Alchip Technologies Ltd

TWSE:3661

Earnings Call Transcript 2020-Q4

[Daniel Wang, executive]

Okay. Dear investors, portfolio managers and analysts, welcome to our Institutional Investor Meeting for year 2020. For this meeting, since we have the investor conference meeting for the first 3 quarters last year, so this presentation, this meeting will [ focus on fourth quarter recap and ] 2020 recap. In addition, in this presentation, I, Daniel; and our CEO, Johnny Shen, will cover -- also cover the 2021 outlook.

So after that, we will enter the Q&A session. [Operator Instructions] And then I guess -- by the way, the video and the audio content of this meeting will upload to the MOPS after the meeting closed. It may take several hours. And you may watch the video or listen to the audio of this meeting maybe by -- in the evening today.

Okay. I'll invite our CEO, Johnny Shen, to do the opening [ routine ], starting from the company update.

[Johnny Shen, executive]

Okay. Ladies and gentlemen, again, thanks for coming to the investor conference meeting again. This is Johnny Shen. I'm CEO and President of company. Yes, once again, allow me to do the quick company identification and update again.

Our company is founded in 2003 and listed in 2014 in Taiwan Stock Exchange. Current headcount, we have 450 employees. More than 75% are engineer. Yes, since we found this company, we've been successfully tape-out more than 450 designs. For last year, we are achieving $240 million revenue, and we are taking more than 30 design in parallel, yes, including project already tape-out and also projects still in the middle of design.

Our company's focus, HPC and AI. Yes, I'll do a quick calculation. The revenue from HPC and AI are more than 3/4. Yes, we are one of the TSMC VCA member. Yes, in fact, only 2 VCA member has worldwide TSMC distribution -- PDK distribution, right? I think only Alchip [ and TUC ] and can do the design service and production for all over the world. Next slide, please.

Yes, I don't want to spend too much time to expand the ASIC market again. Yes, ASIC market is very big and growing. Our company is focus on HPC and AI area. Yes, we consider the entry barrier for HPC and AI are much higher than most of our application. And design is very big and complicated, and HPC and AI always require the most leading-edge technology. Yes, for this year, as I mentioned before, more than 75% of total revenue are coming from [ these field ]. And we truly believe HPC and AI market will continuously grow and sustain and also with a reasonable margin. Next slide, please.

Yes, since this is the beginning of the meeting, let me do a quick company recap for last year. Yes, obviously, our company is doing very well last year. We have a record-breaking number -- yes, we have a record-breaking year for all numbers, yes, including revenue, operating income, net income and also EPS.

Capital market update. I think our share price performance is also very good. Beginning of last year, it's about TWD 200 something. By the end of last year, it's TWD 621. One of the major achievement, GDR, we successfully attract many famous ideal investor. Overbooking -- oversubscribe [ rate ] are very high, 6 to 7x. Currently, we are achieving more than 2 billion market cap right now, and sell-side coverage is also quite promising. They give us a very aggressive target price already from TWD 780 to TWD 1,180.

We have a great position in high-end, leading-edge business. Last year, highest MP volume with strong -- also with a strong forecast. We have the most new logo won and also ongoing projects. Multiple 7-nanometer in production right now. And one of the biggest achievement, by the end of last year, we signed a big -- we signed a contract to win our first 5-nanometer full turnkey design. Next slide, please.

Yes, last year, everybody suffered the trade war tension, and -- but we always consider that's one of the advantage for the company. Trade war does exaggerate the China IC industry. All the China IC maker has a higher dependency for -- from the proven channel like Alchip. We have more opportunity and demand from existing and start-up company.

COVID-19. Overall, we have an insignificant impact on project deliveries. The quarantine requirement favor the local supplier. Yes, to be honestly, we see last year where we encounter much less competition from outside. We rarely see any competition from India, from U.S. or even from Taiwan in the China market. But travel constrain does slow down our Japan and U.S. business.

China capital market, we call STAR market, [Foreign Language]. That's a quick and easy exit for China IC companies. Star Board stimulates emerging IC market. We all know that. And we have a lot of more opportunity due to the STAR market.

That's a quick recap for last year. Yes. So Daniel, please go over the statement?

[Daniel Wang, executive]

Yes. I will take over from the financial numbers on this page. On the page is our fourth quarter quarterly income statement. For the fourth quarter, our revenue reached $66.4 million and -- while the operating income is $8.3 million. In fourth quarter last year, we have quite good nonoperating income, which is mainly contributed by the China government subsidy to our China [ subsidiaries ] across Shanghai, Wuxi, Hefei and Guangzhou.

So the net income for the fourth quarter is $8.1 million -- almost $8.2 million. Translating into EPS is TWD 3.8. You may notice that the net income for third quarter and fourth quarter are different but the EPS is the same. That's because of the FX exchange rate impact to the EPS Taiwan dollar numbers. So the total revenue for last fourth quarter, for the fourth quarter last year, the year-on-year growth is 41% (sic) [ 4.1% ], and the net income year-on-year growth is 77%.

For the yearly P&L performance, you may notice that we have a very successful year last year. The total revenue is $239.5 million, is 71% year-on-year growth. For the operating income part, the operating income is $33.4 million, which up 138% year-on-year. And then for net income, it's almost -- it's more than double growth for last year. It's $28.2 million for the whole year 2020, translating into EPS of TWD 13.6.

And for the revenue breakdown by application, you may see the HPC has become a dominant portion to our revenue. Last year, the HPC-related revenue contribute about 76% of our total revenue, while the niche market contribute 11% and the networking 8% and the consumer down to 4% from 9% in year 2019.

For the process node, process technology, we are very proud that in terms of the process node breakdown, Alchip should be the leader within this industry. You may see for last year, the 7-nanometer-related revenue already accounts for 27% of our total revenue, while the 16- and the 12-nanometer combined contribute 53% of our total revenue. And of course, maybe most of you already know, the major driver for the 16-nanometer node is our biggest customer [ data, the ] CPU for both PC and the server. And combined, 28% -- 28-nanometer, 16-nanometer, 12-nanometer and 7-nanometer, those related revenue already account for more than 90% of our total revenue.

For the regional breakdown, you may see China now is our most important market. Last year, revenues from China customer accounts for 64% of our total revenue. Japan, last year, although the absolute number of the revenue grow at certain level, but the percentage contribution down to 15% from 18% in year 2019. And for the others, the others include the Europe and the North American market, which combined contribute 17% of our total revenue. We expect this breakdown -- the regional breakdown this year to relatively [ 10% ]. The Japan contribution in terms of percentage will go down certain level, and for the contribution from North America and Europe will go up. So that's the revenue breakdown by region.

For the business review. For last year, China CPU shipment is the major growth driver for our revenue growth. We keep on seeing strong demand from our major customers extended from the second half year 2019 to the whole year 2020. China and the U.S. tension, as Johnny mentioned, even stimulates greater demand of China IC designers to release their orders to Alchip or to start even more projects, which will create more design opportunity to us. In addition to the strong CPU shipments, I think the 7-nanometer project flow was another reason for a very good 2020, and those projects came from both the existing and new customers.

The second part is the AI. The 7-nanometer AI projects, many of them are in design phase in 2020. We expect we may see the shipment to start in the second half this year. Actually, 2 of our major AI chip customers will sequentially kick off the production starting from, I think, late third quarter. The 16-nanometer AI shipment last year is a little bit disappointing because according to our customer, they considered these technology node to them will be a trial and test purpose. So the overall shipment volume for those chips are not quite significant. And then for last year, as Johnny mentioned, we won new AI 5-nanometer full-mask turnkey project from a Europe customer. And we already taped out the 5-nanometer test chip project for our North American customer. Sorry. I guess I missed...

[Johnny Shen, executive]

Yes, that's okay. Daniel.

[Daniel Wang, executive]

Johnny, can you see the slides?

[Johnny Shen, executive]

Yes, I can see the slide, but you don't -- we don't have [ that ] page. But anyway, yes, we can go to the Q&A section.

[Daniel Wang, executive]

Okay. Let's go to the Q&A session. [Operator Instructions] CEO, please, you can unmute...

[Johnny Shen, executive]

So once again, you're welcome to raise your question in English or Chinese. [Foreign Language]

[Szeho Ng, analyst]

Congratulations. Yes, 2 questions from my side. Regarding the GDR, so very good response. So you [ turned out ] around 95 million -- 96 million more than your target, right, 100 million initially. So what's your latest plan for the use of proceed? That's my #1 question.

[Johnny Shen, executive]

Okay. [ That's the answer ], sequentially. So Daniel, won't you answer first?

[Daniel Wang, executive]

Okay. Well, the GDR, actually the use of proceed are already announced last Board meeting and the last institutional investor conference meeting. We will use the money we raised from the GDR, first of all, to purchase the mask because the 7-nanometer and the 5-nanometer mask is quite expensive. So we have to prepare the cash. Secondly, we want to use cash to try to secure the 7-nanometer capacity. And until now, I would say, for the first quarter and the second quarter, we are -- we have been doing pretty good. You may notice from our January revenue performance. For the first 2 quarter, I would say the 7-nanometer capacity is pretty much satisfied by both Alchip and our customer.

And for another purpose, there are a lot of opportunity on the market, especially in China right now, that we want to use the money to acquire some design capacity. As I mentioned or Johnny mentioned many, many times that currently, we are short of design resource. The business inflow is much, much greater than our capacity. We try to expand the capacity organically and maybe through some acquisition. And of course, we want to create a maximum design flexibility. It may also require some fund to do such kind of things to maximize our resource flexibility.

[Szeho Ng, analyst]

I see. Okay. Yes, second question, yes, so this year, the focus will be more on the 7-nano and 5-nano, right? So the NRE, the contract value will definitely grow a lot. But in terms of the design engineering input, would that grow in tandem with the contract value? Or would you be able to generate more efficiency when you migrate to this [ modern node ]?

[Daniel Wang, executive]

Johnny?

[Johnny Shen, executive]

Yes. Let me try to cover that. Increased design efficiency is always one of our engineering goal. Yes, currently, leading-edge technology require a lot of resources. Let me just give you an example. For 7-nanometer design, in usual case, we need about 30 to 40 people. Yes, concurrently, I think the design cycle is about 8 to 12 months. So we can -- starting from last year, we can -- we successfully we can shrink the design resource and also shorten the design period. I think that increase our design capacity is always our -- increase our design efficiency is always our goal. But as you know, the demand is much, much more than we can supply. So we still aggressively try to hire more people.

Yes, our current headcount is 450 people. By the end of this year, we are shooting for about 600. Hopefully, we can achieve. Yes, as you know, the resource -- the fighting especially in China is -- resource situation is very tight. So luckily, Alchip is doing very good and being very famous. And we are the only company -- one of the only company doing a 7-nanometer and 5-nanometer design in China. So we're attracting talented resource. I think compared to most of our company, we are doing okay.

[Szeho Ng, analyst]

Great. Yes. [ Maybe if I ] squeeze one last question, a quick one. In terms of the revenue mix, NRE and also the production, [ are we expecting to change that ] from last year to this year?

[Daniel Wang, executive]

Okay. Szeho, let's say it this way. For last year, the NRE revenue accounted for about 44%, 45% of our total revenue. For this year, because the NRE revenue is also strong, so we don't expect too much decrease from the NRE in terms of percentage. We expect the NRE revenue will still account for like 40% -- around 42%, 42% to 43% of our total revenue. And of course, it also depends on the production revenue growth. For the forecast, I think there is still a swing factor for our production revenue is the capacity allocation in the second half this year. So for now, we project the NRE accounts for about 42% of our total revenue this year.

And Charlie, please.

[Charlie Chan, analyst]

Yes. And also, congratulations for a great execution again. So my first question is about the China local AI and GPU trend, right. You can see that your U.S. customers in the 3Q is going to mass produce those AI chips. I think the next [ like should be ] coming from China. So I'm curious, right, because from the recent news flow and research, for example, Baidu, they want to do an AI start-up. Their Kunlun chip is 7-nanometer. And several start-ups like [indiscernible], right, they are doing like 7-nanometer chip. So those are all your clients. Do we miss anything? Because, Johnny, you just said that Alchip is probably the most competitive and probably the only design service provider in China. So how do we think about the future opportunity for the China AI and GPU chip design?

[Johnny Shen, executive]

Okay. Yes, let me try to answer it, and Daniel, you can add later. Yes. So yes, Charlie, you are right. We see a lot of opportunity -- new opportunity in China, people doing the AI, people doing the GPU. Yes, a lot of company acquired talented people from NVIDIA, from AMD, from Qualcomm. And they start to make a start-up company and doing the AI chip or doing the GPU chip and target for the most leading-edge technology, 7-nanometer.

These opportunities keep coming to us. To echo your statement, you are right. In China, we are pretty much the only proven service provider who has a 7-nanometer or 16-nanometer [ big chip ] design experience. So by default, when they establish this company, they only have 2 choices: go to Alchip or try to do it by themselves. Most likely, they choose the first one because put together the whole team, including front end and back end, I think is very difficult for them.

So biggest challenge for us, because they are new, so the design cycle will be relatively longer. And also the production -- their forecast, I think [ RFQ ] compared to reality, that will be -- must be certain gap. So we'd be hesitate to take 2 many of this kind of new project. But overall, since their funding is not the issue, so we're always receiving very reasonable NRE from this start-up company. Yes, so that's why we are aggressively hiring more people, try to take -- and try to satisfy all of them. In fact, almost every month, there's a new company appear.

[Charlie Chan, analyst]

Oh, really? Okay.

[Johnny Shen, executive]

Yes. So Daniel, do you want to add?

[Daniel Wang, executive]

Okay. Charlie, the situation is like this. For the hyperscaler, you mentioned like [ ADI ], like Tencent, like ByteDance, I think Baidu is relatively weak recently. For those service providers, we do have [ comp we spend ] for the projects. But as Johnny mentioned, for those players, they want to do the most leading-edge technology design. But at this time, we don't think their technology is 100% ready. And we really don't think the project they are talking to us, the production revenue is -- will be good.

We -- most of the project we consider are like trial and test purpose. So we intend -- we have to use our resource smart, and so we intend not to take too many of them. And because we do believe given our position in China market, as long as they have certain technology breakthrough, there is no doubt that they will look for us. For the current stage, I think those hyperscaler service providers are trying to do their chip. You may see quite long design turnaround time for them -- for their chips.

[Charlie Chan, analyst]

Okay. Thanks for all of those information. So yes, I appreciate the company being selective, have some discipline in taking projects, right? But it seems like according to those public information, for example, Baidu Kunlun to use Broadcom's design service, right, and Cambricon, they also use Broadcom, so I think there are still some competition in the radar screen. And like [indiscernible], they just want to design by their own, right? And [indiscernible], they seem to work with an [ EDA tool company ]. So I mean the opportunity is big, right? But I mean does the company underestimate the potential competition here?

[Johnny Shen, executive]

I think there must be competition, and we welcome the competition. Without a -- without competitor, a company cannot grow. I truly respect Avago Broadcom, [ this kind of service ]. But I think different company facing different challenge. Like I mentioned before, a start-up company, it's very difficult to complete their first project with international service provider. We've been working very closely to our customer, almost shoulder-to-shoulder every day. So can you imagine the other service provider who has most of resource outside of China? Yes, it's very difficult to understand and satisfy with the local company, especially they are first time or second time doing a chip. Yes. And also, our support pricing structure, I think, is compared to this a company in U.S., we are much, much competitive.

[Charlie Chan, analyst]

Okay. Yes, that's very helpful. So Johnny or Daniel, put all those together, do you see any revenue contribution from those China AI, GPU start-up in 2022? Can you quantify a little bit and also your global AI revenue in 2022?

[Johnny Shen, executive]

Okay. Daniel, why don't you go ahead?

[Daniel Wang, executive]

Okay. Sure. For the GPU product, actually we already -- we are doing 3 GPU projects. 2 is the generic GPU. These 2 projects are all from China customers. And another one is the AI GPU, also from a China customer.

[Charlie Chan, analyst]

Oh, really? Okay.

[Daniel Wang, executive]

Yes. Currently, we have 3 projects in pipeline. And of course, we will receive the NRE revenue this year. And hopefully, the production will start next year.

[Johnny Shen, executive]

Yes, to make a comment on the AI chip, I think revenue for this year, yes, the -- Daniel's presentation also mentioned about we have a very high expectation from our North American customer. That 2 particular customer give us a very aggressive forecast. Yes, so hopefully, by the end of this year, Q3 and Q4 time frame, we will see the very significant AI revenue.

Yes, good thing for this particular customer, we are facing less challenges on capacity support because they also has -- they are also very strong, and they can come to TSMC directly and fighting for the wafer capacity together with us. So once they have a demand, I think we have -- I have a high expectation from AI, especially for that H company, right now maybe I company. They give us a very strong forecast.

[Charlie Chan, analyst]

Okay. Yes, so just last one. This should be very, very quick, and I will be back to the queue. So since the company is so confident, right, so can Johnny or Daniel give us the full year revenue target guidance and also gross margin? I know your NRE mix is going to be similar, right? But given the full mask cost is going to be higher, do you see any margin dilution? So revenue guidance, gross margin, OP margin guidance would be great.

[Daniel Wang, executive]

Okay. For the whole year revenue guidance, [ we said ] in our -- in the last investor conference meeting, we told that the market consensus, which is about 320 million, 330 million, we are pretty confident that we can achieve it. For now, our expectation is higher. We have strong confidence that we can overachieve such kind of revenue forecasted by sales analysts. But how high we can go, it really depends on the capacity allocation in the second half. And since we cannot deliver a very clear picture for the whole year, for the first quarter, as the capacity already booked, we will say we will have a very, very good first quarter. The quarter-on-quarter growth will be, I would say -- okay, Johnny, you can say that [indiscernible].

[Johnny Shen, executive]

Okay. Yes. Okay. So Q1, I think the -- yes, fortunately, we get all the capacity from TSMC. [ So of course ], thank you very much for this major supplier. I think Q1, as Daniel mentioned before, we have a very confidence to achieve another record-breaking number. Yes, and overall, conservatively, I think minimum 50% growth y-o-y.

[Daniel Wang, executive]

Quarter-on-quarter.

[Johnny Shen, executive]

Quarter-on-quarter, yes.

[Charlie Chan, analyst]

50% quarter-on-quarter?

[Johnny Shen, executive]

Yes.

[Charlie Chan, analyst]

50%?

[Johnny Shen, executive]

Yes.

[Daniel Wang, executive]

Yes, 5-0.

[Charlie Chan, analyst]

Well -- so yes, that is very big number. So in that case -- yes, sorry, I promise it's going to be brief, but I kind of think this is important, 50% quarter-on-quarter. Then your full year guidance must be too low because it should mean 2Q, 3Q, 4Q is going to be higher than 1Q then?

[Johnny Shen, executive]

Charlie, we all know that. But I think TSMC capacity, that dictate the revenue number. Without this, I think we are -- we have quite confidence to deliver much higher numbers. Yes, the Q1 capacity already secured. Yes, I think the good thing is one of the giant company in China virtually out of the market. So we get the spare -- a lot of spare capacity on 7-nanometer. Hopefully, this trend will go on. And then we have more -- much, much confidence to deliver a higher number. Q1, Q2, I think the forecast is more accurate, especially Q1.

[Charlie Chan, analyst]

Okay. So -- [ yes, please ].

[Daniel Wang, executive]

I don't want to release too aggressive guidance to all our investors because for the first half this year, as I told to the investor many times, there is a special project. It's a data center project. It's a 7-nanometer shipment. So the total size of this contract is very big, but the shipment will concentrate into the first quarter and the second quarter. So if you are talking about a quarter-on-quarter pattern, I think the quarter-on-quarter growth for the first quarter would be very good. But for the second quarter to the fourth quarter, I think the quarter-on-quarter growth will slow down, will not as high as to like the first quarter. You know what I mean?

[Charlie Chan, analyst]

Yes.

[Daniel Wang, executive]

Yes.

And Haas, please.

[Haas Liu, analyst]

Okay. Thank you, Johnny and Daniel, for the very good results. So my first question is probably just a quick follow-up on your first quarter guidance, which you guided to be up from 50% quarter-on-quarter. So I was wondering if you could provide more detail on your customer breakdown in 1Q. And also for your expectation for the HPC customer in China, you just mentioned that most of the shipment will be skewed to first half this year. So could you also give us some sense about compared with 1Q, what do you think about the customers' contribution in 2Q? And for full year, are you still targeting that customer to be a USD 50 million to USD 100 million contribution for this year, or if you would think any upside from the customer?

[Daniel Wang, executive]

Okay. For the particular project, actually for this customer, our second largest China customer, they have multiple projects in design or in production by us. For the first half, it's the shipment contribute from the first project, which is a data center CPU project. And as you mentioned, we guided that the contribution will be ranging from $50 million to $100 million. I think the actual contribution will still be within this range. And the peak of the shipments should be in the late first quarter and relatively flat in the second quarter and then a little bit [ left ] in the early third quarter. So that's the reason why I said the quarterly pattern, the quarter-on-quarter growth were -- for second quarter to fourth quarter were not as high as to the first quarter.

[Haas Liu, analyst]

Okay. So could you provide more detail about your customer or probably the revenue contribution by region or customer mix in the first quarter this year? And also, what about your first -- or largest China customer contribution in the first quarter?

[Daniel Wang, executive]

Okay. I have -- because the first quarter is not finished yet, so I haven't [ analyze it ]. I didn't do the breakdown for the first quarter. But based on my memory, I would say, for [ Feitian ], the contribution in the first quarter and the second quarter will be lower -- will be lower than our second biggest customer because of this special data center project. And let's say this way. You see our January revenue, which is already announced, is about USD 27 million, USD 28 million. And we expect for the February and for the March, we may see certain growth in February and we may see further growth in March.

So if you do a simple calculation, for last -- for the fourth quarter last year, our total revenue last quarter is USD 54 million. I think for January and February alone, the numbers will be already higher than the fourth quarter. So that's why we are guiding a very aggressive revenue performance for the first quarter this year.

[Haas Liu, analyst]

Okay. So just a quick clarification on your first quarter guidance. Is it up 50% quarter-on-quarter or at least 50% up quarter-on-quarter?

[Daniel Wang, executive]

Not up to, I think 50% quarter-on-quarter is relatively...

[Johnny Shen, executive]

Yes, okay. Anyway, one of the highlights for Q1, I think our U.S., I think revenue from North America, I think, increased quite a bit compared to before. Yes, we have -- yes, we should -- there will be a 7-nanometer tape-out happening in North America, I think based on current forecast.

[Haas Liu, analyst]

Okay. That's very clear. And my second question would be regarding your GDR proceeds. I think you just mentioned the purpose for the proceeds from the GDR. It's mainly for the investment in China and also tape-out for the capacity -- and also some for the R&D capacity and also to secure foundry capacity as well. But could you also provide us some update on the Korea ASIC design investment as well?

[Daniel Wang, executive]

For the ASIC investment, actually, it doesn't use too many cash, the total investor amount is not big. We already announced it last fall maybe. And the older terms has come to the end, almost. And we will proceed the investment maybe -- we'll complete the investment maybe by the end of this month.

[Haas Liu, analyst]

So understanding that you can start working with this company for some of the SSD controller business from later this month?

[Daniel Wang, executive]

So far at this point, we cannot say too much because we cannot represent ASIC, we're just a investor and a partner. So based on our information, we can only say ASIC has good opportunity to win high net projects and maybe Samsung projects in Korean market. If those projects are 7-nanometer with massive shipment volume, ASIC current company scale is not capable of handling those production. So for those projects work with us, we provide the technology support for the 7-nanometer. And if the project won't happen, we will handle the production. But for now, we cannot give you a clear answer about whether we win it or not.

[Haas Liu, analyst]

Okay. Got it. That's very clear. And my third question would be regarding -- you just mentioned you have a lot of design opportunity in China and the U.S. where the capacity is still quite constrained. So could you discuss more in detail if you can sign the contracts with your customers on certain amount of the wafer production to ensure you have the operating leverage from the turnkey business instead of just pure [indiscernible]..

[Daniel Wang, executive]

Okay. Haas, for your question, I think I can tell you what I know. And I guess, Johnny will be much more aware of this question. For the 7-nano capacity, it's not like we -- we negotiate -- not negotiate, we talk to TSMC alone about how many wafer we want. We have to tell them the reason why we want so many capacity. And we have to tell them because of our customer A needs how many wafers and our customer B needs how many wafers. And then they will analyze, then they will make their decision to decide the strength of the support. And -- so there's nothing like we just tell TSMC that we want, like, 4,000, 5,000 wafers without know any backup reasons. So that's the mechanism we discussed with TSMC. So I think the things always come down to what application you have and what the customer you have to try to get the capacity allocation.

[Johnny Shen, executive]

Yes, I believe, Haas, your question is towards the design capacity or wafer capacity?

[Haas Liu, analyst]

Design capacity. Because your design capacity is quite constrained now.

[Johnny Shen, executive]

Yes, design -- to answer your question, I think the reason they want to work with an Alchip, I think it's mainly because our design track record and design capability. Of course, if it's production-only type of design, it's totally welcome, yes, just like Daniel mentioned, that's effortless. We don't need a lot of resources, we can enjoy the revenue. But currently, the demand is all toward the design. Yes, we help them to do the entire back-end implementation plus production.

So we welcome the design -- the production only design, but we refuse design only type of a project. If no matter the company, how big it is, if they come to us, they only want us to do to the design without production, we will refuse no matter how big the company is.

In the past, a few North America customer, in order to demonstrate our design capability, we do take a few design service-only type of project. But right now, at this moment, we are -- company already made a decision, using our design resource, it has to use our production channel.

[Haas Liu, analyst]

Got it. So when you sign the contracts -- NRE contracts with your customers. At the same time, will you also sign a guaranteed rate production right as well?

[Johnny Shen, executive]

Yes. We own the mask. Yes, that's one condition. So we have -- they have to do the production go through us.

[Haas Liu, analyst]

Okay. So there is some housekeeping questions, I want to follow up. Could you discuss more in detail about your subsidies from China government in fourth quarter? What is that for? And the expectation for the subsidies for your subsidiaries in China for 2021?

[Johnny Shen, executive]

Yes. Daniel, do you want to answer that?

[Daniel Wang, executive]

Sorry, I just received a message. I didn't hear clearly.

[Johnny Shen, executive]

Okay. The subsidy from China.

[Daniel Wang, executive]

Okay. The something for China, I think is -- there are different subsidy scheme from different local government. We cannot expect how -- and there are many conditions we have to meet to get the subsidy. For last year, I guess my guess is the total subsidy will be around like TWD 2 million to TWD 3 million, I guess. And I want to clarify, I just received -- maybe I'm wrong about the fourth quarter numbers I just told. The fourth quarter numbers last year, the revenue is TWD 66 million instead of I cannot remember it right. I guess I told TWD 54 million. So...

[Haas Liu, analyst]

Okay. So there's no guidance for the subsidies you can potentially receive in 2021?

[Daniel Wang, executive]

Sorry, fourth quarter '20, the total revenue is TWD 66 million. And so forgive my calculation. I would say, for the first quarter, the quarter-on-quarter growth is based on TWD 66 million for the fourth quarter. I would say, the quarter-on-quarter guidance will be...

[Johnny Shen, executive]

Daniel, the last year, you mean quarter-on-quarter -- I mean Y-o-Y is more...

[Daniel Wang, executive]

Yes, sorry, I was wrong. I think the 50% growth should be year-on-year.

[Haas Liu, analyst]

Okay. That's very clear clarification on your sales outlook for first quarter. But...

[Daniel Wang, executive]

As I comment on the comment on the subsidy, the China support, I think last year, we received certain subsidy from China before. We consider the China subsidy is a onetime deal. Right now, based on the policy -- and this is obviously I can consider that's sustainable. So we expect this year, we can get similar or even higher from China because of many tape-out and also the revenue in China subsidiary.

[Haas Liu, analyst]

Okay. So my last question -- and I will be back in the queue for further questions -- will probably be regarding your OpEx outlook this year. You are going to expand your R&D headcounts to close to 50% year-on-year this year. So what should we think about your OpEx growth in 2021?

[Daniel Wang, executive]

Okay. We take a top-down approach. The year-on-year OpEx growth will be like we currently plan to grow 20%.

[Johnny Shen, executive]

Okay. So Daniel, before you're taking the next question, let's do the forecast, I mean, the clarification on the Q1. Yes, we mentioned about last year, our Q1 number is TWD 50 million, it will be over TWD 50 million. For this year, we have a high confidence to achieve Y-o-Y growth, more than...

[Daniel Wang, executive]

Growth of TWD 50 million. Yes.

[Johnny Shen, executive]

Okay. Yes, that's to take our next question, Szeho Ng.

[Daniel Wang, executive]

Okay. Szeho Ng, please.

[Szeho Ng, analyst]

Daniel, Johnny, congratulations on the good results. Yes. This is Szeho. So my question is regarding your -- because I was looking at your employee number again. So by 2019, you have about 404 employees, in 2020, you have 450, as you mentioned. So you only grow your employee by about 10%. But when I look at your NRE revenue, you nearly grew the NRE revenue some 30-something or 40%. So that's like a massive increase in employee productivity, right? So my question is, do you like use a lot of outsourced engineer last year? And that's why you need to expand your in-house engineer by almost like [ 38% ] 2021?

And then regarding the future employee productivity improvement, right? So like 10% employee growth last year result in some 40% NRE growth, that's like pretty amazing multiple. How should we be expecting this multiple to trend in the future? Like if you grow your employee by 20%, will your productivity be even higher that will generate even higher NRE revenue growth? That's my question.

[Johnny Shen, executive]

Okay. Yes, you have very good observation. I think you are paying very close attention to our headcount and revenue. Yes, our revenue increased quite a bit, quite significantly, but headcount doesn't increase too much. That's true. And to be straight, we do use some in-source. I'm not saying outsource. We invite some contractors working under our facility. Yes, overall, but not too many.

Yes, as I mentioned before, our design efficiency increased a lot. Before we needed 40 people to complete 1 project, we can successfully shrink to about 30. One additional thing is it's not healthy to be stray. Our engineer work extremely hard, day and night. Honestly, it's not healthy because our company doing very good. Most of our employees are benefit from stock option, they are eager to work long hour, including weekend, including holiday, but this -- continuously to go through trend is not healthy. So this year, we have to more aggressively do the hiring to see the China cost structure, honestly, the entry-level engineer is not so expensive. Yes, we try to increase more college grad, not increase our OpEx drastically, but we can increase our capacity and increase our design efficiency. Yes. So Szeho, does that answer your question?

[Szeho Ng, analyst]

Yes. So my question is probably going forward, like maybe your NRE revenue growth will be more in line with your headcount growth. Is that like a -- what you're describing?

[Johnny Shen, executive]

Not really. Because as you know the -- that will be linear function, but you all know that the leading-edge technology, we expect much higher NRE. So 30 people were doing 7-nanometer is actually the same thing doing the 7-nanometer versus 5, the NRE will be -- 5-nanometer will be much higher. Even without capacity increase, the revenue will still grow because the technology migration.

[Daniel Wang, executive]

Okay. Jeffrey, please.

[Jeffrey Ohlweiler, analyst]

Yes. Johnny, Daniel, do you hear me okay?

[Daniel Wang, executive]

Yes.

[Johnny Shen, executive]

Yes. Go ahead, Jeffrey.

[Jeffrey Ohlweiler, analyst]

As first question, crypto price is pretty high recently. So the question is, have you seen any orders come your way for that? Or what's your strategy for or not taking any of those orders over the next year or so?

[Daniel Wang, executive]

Johnny?

[Johnny Shen, executive]

Yes. Jeffrey, which pricing?

[Jeffrey Ohlweiler, analyst]

Let me say..

[Daniel Wang, executive]

You mean the Bitcoin, cryptocurrency?

[Jeffrey Ohlweiler, analyst]

Okay. Yes. Yes. So was just curious are you guys seeing orders there? Or do you plan on...

[Daniel Wang, executive]

We -- to be honest with you, we see a lot of demand. But to be straight, because our capacity is limited, we're choosing -- we have -- we set a much lower priority to take a Bitcoin business so does our major supplier. Yes, their 7-nanometer design capacity is very, very precious. So I don't want to give it to the Bitcoin provider. So that's why we are -- we set a very low priority. To be straight and honestly, many customers approaching to us, but we didn't take any of them at this moment.

[Jeffrey Ohlweiler, analyst]

Okay. Great. Understood. And I guess the next question. In terms of your expectations for your AI customer in Europe and the U.S., where do you see the most upside for the turnkey business for -- is that coming from a U.S. customer or Europe customer over the next couple of years?

[Johnny Shen, executive]

All the major giant social media company, they all start to make an AI chip. Yes, we already won and approached some of them. For example, the biggest data center company is already doing a multiple generation with us. The I company just acquired one of our customer. They have a strong forecast and demand on the AI. And the G company, which is completed a test chip, hopefully, we can win their next-generation design. And other -- other than this 3 giant company, we also aggressively approach others. And company F company, yes, they all have a clear road map to making their chip in the future.

[Jeffrey Ohlweiler, analyst]

Great. And then my last question. I know you talked about revenues for this year. Can you talk a little bit more about the margins? Obviously, you have scale increasing, but you do have more 7- and 5-nanometer. So how do you think the gross margin operating margin trends versus last year this year?

[Daniel Wang, executive]

Okay. For the gross margin, we still expect that gross margin should be above 30%, but it should be in the lower 30s.

[Jeffrey Ohlweiler, analyst]

Okay. How about operating?

[Johnny Shen, executive]

Actually, I don't have the numbers in my mind. But as I mentioned, our OpEx growth target is 20%.

[Daniel Wang, executive]

Okay. There are questions from the message. First question is -- [indiscernible] is asking. How should we evaluate the contribution from North America AI chip in the second half or even in 2022?

I'll go for it first, and I guess, Johnny will -- can do some details. For the AI chip things, we have high expectation for those projects to the North American customer. We expect the contribution will start mainly in the fourth quarter this year. And of course, it still depends on the composite allocation, although we don't worry too much, but there are still schedule risk. So to be conservative, I would say, the revenue contribution from these chips in the second half would be not that significant.

The major contribution will happen in 2022 and for the whole year. And the reason why we are so optimistic is because we already received the PO and the forecast from the customer for both the A customer and for H customer. For the [indiscernible] PO, the amount is quite significant.

[Johnny Shen, executive]

Okay. To add to that, our internal goal is remaining the same. Right now, the North America only contribute about 50% of the total revenue. Yes, hopefully, starting from 2021 will be around 30%. That's our goal.

[Daniel Wang, executive]

And another question from the message that is also the thing I want to do. Sorry for the mistake I made. For the first quarter guidance, the year-on-year growth will be over 50%, 5-0, not quarter-on-quarter, the year-on-year growth for our first quarter this year should be 50% or more.

And for the whole year, we don't have a very clear picture because of the capacity allocation. But for the previous guidance, about -- TWD 320 million to TWD 330 million for this year, we think we can definitely overachieve. But how far we can go, I read some of the report that they are -- estimate our total revenue this year to be about TWD 360 million, TWD 370 million. I would say, for these numbers, it's not -- I think it's achievable. It's not impossible. But if it is 100%, it depends on the 7-nanometer capacity allocation. And Charlie?

[Charlie Chan, analyst]

So yes, thanks for clarification and also taking my second round of question. So Daniel, I know it is now 50% Y-o-Y, but I assume it is based on the U.S. dollars, right?

[Daniel Wang, executive]

Yes.

[Charlie Chan, analyst]

Okay. So we should also consider NT dollar appreciation to calculate your NT dollar 1Q revenue, is that right?

[Daniel Wang, executive]

Yes. Of course, if you are -- because our functional currency is U.S. dollar. So we -- internally, we use the -- no matter internally or for the public information, we use U.S. dollar as our functional currency. For the numbers translating into NT dollars, of course, it will be affected by the FX exchange rate. But given the growth percentage, the number is -- I think the impact will not be that significant.

[Charlie Chan, analyst]

Okay. Okay. In terms of the GDR share dilution, when would that start? And by how much?

[Daniel Wang, executive]

The GDR current -- currently, our total shares outstanding now is about 69-something million, that's paid-in capital NT dollars. [Foreign Language].

[Charlie Chan, analyst]

Okay. So that's after the GDR?

[Daniel Wang, executive]

Yes. That's after the GDR.

[Charlie Chan, analyst]

Okay. Okay. And so there's -- switch back to some long-term focus. I mean, I really care about your 5-nanometer test chip with North America customer because your main competitor Global Unichip, also very confident that they can win this design. So just to give us some more details, I mean, why both of you are so confident? Are you competing with the same projects? And what you should improve to secure design? Because sometimes there are still comments about your capability for those advanced packaging for [indiscernible] HBM2, et cetera? Can you give us some color on this competition?

[Daniel Wang, executive]

Johnny?

[Johnny Shen, executive]

Okay. Yes. So, of course, the G company is a very respectful competitor. They are working with that particular company, I think, for quite a long time. And I think because of success of other project, the G company grant us the 5-nanometer test chip. And the result is -- overall during the design, they are very satisfied with our service. So hopefully, we have a good position to win this project. I'm not saying we're -- 100% chance, but I think about we have a great chance to take this opportunity because we already have a 5-nanometer proven testing in our hand.

Yes. So in terms of a design capability for HBM, I don't know why they are -- they are always mentioned about the -- that part, that weak part to us. In fact, we have many proven HBM design ongoing and also in production. And with the Israel company or with the China company. So I think whoever has HBM2, this kind of a question or customer, they can come to us directly, we can address accordingly. We have a production chip. We have a test chip. Yes, so HBM2 this kind of a design capacity is not our weak spot. Yes, in fact, I think we have a many, many pace and can do the demonstration. Yes, 5-nanometer for turnkey design win, yes, we have -- we already have it, and also earlier in the slide, in China, so we have many 5-nanometer design opportunity.

[Charlie Chan, analyst]

Yes. So I would assume there would be kind of major project after this proven test chip, right? So what would that be? And is that totally confirmed? And is it a kind of zero-sum game between you and GUC for that major chip?

[Johnny Shen, executive]

Yes, the test chip will be -- come out very soon, and we are sharing all the result with the customer and -- yes, this particular customer is also very dynamic. And to be straight, they are shooting for 5-nanometer, 3-nanometer, one of them, yes. They even think about 3. Very, very dynamic.

[Charlie Chan, analyst]

Okay. Okay. Got you. And actually, so really kind of a consent from foreign investors, right? I mean we are glad that your customer fee is that growing so rapidly. And maybe 1 day, they would pursue the China IPO, right? But it's just hard to imagine that a company with USD 500 million or USD 1 billion revenue will continue to use design service. Are you confident that [ Feitian ] will continue to use that design service in the coming 3 to 5 years?

[Johnny Shen, executive]

Yes. In the recent year, I think if they are continuously using our service or not. I think I always tell their General Manager actually need to provide the value in order to sustain the business that they all agree. Like I mentioned before, we 2 company grow all together, they are our customers since 12 years ago. The dependency bond is quite tight. And due to a lot of restriction, and they're thinking about is the best way to work with the Alchip instead of a go to supplier by themselves. That's exactly the wording from their General Manager. Yes, I'm not saying we will do their business forever, but in the near years, this year, next year, the year after, I don't think that's a problem. They don't have a direct contact with TSMC, and they don't want to speak.

[Charlie Chan, analyst]

Okay. And not sure if you noticed that it seems like SMIC very likely already got the U.S. equipment supplier again.

[Johnny Shen, executive]

Yes.

[Charlie Chan, analyst]

Yes. So any kind of information you can provide? I mean, this may change the dynamic again, right, meaning if SMIC will be able to do 14-nanometer or even 10-nanometer like what has shared the life time, right? There were some [indiscernible] projects at SMIC's 14-nanometer. Do you think they will go live again?

[Johnny Shen, executive]

Yes. Yes. We also know that SMIC may sustain their are production in 14-nanometer. I think that's maybe the good news for many IC company in China. But for [indiscernible] I think they already -- their 2 major products, HPC and CPU, already start the next generation with a 7-nanometer design. Yes, I think the existing 14-nanometer, they maybe -- they will use some SMIC as a production. I think that's a missionary kind of a task. But major product line, they already recycle to 7-nanometer. I think it will take some time for SMIC to establish their 7-nanometer design flow, and based on my understanding, they don't even have equipment yet.

[Daniel Wang, executive]

Okay. Jeffrey, you still have questions?

[Jeffrey Ohlweiler, analyst]

Not for me.

[Daniel Wang, executive]

Okay. Go ahead. Go ahead.

[Unknown Analyst]

Yes. I have questions about the mask. So how -- what is your current mask portfolio? Is much the largest cost of good for your company?

[Johnny Shen, executive]

No. Let me answer this question. Actually, we mentioned many times that since we own the mask, so our mask and IP, the amount will be capitalized. So the mask will be our asset. So in the COGS, we have 3 competitors. First of all, the tool [indiscernible] and some local [indiscernible] tooling cost, which accounts for -- accounts for which very minimal -- minimum percentage of our total COGS. The other few component is 1 is the wafer cost and another one is the mask and the IP amortization. So the mask and the IP will go to this item.

[Unknown Analyst]

Okay. For 7 nanometers, how many masks you have right now? And how many you need to invest? And how many years you capitalize the mask?

[Daniel Wang, executive]

Okay. I cannot tell you the numbers. We didn't disclose how many masks we have and how many masks we are going to invest, that's because we have so many. And secondly, for the appreciation time period, it will based on the real product life cycle. For the principle, the mask won't depreciate over 18 months. The maximum -- the maximum amortization period is 18 months. For some special projects like what we did, the currency, the Bitcoin, we will amortize the mask in 6 months. So it all depends on the real product life cycle to decide a different -- to decide the amortization period.

[Unknown Analyst]

So is there a minimum quantity that needs to go into production in order to recoup all your -- the cost of mask?

[Daniel Wang, executive]

Now to recoup the cost of the mask, actually, when we discuss with the customer, the contract -- we are discussing the total contract value or less [ NI ] total value. The total value include the mask and IP cost.

[Unknown Analyst]

I see. And for the 5-nanometer and 3-nanometer, how much more expensive than the 7-nanometer?

[Daniel Wang, executive]

We cannot reveal the price because of NDA between us and TSMC. I can only say that 5-nanometer mask price is much, much higher than the 7-nanometer.

[Unknown Analyst]

Okay. And so this year, say, can I ask the total CapEx for like investing in the mask and IP?

[Daniel Wang, executive]

Total CapEx. Actually, I can give you a very clear number. But to be honest, we will announce our yearly report actually, maybe within 1 to 2 days. You may see the CapEx last year. This year will definitely grow quite a bit because of the 7-nanometer project tape-out. For the total numbers, I don't have an idea of the exact number because there will still be scheduled risk for the tape-out.

[Unknown Analyst]

Okay. So my final question is that in your past experience, have you really experienced that some of the masks, you cannot really get a payback. That means that it failed?

[Daniel Wang, executive]

Yes. I won't say it failed, but for some cryptocurrency customer, they do decide to tape out the project, which means the mask is already produced and the customer already paid the mask, but they are -- but there was no production. They decided not to get into the production phase. We have experienced mainly within the cryptocurrency application.

On the other normal project, usually, when -- you can imagine, if you invest so much money into NRE, especially in the mask, you -- usually, you will try to produce as much chips as you can to recoup the investment from the mask.

[Unknown Analyst]

Can you -- can the customer share the same mask or everyone has their own dedicated mark?

[Daniel Wang, executive]

No, no, the mask ownership is belong to Alchip.

[Unknown Analyst]

So you can use that one -- say, particular one for several customers?

[Daniel Wang, executive]

No, no. Because actually, the mask is -- the design is dedicated for a particular customer.

Thank you. And there is a question, and I want to clarify as well. Someone is asking about, in last investor conference meeting, Johnny mentioned the target employee numbers by the end of this year, it's 700. I just -- it's a miss. I would say, by the end of this year, we target -- the total employee number is 600, and we are pretty confident we can achieve it because there are 2 major incoming new employees in the March, April and another time slice like, after the summer vacation graduation. After these 2 massive new incoming, we think the 600 target by the end of this year should be no problem.

And another question from the message. Someone -- [indiscernible] the is asking, there are some talks about potential competition in the China local CPU market from x86 CPU alternatives. If we also consider software capability for x86 CPU, it's arguably still better for now. How does Alchip the competitive landscape in the China localized CPU market? I think this question is asking about what's the competition status between x86 architecture the CPU and Arm-based the CPU.

Johnny, do you want to answer?

[Johnny Shen, executive]

Okay, sure. Yes, I think -- let me clarify. When we think about the CPU opportunity, of course, [indiscernible] is our number one customer, they are using Arm-based kind of design. We are not saying LG is not taking any x86 kind of a CPU design. In fact, we do have a few opportunity. They are talking to us.

So in terms of market share and also the performance competition right now, I think, Arm is the -- is favor a lot. So we all know, the Apple, a few quarters ago, just announced their Arm-based CPU. They have completely outperformed their previous x86 type of a design. So go through the Arm, I think Arm-based design is kind of a trend. Yes, we also say that I think our #1 customer has the confidence to sustain their market share. But in fact, we start to taking some x86 design also. Yes, Daniel.

[Daniel Wang, executive]

Okay. Is there any questions?

[Johnny Shen, executive]

[Foreign Language]

[Unknown Analyst]

[Foreign Language]

[Daniel Wang, executive]

[Foreign Language]

So if there are any questions, we welcome both English and Chinese questions. Okay. I guess there is no -- no one raised their hand. And if that's the case, thank you for your participation in our 2020 Institutional Investor Conference Meeting. Thank you very much.

[Johnny Shen, executive]

All right. Thank you very much. Thank you. Thank you for supporting us. Thank you very much.

[Daniel Wang, executive]

Thank you. Bye.

[Johnny Shen, executive]

Bye.