Alchip Technologies Ltd

TWSE:3661

Earnings Call Transcript 2020-Q3

[Daniel Wang, executive]

Okay. Dear analysts, portfolio manager and investors, welcome to our Alchip Technologies Third Quarter '20 Institutional Investor Conference Meeting.

This meeting we'll conduct in English. And in the meantime, we welcome Chinese when we do the Q&A session. And this meeting we'll divide into 2 part. First of all, we will still do a short introduction of Alchip Technologies and the second quarter results -- no, the third quarter results and our outlook and operational update for fourth quarter this year and next year.

And after that, we will do a session regarding to our recent issue -- not issue, our recent things decided by our Board meeting, this Board meeting, the GDR offering and our investments. After that, we will do the Q&A session. [Operator Instructions]

Okay. So now we would like to have our CEO, Johnny Shen, to start the meeting.

[Johnny Shen, executive]

Okay. Thanks, everybody, for coming. Yes, this is Johnny Shen, the CEO and President and Co-Founder of Alchip Technologies. Yes. Daniel, please go to Page 3.

Yes. In case you are new to our portfolio, yes, let me make a very quick company identification and also update again.

Page 3. Okay. So our company is founded in 2003, IPO in 2014. Our current headcount is over 480 people. Yes, within 480 people, more than 3/4 are engineer. So by the end of this year, we will be over 500. Yes, company is still aggressively increase our headcount and also design resource now.

Since we have found the company, we've been successfully tape out more than 420 designs, all in leading-edge technology. Our last year revenue is USD 140 million. Yes, obviously, we're doing much better this year. Even the first 3 quarter, our revenue number is over $170 million. Yes, we are one of the TSMC VCA member, so-called value chain aggregator, will be officially help TSMC to distribute their technology, yes, even for the most leading-edge technology like 16-, 7- or even 5-nanometer.

Next page, please. Alchip is doing the ASIC business since we found the company. Yes, honestly, a few years ago, not too many people understand or buy in the ASIC market story. Now ASIC become very popular. Actually, the market size is way beyond what we predict before. ASIC market is growing. Yes, demand is much more than supplier and only few limited player has a track record and experience to do the leading-edge technology ASIC design. Within the ASIC market, we are focused HPC and AI area. Yes, we consider these 2 area has a higher entry barrier. And since the design are huge and complicated, always require the most leading-edge technology. Yes, for this year, HPC and AI occupied, yes, more than 70% of our total revenue, and we truly believe HPC and AI application will continuously grow and sustain with a reasonable [ profit ].

Okay. The following -- the later few slide, I would ask Daniel to share more detail with everybody.

[Daniel Wang, executive]

Okay. This is the result of our third quarter '20 quarterly income statement. It is pretty straightforward. Our revenue is almost record high at $67.7 million versus -- for the third quarter, which is 23% quarter-on-quarter growth and 132% year-on-year growth.

For the operating income, because the third quarter we have -- the major growth driver is the mass production revenue, so the gross margin rate, the gross margin is relatively lower to the second quarter. So the operating income is $9.4 million for third quarter. It's 8.6% quarter-on-quarter growth and 174% year-on-year growth.

For the net income, is $7.9 million for the third quarter. The quarter-on-quarter growth is 22.7% and the year-on-year growth is 111%, translating into EPS, our single quarter EPS for the third quarter '20 is TWD 3.78.

So the third quarter revenue can actually -- are higher than we previously expected. The main reason is the CPU shipment to our China customers is stronger than we expected, especially for the biggest customer, both the server and the notebook use, the CPU shipment is quite strong in the third quarter. And for the online, which means the design service part, all the milestone came as expected. So we have multiple tape-outs in third quarter. So for third quarter -- for third quarter's total revenue as a whole, I would say everything is pretty good from the company's standpoint of view.

For the operating income, this quarter, operating income is $9.4 million, and the gross margin is 31.5%, which is a decline on a quarter-on-quarter basis, as I mentioned, because we have higher mix of production revenue in third quarter. The percentage for the production revenue in third quarter is about 60%. And of course, HPC-related obligation remains our major contributors to our third quarter, both top and bottom line, and our operating expense, if you look at the numbers, which is grow smoothly because we are continuously develop our -- expand our people, our design resource and our business.

And for the revenue breakdown, as usual, I'll start from the application. For the quarterly application, you can see the HPC now account for 73% of our total revenue in third quarter and the networking account for 10%, and the consumer now only account for 4%. And the other 12% is contributed by niche market.

For the -- sorry. I guess I'm sharing the wrong file. I'll upload the file again. I'm sorry about it. There might be some issue.

[Johnny Shen, executive]

That's okay, Daniel. We can just use this, not a problem.

[Daniel Wang, executive]

Okay. Let me go back to -- okay. I guess the -- on the right-hand side should be the yearly breakdown. It's the wrong file, I will try to fix it while we finish the institutional investment meeting. All in all, the high-performance computing accounts for the highest percentage of our total revenue, no matter for the quarter or for the year. And for this stage, it's revenue breakdown by process nodes. And you may see -- again, I'll fix the mistake here. There's no numbers on the 12- and the 7-nanometer. But for the third quarter alone, you may see the 7-nanometer and 16-nanometer, these 2 account for about 80% of our total revenue. And you may see the 16-nanometer contribution grow from 29% from second quarter this year to 41% third quarter this year. That's mainly due to the CPU shipment to our biggest China customer.

You can also see for the year, the 2020, this year, the advanced technology node already accounts for 92% of our total revenue. And I think from the process node breakdown, Alchip is definitely the leader among the back-end design industry.

And for the revenue breakdown by region. For the third quarter, again, because of the strong shipment for notebook and server-used CPU, the China contribution account for about 70% of our total sales. And the second largest is Japan, accounts for 13%, while Taiwan accounts for 6% and the other region total combined accounts for 11% of our total revenue.

For the yearly, it's pretty straightforward. For this year until the third quarter, China -- revenue from China accounts for 62% of our total revenue, while Japan accounts for 16%, and Taiwan accounts for 5%. And then the other region, which is mainly the North America and Europe, accounts for 17% of our total revenue.

And this page is about the industry and the company outlook. I would like to invite Johnny to do...

[Johnny Shen, executive]

Yes, let me continue. So China HPC IC demand, we can say remain very, very strong. Yes, demand from major China customer keep increasing their orders, not only for [ mass production ], but also for NRE. Yes, a lot of -- all our existing customer has the next generation. But in additional to that, we do have a lot of new customer, yes, from U.S., from Europe and also many start-up company in China who are doing the leading -- who are using the most leading-edge technology like 7-nanometer.

China CPU shipment. Yes, for both 16-nanometer and 7-nanometer, still expected to be the major driver for the next quarter and also for part of next year. So next year, in additional to this CPU design, we expect to see some AI chip shipment from China and also from U.S.

All right. [ EDA ] right now is -- we can say is strong than ever. Yes, our design capacity is more than 100% now Yes, in fact, it's 120% or more. So to capacity become our biggest business dependency. One is our design capacity. The other is TSMC 7-nanometer wafer capacity. AI chip sequentially tape-out Q4 and also first half of next year, we have a lot of design win recently. So based on current schedule, most of our design will be taped out starting from Q4 this year. So in the other word, i.e., for -- starting from Q4 and Q1, Q2 next year, will be -- in terms of the total NRE amount, I think we expect grow quite a bit compared to this year.

So in addition to CPU project, our AI tape-out in North America and Europe, and some of them already has a mass production plan. So another significant win, we have another 5-nanometer design won in Israel area. I think the contract already signed. They also have a very aggressive design and mass production schedule. Yes. Next page, please.

[Daniel Wang, executive]

Okay. I guess, many of the investors were very interesting in our just announced GDR offering. And I would like to tell everybody what's the company's thinking about the reason why we want to raise money through GDR offering. We plan to raise about USD 100 million by 10%-plus issuance of the new shares. Yes, again, the purpose because this involves in the company's strategy, I'd like Johnny to explain this part to the investors.

[Johnny Shen, executive]

Okay. Yes, last Board meeting, on Friday, we approved the GDR and 2 other investments. Yes, let me share more detail and purpose of the GDR fund raising.

Yes, first of all, we will press the trigger only when we can see the share price is within a reasonable range. Yes, currently, based on how we all think we are under value, but GDR entire procedure takes some time. Board meeting just approved. We're going to have a shareholder meeting. And later on, we are going to get the agreement from Taiwan authority. So the actual time frame, I think, will be Q1 or even Q2 time frame. So we hope during that time, we will have a more reasonable share price in order for us to press the trigger.

Another purpose is to attract long-term and more stable foreign investors. The banker we are working with has good confidence to attract few ideal target investors. So this GDR offering will be very quick and specific with little discount. And we have a -- they have confidence to attract a few of them.

And another purpose to raise the funding is to prepare strong mass production growth in the near future. Yes, as we can say, this year, our mass production closing rate is a lot. To be honestly, next year, the momentum is even stronger. So we need to prepare enough cash, yes, in order to take capacity. Yes, our debt capacity is very limited, not only from TSMC's standpoint of view, but also for the substrate. So you need to show, you need to prepay, you need to put a -- you need to prepay in order to get the capacity. You need to have more realistic plan with the prepay. So this year, we encountered some form of [indiscernible] customer. We are waiting for customers some money before place wafer order to TSMC. And starting from next year, we have a -- we are working with TSMC more closely, has a more realistic plan once we can do the prepay.

And also, right now, as I mentioned before, the supplier is much, much shorter than demand. Yes, we need to aggressively expand our design resource and solution. Our design resource need to increase aggressively, yes, in order to keep our winning position And some customer, especially the new start-up company and some customers in U.S., not only need the design resource but also need some solution, like a front end, like an IP. Like I mentioned before, I don't intend to do this in-house, but we can do invest, working with suitable target partner, yes, in order to complete our solutions.

The other purpose is to secure and forming a strategic partnership with the customer and supplier. Yes, nowadays, the relationship between customer and supplier is very dynamic, not only for a pure business standpoint of view, but a strategic partnership is very important. They consider you are the good service provider. Sometimes they also consider you are their investor or partners. So with a little investment, can make our relationship with the supplier even tighter. So that's another purpose for this GDR offering.

[Daniel Wang, executive]

Okay. As Johnny mentioned, to solidify our relationship with customer and the suppliers and to explore more business opportunity, I will take our investment cases for an example.

First of all, ASICLand. We are intended to invest ASICLand. For this company, this company is TSMC's VCA member in Korea. Actually, currently, there are 2 VCA members in Korea doing the back-end design service business, one is ASICLand and the one is Global Unichip. And we had done business with Korean customer before, but the Korean market is very special. The local IC makers or system providers intended to work with the local suppliers. But for the leading-edge technology node, there is no local suppliers, no local back-end design suppliers who can provide such kind of service.

And through this investment, it makes Alchip to tap into the Korea HP market through this strategic alliance with the local sales channel and the engineers of ASICLand. And we expect this joint force will win leading-edge technology node projects from very notable customers. Maybe most of you already understand that there are business opportunities in Korean market from not many, a few very note -- very sound NAND customers, such as Samsung, such SK hynix. So I do -- by doing this investment to form the alliance with ASICLand, we are in high hope to win the business from those customers.

The second thing is the KQ [ Kun Qiao ] Semiconductor technology industrial fund. For leasing investment, the main purpose is there are many notable and sound names of the semicon and the IT companies are the general partner of this fund. And of course, for the limited partner, for the LP, there are also many strong IC makers and mega VCs on the list. Through this investment, the amount of the money is not big, but we can get closely relationship with those people and those companies. And to build out this strategic alliance with these professionals, we may explore more business opportunity from our customers, especially in China.

And as Johnny mentioned, for now, the market is quite dynamic. With lease strategic alliance with -- by building out the capability of doing the investment and other things, we think for Alchip's future business development in China is quite -- for doing this is quite important. So those are the examples for our capital-raising purpose.

And then I guess this is the presentation session for today, and thank you for joining us.

And now we are open to question. [Operator Instructions] And one more thing, we are very welcome the questions in Chinese. So don't be afraid to ask questions in Chinese. We can do both English and Chinese without a problem.

Thank you.

[Daniel Wang, executive]

Please, yes, is there any question? Okay. Charlie from Morgan Stanley.

[Charlie Chan, analyst]

So I guess 2 question from me. First of all, is about your next 2 to 3 years outlook. It seems like your capacity is now very constrained, if not considering the potential capacity increase from your future strategic investments. So with that, how big you can grow your business in the coming 2 to 3 years? Can you -- or at least can you give us some revenue guidance or targets for 2021?

[Daniel Wang, executive]

Okay. Johnny, do you want to answer this question?

[Johnny Shen, executive]

Let me take it first. You can add on later.

[Daniel Wang, executive]

Okay. Sure. Okay.

[Johnny Shen, executive]

So Charlie, to answer your question, yes, first of all, I have confidence to meet or beat the major analysts' forecast, like you did. Yes, the reason we need to -- yes, we need to increase the capacity because we see in the demand is much more than we can digest. Yes, we have a chance to grow even bigger.

So after consideration and all the current situation because before we are only thinking about existing customers' product and their next generation. But right now, so many opportunity. Yes, it's much more than we can digest. And all of them are -- we can -- all of them are very sexy. So aggressively increased our capacity, I think, is one of our main goal. So the GDR offering will help and also working with a partner or even more aggressively working with potential partners, I think, is absolutely we should do.

To answer your question, we are quite in sync with your -- in terms of revenue, we have a confidence and quite in sync with your our forecast, and we have a competency to meet or beat.

[Charlie Chan, analyst]

Okay. And then -- and my next question is about the recent markets' concern about more competition for the China CPU market. I know you have very strong partnership with FEITIAN and their company seems to gain a lot of traction in the China local markets.

But if you look at not just next year, right, but for 2022, I guess, 2 questions. First of all, whether the China CPU addressable market can continue to expand. Meaning, today, it's only for the government PC or is it SOE PC? Do you think 2 or 3 years later, you can penetrate to the consumer markets in China?

And secondly, there seems to be more projects going on. For example, VIA Technologies sold this -- some IP to the Shanghai Zhaoxin. And also a Taiwanese company called [indiscernible] claim they can gain a lot of China CPU projects. So I wanted to know your observation about the future market size and also the competition. It would be very helpful.

[Johnny Shen, executive]

Okay. Again, Daniel, let me go first. You can add on later.

[Daniel Wang, executive]

Yes.

[Johnny Shen, executive]

So in terms of China CPU market, yes, based on our customer and also the market analysis, the unrelated CPU is still the major player before Huawei, HiSilicon and also our customer, I think, occupy about around 70% of the CPU market. Recently, as we know, the Huawei we can say is out of this market. So that's why our customers' demand getting increased quite a bit. Yes. But China also sense if they put down -- if they put their future, their CPU demand all on, on base, they will bring certain risks. So a lot of -- so other opportunity increase and start to kick off. But based on our customers, as mentioned, they can -- they have confidence to win most of Huawei, which is their competitors in the market. So next year growth, I think, be our case, I think is another opportunity. I think other than on other CPU opportunity also will increase, yes, as expected. Daniel, do you want to add?

[Daniel Wang, executive]

First of all, Charlie, to answer your question about the outlook for the fourth quarter and next year. For the fourth quarter this year, which is the next quarter, we believe we can still post quarter-on-quarter growth. But since the base, the third quarter revenue is quite high, higher than we previously expected, the fourth quarter quarter-on-quarter growth will likely be single digit. High single digit or low single digit depends on, first of all, the milestone kicking in and the, of course, the capacity support. And I would say this way, the fourth quarter will still be a good quarter for us.

For next year, I know that the market consensus about our top line is about $300 million to $320 million. Based on our current understanding, we are pretty confident we can beat this forecast, this expectation for sure. But for the bottom line, I would say the operating income for the next year, we are still doing the budgeting. And yes. But all in all, of course, you understand our business model and our financial structure. The revenue increase for sound revenue, for example, production revenue, which is [ effortless ] revenue, the profit was to go down directly to the operating income. So I would say our operating income growth rate was even better than the revenue growth rate.

And for the CPU market, I will say this way. For the self-control and the self-made market, CPU market, it will still be dominant by FEITIAN because for the x86, again, the issue remains x86. This architecture function is still controlled by a U.S. company. And for those companies doing the x86-based CPU, I will say it can penetrate into some government market, but for -- I would say this way, the low level, like the county government, like the village government, they may not have the ability to acquire enough the ARM-based PC. So I would say this could be the market for the x86-based machines.

But for the normal or ordinary individual market, I would say, for example, you and me, we won't buy Zhaoxin CPU inside machines. Since we can buy Intel Inside or AMD inside. I would say for this market, it was still dominant by the U.S. players. But for the future, yes, I think the trend is good for the local suppliers. But for the government, which means the self-controlled market, I guess, FEITIAN will still be the #1 supplier.

And another question from Ng, Szeho.

[Szeho Ng, analyst]

Daniel, it's Szeho from China Renaissance. And my question would be regarding for next year, the company will be facing some bottlenecks back into [indiscernible]. One is the engineering resources and the other would be the wafer capacity support. So how are we going to deal with these challenges next year?

[Daniel Wang, executive]

Johnny?

[Johnny Shen, executive]

Okay. Yes, capacity. First of all, there's, yes, many ways to increase our capacity. Of course, we continuously try to improve our design efficiency. Yes, trying to use less people to support more projects. But in the meantime, we're doing a lot of aggressive hiring, and we are working with partners. Sometimes we're doing some in-source, inviting some customer, yes, inviting some partners engineer, working some projects in our facility. Yes, we can use that to handle the search business. So gradually increase our headcount, improve the design efficiency and working with a partner. I think that's the way handle the resource.

Yes, in terms of capacity, we need to aggressively, very closely working with TSMC, better understand our customers' outlook and forecast. And also, the most important thing is when TSMC provide the capacity, we need to take immediately. So that's why GDR also helped a lot. Once they have a capacity, we have enough funding to grab, yes, immediately.

[Szeho Ng, analyst]

Okay. Good. Also, regarding the GDR proceed you get, when are you going to put it into the investment of ASICLand?

[Johnny Shen, executive]

Yes. Daniel, do you want to answer?

[Daniel Wang, executive]

Sure. Okay. Szeho, money is money. Money doesn't have [ names ]. So, of course, we have -- currently, we have cash, and we have certain money in hand. And after the GDR offering, we will have more money. And for the investment amount, putting ASICLand is not significant. So there will be no issue for this investment.

[Szeho Ng, analyst]

Actually, got you. Yes. And last question from my side, yes, before I jump into the queue again, will be the engineering headcount you are growing to grow next year and any update?

[Johnny Shen, executive]

Okay. By the end of this year, for sure, 500. And next year, based on our hiring plan, I think we try to meet -- current target is 700.

[Szeho Ng, analyst]

Okay. And the majority will be from China, right?

[Johnny Shen, executive]

Majority is still in China region. Yes, you are right.

[Daniel Wang, executive]

Okay. And the next one is [ Yvonnne Tsai ]. I guess she's from UBS.

[Unknown Analyst]

Can you hear me?

[Daniel Wang, executive]

Yes. Yes, please.

[Unknown Analyst]

I just want to -- because I think Johnny just mentioned that HiSilicon used to have 70% market share in -- is it local -- China local CPU market? Is it like for public service market? Or does that include commercial market? And are you, say, like have confidence to -- I mean, for fighting to take over HiSilicon's share in this market?

[Johnny Shen, executive]

Yes. Yes, let me correct a bit. Now HiSilicon is 70%, combining FEITIAN and HiSilicon have 70% of the market. Yes, of course, the distribution between FEITIAN and HiSilicon is a majority.

[Unknown Analyst]

Okay. And then you're quite confident that FEITIAN can get this -- like the 70% market share?

[Johnny Shen, executive]

Right. Right. Yes. Yes, based on FEITIAN's justification, the market already there, what -- they are lacking of the supplier. So it's by natural, it's very easy for them to take over this market. The only dependency is the software. But right now, I think their software is -- can -- both on software are Linux-based. They can be, yes, adopt like very easy. So they have a confidence to grab this market.

[Unknown Analyst]

Okay. Okay. And this, just to clarify, so this is the like public service, like government and SOE market, right?

[Johnny Shen, executive]

Government -- or even commercial, they have some. On the laptop and notebook business, they also put the very high expectation because the ARM-based laptop can work with mobile device very closely. As you know, the -- all the cellphone are made by ARM-based CPU. So the handshaking between ARM-based notebook and ARM-based cellphone, it will be very good, very smooth. Yes, that's why you can see even Apple start to make ARM-based Mac computer. So you can see a lot of mobile application can rounding in the laptop. Yes, even right now, FEITIAN's notebook -- cellphone application, including [ game ], including ByteDance, a lot of application. I personally, I try, it's quite amazing, yes. I expect this market will be very good.

For the Intel base, they cannot run. They cannot translate the mobile application. Yes, but ARM-based, I think, there's no problem for them to do this.

[Daniel Wang, executive]

Okay. Next one is Chi Ho Wong.

[Chi Wong, analyst]

This is Chi Ho from Dymon Asia. Congratulations on the good quarters. Can I ask a question about if you look at the -- maybe give us some overall statistic about this 5-nano, 7-nano tape-out -- number of tape-outs you have done in 2020 and maybe some of the pipeline you can see for 2021 in terms of the more advanced NRE project?

And then the second question related to that is the -- maybe for the tape-outs that you have done this year, what kind of mass production volume can we be expecting in 2021 for these newly tape-out projects. Would you have any color for like a contribution in 2021? Because I feel that from your opening remark, there should be quite a lot of volume coming from the newly tape-out projects this year. But I guess it's also skewed to the FEITIAN. So maybe I think the FEITIAN situation is well understood. Maybe talk a little bit about outside the FEITIAN situation.

[Johnny Shen, executive]

Okay. Daniel, want to go ahead first?

[Daniel Wang, executive]

Okay. Sure. Chi Ho, for the latter, we are not discuss the project by project into detail because our customer has some issue with the confidentiality. So we no longer provide the exact tape-out schedule for our projects. The thing I can say about our 7-nanometer and 5-nanometer project, I would say this way. And for now, we haven't taped out any 5-nanometer projects. The tape-out schedule for the current project we won, which are 2 projects, one is for test chip, one is for full mask, the tape-out schedule will be, I would say, the end of last year or the early year 2022 because the design is quite complicated.

But for the 7-nanometer project, actually, we tape out many 7-nanometer project this year, and you may see it from our contribution since the 7-nanometer production revenue until now is not that high because all of them are just the tape-out or just into the mass production phase.

For the next year or fourth quarter, we will sequentially tape out 7-nanometer projects, both in artificial intelligence, your AI chip, and the HPC. For HPC, we have several projects from our biggest customer, FEITIAN. And the second largest customer, and for the second largest customer, the first 7-nanometer project enter the production phase now. So you may see the contribution in -- starting from fourth quarter and very good contribution in the first half next year.

For the AI chips, I guess the tape-out schedule will still in the late this quarter or the first half next year. So we expect some of them, the production revenue will start from the second half next year. And we also have high expectations on some AI chip projects because of the customer is very aggressively to promote to try to sell their AI chips onto their system or onto the general purpose AI market. So that's what we can say about the tape-out schedule and the tape-out situation right now.

Okay. And next one is Jeffrey, Jeffrey Ohlweiler. Yes. Jeffrey, we cannot hear you.

[Johnny Shen, executive]

So in that case, with that, next to ask the question first.

[Daniel Wang, executive]

Okay. Charlie, please.

[Charlie Chan, analyst]

Okay. Yes, maybe I try to fill the void. And sorry for the belated congratulations to your company. Very good execution So my follow-up question is about your GDR. One is more kind of concern because you also mentioned that there could be some strategic investment or partnership with customers. But are those, no matter for foundry or design service business model, you want to be as independent as you can, right? So how do you deal with this kind of a deeper relationship with customers, but they could trigger some conflicts of interest with other customers? I guess there's a lot of question. And second question is easier. Your timing for the GDR.

[Johnny Shen, executive]

Okay. I think, Charlie, that's a very good question. We've been discussed that internally many times. Yes, obviously, we will avoid any conflict of interest. Yes, to be straight, our investment amount will be very insignificant compared to our customer and also suppliers situation. Right now, China, as you know, so many people has -- the customer fundraising is not the issue. They can -- a new start-up company can do the fundraising, I think, immediately as long as they are willing to issue their share. The reason we're working with the customer and supplier, the reason we need to have -- we need to involve for this kind of investment because people are building some alliance. Yes. And once they see you have a solution, you have a power to make an investment. I think our winning chance will increase quite a bit.

Just for your information, one typical example. When we compete one of the cases and once they know, they sense we have a GDR opportunity, they immediately give us the opportunity. So this kind of a situation, I think, is quite common in China. I think to be specific, I think I can share more specific example. Yes, of course, for sure, we will avoid, our company scope still remain independent. Our independent position is the most important and the highest asset for the company. We won't compete with our customer. Customer can do the chip -- we say Alchip to compete each other, but Alchip will not be their competitor. Daniel, do you want to add some?

[Daniel Wang, executive]

Okay. Charlie, for sure. First of all, I would say, as Johnny mentioned, the investment amount will be not significant compared to our customers' market status. And secondly, the major principle is business as usual. And I would say this way. Our investment could be a signal to comfort our customer because, as you know, currently, the back end -- the back-end resource is kind of a treasure in the market. Our customer also worry about they cannot secure the design resource from us. So I would say it's a mutual understanding, is doesn't involving any like we choose the customer, we stand on some certain customer's side. And that's our main purpose.

[Charlie Chan, analyst]

Okay. Got it.

[Johnny Shen, executive]

To answer your second question, based on our estimation, the fastest way to execute the QDR without any problem, I think you need approximately 60-day minimum. Yes. So I think, like I mentioned before, I'll wait for the right timing when we are -- when our share value return to certain reasonable range to press the trigger. So right now, the fastest time frame is 60 day later after the shareholder meeting, assuming everything is go smoothly. So I will say Q1 or maybe Q2 time frame.

[Charlie Chan, analyst]

Okay. Okay. Fair enough. Yes. So because I think over the weekends, I think some investors are also asking why not just use the credit line because now interest rates are very low. And maybe if you want to introduce some strategic investor, you can just use the private placement. So yes, I mean, maybe you don't need to foresee it, right? If the market is not right, maybe you can just use other vehicles. Yes. Because I kind of agree with your kind of -- most part of the strategic investment. But GDR sounds a little bit complex here.

But anyway, so just a small question, and I will get back to the queue. Now the company's revenue profile seems to very skewed to China CPU customers. So another big growth potential segment should be AI, right? So I think other analyst also asked about this topic. But can you quantify this year how big is the revenue contribution from AI in terms of percentage? And how big that would be for 2021?

[Johnny Shen, executive]

Okay. Yes, Daniel, why don't you go first?

[Daniel Wang, executive]

Sure. For the AI contribution, I would say, roughly because it's for some project, you are quite difficult to identify if it is on AI or it is on HPC because HPC, some HPC chips can also do the AI thing. So lastly, I would say the current contribution from AI-related chips accounts for about 15% of our total revenue. And for last year, and of course, I will say this way, both our customer and Alchip believe for the 7-nanometer AI chip, the shipment volume will grow significantly compared to the 16-nanometer ones because we all understand that for the AI topic, the -- this topic has been hovering around very hard for quite a while. Like past 1 to 2 years, everybody is talking about AI. But except NVIDIA, there is no major AI chip suppliers on the market or the shipment model is not significant.

I think that's because for the AI users, when they use the AI chip to build out their system, they have to be -- they have to make sure that the system is stable, the system is work, the system is efficient. So they have done a lot of testing. So for the 7-nanometer, which I think for -- okay, for the 16-nanometer, I think, is mainly for the testing purpose. For the 7-nanometer, we expect the volume will be quite good compared to the 16-nanometer. Otherwise, maybe AI is not -- will not be a hot topic anymore.

[Charlie Chan, analyst]

Okay. So in terms of revenue percentage next year? Could it be more than 15% or...

[Daniel Wang, executive]

Definitely. Definitely, yes.

[Johnny Shen, executive]

Yes, the growth rate is quite promising. But I don't know. Maybe the CPU also grow very fast. But in other -- so other than China, we expect AI shipment in U.S. will increase significantly. Yes, 1 of our 2 major customer, I think they put a strong forecast. Yes. And as you know, one of them just acquired by another giant company. And the 7-nanometer is in mass production right now. So we expect good shipment on that specific chip.

[Daniel Wang, executive]

Okay. Jeffrey, can you hear me?

[Jeffrey Ohlweiler, analyst]

Yes, I can hear you. I changed my microphone. I hope this helps. Is it better?

[Johnny Shen, executive]

Much better.

[Daniel Wang, executive]

Yes. Yes, I can hear you clearly.

[Jeffrey Ohlweiler, analyst]

Okay. Great. So my question is you talk about potentially prepayments to foundry and maybe substrate. I guess, can your customers actually do the prepayments? Or why does that fall upon you guys to have to do those prepayments for capacity?

[Johnny Shen, executive]

Of course, they can prepay, but there's a gap. For example, TSMC recently just have, for example, 500, 1,000 pieces of wafer. They asked us, do you want to take? If we want to take, we need to place order immediately. But in the meantime, of course, the customer is willing to prepay, but there will be some gap before place the order to supplier and receiving money from customer. So because of that, we are losing some capacity. And looking forward for the next year, we see a lot of opportunity, especially in 7-nanometer capacity fight between all customer, I think, will continuously sustain. So we need to prepare ourselves to receive the order from...

[Jeffrey Ohlweiler, analyst]

Okay. Great. One more question...

[Johnny Shen, executive]

Please.

[Jeffrey Ohlweiler, analyst]

Yes. Just on 5-nanometer mask, I guess, you also include 7 there, but what's the approximate cost of a 5-nanometer set of mask? And when do the customers pay versus when do you have to pay for those masks?

[Johnny Shen, executive]

Based on our business model, customer will pay for the mask. Yes. Yes. Regarding the price, I think we are not -- we are on the NDA, we cannot share the -- any specific pricing information. But based on our business model, customer will do the -- will pay the mask.

[Daniel Wang, executive]

Thank you. Next one is -- let's know -- the name on the system is Zoom user. Can you state your name and the company you are working with?

[Unknown Analyst]

Management, could you hear me?

[Daniel Wang, executive]

Yes.

[Unknown Analyst]

This is [ Hans Li Tong ] from Public Mutual. So firstly, regarding to the GDR. Has Alchip identified any targeted investors? And if the answer is yes, could you share like which country these investors are coming from?

[Daniel Wang, executive]

Okay. Like I said, we do identify some potential investment target. But for now, we are not allowed to reveal it. And the example is other than China, of course, we do have some potential investment target. For example, ASICLand, which is one of the investment target. But for the other targets, since all the investment case need to be -- need to get permit by the Board, so, so far, we have candidate, but we don't have -- we cannot reveal the name.

[Unknown Analyst]

All right. Then my second question is that I think in the beginning of the presentations, you have mentioned that probably might be a good opportunity to tap from the Korea market. Has the company have targeted any revenue percentage of contribution from the country in the next maybe 2 to 3 years?

[Daniel Wang, executive]

I will say this way, and Johnny can add some colors on it. The thing is we do the strategic alliance through investment in the technology corporation. So for the design -- for the design portion of partner, which is ASICLand, will take the design and we share the profit through the investment. But for our -- but for some projects, like the 7-nanometer with huge shipment volume, our partner may not have the capability to handle it. So for this portion, those production business will go directly to Alchip.

[Unknown Analyst]

All right. And the foundry -- sorry, go ahead.

[Johnny Shen, executive]

Go ahead. Go ahead.

[Daniel Wang, executive]

And of course, for the foundry part, we will still be in the TSMC.

[Unknown Analyst]

All right.

[Johnny Shen, executive]

Okay. So in terms of revenue expectation, I think Korea is very specific. Unlike other region, they have [indiscernible] customer are very big. Once you win, we expect the very significant revenue growth. So I think after consideration, Korea is definitely a place we shouldn't give up. I think the -- but if we want to win the Korea business without the Korean people, I think, is very difficult. So I think working with ASICLand, the size is quite match. They have a design dependency on us. We have some business channeling dependency on them. I think cooperation between these 2 company, I think, it's quite helpful.

[Daniel Wang, executive]

Okay. Prior to the next one, Szeho, I would like to answer a question from message. An investor [ field ] is asking about the schedule of FEITIAN's project. As I mentioned, I cannot review the details of the single project. But what I can say is for -- currently, we are shipping the 16-nanometer products to FEITIAN for server and for PC. And for the next generation, for the server, we are in design phase. For the notebook CPU, the design is -- hasn't kicked off yet. But the potential by the scheduled kickoff date should be around the first half next year.

Okay. Then we go back to Szeho, please.

[Szeho Ng, analyst]

Yes. My follow-up question would be on the company's business model going forward. Would the company consider or be more open for those design-like business or even a production-only business?

[Johnny Shen, executive]

Yes. Of course, we are open for production only-business for design life or -- we have 4 different business model. Yes, we can take design completely and we can do partial design. And we can even do the production-only business.

[Szeho Ng, analyst]

Going forward, would the rating for production-only business to go up?

[Johnny Shen, executive]

Honestly, yes. I think, other than China, yes, in other regions, for example, in U.S., we did see a lot of production only or partial design opportunity.

[Szeho Ng, analyst]

Okay. So basically, you are indifferent, you're open for all kind of business opportunities?

[Johnny Shen, executive]

Yes. Yes.

[Daniel Wang, executive]

And the next one is Oliver Lee from BlackRock.

[Oliver Lee, analyst]

Can you hear me?

[Johnny Shen, executive]

Yes.

[Oliver Lee, analyst]

Yes. I'm just wondering, if I look at your balance sheet, you have about roughly USD 140 million in the second quarters, and then that is kind of support roughly $250 million revenue this year. So does that mean if you have additional like USD 100 million, so that would be somewhere like 240? Can I use that ratio to project the size of your potential order?

[Johnny Shen, executive]

Yes, Oliver, I think our money is not 100% for production, right? The ratio is varied, but you need to discount a bit, right? We need to reserve the cash for operating and for operation, yes, and for the investment, for the employee salary and benefit.

Yes, you can roughly -- yes, we expect mass production volume increase a lot, so with additional funding. But is precise ratio, please try not to do that.

[Daniel Wang, executive]

Okay. [Foreign Language] [ Mr. Go ] is asking through message, that he is asking about market ignoring that FEITIAN has a military background. How does Alchip view these things?

[Johnny Shen, executive]

Okay. That's a tough one, but expected. Daniel, why don't you go first?

[Daniel Wang, executive]

Okay, sure. First of all, Yes, it is true that FEITIAN is originated from the National Defense University. It is a fact. But many, many years ago, FEITIAN already separate from National Defense University and move the company to Tianjin. Originally, this company is located in Changsha, where is -- where the National Defense University located. So since then, FEITIAN is focusing on nonmilitary purpose ICs. And for the military purpose ICs, it will be -- it is -- it will be done by the National Defense University itself. So the separation is very clear.

So we cannot stop the rumor or we cannot -- we -- I think for those rumors, everybody has their own opinion. We are not in a position to change their mind. But to us, we will say this way. FEITIAN now is 100% normal company. They are doing CPUs for server and PC. And even though in the -- for the trained scenario, even FEITIAN is considered to possibly on the essential entity list, I would say, for the sequence, it will be on the EDA tool and the IP [ banned ] list first. If FEITIAN is on the list, we don't see too much impact to the business for FEITIAN or for the business we are doing with FEITIAN because for EDA tools and IP, there are solutions. And especially for EDA tools, so we think the dependency from FEITIAN to Alchip will becoming even bigger, which we consider it is a good thing.

But honestly, we do think FEITIAN on essential list is low possibility thing. Otherwise, if FEITIAN is really doing the military background ICs, they will be on the essential list already.

Johnny, do you want to add some color on it?

[Johnny Shen, executive]

Okay. Sure. Yes. Before, FEITIAN spent a lot of time to separate their business. The company is registered in Tianjin is doing a pure commercial and specific usage. They will not touch any military-related application. And I think this one already, between us, TSMC and FEITIAN, we have a consensus agreement. Whatever they do is a pure commercial, and we know their customer, we know their usage. They shouldn't -- they were not doing the military related using that for CRC using FEITIAN for sure.

Yes, just also add on to the Daniel's potentially people are thinking about since they are doing a CPU, high-profile design due to any kind of reason, maybe U.S. will put them on the essential list. Of course, it's not because of military, because of any strategic reason. Yes, once FEITIAN in the essential list, we consider the opportunity and also dependency to Alchip will be even higher. Yes, because essential list they cannot acquire EDA tool nor they cannot acquire IP. That's our fundamental business. We can help them to do the design. We can help them to prepare the IP and tape-out.

So If we're thinking about eventually FEITIAN will be like HiSilicon, I think that will be a long way to go. Yes, I don't think other than the HiSilicon, none of them are start production from TSMC. So I think even FEITIAN on the essential list, I think that's good news for Alchip and because they have a much, much tighter dependency for us. Yes, we all make sure whatever we're doing on the NDA, there will be no military-related application involved for sure.

[Daniel Wang, executive]

And for now, there is no raise hands, and we leave one more question for the Q&A session. [Operator Instructions] Okay, I guess there is none. Then thank you, we are closing our third quarter '20 institutional investor meeting right now. And thank you all for your participation. Thank you.

[Johnny Shen, executive]

Okay. Thank you very much. Thank you for your supporting. Thank you.