesting growth driver going forward.
Yes, definitely.
Great. Thanks a lot and yes, sounds very interesting. Thanks a lot.
Well, all the best. Any further.
It appears that there is no further question at this time. Mr. Speaker, I'd like to turn the conference back to you for any additional or closing remark.
Well, let us thank everyone for participating in this call. If you have any further questions, don't hesitate to contact us. Thank you. Bye-bye.
That concludes today's event. Thank you for your participation. You may now disconnect.