Alchip Technologies Ltd
TWSE:3661
Earnings Call Analysis
Q4-2023 Analysis
Alchip Technologies Ltd
Greetings from CEO Johnny Shen and CFO Daniel Wang marked the onset of Alchip's '23 fourth quarter investors meeting, where they celebrated a record-breaking quarter with momentous highs in revenue, operating income, net income, and EPS. The past year was exemplary, characterized by significant growth in revenue, reaching $978 million, which more than doubled year-over-year, alongside a net income leap to $106 million. Share value soared from TWD 988 to TWD 3,275, marking the company as having the top share price among all Taiwan's listed companies. AI market focus has been a boon, securing about 85% of total revenue, and showcasing dominance in ASIC revenue and market share, allowing Alchip to emerge as the premier pure-play fabrics ASIC company globally.
Projections for the North American market remain bullish, mainly fostered by Alchip's largest customer, who is expected to see a more than 30% year-on-year AI chip shipment growth. Notably, another large North American IDM customer has shifted their AI shipment focus to Alchip, signaling a trust in Alchip’s capability to deliver. Revenue from AI and HPC applications rose to 84% of total sales last year, highlighting the specialization in these advanced and lucrative domains. Meanwhile, revenue from China is scaling back to under 15%, owing to efforts in diversifying the business geographically.
Despite the financial victories, there's been a discernible dip in blended gross margin, dropping from 32.3% in 2022 to 22.3% in 2023. This mainly resulted from the increase in production revenue’s contribution to total revenue, which was between 70% and 80%. The average gross margin for the products from the largest customer is relatively lower, thus affecting overall figures negatively. However, moving forward, Alchip expects its gross margin to receive aid from a more diversified revenue stream and better pricing negotiations with suppliers, albeit cautious in projecting significant uplifts.
Despite macroeconomic pressures and previous performances not meeting expectations, the company has reassured investors of recovering NRE demand from China and promising project inflow. This indicates a growing trend, especially with ongoing process node migration contributing to NRE growth, signaling robust NRE revenue growth expected this year. Additionally, the AI and HPC demand remain resilient, insulating Alchip from consumer market fluctuations.
Alchip’s strategic focus for headcount growth emphasizes on the engineering sector outside China, aiming to equally distribute headcount between China and non-China engineers by the end of the year. This plan reveals an aggressive hiring strategy projected in Japan, Taiwan, Southeast Asia, and a significant increase to about 80 people in Malaysia, indicating a push towards building a diverse, global talent pool. Furthermore, market focus areas for growth include automotive and CPU sectors, notably capturing a much-desired Chinese OEM automotive client, hinting at vast potential for mass production revenue in the near future.
Alchip articulated a confident stance amidst competition, underscoring that competition drives Alchip's progress and innovation. It also remarked on the recent news about competition – especially from NVIDIA’s consideration of ASICs – suggesting that navigating this space demands careful strategizing, given the potential conflicts between standard product markets and ASIC developments. Regardless of such challenges, they affirm a strong and long-term product plan that will keep them in good stead, especially for enduring customer relationships and continuing shipments of successful products like their 7-nanometer offering.
In terms of housekeeping, Alchip plans a conservative operating expense budget, aiming for an estimate between 105 to 110 million for the current year.
Alchip continues to advance in technology, with 89% of their revenue flowing from 7-nanometer or more advanced nodes like 5-nanometer and 3-nanometer products. This indicates Alchip’s commitment to staying at the forefront of technology and catering to the ever-evolving requirements of high-performance computing and AI-intensive projects.
You don't have any saved screeners yet
Okay. Dear investors, portfolio managers and analysts, welcome to our fourth quarter '23 institutional investors meeting. This meeting will be hosted by me, Daniel Wang, CFO of Alchip; and our CEO, Johnny Shen.
So follow the next page. Okay. So routine safe harbor disclaimer. And next page. This meeting will be in English. If you need Chinese presentation slides, please go to the MOPS to download the Chinese version. And at this moment, I believe you should be able to download the version already.
[Operator Instructions] We can answer your questions during the QA session. And again, Johnny and I are all Mandarin speakers. So if you don't speak English, you can ask questions directly in Mandarin. And this video and audio content will upload to MOPS about 2 to 3 hours after this meeting. Okay. Next page.
So again, in case that some of you may not know Alchip too well, Johnny will start with a very brief introduction and then its messaging to investors.
Okay. Good afternoon, ladies and gentlemen. I'm Johnny Shen, President and CEO of Alchip Technologies. Thank you for joining the Investor conference meeting. We appreciate the opportunity to share our Q4 and 2023 result and provide guidance for our future business outlook.
So let me provide a brief company update as usual. Our company is founded in 2003. IPO in 2014. Our current market cap is a little bit over USD 10 billion. Current employees, around 600 people. Since we founded the company, we've been successful, tapeout of 500 designs in leading-edge technology. Within 500, more than 55 are FinFET technology, and we have more than 15 CoWoS tapeout already.
Last year, our revenue is $978 million that compared to the year before, we have a significant amount of growth. And 85% of revenue directly coming from HPC AI area. We are the TSMC 3DFabric Alliance member and also joined the VCA since 2009. Our current capacity is 20 to 30 tapeout -- new tapeout annually. Market focus is HPC, AI and also automotive.
Let me provide some update. For Q4, we are pleased to announce another record-breaking quarter, historical high for all number, including revenue, operating income, net income and EPS. The detailed quarterly breakdown comparison will be presented by CFO Daniel in the later section.
As for the 2023, allow me to provide us a summary/recap. Last -- the past year was a -- is truly exceptional for Alchip. Let me highlight some major achievements. As for the financial performance, our revenue reached $978 million, more than doubling compared to previous year, with net income, $106 million; and EPS, TWD 45.5. This figure represented a historical high and significant growth year-over-year.
In terms of share performance, our share saw extraordinary growth from TWD 988 at the beginning of the year to TWD 3,275 by the year-end. It's a TWD 2,500 appreciation, sharper than TSMC -- sharper than Taiwan Stock Exchange record for biggest stock dollar amount when we think of fiscal year. It also made the Alchip share price the highest among all listed company in Taiwan. As for market leadership in HPC and AI, most of our AI -- most of the AI-related market segment performed very well last year. In terms of HPC and AI concentration and purity, Alchip is the highest among all players. Approximately 85% of our revenue is contributed directly from AI market.
ASIC dominance enhanced the ASIC revenue and market share. Alchip has surpassed all other pure-play fabrics ASIC company, become the world's #1 in this category. Moreover, when compared to several standard product companies for whom doing the ASIC business as a secondary endeavor, Alchip still rank as the leading contender.
Successful GDR fund raise. Earlier this year, we completed GDR offering, raising $400 million out of overwhelming demand, 3 billion orders. These achievements set new records in terms of discount range, overbooking demand and percentage of loan-only fund involved. The foundries position us strongly to support and expand our current business conditions. As for the future business outlook, demand for mass production in HPC and AI application remain robust.
The PO from a couple of our largest customer continuously increase. And this year's forecast significant surpassed last year's figure. Additionally, several other customers have drastically increased their demand as well. We anticipate substantial revenue -- production revenue growth for the whole year, unfortunate capacity-related challenges much less than last year and the year before. As for the new potential business, we have worked very closely to major CSP companies. And we have several design wins from many AI-related companies who are using leading-edge technology like N5 or even N3. Overall, we are expecting much strong NRE revenue growth for this year. As for the geopolitical risk management, we have successfully diversified our business concentration beyond China to other regions. In 2023, less than 15% of the revenue originated from China. But our future business direction to China remain unchanged. We are continuously delivering and support business in China as long as they are financially healthy and complies with rule and regulation.
In terms our workforce, we have embarked a very aggressive hiring plan to strengthen our engineering support in Japan, Taiwan and Southeast Asia. Currently, we have around 80 people working in Japan. Our Malaysia office is already staffed, more than 20 employees. The workforce in Southeast Asia is planned to be around 60 people by the end of this year. This strategic expansion is designed to offer more flexible and cost-effective solution to meet our customers' dynamic requirements in the global market.
As for the conclusion, yes, we are confident that our business in an excellent stage. We anticipate very strong growth for this year. 2024 is going to be another outstanding and record-breaking year for Alchip. Thank you.
Okay. And then that's the easy part, the numbers. For the fourth quarter last year, as Johnny mentioned, we had record-breaking revenue and profit. So for the revenue, in fourth quarter last year, the revenue is $291.1 million, which is 21% quarter-on-quarter growth and 96.3% year-on-year growth. In terms of the -- okay. Please don't use the Zoom functions to disturb the meeting. Thank you very much. And for our operating income, last quarter is $39.4 million, which is 23.3% quarter-on-quarter and 96% year-on-year. And for the net income, it's 35.4% (sic) [ $35.4 million ] for the fourth quarter last year, which is 26.2% quarter-on-quarter growth and 115.7% year-on-year growth, and translating into EPS of TWD 15.2.
Okay. For the yearly income statement comparison, you can see for the revenue part, we enjoy more than doubling growth last year. And for the net income part, we enjoyed 73.3% year-on-year growth. Next page. For the breakdown by application, you can see for the fourth quarter, our HPC-related -- the HPC here include the AI application, already surpassed 89% of our total sales. And the other networking, niche and the consumer makes about 11%. And for the yearly, last year, the HPC-related application is accounted for about 84% of our total sales.
Next. And for the process node, you may see that 89% of our total revenue belongs to 7-nanometer or even more advanced process nodes like 5-nanometer and the 3-nanometer. And we did have 3-nanometer revenue starting from last year, and so there might be something from here. For the yearly breakdown, I mentioned all of it. For the yearly breakdown -- sorry, next page, first. Our outlook for the correct yearly numbers.
And for the revenue breakdown by region, as Johnny mentioned, North America market already become our biggest market among our overall revenue exposure. 67% exposure to North American markets in the fourth quarter last year. And for the Asia Pacific, it accounted for 20% of our total revenue in the fourth quarter last year. And for 2023 as a whole, 63% of our revenue contributed from North American region, and the 22% of that coming from the China -- coming from the Asia Pacific region. And for China, in 2023 last year, the China exposure already down to below 15%.
For the business review, again, record-breaking revenue. And the reason for this strong revenue growth was because the strong AI chip shipment to the company's biggest North American customer. I believe most of you already know the name of this customer. And another reason is starting from the late 2023, our second largest North American customer who is IDM gradually amplified their AI shipment focus to us. And we expect the shipment and the revenue contribution from these customers will keep on growing significantly in 2024. In the following line, last year, the NRE revenue -- the design revenue accounted for around 20% to 30% of the total revenue, while the rest was the production sales. And for the gross margin, I know many people pay a lot of attention to our profitability.
Our gross margin -- blended gross margin last year did went down from 32.3% in 2022 to 22.3% in 2023. The gross margin went down due mainly to the surging weight of the production revenue. For last year for 2022, our production revenue accounted for 70% to 80% of the total revenue. For last year, the production revenue accounted for 70% to 80% of the total revenue. And another reason for the gross margin is for our biggest customer because the shipment volume is huge, so the average gross margin for this product is relatively lower to the other products and the -- lower to our average production gross margin. I think those reasons combined to -- combined cost, the gross -- blended gross margin to 22.3% in 2023.
Next page. For the business outlook, we believe the AI shipment to the North American customer and the overall demand, the overall AI chip design demand or production demand from North American customers keeps on growing. The biggest customer of us gave us year-on-year shipment forecast for its AI chip, a strong year-on-year shipment growth. And for the AI chip shipment -- okay. All in all, like I mentioned, the biggest one in the two customers all delivered very promising outlook to us. I can remember in our last quarterly investor meeting, we guided to our investors that for 2024, the product -- the AI chip shipped to our biggest customer in 2024 will be flat or a little bit grow up on quarter-on-quarter -- on year-on-year basis. But for now, we believe the -- I would say we revised up our guidance from flat to slightly grow to more than 30% year-on-year growth for this product.
And for the second largest customer, I think the demand is not the issue. The issue right now is the CoWos allocation by our third partners because this customer is using the CoWoS-S, which is a very tight capacity status right now. So to see if there is any upside for our unit revenue target, it depends really -- it depends heavily on the CoWoS capacity allocation for our second largest customer. And therefore, the NRE demand of North American market and China, and I would say the NRE demand for AI-related budget remain very strong for the North America region, for both start-ups and the mega customers.
And for the NRE demand of China, I think the NRE revenue suffers -- suffered a little bit last year because of the macro environment, the weak macro environment in China. But we feel the overall demand is recovering starting from the late 2023. And we believe for this year, for the NRE revenue, we may enjoy quite significant growth because of these two markets.
And another reason for our NRE growth is the process node migration will continue. The process node migration from 7-nanometer to 5-nanometer and even from 5-nanometer to 3-nanometer chips are moving. It benefits our NRE growth. And you -- we will see very promising growth for our 3-nanometer-related revenue growth for the year-on-year basis in 2024, which means we will have a significant NRE milestone to happen this year for the 3-nanometer projects. And I believe that's the outlook for the 2024. And now we are going to have our Q&A session.
[Operator Instructions] The first one, I will invite Charlie.
Thanks, Daniel and Johnny. Congratulations for a very strong results. So my first question is about the industry competition, right? So no doubt, the addressable market for AI is growing. But recently, you probably noticed that there's a Arm Total Design alliance and also, some Taiwanese IC design companies joined the alliance. So my question is, first, why Alchip does not participate this opportunity? And secondly, would the company consider to build from and design a service team in the future?
Okay. Let me take this question, and Daniel can add in later. Yes, we all know -- have a kind of a alliance try to put together a subsystem before. We've been thinking about because the unrelated IP, we consider as a software IP. In usual case, our company involve architecture design not too much. So we are not thinking about to join this alliance. But later on, after when we talk to Arm, Arm also approach us very aggressively trying to plan out some of the partnership.
We consider that it might be a benefit for our own customer before. We have a very strong alliance to all physical IP provider. But for software IP, in order to save some time, and we may -- we consider to join this alliance directly or indirectly. As you know, we have a few portfolio companies doing the -- also the partner company doing the front end. And we may consider using them to join this subsystem program.
And if we consider there's a certain benefit for both Arm and Alchip, we can also consider joining direction. But right now, to be straight, we don't see any end customer and also potential customer complain about we are not joining this associate [ with Arm ]. But again, we have -- because Arm, I think, is a long-term partner, we don't against it. We can consider to join. That will be sooner than you expect.
Okay. Thanks, Johnny. So I think attached to that question. For example, some IC design peers like Novatek are getting some AI start-up company in the U.S. So I'm wondering, besides the bigger CSP you want to engage, do you consider to also allocate some resource for those AI start-up company? And if that's the case, how are you going to differentiate yourself to those potential competitors like Novatek and Realtek?
Yes, I think for CSP and other start-up company, we have a different approach. Yes, for CSP, to be honest, I don't worry about the competition at this moment because the selecting criteria for CSP is very strict. You need to have a track record. You need to have a reference. You need to have a production experience. You need to have a CoWoS-related design experience. And also, you need to have enough funding. Those kind of challenges will go over one on one -- one by one. For a smaller company or the new company, just to consider to do the ASIC for CSP, I think the entry barrier will be much higher than they expect.
But for the start-up, yes, if the start-up company has some fund raising or funding challenge, they may consider to use a unproven solution from a newcomer. Now I think that's possible. And we also have a lot of start-up -- we also win a lot of start-up company. For a start-up company, to be honestly, the chance for them to go to production at one shop will be very slim. The reason we are trying to -- we have select some start-up company as our future customer, there is a few selecting criteria.
For example, if they're using a very unique technology, either it's the D2D side or on the interface side, we consider they have a [ list of trend ] we have for a brighter future. We've been thinking about, from an R&D standpoint of view, we want to take. Another example, if the start-up invest, by example, for example, CSP customers directly, we consider the production change will be higher. I think we have a few significant win for start-up. I think later on, once both company agree, we can share accordingly. I think that some start-up company, even they are new, are very, very famous in the industry.
And Haas, please. Haas of UBS.
Johnny and Daniel, congratulations on the solid results in 2023 and business outlook for '24. I guess, my first question would be regarding your sales pattern for this year. In the past few years, as you have been ramping up the chipset production for your customers through the year, and therefore, the sales has been more second half weighted, based on your current order visibility and...
Haas, you can slow down. We cannot get your speed.
Okay. Yes. What I mean is that basically -- okay, sure. In the past few years that your sales has been more second half weighted because you have been ramping up the chipset production throughout the year, so maybe you could provide your view on the quarterly sales pattern for this year. Would it be the same as the pattern in the past few years, to be second half weighted? Or do you think this year could be a little bit different? Yes. That's my first question.
Okay. To be honestly, the first quarter, we were -- we previously were expecting a very strong first quarter. However, because of the manufacturing issues and the Chinese New Year, the first quarter actually will be still strong. We believe for the first quarter, we can still close the quarter-on-quarter growth. And for the whole year's seasonal pattern, right now, we expect quarter-on-quarter growth for each of the quarter sequentially for the rest of this year.
Okay. So you mean, throughout the year, you expect Q-on-Q growth for the fourth quarter this year, okay. So what are the positive and also negative factors for your outlook in second -- especially in the second half of this year? And I think a separate question is that do you think NRE mix will be higher compared with the previous few years for this year?
For sure, for sure. As I mentioned, the NRE revenue because of the reasons I just mentioned, we do believe our NRE growth could be significant this year. To be honestly, for last year, our NRE revenue is not like what we expected previously because the China market status is -- was really, really bad due to the funding, due to the macroeconomic environment, due to the geopolitical restrictions. But for this year, as I mentioned, the demand, the NRE demand in China since recovery, and we do see a pretty good project inflow this year.
And then for, first of all, production, the demand is absolute no issue. The challenging side is still from the capacity. Unfortunately, one product line's capacity already -- we already allocated to a certain degree. That ensure our production revenue growth. And our upside will be coming from additional capacity for our second-largest customer. Demand is also very strong. But capacity is more challenging for the second-largest customer because using different solution. I think it's very much better than the previous years. The NRE, I think, is very much better than the previous years.
Okay. Got it. As a follow-up to my first question, I think I want to dig a little bit more detail about your China projects. Could you share what kind of applications and also which nodes these new China projects are going to use for 2024 and beyond?
Okay. I go first and then you. Yes. For -- in China, we try to -- we try not to touch like AI or GPU-related project if they are using the most leading-edge technology like a [ 503 ]. But right now, our market focus will be automotive, which is going pretty well, and also CPU. And CPU, even they are using the most leading-edge technology like the N5 or the N3, but the total compute power and also the fast transaction speed will be within the requirement. So these are 2 area we are focused in -- we are focused. As for the more consumer side, I think there's a few opportunity on AR/VR because the Vision Pro. Our remaining opportunity, I think, is here in China. So we are taking some of that.
Daniel, do you have any additional comments?
No. As Johnny mentioned, the majority will still come in from the HPC sector.
Okay. Got it. That is very helpful. And my second question would be about your growing opportunities on the advanced nodes across high-performance computing and automotive applications, as you just mentioned. And we have saw your sales been growing at around a 66% CAGR in the past 3 years. And I was wondering if you could share your sales CAGR outlook or view in the next few years following the strong ramp in the past few years.
Okay. I knew that the market has a very high expectation of us. But to be honestly, I will say, first of all, the revenue base right now is already big, as Johnny mentioned. If you take a look at the pure ASIC player, Alchip is already the #1 among this marketplace. And even comparing with those mega companies like Broadcom, [ Avago ] and MediaTek, by doing the apple-to-apple comparison for the ASIC business, actually, we were just behind Broadcom. So for next few years, first of all, for the 2025, we believe the growth momentum will not be as strong as we had in 2023.
However, we don't see any demand decrease. The reason for the momentum -- for the relatively weaker momentum in 2025 is because we may encounter some generation migration fit. But for sure, for 2026, it will definitely be a very, very good year of Alchip. So I think for the revenue pattern, it will be very, very good this year, okay, slowing down a little bit next year, and we'll begin the momentum, really good momentum, in 2026.
Let me share a general guideline for the company. In usual case, we are targeting like a 30% -- minimum 30% Y-o-Y in each of our category. For the good year, we can outperform like last year, like this year. But sometimes, the -- our fundamental requirement, as in target, will be minimum of 30% Y-o-Y.
Okay. Yes. Got it. If I can squeeze in one more question before getting back to the queue. Would you be able to discuss the hyperscaler opportunity you are seeing now in addition to your existing projects? And would those new potential hyperscaler opportunity to be this diverse in 2026 to sustain your lower-term strong growth?
Okay. It is -- Haas, you are so good. You never give up. Okay. For the other mega accounts in North American market, to be honestly, the confidentiality requirement right now is through the roof, off the charts. We cannot talk too much about the progress for the project. We are in talk with other mega accounts. However, I will say this way that given our good position within the AI ASIC marketplace, I believe no matter which of the mega accounts or the hyperscalers in North America, if they want to do an ASIC, they will give us RFQ. So if you are asking if we are talking with those hyperscalers or mega accounts, yes, we are in talk with them. But win or lose, that's another story. That's what I can tell you the most about those mega accounts.
Yes. Yes, let me add a little. I think Daniel covered it very well. First of all, we've been, [ one ], not mention about anything so many times on the customer or direct potential customer directly. But right now, I can share each of our CSP are very aggressively trying to minimize their dependency from current single supplier. So the only solution is to go and consider ASIC. There will be very limited ASIC player, as a experience. So if our market trend like that, our CSP continues to growth, we will have -- we definitely will be one of the winner. Yes, as you know, we have a very aggressive year-over-year growth target. To win a start-up, those kind of companies maybe help the R&D story a little bit. But fundamentally, the revenue need to grow every year. Our target is very specific. There is no hidden customer. We have to win in order to grow. And so far, I think the chance is coming to us.
Prior to go to another question. I will say, from the message board, I received a question maybe most of you are interesting about. Let me ask that first. For the NRE, what is the expected growth outlook for 2024 and which quarter? What is the mix percentage of tapeouts from CSPs versus start-ups and the gross margin impact in very strong growth in production as well?
I would say for the 2024 outlook, I already mentioned. And forgive me that we are not going to going to give you such detailed information about the status of our business with the cloud service providers in North America. But I believe for this year, the revenue breakdown, the revenue mix between NRE and the production will be pretty similar with what we did last year. So for the gross margin, I would say we'll be still pretty similar. The good thing is, first of all, for this year, there will be another big contributor to our production revenue. Unlike last year, almost the majority of our production revenue contributed by one customer. And for this year, it will be more diversified. So we expect it will be help a little bit to our gross margin.
And secondly, we are working very hard to talk with our suppliers in order to get better pricing and better margin. That's for the margin side. And the second question is, Johnny, please share your thoughts on the competitive landscape, given [ immediate ] new ASIC business unit in Arm's total design platform, in particular, on competition and the pricing.
Okay. Let me try to take that. For competition, I think like I mentioned before, if we're -- even for the CSP customer, I don't worry about the small or new competitors try to win this market. This -- a few player, we consider this will be the major competitor, Broadcom, [ Avago ]. MediaTek is also a very respectable company. They try to get into recent market. They do have some IP solution, and we do have a very good supplier relationship. When thinking about competition from MediaTek may influence -- may produce another choice for CSP. I think that's possible.
But if you look at the design side for CoWoS-related tapeout, for all the track record, we -- I think everybody still consider, even the potential customers still consider Alchip as the one holding the most track record. And also, even different technology, different form factor, we already have a solution for them to select. Yes. For recent year, I know there's a lot of company, the IC company, all kinds of IC company, even they have said that current markets are saturated or they are not doing very well. They have considered to do the ASIC.
For those companies, I think they may have a chance to win mainstream consumer or IoT-related business. I think as long as they have a strong design team, they do tapeout, they do production before, they will have some market. I think for those ASIC business in the target consumer IoT, those types, that I think they may have a chance to win. But from new -- again, for newcomer, if we want to win the CSP, that will be very difficult. So overall, let me emphasize it again. Alchip will never afraid of any competition. Yes, we don't -- we always sharpen ourselves very, very strong. We need a competitor in order to position ourselves. I'm not enjoying dominating the market. So I think for the past 20 years, we're going slowly step by step with older competition. And at the end, we are winning. Thank you.
And then for the total design platform.
On design platform, I also -- yes, I think I already addressed that during Charlie's question. This -- we are working right now aggressively. And in a short period, we are considered joining this program directly or indirectly from our portfolio company.
Okay. And Robert, JPMorgan, please.
Thanks, Daniel. Thanks, Johnny. So first one is on your existing projects. So can you help us understand about your biggest U.S. customer, the growth momentum for [indiscernible] sales there is softer. Is that due to your conservativeness? Or is it just we're not sure whether there'll be a kind of transitional product launch or not? Because I remember same time last year, your guidance for this project was to decline 30% in '24. And then over a period of time, you revised up the project revenue for this year. Will we see a similar trend in the next few years that you could revise up the revenue for '25 as well? And then for this project, it seems to have won the next generation that will ramp up from 2026. Can you help us understand how should we think about the gross margin for this new product generation starting to ramp in '26?
Okay. Robert, your question is too long, but my answer will be very short. We are not allowed to talk about this customer and base project. And I would say this way, for this year's growth, which is coming from the same product we shipped starting in 2021, actually in 2021. This chip is very efficient and there is a very big success for both Alchip and our customer. And for the upcoming projects or business, we can only say we will continue our relationship with our biggest customer going forward. That's the message we can say right now.
Okay. Second, on your automotive project. So I think you have won a big automotive project with a Chinese OEM and it sounds a little bit quite positive on the China auto customer for this. So can you help us understand about what should we expect? Should we expect another big project win sometime later? And for the current project win, when should we think about the mass production revenue to kick off?
Okay. I'll go first, and Johnny may give you some -- may have some color on it. For the automotive customers in China, we believe the customer is actually is the jackpot. This customer is, I would say, the most wanted customer among all the China automakers, even include [ BYD ] so the outlook and the forecast for this project is really promising. The only concern, all I'll say, the only worry is scheduling because, as I mentioned to the investors many times, that for the automakers, they are new to doing chip, so they need a lot of support from front end and the back end. So if the schedule is perfect, we will have the production revenue contribution starting from 2025. However, for the scheduling, if they started late, we may have limited revenue contribution in 2025. The massive contribution will be in 2026. So if you are asking me about the schedule, yes, we do worry. But if you are asking me about the model, the forecast and the extension, they are all very positive to us.
Okay. Sir, my last question is about your second biggest customer for [indiscernible], the IDM customer. How should we think about the future outlook, say, in 2025 and 2026 and beyond because this IDM customer seems to have some in-house AI projects that based on their product roadmap. Yes, so can you help us understand about this project outlook in the longer term?
Okay. For the IDM customer, I would say, this year will definitely be our second largest customer and it contributes a lot compared with the contribution in last year. And for this customer, we support their 7-nanometer and 5-nanometer chips. And we do believe that for the 5-nanometer, AI chips, the shipment will extend into 2025. For now, the information we received is too good to be true. And, yes, another too good to be true. But we believe, given the profile of this customer, the shipment, the message we received, yes, there are certain chance we can. We may have it in 2025. If that happens, 2025 will be another good year to us. And like you asked that this customer, it is an IDM customer. Will they bring back the project, the future project in-house, I would say, possible, but we will try very hard not to let it happen. We will keep on keep on delivering our value and show our value to the customers, yes.
Okay. Let me add some color for this. Due to a reason, a specific reason, I had a chance to meet division leader from this chip maker. I think he has a very high confidence on the result, this 5-nanometer design. They say, in the current market, it's very difficult to find any chip that has a higher performance than this. In terms of interface, even in terms of performance. So the forecast they give us is just -- the capacity we received for this year, just maybe a little bit over 20%. I think, eventually, according to them, the volume will be very significant. That's it. Most of the volume will be -- will happen on 2025. This year, I don't think the upside for process will be too much. I think that's what it is.
And then we go to next question. Szeho, please?
I have a few questions for you. First one regarding our IT collaboration with Synopsys. When will the IPs be available for customers and when should we expect the product starting to come out?
Szeho, are you asking me about the partnership with our IP supplier?
Yes, the Synopsys, the IP co-development.
Okay. I will say, it echoes to the question or the message before that there's a question could you talk about how Alchip's partnership with Synopsys? What is the value position here that differentiate Alchip from the other in winning projects? The solution is very unique, very large sized, very large dye size, 8 HDM modules, 224 giga -- [ 224 giga service ]. I don't know what is -- how come you know the numbers, but I would say our partnership is kind of, right now, it's -- I won't say conceptual, it's already validated, but not connected to any specific project so far. And for the partnership, I would say, Johnny is a better speaker than me.
Yes, I had a meeting with their new CEO and all the IP leader quite frequently. I think, right now, we can say the 2 companies are complementary to each other. Why? I think it's very easy for the center product company in usual pace, they're doing their IP back end. They're doing their IP by themselves. The common interest for Synopsys, the CSP provider and also system provider. But CSP and system provider not only need IP, but also need the ASIC provider. So quite natural, if you put us together, I think very closely try to win together. As you may say -- as you may see on the news, starting from last year, both Synopsys and Alchip joined a lot of event, for example, TSMC, OIP, RCTO and also the Synopsys IP leader, we have a co-presentation at every single event.
So looking forward, I think there's a lot of things will happen. We have -- we never compete each other, it's quite complementary. And according to them, that's the same wording, they consider Alchip would be the best partner for them.
I see. And in terms of investment, can you talk about the event the implications of how much we would contribute to that IP development project?
You are saying, how many projects?
No, no, no, no. I mean for the IP co-development, how much money we need to...
Co-development. I think we are not trying to involve for IP development. We're just working with Synopsys very closely, how the potential customer needs. Yes, in terms of our IP pass, the cooperation, every year, majority of our paper using Synopsys IP. In terms of IP user, we are the -- one of the biggest customer for Synopsys for years. The new IP design, again, we are not yet -- we try not to get involved for any mixed signals IP design, we are just working with the Synopsys and potential customer identify what they need and what their roadmap, and Synopsys we are doing their design toward that direction.
I see. All right. Second question, I think earlier you mentioned that for every year, the company will focus on around 20, 30 NTO project, right, per year. So going forward, should we expect the number of projects to grow over time? Or the company will be more selective in picking the project with the biggest potential, biggest benefits to the company?
We don't intend to grow largely the number of customers or papers or projects we are doing, which means it will -- it requires considerable design capacity expansion. We want to choose the project wisely because you can see only 1 to 2 projects can constitute the majority of our revenue. Given our position right now is really good, we want to win those so-called important projects rather than those projects that only deliver very small amount to our revenue and profit. I think the focus to maintain the number of annual project we can do is to maintain our technology, is to keep our engineers busy in order to take important projects. That's our intention. The goal is to remain the same. We plan to take out 20 to 30 design every year. We all know every single company is quite similar. Customer can dictate more than 80% of total revenue. But for us, I think in order to bring continuously position ourselves as a leading AC provider and also bring the ecosystem and also getting more design reference, we will take out -- we will maintain the capacity for 20 to 30 people. The headcount will be increased accordingly. For leading-edge technologies for your information, 5 nanometer, we may need about 60 to 70 people. For 3-nanometer, usually, we need about 80 people to complete. So I think in order to maintain our exact capacity, our headcount will be grown accordingly.
I see, very good. And last question. Yes. Right now, our foundry partner is TSMC, but would some point time we would consider to explore the engagement with some other foundries, let's say, the ones in the U.S. focusing on leading edge?
So far, we don't have this trend. Yes, the partnership with TSMC is still very strong. Yes, we are part of the guideline for BCA status. If the customer really want to go for the other foundry and that will be the prior notice and last book [indiscernible] for TSMC. I think, in fact, we are happy with their support and also capacity. Okay. Next one.
Just 1 quick follow-up on what Johnny just mentioned about the capacity, the HR, have you been expanding for the R&D going forward given that we have been getting more 3 and 5-nano technology going forward? And what kind of the expansion are we looking at? And would you be able to break down in terms of geographical, like, say, Japan, Taiwan or outside of EMEA or Southeast Asia. So any particular project you could actually share with us? That's my first question.
I think the Malaysia subsidiary is, one, we want to grow our engineering number aggressively this year. For now, we have about 27 people there. And by the end of this year, we are planning to have like 70, 80. Yes, that's the target for our headcount growth in Malaysia. For Japan, Japan, I would say, the majority of the headcount increase is from the engineer relocation from China to our Japan office. And of course, we will also do the organic growth for the engineers in Japan and Taiwan, but the overall pace will be much slower to the pace in Malaysia. And then for this year, we will kick off the B&M plan. We also plan to have a subsidiary in year-end to expand our engineering pool. Yes, in terms of geopolitical breakdown, my goal remain the same. By the end of this year, I hope China and non-China engineer will be 50-50, the non-China one, including the engineer willing to work outside of China. I think, right now, I think, so far so good for that direction.
Got it. And just a quick follow-up in terms of the total RD, what's the number right now? And what are we planning on for this year and next year?
RD, are you saying the designer?
Yes, the engineer for your project.
Doing a design. Total engineer, we have 600 people, 75% of our employee are engineers.
Okay. Got it. And my second question will be on your EV or the auto-related, because we've been saying this could be a potential business for us in the next few years to come. And do you have a forecast in terms of the revenue for this year and the next year or even the year after in terms of the rough number, because we have been hearing mostly about the HPC. Just wondering is there anything outside of that, that we'll be able to see some meaningful growth.
In EV, okay. I would say not limited to EV, I would say, automotive chips, [indiscernible], obviously, we believe will be another revenue contributor in 2025 to 2026. And of course, we will receive quite good amount of the MI revenue this year. And beside of that, I would say the -- regardless of the applications, the leading edge and the picture is our focus. So for any application, the leading edge process now and large scale design, it fits our business model. And Joey, please?
So I have 2 questions here. So first, Johnny, I think for your opening statement, you mentioned about the order size is around $3 billion. Did I hear that correctly?
Yes. Order size?
Yes, like -- I think, like Johnny talked about like $3 billion order or I might misheard?
$3 billion order, that's GDR. We are raised -- we have already raised $400 million out of $3 billion orders.
Understood. Understood. Yes. And the second question is on the gross profit margin. So Daniel, you were saying that the 2024 will be similar to 2023. But like given chips like second to Broadcom now in ASIC, like should we expect like any like gross profit margin benefit like from Alchip's foundry partner in 2024? Or is the 2024 margin upside mainly come from the NRE increase?
Yes, you can expect that. But to be honest, we want to be conservative because we don't want to deliver too optimistic information to the investors. To talk with suppliers and customers, especially for pricing is really difficult. For now, we are more confident that, for the supply side because our position within our foundry partners is getting better and better. So we are pretty confident that we can get some room from our suppliers. But for customers doing business with -- especially for the North American [indiscernible], especially for the North American customers, it is very hard for you to expect the margin, especially from the same project. If we are talking the new project, mainly for existing projects, it may be difficult to extend the current gross margin for the existing projects.
Understood. So the margin bump for new project, that's from the pricing increase you mentioned earlier, I assume.
Yes. Thanks, Joey. And there is a question on the message board. Do you have more than one auto customers? Do you have auto customers outside China? Yes, we do have. And the question is, can you do a project for 18a or 14a? I don't really understand what is the question. So we will jump over it. No, sorry, we don't really understand the maybe 18 and 48.
On Intel. No, we currently don't have the plan to go to Intel for the projects. And Charlie, please?
So first of, you kind of revised up your revenue growth outlook for the major North American customers. So may I get a sense what is the reason behind? Is it because of the more CoWoS capacity, as you said? And also after this kind of CoWoS supply increase, do you feel like these U.S. customers demand totally get fulfilled? I mean, whether there is further upside for this customer in the rest of this year?
Yes. The demand for this specific customer, we mentioned about many times this year is surpassed last year's figure. Honestly, I don't quite remember what size the number percentage, we are -- we disclosed on the previous meetings. But to make it simple, the number -- the forecast we are receiving keep increasing. We never revised up any numbers. So I think, this year, is going to be a very, very good revenue produced by this business specific project. 7-nanometer is [ people ] consider is the most cost-effective solution. So I think yes, the capacity issue is soft. I think revenue expectation from this project will be very high for this year. Please believe me last, the guidance, the number of guidance for the shipment for everything we gave is conservative.
Okay. Okay. Yes. I will believe you. So -- and Daniel, just a housekeeping question. I remember you right, there was some NRE from Japanese customer in last 2Q. Have you fully write back those receivables?
Not yet. Not yet. We did receive some, but not in the 4 months, we buy it out last year, but no more bad debt from this customer [indiscernible].
Okay. So what was the financial implication to your margin or OpEx on these events?
If we can receive the amount, of course, it will go to the profit. So it's -- there will be only up, no downside for these customers expect that.
Okay. Got you. So I saved the -- one of the interesting topic very last. Maybe Johnny or Daniel, if you can share your view on this topic. I believe you see the news about -- from Reuters, I guess, they said MVS they are building a team to do the custom chip for CSP customer. So I wanted to know from your industry source, whether this is reliable and whether that will impact your future market share as well? And also from NVIDIA's recent results, right, one surprise is that they said 40% of their GPU is now for the inferencing. So do you have a concern that NVIDIA's GPU will continue to be very competitive and they will kind of constraint the future growth of ASIC market?
Yes. I also see the news. I think everybody does. Once these news announced, I received a lot of text message. So, overall, I don't think that's -- yes, to be honest, I don't think it's the right thing for NVIDIA to do. There's a big conflict between standard product and also the ASIC. If the target market is the same. Yes, as you know, the -- you say the product owner is NVIDIA, if an ASIC owner is their end customer. So by natural, there is a big conflict if they -- because NVIDIA start to do the ASIC. They will -- they definitely will compete with their standard product customer. I don't think that's the right thing to do. And also, in terms of gross margin, standard products' gross margin are much higher than the ASIC. So I think for a lot of company. Product company considered to do the ASIC because their markets are saturated because they are trying to change the gear. But for NVIDIA, they are the winner. I don't think they are -- they try to sacrifice, it's not the right move to sacrifice winning standard product market to the ASIC.
And also the developing support effort will be different. By definition, basic only serve one customer. Standard product can serve multiple customers like 100,000 is if they want to tailor make, spend a lot of effort, tailor make their product to serving 1 customer, I don't think that's not so work to do. And also, the main reason for CSP developing the ASIC, they try to reduce the dependency from NVIDIA. Yes, if they use NVIDIA, if everyone use NVIDIA's ASIC solution, eventually, I think the market is still dictated by one company, [indiscernible] the whole purpose customer want to do their own stuff.
Yes, thanks, Johnny. Yes. So yes, I saw the ASIC would be very competitive for a super insightful. So just a follow-up, right? I thought the basic will be very competitive for the inferencing market. But it seems like NVIDIA is also -- has a very strong kind of presence in these influencing markets. Any concern we should have that GPU will gain share from ASIC in the inference?
I won't say the influence or the training are the key issue because it all depends on [indiscernible] complete the cost of computing power. Obviously, ASIC is much more efficient than product, especially under the current market status that the NVIDIA product is very, very expensive. And another important thing is the capacity is limited and for those peak, customers providers. They have their own plans and these plans cannot be jeopardized by the shortage of the chip supply.
So to go to ASIC, first of all, to reduce the cost; secondly, to extend the flexibility and -- for their infrastructure building or their visual plan. So I think, yes, many, many reasons for that, not because of the applications, because you can see that [ total ] company is training and, therefore, the others like the database, doing inferencing, and so I don't consider application is a key issue for the future plan of those customer providers.
And think about positive side when I see the NVIDIA news, I consider even almighty NVIDIA start to care about ASIC, because everybody can see the -- how powerful the AC chip is. I believe they will have a much brighter future on this area. Thank you, Joey. Okay. Prior to go to Joey. There is a question from [indiscernible] on the message board management. Could you please share a little bit more color about why you revised up guidance for your largest North American customers growth this year recently. Is this because of new project gains, new project gains or because of longer than expected lifespan of the existing product lines? Thank you.
I think the answer is really straightforward. First of all, is the existing product, we will keep on shipping the 7-nanometer product to this customer. And the shipment made we enjoy pretty good year-on-year growth. And the reason, to be honest, we are also doing a guess that, I do believe the end events drive the NVIDIA drive the shipment up because, as I mentioned to investors before that this should -- success is much, much, much better than else's expectation and also our customers' expectation. So yes, I would say the lifespan and the demand increase simultaneous. And for the future projects, forgive us that we cannot disclose the whole story for these customers anymore. Yes. As I mentioned, we will keep on working with these customers for sure. We are long-term products. And go back to Joey.
I also want to follow up on the ISP project. So I seem like back in '23 Q2, I think you, guys, mentioned that the ISP projects should have very high volume shipment in early '24 Q1? Is that stack? And another thing is, I think you, guys, also mentioned that for a main chip win for North America handset, which will start in mid-'24. Is that still on track?
Okay. I guess there was information received that the ISP chip is not for North American customer. It's for Japan customer.
Yes.
Yes. This customer, I would believe the shipment will begin this year. However, we don't see very slow order inflow because maybe the end application market is still weak, I think. So like I mentioned, for consumer the single chip the single range price is not high. So indeed realize really on the end demand of the market and the consumer market is really difficult to predict. So we intend to be conservative on the revenue contribution from this product. However, for HPC, or especially for the AI chip, I would say it's not that evolved or related to the consumer sector. It's more related to our customers' costumers. And obviously, for those AI chips, they are dealing with so many different things like the show video, the social media content and the demand are still really strong. So we don't worry too much about HPC-related demand, and we still believe this marketplace will be the major application for our future growth, for sure. from that, sorry, I can't use the raise-hand function, but I appreciate it if management can take my question. If the supply constraint alleviates on the partners side, in the future, especially for the advanced packaging, [ where the still ] need to do the fundraising for backing up the working capital? Mike, I would say, it really depends. Depends on how big the jackpot we could win. For example, if we win a very, very big project, from those mega accounts, maybe we need another fundraising to support the project. Actually, anything like that happens is really good news to us. The project scale, we are talking about more than $1 billion revenue contribution per chip, annually. So yes, it really depends. Yes. So far, as I mentioned before, we have been in for funding to support our current project even with the upside. Yes, unless we win another jackpot, I think the [indiscernible] think my question is asking for the constraint on the core side and [indiscernible]. Okay. Yes, let me try to tackle that. Yes, right now, we all know the biggest constraint on the capacities package [ cross-related ] capacity. When we talk to end customer and also the TSMC so much of the TSMC can provide sufficient capacity, we will not go through other alternative solution. But on the flip side, if capacity continuously facing some challenge, customer may consider to use the another type of solution. The cheaper wafers, they are doing TSMC, we will structure that they can continue to use solution, if TSMC cross capacity continuously facing a certain challenge. Okay.
Okay. And the time is late, and I see no questions. Let me do the continuation part. For the breakdown of our -- yearly breakout -- for 2023, I would say, 89% of our total revenue is both to 7-nanometer or [indiscernible] like 5-nanometer and the 3-nanometer, 89%. And 4% to 16 and the 12, and 5% for 28, and only 2% to 14-nanometer and more legacy nodes. That's the breakdown and the updated slides already to the MOPS, so you can download it. That's correct slides. Sorry for that. And okay, please. that's the last few questions, okay -- now for this today's investor meeting.
Just a housekeeping one. So how should we think about the OpEx dollar for this year?
Okay. I would say conservatively, we plan to have about 105 million to 110 million for operating expense this year. And that's it, I think. Thank you very much for joining us today and thanks for all your support to Alchip these days. Thank you. All right. Thank you very much. Thank you. Thank you for your support. Thank you.