Okay, good afternoon, shareholders, analysts, and the fund managers. I'm Daniel, CFO and the spokesman of Alchip. Welcome to Alchip's third quarter 2023 earnings call conference. This meeting will be hosted by our CEO, Johnny Shen, and myself. It's already 2:30 P.M., so we are starting this meeting. This meeting will be in English. If you need Chinese presentation materials, you may go to MOPS [Foreign language]. Besides, during the Q&A session, you are welcome to ask questions in Mandarin. Johnny and I are both native Mandarin speaker, so we can answer you and understand you without a problem.
This meeting will start with company brief and the CEO's message to investors, followed by routine third quarter P&L business review and outlook. And then we will announce our fundraising plan, and then the Q&A session. Prior to Q&A session, please mute your speaker to ensure everybody's right to listen. During the Q&A session, please use the Raise Hand function of Zoom. We then can call your name and for you asking questions. Thank you. Okay, this page is again a routine standard offer disclaimer. Okay, I already told that this meeting will be in English, and our participants, please ask questions through the Raise Hand function, or you can write down your question through the Zoom's message function.
I can read it, and I will answer your question immediately. Sorry. The video of this presentation will upload to MOPS [Foreign language] after about one to two hours of this meeting. So if you want to review or re-listen to the meeting's content, you can download it from MOPS page. Okay, I will invite Johnny Shen to start with the presentation we prepared.
Hi, good afternoon, ladies and gentlemen. Yeah, I'm Johnny Shen, President and CEO of Alchip Technologies. Thanks for joining our investor conference meeting. Yeah, we appreciate the opportunity to share our Q3 result and provide a guidance to our future business outlook. Yeah, for those not too familiar with Alchip, yeah, let me provide a brief company update once again. Our company is founded in 2003. We are IPO in 2014, so current market cap is a little bit over $6.5 billion. Employee headcount is 600 people, so obviously we are very healthy on the revenue to employee ratio. So next, last year, our revenue number is $460 million. Yeah, for this year, we'll grow quite significantly.
In the first three quarters, we already achieved more than $ 687 million. Our market focus is HPC AI, and also recently we put the automotive to become one of our main focus as well. Since we found the company, we've been successfully tape out more than 500 designs. And in for the designs, we are tape out more than 60 FinFET-related technologies. Even for the latest CoWoS technology, we are tape out more than 15 already. Our capacity is around 20-30 new tape outs every year. And for the recent year, majority of our revenue, more than 80%, is contributed from HPC and AI area. We are the TSMC VCA member since 2009, and last year we just joined TSMC 3D Fabric Alliance. Okay, yeah, allow me to continue.
For Q3, let me give you everybody a quick recap. In Q3, we are pleased to report a good result. Our revenue number reached $ 241 million, with operating income $ 32 million, and net income $ 28.1 million, resulting at EPS TWD 12 , TWD 12 . While our Q3 revenue is slightly lower than Q2, but we saw an improvement on our gross margin, resulting a record-breaking income figures and also historical high quarterly EPS. The primary reason for lower revenue was due to the shortage in CoWoS. Fortunately, we are making steady progress in improving the capacity. We expect smoother production of CoWoS-related products starting second half of this quarter. For future business outlook, the demand, the mass production demand for HPC and AI application remains robust.
Our largest customer demand continuously increasing, and their next year's forecast is significantly surpassing this year's figures. Additionally, several other customer has drastically increased their demand as well. We anticipate substantial revenue growth for the next three to four quarters. On the flip side, we are facing a operating cash challenge due to the high production demand. Even though our cash flow is at very highest, high level, it is still not enough to support the future production demands. As a result, we are planning another GDR offering to meet our customers' strong demand and also suppliers payment term. For geopolitical risk management, we have proactively diversify our business concentration beyond China to other regions. In Q3, less than 15% of revenue coming from China region.
Due to the new rule and regulation from U.S. BIS department, some of our Chinese customer has directly or indirectly impacted. Fortunately, the potential revenue and income impact is very minimal compared to the total figures. Our future business direction in China remains unchanged. Yeah, we will maintain a cautious approach by close cooperation with IP and foundry partners. That include detailed review for potential customers' background, their end user, design spec, before accepting projects. We continuously to believe in and support business in China, as long as they are financially healthy and comply to all the rule and regulations. In terms of workforce, we have embarked on a very aggressive hiring plan to strengthen our engineering support resource in Japan, Taiwan, and Southeast Asia. Our newly opened Malaysia office already staffed more than 20 engineers.
This strategy, strategic hiring, is aimed at to provide more flexible and cost-effective solution to meet our customers' evolving requirements. In conclusion, yeah, we have a confidence that our business is in excellent stage. We anticipate a strong finish for 2023, and hold a very positive outlook for 2024. Thank you very much. Mm-hmm.
Okay, then we enter the routine, the financial numbers. This page is for the third quarter P&L. For the third quarter, I guess you already knew the revenue numbers. I'll still go through the P&L. The revenue in the third quarter was $240.6 million, which is a 7% quarter-on-quarter decline, but a 104.7% year-on-year growth. And, for the third quarter, we have slightly, profit margin improvement. So, the operating income for the third quarter is $31.9 million. It's 12.6% quarter-on-quarter growth and a 71.6% year-on-year growth. And, contributed by the strength of the U.S. dollars and our interest income.
The net income for the third quarter is $28.1 million, which is a 17.4% quarter-over-quarter growth and a 92.9% year-over-year growth. The net income for the third quarter translated into EPS is TWD 12 . This is a breakout, breakdown by application. You can see obviously, HPC is still our focus. In the third quarter, HPC accounts for 85% of our total revenue, followed by the niche market of 4%, networking 10%, and 1% for consumer. For year 2023, until today, 83% of our revenue exposed to HPC application. For the regional distribution, again, we have been shifting our focus from China to North America for the past one to two years.
To be honestly and proudly, we have done a pretty good job. For the third quarter, the North American revenue account for 59% of our total revenue, followed by the Asia Pacific, 27%. Japan and the other area makes the balance. For this year, until today as a whole, the North American exposure accounted for 61% of our total revenue, and the Asia Pacific region accounted for 23% of our total revenue. Okay, for the process node distribution, again, Alchip has been focusing on the most leading-edge technology node since day one of this company. For the third quarter alone, more than 92% of our revenue exposed to 7 nm or even advanced process node.
For the yearly distribution, 88% of our revenue exposed to 7-nm or more advanced process node. We can proudly say that by comparing with our industry peers, in terms of the revenue concentration, we should be the number one, the leader within the most leading-edge technology nodes. So for the third quarter business review, revenue, as we guided in our second quarter earnings call. The second quarter revenue came in mildly decline quarter-on-quarter, but it meet our expectation. We already guided to the outsider that we may suffer a little bit from the revenue side in third quarter.
The reason for the slightly decline quarter-on-quarter revenue is because, first of all, the packaging engineering, and then secondly, the high quarterly revenue base in second quarter. For the margin side, for the profit margin, in third quarter, profit margin recovered mildly. I'm not saying a very significant improvement, but we do improve the gross margin and especially the operating margin. For the third quarter, gross margin came at 24%, recovered from 21% in second quarter this year, because of more favorable sales mix. And the operating margin, because we have a very good operating leverage, our operating margin improved quite a bit. Coupled with our OpEx back into normal, you can see, we have quarter-on-quarter decline OpEx third quarter.
It also contributes to our operating margin. As I mentioned, the non-operating income came in at $4 million in the third quarter, due mainly to the strong U.S. dollar, coupled with the interest income we received in the third quarter. That's the review for the third quarter. For the business outlook, as our CEO just mentioned, we are having increasingly confidence toward the 2024 revenue performance. First of all, the AI chip shipment to our North American server customer, not only to sustain, we right now expect we may have year-on-year growth potential in 2024 for this product. We expect the shipment for the training chip to another major North American customer to gradually heat up, starting in late fourth quarter, 2023.
The order from customer. The order and the forecast from customer are getting better and better. I will discuss it into later, it could be another reason for us to have the fundraising activity. In addition to these two customers, we will have multiple high-volume projects to start production in 2024. Some in the first half, and some in the second half. So in terms of the revenue growth, right now we are very confident for the year 2024. We remain our positive view on the AI ASIC growth. We can sense that the major CSPs, especially in the North American region, trying to gain control on their chip and the system technology. So almost every of them are trying to build up their own chip.
Secondly, as you may know, that the current ASSP, the product supply for AI chip is in shortage. We believe Alchip enjoys a very good position within this trend. We believe there are not too many chip vendors who are capable to provide the most leading-edge technological design and the experience of handling large-scale production. For the recently released new BIS rules, we consider it has limited impact to our current operation. First of all, for the Customer B, I believe maybe some of you already know who it is. The revenue exposure to the Customer B is around 1%-2% of total revenue this year. Because this customer was being put on the entity list, the full entity list.
All the business activities for both production and the NRE currently put on hold already. Our current revenue exposure to China market goes down significantly from last year. We expect the revenue contribution from China market this year as a whole, including the first quarter, will go down to 15% or lower. So, we believe the geopolitical risk between China and the U.S., we have been doing pretty good to trying to alleviate the potential risk from the dispute between these two countries. Okay, GDR offering, as Johnny mentioned, we plan to have another GDR offering. This is the second time we do the GDR offering. We plan to raise around $350 million through GDR offering. It is about 5% dilution.
We care a lot about the dilution, so this time we plan only to dilute 5% versus about 12% last year, last time we do the GDR. The reason, actually, we have surging demand from existing North American customers. It consumes a lot of our working capital. The majority of our production are applying the CoWoS packaging. The chip with CoWoS packaging require much longer processing turnaround time than the ordinary project. So it also sink a lot of the cash into the production phase. The current cash position will be tight, based on our calculation, as multiple projects enter production phase last year, or let's say, the late this year, we may see pretty tight cash position on our book.
Even, yeah, we are a cash-rich company. To be honest, I never think about that, with a cash-rich company like us, we'll be short of net working capital. But because the demand is so strong, we are facing this tight cash position. Another big reason for the GDR offering is we want to strengthen our financial structure in preparation for taking future huge volume AI projects. Because the overall demand from DC industry, from those cloud service providers, is getting stronger and stronger, based on our information, any given project from those cloud service providers in North America region, if the project is for the data center main chip, the single chip production volume will be huge annually.
For our current cash position, we may not be able to take the projects, so we want to strengthen our book in order to convince potential customers to grant us more project. I believe, to do the GDR offering, although there will be 5% dilution, we consider it is a very positive signal to our operation. And, hopefully, we will place this GDR offering successfully, when we do the placement, next, early next year. Thank you. And, we are going into the Q&A session.
If you have questions, please raise your hand. First one is, Szeho. Please unmute your speaker. Szeho, please.
Oh, hi. Hi. Hi, hello, Johnny, Daniel. Thanks for the time. Yeah. Regarding the GDR offering, who would be the target investor you're looking for? Are you talking about the financial investor, or supply chain, or your customer as your potential investor for the GDR?
Okay. It is just a normal GDR. We of course want to attract some long-term investors to purchase the shares. But we don't have specified targets for giving them the shares.
I see. All right. Okay, no problem. And my second question regarding the bad debt provision you made in Q2, and last time you mentioned that there will be a provision made in the second half this year. Just wonder, is this still the case?
Excuse me. Can you, can you say that again?
Oh, yeah. The bad debt for a Japanese customer that you made in Q2. Yeah. Last time you mentioned-
Oh.
There will be a reversal in the second half. Yeah. So what's the status? Yeah.
We have been recruiting, recruit a little bit of the bad debt from customer, but not all of the bad debts. We don't recover. We didn't realize any AR in the third quarter. You may see the operating expense a little bit high compared to the first quarter. That's because, first of all, we make more money than we expected, so we have to accrue the bonus. Secondly, we have relatively more activities in the third quarter, like the setting up subsidiaries in Malaysia. Compared to the first quarter, the operating expense is a little bit high.
I see. And then the last question from my side is, we'll bring back to the queue. Yeah, and for the production business, the gross margin tends to be lower. Do we have any plan to raise the market percentage in future? How easy or how to go would that be the case?
Okay, I'll target it first and maybe Johnny can give you more.
Sure.
Give you more information. For the production gross margin, yes, we understand it is low because we are facing customer with very good bargaining power. To be honestly, if we do the generation by generation project with the same customer, it is really difficult to expand the gross margin from the same customer with similar generation by generation project. However, we do try our best to trying to gain some margin from the supplier. We already have some achievements. Hopefully, hopefully, we can expand our overall gross margin a little bit by doing that.
Okay, let me add to that. Yeah, Daniel is right. Right now, I think the single customer in the U.S. dictates a majority of our production margin, production revenue. You said, and for the future business, I think for this customer alone, it's kind of difficult to increase the profit margin. Right now, I think starting from next year, we have, we've been diversifying the business to other accounts. I think this customer will no longer dictate total revenue, and they are also seeing our value. Starting from the next project, we're going to have more margin improvement. In parallel, our supplier also realizes our value. I think this is a good talking to very high-level people.
Yeah, starting from next year, yeah, we're going to enjoy a more margin, yeah, contribute from the supplier side. So overall, the production cost margin will be still much lower than NRE, but we will see a certain degree improvement. Yeah, we continuously to do so.
Okay, thank you very much, and congratulations.
Mm-hmm.
Thank you, Szeho .
Thank you.
Robert, JPMorgan, please.
Okay, thanks, Daniel. Thanks, Johnny. So, I have a few questions. First is on your largest U.S. customer. It seems like a little bit more confident this time around about their 2024 growth. So, do you think that... I mean, could you give an indication about the 2024 growth outlook? And regarding this solution, we're now in Gen 2, so how should we think about previous leaders and talk about there could be an interior solution, kind of 2.5. Does that will have any, will that have an impact on your 2024 business? That's my first question.
Okay, Robert, let me say it this way. We are not allowed to give any numerical guidance in the earnings call. We have been educated by the authority a lot last time. So for next year, as mentioned, we have very positive view on the revenue growth. First of all, the orders for the existing project is really, really strong, stronger than we previously expected. And of course, the advanced packaging capacity allocation will be a key deciding factor for our revenue growth. But fortunately, our supplier, TSMC, gave us a very good support. Even as Johnny said, we already show our value to TSMC, we already show our value to our customer, and we are already a very important customer to TSMC for its advanced packaging.
So the support effort from TSMC is best than ever. So, still, we worry a little bit about the CoWoS capacity, but all in all, that's already we are more than satisfied with TSMC's support this time for our CoWoS packaging. So for next year, for the existing product, we don't see a big problem from packaging capacity allocation. For the other customers, we are still fighting for more CoWoS allocation to them.
Mm-hmm. Okay, Robert, yeah, let me add to that. So basically for next year's revenue growth, I think, just like Daniel mentioned, we have a very high confidence. Yeah, to be honest, yeah, if a revenue growth is similar to the last time we forecast, we will never think about the GDR. But right now, I think it's, we have to have a GDR in order to support the revenue numbers.
Got it. Got it. Thanks. And about your another North America AI customer, I think it's the IDM customer. It seems like the project will heat up from late this year with much stronger growth in 2024. So, could it become a second largest customer next year? And how should we think about the production growth margin for this customer? Also, I think this customer has been indicating that the design pipeline has been doubling over the past three months. Are you seeing a same similar trend? Thank you.
Okay. Yes, this customer will become our second largest customer in year 2024. For this customer, we have two projects to be in the production. Actually, one already in the production for next year. The 7-nm one, the production will go up quite a bit in the first half, and then for the second half, the 5 nm , the shipment of the 5-nm project will heat up as so, once we finish all the testing, all the testing process. For next year, the second largest customer may enjoy the biggest growth year-on-year compared to the revenue contribution this year.
Yeah, and the margin stuff is better.
Yeah, I would say this way, this customer also has a very strong bargaining power. I believe you all know that.
Mm-hmm.
A little bit gross margin improvement, but not improvement, a little bit higher gross margin compared with the other North American customer. But yeah, still much lower than our NRE gross margin.
Mm.
Yeah, yeah. Got it. Thank you. My last question will be on the automotive. So I think you, back in a few quarters ago, you talked about that you have a design win with the Chinese auto OEM about autonomous driving project. I think this is a spec- in project because you need the front-end IT and to integrate. So how do we think about the production GM? If it's a spec- in project, then the gross margin should be much higher. And this customer also uses some merchant solutions, like from NVIDIA for the Orin. And how should we think about the adoption of these self-developed chips when it comes into production in the initial years?
Okay. So, yeah, Daniel, let me answer that. So I think this project, yes, you are right. It's a spec- in. The design cycle, we estimate, is during our time, is about 1.5 years. So far, in the design stage, I don't think there's a major problem. And so far, so good. We are about to tape-out in the middle of next year. Yeah, in terms of production margin and production forecast, to be honest with you, for this customer, they give us a very, very high forecast. But I don't know, the automotive stuff is still unproven to us. Yeah, we'd be conservative to estimate their production volume.
But in terms of a margin, I think it's much better, yeah, than the HPC business and from North America.
Okay. Sorry, let me squeeze in one last question about back to the HPC side.
Mm-hmm.
So other than this existing U.S. cloud services provider, so how should we think about your breakthrough in the other three CSPs? Are we seeing some progress in the next probably 6-12 months?
Okay. Yeah, let me try to answer first, and Daniel, you can add more. Yeah, obviously, there are only few CSP service provider has a, the, out there, has a huge volume. Obviously, none of them are satisfied with their current solution, yeah, because the current supplier are too expensive, and all of them has a plan to developing their own ASIC. Some of them already start, some of them has in their plan. Yeah, from our side, we can see that there are only a handful of company, no more than 5, maybe only 3-4, service provider has such a track record and also capability to do the design. I think the trend is the entire revenue will continuously grow, and they try to minimize the dependency from the current chip provider and also service provider.
I would say everybody has the chance, yeah, to win those projects. Daniel?
Okay, pretty much the same. Again, for this question, I would say I actually be in a very good position to potential, cloud service provider customers in North America.
Right.
Okay.
Yeah, let me say in this way, so other competitor in U.S., their gross margin is very high. On the book, it's more than 50%. Yeah, so it's a lot of incentives, yeah, for those customers to diversify their concentration.
Thanks, Daniel. Thanks, Johnny. I'll go back to the queue.
All right.
Okay, and Charlie. Charlie of Morgan Stanley, please.
Yes, Johnny, Daniel, good afternoon, and also congratulations for very, very good results. So my first question is about GDR. So first of all, why, why not doing a, you know, a corporate bonds insurance? Are you planning to introduce a strategic measures? Can you, can you explain to us why, why, why you need equity funding? Thanks.
Okay. For the GDR, actually, we talked with investment bankers for multiple ways of raise money, like convertible bond, corporate bond, GDR, and so on. But for the current condition, because if we do the borrowing, we want to borrow U.S. dollar, and the U.S. dollar interest rate is really high. So if we do the convertible bond or the corporate bond, the cost is high. And for the GDR, we do consider the potential dilution. So that's why, in the end, we conclude that we only dilute 5% of our shares to raise the extra fund to support our operation. So that's the whole consideration process.
Another thing is, for the GDR, since we've already done it, since we already did it before, everything could be done smoothly, 'cause, according to the management plan, we may suffer tight cash position in the beginning of next year, the first quarter of next year. So we want to do it quick, with limited shares dilution and with the lowest cost. That's the reasons behind this GDR offer.
Mm-hmm. Okay. Yeah. Charlie, let me add, add to that. The GDR offering, we are thinking about many times, how, how to prepare the cash for the operation. Yeah, like Daniel mentioned, right now, the interest is so high. Yeah, given our profit margin is not so high, I don't want to give a certain portion as a, to the bank. Also we consider this revenue, and also production trend, is not short-term. That will be sustainable. Yeah, we need enough cash, instead of a continuous in borrowing the cash from the bank. Another important reason Daniel also mentioned during his presentation, when we talk to other service provider in the U.S., the cash flow is always their consideration. Yeah, we all know we are handle.
Eventually, if we win, that will be multi-billion type of a business. Based on our current cash flow condition, they will not have a confidence to work with us. Right now, I think company grow quite a bit. We are no longer compete with some vendor in Asia. We are directly compete at Tier 1. As you know, our competitor, Qualcomm, Broadcom, their cash flow condition are much, much healthier than us. Yeah, we need to prepare ourselves in order to win big.
Got it. Thanks, Johnny and Daniel. So, yeah, my next question is also associated to this GDR, right? So I can understand that next year definitely see a revenue upside, so they require more working capital. But besides that, is there any change of the kind of payment term by your vendors? And also, I know it's already a done deal, right? But your vendor, TSMC, right? During this kind of very tight supply or strong demand, they ask customers to prepay or sign a LTA, right, to share the cash flow burden. So would you consider to ask your customers to give you some down payments to share the burden of cash flow?
Daniel, do you want to go first, or I can answer?
Charlie, very good question, and, I, yes, we already, we already did. We talked intensively with our customers, but, as I just mentioned, when you do the projects with generation by generation, to change the terms is really difficult. For some other customers, we already get some progress that, for example, they may pay certain percentage of the total payment when the tape-out is out. We don't need to wait until the whole CoWoS process to finish and then to get the money from customers.
Hmm.
I think gradually we will try every way to shorten to improve the current payment term for every customer. But to be honest, when you're facing customer—that's with significant bargaining power. But yeah, it's really stressful for us.
Okay. Let me add more. Yes, when we encounter some cash flow issue, personally, I go to supplier and customer a few times directly. Yeah, we provide such a high volume to the customer. We should provide the service, provide a good design to them. We show them bank letter. I think that's so obviously. But somehow, when we talk to supplier, right now, I think supplier is very strong, and the CoWoS is in shortage now. Without the top die sitting there, they, we will not get any CoWoS allocation. But top die, once top die go out, we have to pay supplier. Not only us, all other customer, even much bigger than us, follow the same route.
Yeah, our supplier also has a little also facing some cash flow challenge. So I think this due to the CoWoS shortage, it's very difficult to negotiate with the supplier. But when we talk to customer, I think, to be honest with you, customer is very supportive. They give us a lot of proposal. Yeah, since their revenue, I think, is 3x-4x higher than their original estimation, they totally appreciate our design work. They want to help us. There were many, many different models. We can convert some of them to a royalty base. We convert some of them to the revenue base. And also, we are going to take it. We are shooting for to win their next generation.
So after all the consideration, we're thinking about, to raise another GDR will be the easy solution. Yeah, when, whe n we have a more customer, then entire revenue will be if it's more diversified. I think we have many way to control the cash flow and also gross margin.
Got it. Yeah.
Mm-hmm.
So, yeah, thanks for all the explanations. Yeah, I'm not trying to be picky, but it would be great for investors to understand all the consideration. Great job. So, just a last question on this, the GDR-related parts. I did some calculation, right? You probably, by the end of this year, can end up with $3 billion cash on hand. I'm not sure if that is the right estimates. And you are raising around $350 million in addition , right? So, yeah, I'm just doing some back-end level of calculation. Assuming you keep, like, $1 billion as a kind of working capital, I feel like you probably need an additional 15% revenue upside for this year. Is this a kind of right calculation? I mean, this is the $350 million additional working capital imply.
Okay, we are unable to give confirmation to your implication. Like I said, we were educated by the authorities. But yes, you will see by the end of this year, our balance sheet, you will see a lot of inventory. The majority of them will be the wafer, will be wafer waiting for CoWoS capacity. And yes, the current growth outlook is really, really good. You can see the picture that we may have pretty good upside growth outlook for the existing project. It is already our biggest revenue contributor. For next year, this shipment won't decline, and the upside potential could be up to 50% or even higher, depends on the CoWoS status.
Mm-hmm.
Plus, we have the other customer to become the second largest one of our customer portfolio. So the growth outlook is really, really good in terms of the revenue.
Thank you. Yeah, so next one, and a quick one, I will back to the queue. It's about actually the near-term trend. So can you give us some sense about fourth quarter revenue trend? I know you cannot give a numeric number, right? But can you compare this quarter to any quarter in the past, right, to give a sense about the revenue level and do you see upside to your previous full year revenue guidance?
Okay. For the fourth quarter, fourth quarter, no matter, I think no matter the top line or bottom line will be the highest quarter in 2023. For the percentage-wise, I would say, higher than the second quarter, which is right now the highest quarter among the first three quarters. How high it can go, it still depends on the status of the advanced packaging. That's what I can tell to the investors for now, but definitely the fourth quarter will be the highest quarter this year.
Okay, got it. Much appreciated. I will be back to the queue. Thanks.
Thank you, Charlie.
Okay. Okay, next is, Laura, please proceed.
Yes, hi. Yes, hi, can you hear me?
Yes.
Yes. Thank you, good afternoon. Thank you for taking my question. My first question is about, like, given the strong growth outlook and also so many projects on hand, I'm just wondering that, what would be the OpEx, longer term OpEx ratio you are looking for? And, what will be more reasonable estimate, in terms of the, your, R&D force planning all over the place. So can you give us, some idea about, your, OpEx trend looking forward? That's my first question.
Okay. First of all, for the OpEx, usually we take the top-down approach to do the OpEx budgeting. So for now, we are still in expansion, especially the expansion in the overseas sites, like Malaysia and maybe early in the future, the VN project. So for now, I plan to expand the OpEx by up to 15%, 1 5, 15% of our current yearly OpEx. That's my trend.
Mm-hmm.
Okay.
Uh.
Thank you. So sorry, please. Please go ahead, sorry.
All right, Laura. So I think the plan doesn't change. Our OpEx increase every year, I think is insignificant compared to other competitors. Yeah, I know recently our trend, our plan changed a bit. Before, majority of headcount is in China, but right now we are diversified quite a bit. We compare the cost between China and Malaysia. I think we see some cost saving out there. But in the meantime, since we need to support some project by sending some engineer to Japan or to Malaysia to do the engineering work, that will be about, you know, 10%-15% overhead on their salary level.
I think everything, if I add all together, I think it is the OpEx estimation for next year. I think it will be very reasonable increase.
Okay, thank you. Can you give us a rough idea what's the ultimate allocation on different regions like China or Japan, Malaysia? What will be the shape, maybe in two, three years perspective?
Mm-hmm. Yeah, I already give out the proposal few calls before. For this year, yeah, we plan to prepare, like, 100 engineers to outside for China. By the end of next year, I hope between China and non-China will be 50/50. Yeah, so that percentage in include the engineers sending out of from China. So by the end of next year, it's a 50/50. And just for your information, right now, we just do the calculation. Less than 15% of customer has a specific requirement to do the design outside of China, but we believe this number will increase. But right now, it's less than 15%, customer has such kind of requirement.
Okay, got it. Very clear. And, my second question is about, again, the CoWoS capacity planning. We understand that you are working with TSMC on CoWoS-R. At the same time, as you mentioned several times about the strong demand and also the high capacity on the backend advanced packaging side. So my first one is about, like, if we goes to CoWoS-R, how much benefit of the growth margin or cost can give us to, like, support the better gross margin, go, upon the longer term perspective? And also, at the same time, since probably your smaller clients may still have some tightness on the CoWoS allocation.
Mm-hmm.
Are you also working with the old suppliers to looking for their capacity support?
Okay, Laura, let me answer you first for the packaging change. As I mentioned many times, that for the existing budget to North America, our biggest customer, is a cost-plus margin scheme. So cost change won't affect our growth margin. So there will be no difference for whatever the advanced packaging is. And for the future project, yes, we do have the program, but because these type of things related to our supplier, we respect our supplier, so we cannot discuss into detail about the cost things here or pricing things here.
But all in all, at the end of the day, you can think this way, that given our position is getting stronger and stronger, we have more room to talk with the suppliers and the customers. So I believe there will be some room for us to expand our gross margin.
Mm-hmm.
And as for the offset, saying, if we can find other offset for the advanced packaging, it really depends on, first of all, the customer we serve, our customer. Secondly, the foundry partner. The strategy our foundry partner takes, we will work closely with them.
Yeah. Yeah, let me add to that. So for CoWoS-R or CoWoS-S, I think the gross margin is not, we don't see bigger difference, but CoWoS-R will give us more capacity. I think that's the biggest incentive when we try to introduce the CoWoS-R, working with TSMC very closely. Yeah, if CoWoS-R can go to production, I think that will solve our capacity issue for the next year. And also, there's a good demonstration for CoWoS-R to other customer. Yeah, for this foundry. Yeah, the gross margin stuff, I think we are talking with the supplier separately. I think no matter it's a CoWoS-R or CoWoS, CoWoS-S, I think we will get some support from foundry. Yeah, next year.
Thank you very much. Just a very quick follow-up. I mean, during the transition from, like, CoWoS-S to CoWoS-R. Do you see, the output, the revenue may get some fluctuation, even like, probably the CoWoS-R is just initial ramp stage?
Mm-hmm.
Will this give some, like, a sales volatility then?
Already happened. That's why we see a little bit revenue fluctuation between Q2 and Q3. But thanks for the foundry's support. If CoWoS-R encounters some problems significantly, they will give us more CoWoS-S to cover it. Yeah, but starting from next year, later this year, I think the CoWoS-R will become the majority.
Okay. Thank you. Very clear. Appreciate it.
Thank you.
Thank you.
And, Bruce, please.
Okay. Thank you for taking.
From Goldman.
Yes.
Yeah.
Yes. Can you hear me now?
Oh, yes. Yes.
Okay. I still want to ask about the GDR things, because I understand as for the working capital requirement, which is, you know, basically is the top type waiting in the queue, as well as the production time for the CoWoS. However, you know, TSMC is meaningfully increasing the CoWoS capacity, and you will also do a lot more CoWoS-R in the future. So i.e., the CoWoS problem will be a lot better, you know, in second quarter, after second quarter next year. So your working capital requirement is most like a short-term thing. You know, why do you need to do a GDR instead of bank borrowing or bank borrowing or anything else? I mean, this for me, this is more like a short-term working capital requirement, right?
Bruce, I already mentioned, it is not, actually, for six months or nine months, is not considered a very short term. We are an asset-light company, so the ability to borrow money from a bank is low. We don't have deposits for us to borrowing money. If we do the credit line borrowing, the interest rate will be very high, especially, we are trying to borrow dollars.
Right.
Another big reason is, like I just said, we want to strengthen our book in order in preparation to take other projects from other U.S. major CSPs. So the purpose is not only to solve the short-term cash position problem, but in the long run, we have to be prepared to take projects to make us to be more suitable to compete in the North American hyperscaler market.
Mm-hmm.
I see. Thank you. I think my second question is for, you know, you heard from MediaTek saying that they will be very flexible in terms of business model. Do you consider them as a competitor? Are they gonna be flexible enough to do the design services and to compete with you in the design service business?
Okay, Bruce, actually, first, first, first of all, I don't know what the definition of flexible or flexibility. MediaTek has been in this business for a while. So in some parts, we are competitors, we are competitors for sure, but we don't avoid to cooperate with MediaTek, depends on the project. We are... Actually, two, the management, the managements in two company are familiar with each other.
Hmm. Okay, so those projects.
Mm-hmm.
I'm sorry, I'm sorry. Please.
So, I'm not understanding. I'm not getting what's your question you are asking? If MediaTek is a competitor to us? If that's the case, yes, MediaTek is a competitor to us.
Do you face MediaTek's competition in the current project or the potential project bidding at this moment?
Okay. Let me say this way: We don't encounter with MediaTek for many projects. Usually, the competitor for the project right now in North America market is Broadcom or Marvell, and sometimes QC. We rarely hear about MediaTek, but I don't know. I'm not MediaTek person. Maybe they are doing other projects. We don't know. Yeah. I see. Mm-hmm.
Okay. Thank you. Okay.
And, Charlie, please.
Okay. Thank you. So may I circle back to the China restriction issue? Because you're another U.S. customer, which I believe is Habana, also have a very high performance GPU products, right? So since the new restriction rule, do you see any change for your forecast? Why and why not? Can you give us some more comments?
Okay. Yeah, we are not allowed to mention any names of our customer. So, we do worry, but based on the current information we received from our customer, everything will be as slow as it gets. And, to be honest, we don't want to give too optimistic view to investors about the North American customer. But, for now, the PO they placed to us, the forecast they gave to us, is really, really good. So, we want to be conservative, and, for now, what we can do is only to trust our customer. They say, no problem, and, yeah, that's my answer for this question. Mm-hmm.
Yeah. This specific customer, according to them, they are not just selling the chip, they are selling the system level with the software, so they can, according to them, they can dynamically to adjust the performance. So since after the incident, I didn't see any decreased demand, but also they continuously adding more orders. Yeah, hopefully, everything will be go smoothly.
Hmm. Okay. Interesting. Thanks. Yeah, so another question is about also related to this accounts, right? So assuming this customer is making $100 revenue, how much can Alchip get from the design service side?
Sorry. Sorry, Charlie, I don't get your question.
Yeah. So, so, so then let me go straight, right? So if this customer... Yeah, if, if this customer is making $1 billion revenue f rom this chip, how much revenue can Alchip book from a design, design service vendor's perspective?
Okay, our revenue is our customer's cost.
Hmm.
We don't know the price our customer sells to its customers. Honestly, we don't know. We don't know. But, if we, if we compare our price to the current, product solution, the ASSP solution, the price of us is very competitive.
Hmm. Very cheap. Hmm, okay. Can I assume this customer will be, like, 20% of your revenue next year?
Hopefully. Hopefully more.
I'm lazy. I haven't done the very cold forecast for now, so I have to take responsibility for my words. So, yeah, I'm sorry that I can give you the answer.
Okay, okay, okay, that's fine. Yeah, we know it'll be strong, it will be fine. Second largest customer, I think, that should be good enough. So, my last one should be, not again, not very specific, right? And you can call it in the coming three years, right?
Mm-hmm.
Do you see, N3, right, the 3 nm , no matter NRE or, or turnkey, will become a, a major growth driver?
Mm.
Because, I think now, your several process at the prime time, even 7 nm, can contributes more than half of your revenue. So I just want to get a sense of when Alchip will migrate to the 3- nm era.
Okay. Yeah, Johnny.
Yeah, let me take that. So right now, I, we can say for this year, we consider 7 nm is kind of, typical for, for us, and 12 nm, we can say, is mainstream, and, and 5 nm is more advanced. So starting from next year, we can consider is a 5 nm is our typical case. We will have a, we have a few 3 nm and few 7 nm as well, majority of the design will be, sitting at 5. But we already have a 3 nm, 2 - 3 nm, 3 nm design win. So I think starting from 2024, 2025, and 5 nm, 3 nm, I think will become the key revenue contributor.
Hmm. Okay. So for those 2-3 nm, 3-nm design, should I assume, or, or those are hyperscalers or, or some of them are.
All hyperscaler.
Yeah.
I think the only few, only few application will use the three. Either is a cell phone or any CPU, GPU, HPC, compute power related project. Yeah, we are not touching the cell phone. I think majority, everything will be HPC and AI related.
Hmm. But I remember the company said before, that you probably will win some U.S. EV customers at N3A. Is that still the case?
N5A, I think is the c ompany know they are using, or N3AE.
Hmm.
Yeah.
Okay.
N3, so, go ahead.
Yeah, so do you set to win any N3AE for the automotive customers in the coming two to three years?
Yeah, we are competing. We try to win, but it's no winning confirmed yet. But N3AE will come later than N5. N5A.
Okay.
Mm-hmm.
So the 2-3 nm, 3-nm design wins, those are all for hyperscalers for the time being.
Yes.
Okay. Okay, got it. Thank you.
Thank you.
Okay, before answering the-- not answering, before Robert to ask question, I want to ask some questions in the message box. First of all, someone asking about the revenue breakdown between NRE and the production. The quarter, in second quarter, the NRE and the production split, is the NRE revenue accounted for about high 20-something%, high 20s%, and the other is production. That's the split, that's the split for the sec-- for the second quarter. And someone is asking about the [Foreign language], the biggest customer, year-on-year growth rate, is 50% or 15%. I'm not trying to emphasize the numbers. Let's say this way, it is definitely not 5%-10%, those kind of growths.
We have pretty good upside potential year-on-year for the same product we are shipping right now. But is it 30%, 40%, 50%? Because the situation is changing every day. So I can just say 50% is not unachievable. The potential for year-on-year growth is pretty good. Not that 5%-10% range. It's a much bigger percentage than that. So that's the answer to the growth rate for our biggest product. And another is, given the higher than expected OpEx in the second quarter, what OpEx level would you expect this year? Okay, I would say, by... because we wrote down some AR in the second quarter.
So for the whole year, I believe, by factoring the AR write down, the whole year operating expense will be like in the middle of $ 90 million-$ 100 million. Okay, and then, Robert, please.
Oh, thanks. Thanks, Daniel. So I just have a few follow-ups. So on the NRE revenue growth, so since I mentioned that customers are migrating down on probably 5 nm more next year and probably 3 nm beyond, how should we think about your NRE revenue growth CAGR? Could it break beyond kind of 20% CAGR that you had in the past few years because of the higher mask, higher tape-out, higher software and verification costs? Yeah, so just to get your thoughts about the NRE revenue growth outlook.
Okay, Robert, to be honest, for the NRE, the visibility is not that low. As you can imagine, if we are discussing a project with a customer, the time frame will be the next half, next six months to 12 months. So, based on our experience in the industry growth, we usually give ourselves target to grow our NRE 20% year-on-year in a normal year. Of course, in a good year, like what we are currently, we always expect higher growth rates for our NRE, i.e., more than 30%. But for these two years, it's really complicated. Very good opportunity in North America, but for China market, the whole IC industry is kind of sluggish.
So, but anyway, we still managed to deliver the growth rates we consider is good to this industry. So if you are asking about the long-term CAGR, I will tell you 2 0% is our target.
Okay. Thank you. My last one will be back to the, yeah, other CSP customers. So I think you sound confident to compete with, Broadcom, Marvell, to win other CSP customers for AI projects. Does that mean that you're stick with the current, project type, which is the full back-end design plus the back-end design, or you also become more flexible to do the only, the COT design, package design, or in your, term, the PD3 kind of project?
Of course, we have to be confident. Otherwise, do you want me to tell you, "No, we don't have confidence to win project?" Of course, we have confidence, because we are one of the few that are capable to provide leading-edge, large-scale design, and we track record of large scale production. And, for our advantage to the competitors, is very straightforward. We are more flexible, we are quicker, and, of course, our pricing is better to our competitors. And of course, our competitor, our competitors have the reasons to become the industry giants. They have everything. They have front end, they have software, they have IT, they have almost everything. So I think it really depends on the customer's preference of the project.
We do see opportunity emerging in recent years that we have the room or window to tap into the North American HPC chip supply chain.
Mm-hmm.
Okay, let me add to that. So, yeah, to answer your question, our business model also very flexible. Yeah, if the Tier 1 customer come to us only asking for the packaging service to prove ourselves, of course, we will say yes. But for our side, I say, we all know what's our expertise area. We're doing the design much better than most of our competitor and also customer itself. Yeah, we've been taping out 30 designs every year, generation over generation. We can control the schedule, we can make the die size smaller, much higher performance. So that's our, that's the key area customer like us a lot. If we're just doing the production, handling the packaging, we become. Later on, we'll become a DC of a foundry or supplier. There's no value.
So I think winning, we consider doing the PD1, doing the design, will be is a way to sustain the revenue, or to sustain the customer. So we are continuously focused on that area. I think, if they want us to do some piecemeal or any flexible business model to prove ourselves, of course, we will say yes, but eventually we will win the design, become a true partner of them in order to sustain the business.
Thanks. Thanks, Johnny. That's very clear. Thanks, Daniel. Thank you.
Yeah.
That's all.
No problem. Okay, since the time is late, I'll answer the question that the last question from the message box. What is the cash conversion cycle for your production work with HPC customers? It really depends. Like Johnny mentioned, the whole process right now, and the majority of production revenue, the majority of our production revenue comes from projects with CoWoS process, CoWoS packaging. So for CoWoS packaging, you have to be top die ready and then waiting, and then wait for the CoWoS to start. For now, the gap between wafer ready and the CoWoS start is not under our control. So yeah, but all in all, the cash cycle for projects applying CoWoS is much longer than the normal project.
Okay. Yeah, let me add to that. So the GDR offering, I don't know, people consider is a positive or negative, but for me, I think this is positive because our companies are growing. Yeah, I agree with Bruce. If the CoWoS situation, payment term, I think eventually will be less and less burden. But if our revenue just stay into a certain level, we don't need a GDR. But fortunately, our company grow very fast, yeah, in order to sustain the future business and win other key customer. I think the GDR offering will definitely help us bring, upgrade company to another level. Yeah, so overall, I think personally, I'm very positive, and dilution is insignificant.
You know, I see the, with this GDR, company will, in a very confidence level, yeah, for a while.
Okay. I guess, that's it. Thank you for participating to our third quarter earnings call today, and thank you for your interest in Alchip Technologies. See you next time.
Okay.
Thank you.
Thank you very much. Thank you.
Thank you.