I guess now it's about time to start our second quarter institutional investor conference meeting. Again, please keep on mute your speaker to better process this meeting. Then for today's meeting, first of all, welcome to our meeting. Today's meeting will be hosted by me, Daniel Wang, and our CEO, Johnny Shen. As usual, we will start from the second quarter presentation, which includes our second quarter financials and then the industry view and the future outlook for the second half of this year. Now we would like to welcome our CEO, Johnny Shen, for the opening. Before that, again, here is the safe harbor disclaimer as usual. This meeting will be in English. If you need the Chinese slide, please go to the [Non-English content] to download the Chinese version.
Johnny and me, actually Chinese is our mother tongue. If you have any question you want to ask and be answered in Chinese, please let me know. Don't hesitate to just use Chinese for your question. Johnny, please.
Okay. Good afternoon, ladies and gentlemen. This is Johnny Shen, President and CEO for Alchip Technologies. Thanks for attending the Q2 investor conference meeting. We appreciate the opportunity to share our Q2 result and also provide the guidance for future business. In case you are not too familiar with Alchip, allow me to make a company quick update and identifications. Our company is founded in 2003, IPO in 2014. Our current headcount is 485 people. Majority of our employee are engineer, about 75%. Since we founded the company, we've been tape out more than 470 design in leading-edge technology. Last year, our revenue about $240 million, currently we are supporting more than 30 design in parallel. Majority of our revenue are coming from AI and HPC area, we are one of the TSMC VCA members.
A quick update for Q2. Overall, we can say we have another good quarter with the record-breaking numbers in almost all category, including revenue, net income. Q2 EPS NTD 5.54, slightly lower than the Q1 number due to fractional outstanding share increase and also the weaker NTD exchange rate. All detail number analysis will be provided by Daniel in a later section. We had a multiple tape out and design won, mainly 7nm designs. Leading-edge technology businesses remain very strong, especially in GPU, CPU, and AI application. Quick update on Phytium status. Since the incident happened on April 8, we are working with the customer, U.S. attorney, and communicate with our major supplier very closely. Finally, we have achieved a memo from U.S. attorney clarify the Phytium product are not subject to EAR. I think that's a very good progress since the incident. Okay, Daniel, please.
For the second quarter financials. The second quarter total revenue, I guess we already announced the revenue. The total revenue in terms of the US dollar is $97.4 million, which is 3.9% quarter-on-quarter, and 77% year-on-year. For the operating income. The operating income for the second quarter is $70.5 million, which is quarter-on-quarter declined 1.2%, but for the year-on-year basis is a 101% increase. The net income benefit from a little bit higher non-operating income. The net income turns out to be higher than the first quarter. The second quarter net income is $13.9 million, which is 1.7% quarter-on-quarter and 117% year-on-year.
As Johnny mentioned, due to the increase of our total outstanding shares because of the realization of option and the strong NT dollar performance, the second quarter EPS is NTD 5.5. This is the revenue breakdown by application. You can see we are still concentrated in the HPC area. In the second quarter, the revenue from HPC accounts for 87% of our total revenue. The other three, the consumer, networking, niche, combined accounts for 13% of our total revenue. For this year, the first half, the HPC revenue accounts for 89% of our total sales. You can see we are very concentrated in the HPC area. For the technology node, I can very proudly say that our chip should be the leader in terms of the process node. For the leading edge technology node, especially for the 7nm or even higher, 5nm or 6nm.
In the second quarter, combined, this revenue from 7nm or higher account for 74% of our total revenue. If you combine the 7nm, 16 nm, and the 28 nm, which already accounted for more than 95% of our total revenue in the second quarter this year. For 2021, the first half, you can see the 7nm accounts for 67% of our total revenue, combined with the 16 nm. These advanced technology nodes already accounted for more than 90% of our total revenue. For the regional breakdown. In the second quarter, again, China is still our major revenue contributor, which accounts for 64% of our total revenue in the second quarter. You can see the others increased from 5% to 27%, which is contributed by our North American and the Middle Eastern customer.
We have a 7nm tape out for our U.S. customers in the second quarter. I think this trend will continue to the second half or even 2022, as the contribution from non-China customer will increase in the future. For the second quarter 2021 operation review. Again, China HPC chips remain the main growth driver of our total revenue. Production contributions from our major China HPC customer, which formerly is Phytium, now is what we called the second-largest customer before. Its 7nm HPC chip, shipment, is the major reason and the major growth driver for our second quarter performance. The gross margin in the second quarter went down slightly because of higher mass production revenue mix. Our second quarter, the mix between design and the production is around 40% to 50%. The design accounts for about the high 40s, about our total revenue.
Again, our revenue kept on concentrating in leading-edge process node, which is a very obvious trend for our revenue mix. The operating income for the second quarter came at NTD 17.5 million, which, we are very happy that's up 101% year-over-year. I think the promising result was mainly due to, first of all the projects went on schedule and as expected. We get very good revenue support from our suppliers, of course, thanks to TSMC's fully support. The operating expense went down quarter-on-quarter. As I mentioned in our last conference meeting that for the first quarter, the operating expense, which include some one-time expense, such as the GDR expense, such as the other fractional one-time expense. For the second quarter, for those one-time items, they are less than the first quarter, so our operating expense went down slightly quarter-on-quarter.
The net income, again, hits the record high at NTD 13.9 million for our second quarter. For the second half outlook, I will roughly talk about this outlook. First of all, I have to say sorry to all the investors because the authority warned us that we cannot give some guidance in a very clear numerical way. We cannot give out the numbers anymore. I would like to try my best to give some directional guidance to you. First of all, for the industry, the industry-wide demand outlook remains bright. I guess many of you already know that there are massive investments into the semiconductor industry, especially for IC design in both China and North America. The design opportunity keeps on increasing. For now, winning project is not a concern anymore.
There are plenty of projects out there waiting for the design resource. The dilemma for Alchip right now is to decide which projects we are going to take. For the business wise, the industry is very hot. For HPC related design, the demand keeps strong for both the computing and the AI applications, which means the CPU, GPU, which is computing, and for the AI inference chip, training chip, all across the board, the demand is very strong. For the 7nm AI chip shipment, we believe it will become a very important growth driver for our second half this year and for 2022, the whole year. We plan to kick off the shipment for our first 7nm AI chip to very important U.S. service provider customer.
We think the revenue contribution will be starting from middle of the third quarter, and we have very high expectation on the future revenue contribution because it is a high volume project. Our second and the third AI chips to another U.S. importantly customer will sequentially begin shipment right after the first one. We believe these two chip, the revenue contribution, will become very significant in 2022. For the second half, the China HPC customer, which I call it our second largest China customer, its seven nano shipment tapers, because its product life cycle near the end in third quarter. We don't expect very significant shipment for this project, starting from the third quarter. The good news is, we expect another future demand for the same chip to kick in 2022 because this customer wants extra orders for the system.
The demand would temporarily ease, but we expect the revenue contribution to grow again for this chip in next year, 2022. That's the rough guidance for the second half. I know many of you want to know some revenue or profit guidance. In our previous institutional investment meeting, we told that we are aiming on 20% year growth for this year, which means close to NTD 300 million. For now, we are becoming more and more optimistic towards the second quarter performance, especially towards the second half this year, especially for the third quarter. In the past, previously, I think many of you already heard about, I was worrying about the third quarter performance, which will be a very challenging quarter for Alchip. For now, because the mixed result of, first of all, the NRE inflow is very strong.
The design inflow is very strong, and some customers are pulling in the production shipment, and then many other reasons. For now, the third quarter outlook is much, much better than our previous thoughts. For the fourth quarter, again, we don't think fourth quarter will be a weak quarter. Usually, fourth quarter is the high season for our design. I think this pattern will also apply for this year's sales performance. Another thing is, I guess everybody knows the pattern. We do believe as long as we can have some breakthrough for our customer, we may have upside potential for the second half and year 2022. I guess that's pretty much all the content in the presentation for today's conference meeting. Now we are entering the Q&A session. If you have questions, please use the raise hand button, and unmute your speaker.
We will take the questions. Okay. Jeff Lowe, Macquarie , please.
This is Jeff Lowe from Macquarie. Thanks for taking my question. I know Johnny mentioned that the 5nm products are not subject to EAR. What exactly does that mean, and what do you have to do beyond that to start shipping again?
Okay, regarding the Phytium status, we can only say that we currently have positive results from the lawyer, which is a U.S. attorney. Those projects are not subject to EAR. The result we are receiving is positive right now, and we are processing the BIS classification right now. Other than that, in order to serve the best interest of our customers and the suppliers, we are unable to further state more or comment more on patent. Please understand our stance.
Okay. Second, last question, I'll get back in line.
For your second largest customer, which I guess became your first largest customer in second quarter.
Right.
What percent of the sales, and can you talk a little bit about what exactly project is it that they're doing? Just a little more specifics on their product or end product. Thank you.
Okay. The end product is for HPC application. The contribution percentage, I don't have an exact number in my mind, and I don't have the Excel right now. Let's say, the second largest customer maybe accounts for about 30% something of our total revenue in second quarter.
Great. Thank you.
Thank you. Si Ho, please.
Oh, hi, gentlemen. Two questions from my side. The first one regarding the U.S. customer's first seven nano shipment. How long would that last in terms of the volume production?
Okay. If you are referring to the first chip, the service provider in U.S. customer.
Yep.
Far, we don't know when the product will end its life cycle.
At least it will contribute four years next year.
Okay, great. Any chance you can talk about any five nano NRE project we are working on now?
Actually, in previous meeting, I already told that we have one full mask, a one full mask, 5nm full mask project from Israel customer.
Which is AI application. Now we are in the design phase, and hopefully, the tape-out will be in the early next year.
Yeah. Okay. I see.
Yes. In the same time, we are currently engaging in many talks with customer for the 5nm or 6nm projects right now.
I see. All right, sounds great. Also, in terms of the capacity and booking, I mean, for foundry, how comfortable are we getting enough support next year?
Maybe Johnny can comment later.
Okay. Yes.
So far, not comfortable. The supply chain is a risk to our revenue performance next year for sure, because the leading edge technology node, the wafer capacity is still tight, and the substrate capacity is still tight.
For the first half this year, we did get good support from our suppliers. We will keep on working, but at the current stage, nobody can guarantee how much we can get or how much they can give us.
Okay.
Okay. Si Ho, let me try to add to that. Yeah. I agree with Daniel. At this moment, it's not appropriate to comment on the capacity commitment from supplier, not only from wafer, but also from substrate. I think information is quite sensitive. Based on the track record and based on the history, we should get more than most of people. I think for the first half, like Daniel mentioned before, I give a full credit to our supplier.
Basically, our demand is almost all filled. Looking forward for the next year. Next year will be very challenging, but we are working with the customer and supplier very closely. Sometimes if necessary, we can do the prepay, we can do something in order to secure the capacity. Right now, I think the most challenging one is not only the production side. I think the resource is also quite challenging, especially in China region. The design resource.
All right. Thank you. Thank you very much. Yeah, keep up the great work. Yeah.
All right. Thank you.
Okay. Johnny, please.
Thanks. Good afternoon, gentlemen. My question is about your 7nm NRE. Does that include any contribution from Phytium in second quarter or third quarter?
For the second quarter, because the Phytium incidence happened on April 8th, we think there were very minor revenue contribution from Phytium in the second quarter. For the second half, third quarter, we will see. As mentioned, regarding the Phytium status, we cannot talk into detail for single customer too much. If there any chance, we will fully support Phytium for their design.
Okay.
How about those NRE? Do you continue those designs for 7nm? It's the future revenue stream, right?
7nm projects.
Okay. Yeah. Let me try to answer this. Yeah, the Phytium design, I think is ongoing. We never put on hold. The tape-out, I think we are working with the major supplier, whether they agree to proceed with prototyping or not. Design stage, we never stop. That's why we still have a certain revenue, milestone revenue recognized based on the design milestone. Yeah. Fortunately, today they are still a few months away from their original tape-out target, and hopefully everything goes through, we can tape-out on schedule. Right now, we are recognized based on the contract.
Without any problem.
Okay. Thanks, Johnny and Daniel. My second question is about the AI contribution. It seems like your second largest U.S. customer, the contribution get a little bit pushed out, according to my impression. Can you refresh the guidance, the AI revenue contribution by the end of 2021? What do you think that is going to be in 2022? Thanks.
For the AI contribution, I don't have numerical numbers, exact numbers for you right now. I will check with my Excel sheet to get back to you later. As mentioned, we have very high expectation for the revenue contribution, starting from the third quarter this year and especially for next year. This 7nm AI chip is considered a very meaningful chip for our customer. Our customer will use their own chip to gradually replace their current solution. The shipment volume, we do have very high expectation. For the, you say, according to your memory, there's delay for our second important AI customer, U.S. customer. I think the schedule is not off a lot, but there are always some design problems, maybe happens, caused by our customer or caused by us. Overall, the schedule is not delayed a lot.
As I mentioned, sequentially, we will see tape-outs for our second largest customer in the second half this year. We do think the future for these two 7nm AI chip will contribute us in a quite significant manner next year.
Okay. Thanks very much. I will be back to the queue. Thanks, guys.
Thanks, Johnny.
Mm-hmm. Hans Li, please.
Hi. Thanks, Daniel and Johnny for the business update. Basically, I have two questions. First one is that, I know that it's still a bit far, but if you look at next year's sales contribution from U.S. and China, are we safe to say that the growth from the U.S. side would be stronger than China due to the current visibility that we have?
Okay.
Hans Li.
Yeah. Hans Li, let me try to repeat your question. You mentioned about the weight between U.S. and China next year, are you thinking about the percentage from U.S. will increase, right? Is that your question?
The weight as well as the sales growth from U.S. as compared to China.
Sorry, Hans Li, your voice is not very clear. I cannot hear you very clearly. Are you saying?
Sorry.
Mm-hmm.
Could you hear me now? Is it better?
Are you saying the project count?
I was saying the sales growth from the U.S. side as compared to China, if you look into next year.
Okay. Sales growth. Uh-huh.
Yes.
Okay, Daniel, do you want to comment?
In terms of the percentage, the year-over-year increase, of course, U.S. growth rate from the current point of view should be higher than the growth rate of China. The base is different. Our China revenue base is high. There is also another swing factor, which is Phytium. I cannot give you a firm answer right now, but according to our pipeline and our future outlook, I think from the growing pace, North American market, we have high expectations.
All right. Thank you very much.
Okay. Thank you.
Okay. Bill? Bill Huang. I guess you are from Morgan Stanley .
Yes. Hi. Thank you for the time. I'd just like to follow up the question that the gentleman just had. Could you talk about, in terms of the revenue contribution, maybe in the second half or even 2022, between the China and also U.S.? Because I look at the China number in the second quarter, it seems like the numbers still growing quite nicely despite the Phytium being minimal.
Right.
I'd just like to get a sense as to what we'll be looking at into the second half also 2022. That's my first question.
Okay. For the second half, I think, again, if you look at the 2021 overall, China will still be our most important revenue source. For the second half, as I mentioned, our seven nano HPC project to our China customer is tapering off. Coupled with the increase, the takeoff of the first 7nm AI project to the service provider in North America. I think for the second half, the contribution from North America will increase in terms of the percentage.
Mm-hmm. Yes.
I see.
Bill, let me try to add to that. Daniel also mentioned there's many factor dependency for the overall revenue. Capacity allocation, whether we can get enough capacity. The swing factor from the Phytium. If everything goes smoothly, we expect a lot of larger production from China as well. From U.S., if we just look at it separately, compared to itself, I think the U.S. growth rate on the second half and the future year will be quite substantial.
I see. Thank you very much. My second question will be on your R&D capacity, because you have been mentioning about you have to choose the project, basically. I'd like to get an update as to what is the R&D capacity right now. How many R&D that you have? Also, are you looking to expand into this year and also next year? Thank you.
Mm-hmm. Okay. Right now, to be straight, our capacity is 100% full. We do reserve, other than the R&D people taking a project, we do reserve a certain percentage, like 20%, for the future technology developing. We are doing the hiring very aggressively. Like I mentioned before, our current headcount is 485. Ideally, by the end of this year, we can be in the 550 range. That will increase the capacity quite a bit, starting from maybe later next year.
I see. Can I have a follow-up on the R&D expansion part? I believe most of your R&D will be in China. Are you thinking when you expand, will you be still focusing China, or would you be expanding elsewhere? Thank you.
That will be two type. For the physical design, yes, mainly in China. For the package type of design, including CoWoS, including circuit design, COT, I think Taiwan will be the area we are growing the resource.
I see. Thank you very much. I'll go back to the queue.
Thank you.
Okay. Charlie, please.
Thanks
Yeah. This question is more industry-wise, maybe not just Alchip. My question is about you have seen lots of China system companies, internet companies, even module house, like Sunny Optical, et cetera, announce they want to do their own ASIC. I'm wondering whether these potential customers really have that kind of capability to do ASIC design. How do you think those ASIC design can be fulfilled?
Okay. That's a very good question. Right now, selecting the right customer is our number one priority. We have a meeting internally we call beauty contest. In order to choose the right customer. Yes, some of the system house, they decide to do ASIC. In fact, most of the system house right now they have a plan. We have our selecting criteria. Basically, when we approach this customer, we're doing the two-way interview. They're interviewing our design capability, we are interviewing their front end design experience, and also the selling channel, et cetera, and also funding situation. I think that will be a very complicated process. By the end, we can say we do select few of the system house who are making their ASIC.
Right now, majority of our customer we select are still the ASSP, GPU, CPU maker, where we consider they have more experience than most of system house. I hope that answer your question.
Yeah. I think that's very helpful. I just want to get a sense, whether you notify any kind of dynamic in terms of competing landscape. Your industry peers, EDA tool companies, whether they also see the similar demand? Given such a big demand, whether or how you are going to retain your talents or compete with those talents?
Exactly. Mm-hmm.
Okay. Just a follow-up, I think follow-up the 5nm question from previous speaker.
I remember you have more than just one 5nm projects. How do you define those 5nm project in the pipeline? If we want to have a more aggressive assumption, for example, how many 5nm are you engaging? For 2022, how many 5nm project probably you can win? Any thought about revenue contribution from 5nm for 2022?
Okay. Let me try to answer this, and Daniel, you can comment if you have later. Right now, at this stage, we do have a multiple 5nm design won. To be straight, Charlie, right now, when we selecting the customer, design technology is not our number one priority. We're choosing the right technology, which has a potential volume. That will be our number one selecting criteria. Number one, for sure, all our existing customer will go to 5nm sooner or later. They are doing 7nm, they have tons of funding, they have to go for 5nm. Whether we show pushing them to go to 5 or not, I think it really depends on, as you know, 5nm in terms of ROI, in terms of PPA, and there's a big trade-off between five and seven and six.
Choosing the right project, potentially will produce the bigger volume, will be our number one priority. I can say, we do have a multiple 5nm design, but based on the revenue forecast next year, majority of contribution will be coming from 7nm or 6nm designs.
Okay, understood.
Yes.
One more thing. For the 5nm plan, it is kind of complicated because, as mentioned, currently our design resource is very precious. It really depends on the customer's technology. Usually, based on our experience, a start-up company, now I guess you know many start-up company has very strong ambitions to go directly to the most leading technology node. For such kind of customer, we may worry about the design turn long time will be prolonged or delayed because of some technology issue, because the customer is not that familiar with the technology. I think the 5nm plan, another dependency is how we view the potential customer, how we view their technology is ready. I think demand is never a question for now.
Okay. Yeah, that's all my question. I look forward to your next quarter results and guidance, and hope we can get a better clarity about the situation then. Thanks.
Thanks, Charlie.
Okay, Hans Li, please.
Hi. Okay. Is my voice clear enough right now?
Much better.
All right. Okay. I have a follow-up question with regard to the second largest China clients. Previously, you've mentioned that entering third quarter, some of their project will start to see some tapering. Moving into next year, how should we see the contributions as compared to this year from this client? Yeah.
Okay. For this customer, I mentioned in previous meetings that for this year, we have two projects in production.
The first one, the production period starting from last year's December until July this year. We may see the contribution from this project to taper off in third quarter. In the same time, we have second project, but the shipment volume is not comparable to the first one. We expect the contribution from this customer to decrease in the second half. For next year, as I mentioned, the good news is for the first project, what I said the first project, there will be another demand from our customer's customer. In the same time, we also receive other design project from this customer, and we are realizing the milestone right now. Those design will probably tape out in the second half or the first half next year.
This customer is a repeating customer, and their project will keep on contribute our sales for sure. Next year, the contribution, since there is another system, we have to ship a certain amount of the chip to fulfill our customer's needs. If we assume the supply shortage is not an issue, I don't expect the contribution will be lower compare the next year and this year.
All right. This is very clear. Thank you very much.
Okay, is there any question? Again, we welcome questions in Mandarin. If you have questions you want to ask in Mandarin, Johnny and I will be very happy to answer.
Hey, Daniel.
Yes.
[Non-English content]
[Non-English content]
[Non-English content]
[Non-English content]
[Non-English content]
[Non-English content]
[Non-English content]
[Non-English content]
[Non-English content]
[Foreign language] Is there any question to ask? Dai Yi Oh. Please.
Hello, Hi Johnny, can you hear me OK?
Yes.
Great. I just had, I guess a bigger picture or longer term question to ask. I saw an article about how Google had used AI to design a Tensor Processing Unit. I'm just kind of curious. I'll send the link to the article in the chat box. I'm just kind of curious whether this is something you had heard about, and whether you think this could represent a future disruptive threat to you or the IC design industry overall?
[Non-English content]
Wondering whether you had any thoughts on Google using AI to design AI chips more quickly and possible threat to you? I'll include the link to the article here.
I can make a quicker comment on this. I think that's a similar question I've been asked many times. To be honest, I don't worry about the AI doing the design, especially on the future technology. We are doing the AI chip inference or training. The reason that AI is so useful and so powerful because they have enough data. They even influence each other in order to find out the best solution. If you have enough data, yes, I believe AI among all the data, they can choose the best path. For leading edge technology, so-called five and three, they don't have any data, then how can they do any inference?
I think AI is very good to pick the right answer among all the data. First of all, you need to accumulate all the data. You know what I mean? If we focus on more mainstream technology, yes, eventually the design resource will be decreased. That already happened. It's not future, it's already happened. When we do the 28nm design, five years ago, we need 40 people. Now, when we're doing a 28nm design, maybe we only need 10 people. Why?
Machine, they taking care most of result. They can save a lot of unnecessary design effort. Leading-edge technology, when we're doing the 5nm, we need 40 people, need 50 people. It's very difficult to reduce the head count. Unless the design reference become big, data become big, and I can easily tell, 10 years later, when you're doing a 5nm design, you may only need five people.
Currently, if you stick with on the leading edge technology, AI is very difficult to replace the actual design resource.
Thank you very much.
Okay. Thank you.
Tina from HSBC, please.
Hi. Thank you very much. Can you hear me?
Yes.
Okay. I just have one question on your human resources. I think earlier you mentioned you're trying to increase headcount to about 550 by end of this year.
Past two years, I think the increase of the employee count, the headcount is actually quite small.
I think the year before, about 20 people a year or 25 people a year.
Last year, 426 all the way up to 550, that's like 120 people. Can I say you're doing this one just based on your project in hand, that's why you want to increase your capacity by so much?
Let me answer you first. Actually, we have done aggressive hiring for a long time. In China, since our operation is mainly in China, in China, all the engineering, actually, every company is fighting for the engineering. To increase the design resource, honestly, is very challenging right now. The good thing is, we have advantage in recruiting fresh graduate students from campus because they know if they work in Alchip, their value will increase very fast. That's the reason why Johnny said, we are shooting for 540, 550 by the end of this year. That's because there is another wave of new, fresh engineers is going to coming in our company. For the experienced engineers, actually, the salary is hiking very fast, and almost everywhere is fighting for hiring those engineers.
We have tried our best, but the situation, the environment is very challenging. Johnny, you want to add some more?
Okay. Yeah. Based on the history, our headcount increase is not so drastically. That's a reality. For this year, we make a more aggressively offer. Just like Daniel mentioned before, is to new college grad, and also we established a subsidiary in the more remote area. I think those area competition is much easier compared to Shanghai and major cities. Hopefully, hiring is easy. To keep the employee, reduce the turnover, I think that's some area we need to pay more attention. Yeah, the 550 is still my goal for this year.
When you say low cost, you mean China or China, India as well, or Eastern other country you have in mind?
Yes. At this moment, because for the sake of training and China, because we have a majority of our engineer in China, expanding in China will be most efficient. To be straight, yes, we are also thinking about other country, like India or others.
Okay. I think you don't really want to talk too much about the DOJ stuff or the commerce stuff. Can I say, is there any implication about the segregation of the Chinese employee or engineer resources and the non-Chinese engineer resources when coming to the design required by the U.S. authorities, that's why you need to increase the headcount?
I would say, when we do the hiring, we are mainly pay majority attention to our resource availability. For the geopolitical issue, of course, we have planned to, I would say, separate or increase some resource overseas. We don't feel direct pressure from authorities to ask us doing so.
Got you. My last question is regarding the leading nodes on the seven, six, and five. Can you talk about what's the proportion of the CoWoS design now your customer ask you to implement, and how this going to impact your kind of margin going forward?
Actually.
The design resources is to the operating expense item. For the gross margin, I don't think it has strong relations between human resource and the gross margin.
Oh, I'm talking about CoWoS, the package design CoWoS.
CoWoS.
Yeah. What's your proportion now being incorporated into the project for the seven and those six and the five, and what will be the impact on your margins for those projects coming into the volume productions?
Let me comment on the number of CoWoS design. In China or in the U.S., the major CPU maker or the AI training maker, I think they have to use the CoWoS. In terms of a percentage, I don't have a number on top of my head. We have many CoWoS designs. In fact, people consider using 5nm or most of the seven. I think CoWoS occupy at least 50%. In terms of margin, CoWoS capacity is another challenge. Other than the wafer capacity, CoWoS capacity from the major supplier, I think, is even tighter. Margin-wise, I don't think there's a big impact based on our current quotation. I don't think the CoWoS design will produce less margin. It's not the case.
Okay. Yeah. Thank you. That's all from me. Yeah.
All right. Thank you.
Okay. Harry Anderson, please.
Hey. Yeah, thanks for taking my question. Two, hopefully, quite quick questions. I understand this is a sensitive topic, but I just wanted a little bit of clarity around, is it all 5 nm products that you expect to not be subject to EAR, or is it only certain areas, such as maybe PC rather than server?
I would say for now, the results we received from the U.S. lawyer, the result for all the projects are positive. For this case, we cannot do any expectation because it's not controlled by us.
That's what we can say at the current stage.
Okay. That's helpful. Thank you. My second question is just around maybe the longer term, your aspirations, your targets for growth. Do you have any kind of targets or guidance for growth in 2023 or maybe a three to five year growth estimate? Thanks.
First of all, we are not releasing a guidance. I would just say, in a normal year, usually, what the company ask our sales, our engineer, our usual, normal year, we want at least the 20% growth rate. In the next two to three years, we still believe the whole ASIC market and the backend design market is in a very good shape. Usually, we will adjust our sales target according to the industry status. For 2023, 2024, I'm sorry that I don't have the numbers of sales growth for you.
No, that's fine. Thank you.
Hans Li, your raise hand button is still there. Are you having questions?
No, I will just lower down my hand. Thanks.
The same, Tina. Tina Chen, do you have questions? Your raise hand button is there. Okay. Is there any questions? We take one more question because we are running out of time. Last question, or if you don't have a question, we will close today's conference meeting. Okay, I guess that's all for Tina.
Hi. Sorry. I just take the last opportunities.
Yeah. Sorry. I'm so sorry.
That's fine. Yeah. I just want to check with you because TSMC's CoWoS, they actually have the different type of CoWoS. I remember they got the CoWoS-S and CoWoS-R, maybe some other CoWoS.
You'll be able to provide all type of the CoWoS now, given you not only engage in HPC, but also the AI inference or this type of chips?
Good question. Yes, we can provide all kind of service for the TSMC available, the CoWoS solution. Yes.
Okay.
Do you have any test chip or any design now is within the three categories?
Yes, we do. We have one in seven, and we have one in five. Seven is already out there, and five will be coming out very soon. Within a month. Yes.
Is this within the three U.S. chips you're talking about or in addition?
No, this is our own test chip.
Okay. Your own test chip.
Okay. Thank you very much.
Okay. Thank you.
I guess that's it. Thank you all for joining us today.
Have a good weekend. Thank you.
All right. Thank you very much. Thank you for supporting us. Thank you.
Thank you.
Good weekend.