Okay, let me give everybody some update for Q3. Yeah, we can say we have a fairly acceptable Q3 with a record high revenue of TWD 117.5 million. However, the net income TWD 14.6 million and EPS TWD 6.2 are just slightly better than Q2. Mainly due to a few NRE milestones delayed to Q4, plus a one-time expense after we change our stock option exercise scheme. Our mass production demand remains very strong, especially from North American region. The ABF substrate shortage limits our mass production revenue in Q3. Fortunately, capacity-related issue has been resolved. We already received capacity support commitment from major suppliers starting from this month. We are expecting to see tremendous revenue growth starting from November and lasting for at least six quarters. Alchip continuously.
Sorry, Johnny, you are mute.
Sorry, I pressed the mute button. Okay. Yeah, let me continue. Yeah. Alchip continuously hold a leading great position in leading-edge technology business. Our business inquiry this year are enormous. Majority of them are in advanced technology for 7 nm or below. With many design win, we expect healthy revenue growth starting from Q4 and continuing through next year. We have also made a significant progress in automotive application. Many electric car makers start to develop their own ASIC and have approached us recently. Some of these will be kicked off very soon. Therefore, we truly believe, starting from late 2024, automotive application will be another booming revenue driver for Alchip. A quick update on the U.S. BIS extension issued on October 7. Overall, we encounter very limited impact on our current business. The updated restrictions are very specific.
It clearly defines the spec and rule for our supply chain to follow, yeah, including us, our partners, and customers. We expect all business will execute normally as long as we comply with the regulations. Lastly, I would like to emphasize our neutral position and diversified business condition. Similar to our foundry partner, Alchip will never make a product to compete with any customer. We have a fully diversified and well-balanced business from all region, yeah, including North America, Asia- Pacific, Japan and Europe. In terms of headcount, we grew quite a bit recently. In addition to China, we have implemented a more aggressive hiring plan to increase our engineering and supporting resource in Japan, Taiwan and U.S. region. We also have a plan to open additional office in Southeast Asia. Now, this effort will provide more efficient and cost-effective solution to meet our customers requirements. Thank you.
Okay, this is the third quarter quarterly income statement. Actually, we already supposed to announce the third quarter top line and the bottom line, so I just repeat the numbers and go into detail in the next slide. For the revenue in third quarter, our functional currency is U.S. dollars. In third quarter 2022, the revenue is $117.5 million, which is a 16.2% quarter-on-quarter growth and 28% year-on-year growth. For the operating income in third quarter is $18.6 million, which is 3.6% quarter-on-quarter decline, but 21.9% year-on-year increase.
For the net income, the net income is TWD 14.6 million in third quarter, which is 0.5% quarter-on-quarter growth and 12.6% year-on-year growth. Because of the exchange rate in third quarter, the NT dollars to U.S. dollars depreciate a little bit. The EPS is TWD 6.2 for the third quarter. This is the revenue breakdown by application in third quarter. Again, you can see, HPC was still our main focus, which accounted for 78% of our total revenue. The niche market accounted for 6%. The network application accounted for 6%. The consumer product accounted for 9% of our total revenue in third quarter.
In this year, the breakdown is pretty much similar to the quarterly breakdown. HPC accounts for 80%, while the other three applications, consumer, network, niche, accounts for 6%-7% of total revenue. This slide is the revenue breakdown by process node. As always, Alchip is focusing on the leading-edge technology nodes. You can see in the third quarter, the 7 nm or the more advanced technology nodes, revenue from these nodes account for 69% of our total revenue. For the 16 nm and the 12 nm, it accounted for 21% of total revenue in the third quarter, while 28 nm and the 40 nm accounts for 6% and 5% respectively for our third quarter revenue. For the yearly, which is the same, 7 nm or more advanced technology nodes.
The revenue from those nodes accounted for 66% of our total revenue this year, and the 16 nm and 12 nm revenue accounts for 21% of our total revenue this year. This is the breakdown by region. In third quarter, the revenue contribution from North America keeps on rising to 39%. The revenue from Asia- Pacific, which is pretty similar to what we had in second quarter. In third quarter, the two major regions that contribute us the majority revenue is U.S. market and China and Asia- Pacific market. Now for Japan accounted for 40% of our total revenue in third quarter, while the other region total accounts for 9% of our total revenue.
From first quarter to third quarter, the accumulated revenue in the U.S. accounted for 36% of our total revenue, while revenue from Asia- Pacific accounted for 40% of our total revenue this year. For the third quarter business review, honestly, this revenue is a little bit lower than what we expected because of the insufficient supply of the ABF substrate. We mentioned it in our second quarter earnings call meeting. Throughout the third quarter, we still suffer from this issue. We expect this substrate insufficient supply will significantly improve in the late fourth quarter this year. For the NRE, which means the design revenue, the design demands from our customers remain robust, as most of the project milestone in third quarter ending on schedule.
The profit margin down, that is because the higher production revenue in the third quarter and the other one-time off items. I will explain it later. The third quarter gross margin is 31.7%. Compared to the second quarter, the gross margin is lower in third quarter. That's because, first of all, lower NRE revenue contribution. In third quarter, our NRE accounts for about 40%-45% of our total revenue, compared to the second quarter. Second quarter, the NRE accounts for 45%-50% of our total revenue. In third quarter, because we are aggressively adding the substrate supplier, so there are several new substrate suppliers to get into our supply chain.
We have to spend some cutback to do the qualification, which is another reason for the gross margin in third quarter. For the operating expense, well, you can see our third quarter operating expense is about TWD 2 million-TWD 3 million higher than the second quarter. The first reason is we have relatively large hiring near the end of the second quarter. Secondly, there is a one-time, the ESOP option policy change. It causes about TWD 1.1 million-TWD 1.2 million increase to our routine operating expense. That's the major reason for the profit margin in the third quarter.
For the fourth quarter business outlook, first thing is, we like that actually today is already November fourth. Some production revenue shifted to November from October. The October sales will suffer. Again, still suffering the insufficient substrate supply. In addition to that, because of the testing program renewal by our customer, the AI chip shipments to U.S. customers delayed to November. We shipped a very limited chip to our North American customers in November. Of course, which causes relatively big impact to the October revenue. For those products, we are the manufacturing, the production is as normal. Those products will ship onto our customer in November, which means in November we expect a relatively very significant month-on-month revenue growth because of that.
For the substrate supply, as I mentioned, we expect the substrate supply to significantly improve from November. The number of substrate suppliers increased from one in third quarter to three in fourth quarter 2022. Previously, prior to the end of the third quarter, we only have one supplier, who is Nan Ya PCB, but in fourth quarter, we will have Nan Ya PCB with better capacity support. In addition to Nan Ya PCB, we'll also have Unimicron. In December, we will have another new supplier. As I have mentioned many times, this new supplier will have a dedicated line for this product. The substrate shipment from this new supplier will begin in December this year.
We have very good support commitment from new suppliers for late fourth quarter 2022 and the whole year next year, 2023. To us, if we look into 2023, we don't consider substrate supply will be an issue anymore. In addition to substrate supply, the CoWoS capacity, TSMC gave us very good support. We also don't see the CoWoS capacity as an issue anymore, again, in 2023. For the design parts, the overall design demand remain, I would say, good. I know everybody have concerns about the new BIS regulation. I will separate the session later.
Based on the current rule and the situation, the reaction by our vendors and our understanding to the regulation and the reaction of our customer, the new BIS regulation for our ongoing projects will be very limited. That's the BIS part. For the automotive-related application business, as we mentioned last time, that we have very good chance and the position to win some automotive-related projects in both U.S. and the China markets. We already have good progress, and we expect business from this application will become larger and larger in the fourth quarter and the whole year next year. For the process node migration, we still expect the process node migration from 7 nm to 5 nm to continue in 2023.
For 3 nm, we don't see relevant revenue contribution from 3 nm, this year and the next year. We expect the revenue contribution from 3 nm would likely to have a meaningful increase in 2024. This slide is for the BIS new regulations. First thing is, I would like to say our understanding on these regulations. Based on our acknowledgments, instead of using Entity List, BIS this time is by defining the specs to be the parameter for chip design restrictions on China IC vendors. Another approach from BIS, based on our knowledge, the end users, or let's say KYC, know your customer. As regulations for identifying the MEU, the military end- user.
I think these two directions are the major change for the new BIS regulation. In the past, the BIS used the entity list to achieve their target. For this time, they seem to change its approach. In conclusion, we view these new regulations as positive to Alchip. Let me explain to you why. First of all, we see limited impact. As mentioned, and as stated in our previous slides, our China exposure this year down from 70%-something in 2021. For this year, the first quarter to third quarter cumulative revenue from China customers already down to 26% of company's total revenue.
By sorting out our ongoing projects, only one project may need to go for ECCN classification, which is a GPU project from our China customer. This project, the projected revenue in fourth quarter is estimated to account for about 1%-2% of our total 2022 revenue. Of course, it is a projection, it is a projected revenue. The reason why we say we see this new regulation as a positive thing to us is because from now on, it seems everyone, every party has clear game rule for the BIS thing, for the China geopolitical thing. In the past, honestly, we don't know which of our customer, our China customer will be putting on the sanction list.
We don't have a clue. Right now, the targeting is on project itself. For example, for the customer I mentioned that maybe have a project falling into 3A090, which is a new class of the BIS classification. The other project may be from the same customer, which is not into the class of subject to EAR. They can still go and the vendor have very clear game rule for what project could support and what project could not support. To us, it makes us or our supplier easier, much, much, much easier to avoid the potential risk. I guess that's the whole presentation for today. Right now we are going to the Q&A session. Again, please use your raise hand function through Zoom, and I will invite you for your question. The first one is Charlie, Morgan Stanley, please.
Thanks, Daniel. Good afternoon, Johnny. First of all, congratulations for your great execution, even during this very tough geopolitical environment. I do have a few questions around this issue.
Mm-hmm.
First of all, you mentioned that the new BIS rule is very specific, which you consider as a kind of a good news in the long term. I am wondering, you know, first of all, how do you ensure the KYC know your customer that their products wouldn't be for military usage, maybe after several layers of integration and redistribution. That is the first question. Also second question to Johnny.
Mm-hmm.
How are you going to manage your China business in the long term, right? Meaning, for example, if some China projects margins is super attractive.
Mm-hmm.
Whether you would prioritize those China projects? That's my first question. Thank you.
Johnny, do you wanna go?
Okay.
Or-
All right. Yeah, for us, for the KYC, I think we work with all the supplier. We define a kind of a game rule. First of all, before we sign up, for example the project, getting started. During the NDA stage, there's a lot of documents for our customer to sign. We need to understand their market segment, we need to understand their application and also their end customer as well, in which category. Both us and also major supplier will review. Once we feel comfortable, we will prepare all the documents for our end customer to sign before we sign the three-way NDA. This document will be signed twice. Project start and project payout, during the project payout. That takes care of the KYC portion. Yeah, for our end customer, same thing.
They need to sign the KYC with their potential end customer. On their list during the shipment, we will review that as well. That's again a rule we define. For all the high compute power related project, we will apply for the classification. Make sure that from the third party, they won't violate the 3A090 category. Before payout, all the classification need to be done. KYC related document need to be signed. I don't know, Charlie, does that answer your first part of your question?
Sure. Yeah. I'm sure that
Mm-hmm.
Your vendors, partners want to ensure-
Mm-hmm.
The KYC. Just want to know more details. My second question also.
Yes.
About your-
Business.
China strategy. Yes.
Yes. Yeah, like I mentioned before, I will make a normal operation business as usual. Yeah, as long as all the customer and us comply with the regulation, and we will take the business. We will continue to take their business. By nature, to be honest, we will focus non-China business more. The China business will be selective. If it gives us, we consider has a great potential, reasonable profit, and we will continue to support. Internally, we have this kind of game rule. We need to make sure every single account contribution will be limited. For example, we will not allow any account contribute more than 10% of the total revenue.
Yeah, if it's over that, and we can easily change the ASIC back to COT or royalty model, yeah, in order to prevent the big impact. We, to answer your question, we are continue to support China business selectively and focus on non-China business more.
I see. Thank you.
Mm-hmm.
My second question is associated to that actually.
Mm-hmm.
Since now you want to focus more on non-China business, and the U.S. is a kind of big growing market, whether you are hearing some customers demand that they don't really want their design based on by your China operation or even they don't want their chip to be produced in either Taiwan or China, how are you going to manage those requests if that is the case?
Okay. Let me try to answer your question. Yes. To be honest, three months ago, when we have an investor conference meeting for Q2, I do some calculation. Only 20% of our customer has some requirement or he mentioned about do not use China resource in China. But recently, you are right. I think I get more and more this kind of a request. Precisely how much, I don't know. Still insignificant. But looking forward, I think this number will increase. That's why we have a more aggressive hiring plan in other region, in Japan, in Taiwan, and also the Southeast Asia. Similar to other U.S. company, we do have a plan to send our engineer working in different region.
Now, to be honest, I do some internal inquiry asking the senior engineer, are you willing to work in Japan or in Singapore or Malaysia? I think the feedback is positive. All of them. I think most of them are not against working in other region for, like, to to three years. That's the work around. We will continuously hire, more aggressive hiring in other region. Yet in the meantime, we can send our existing China resource to other region to work on the design.
Got it. Thank you. Yeah. I just want to make sure, those request or requirement doesn't limit your ability to win the future projects. For example, we know that, Marvell, Broadcom, they are super aggressive.
Mm-hmm.
To compete with your future projects. Thanks for your explanation. I'll be back to the queue and raise some follow-up questions later. Thank you.
All right. Thanks, Charlie.
Mm-hmm.
Okay, Szeho, please.
Hi, Johnny and Daniel. Thanks for the time. Yeah, first question, relating to a question similar to Charlie's ask earlier. What percentage of the engineers right now is based in China, and how do you think the percentage change in the next two, three years?
Okay. Right now, about 70% of our engineering resource are located in China. The other region, we have two teams in Japan, which accounts for about 20%. In Taiwan, we have one team, which contributes about 10% of our design resource.
Mm-hmm. I see.
Yeah, looking forward, by next year, we will increase more aggressively. Near the 2024, I have a plan to have about 50/50 kind of weight between China and non-China.
All right. I see. Would the company's cost structure change meaningfully as a result of this kind of engineering allocation? Yeah.
We do some calculation. I think the increase will be very, very insignificant. Yeah. As you know, right now, China itself, the cost is not so low. Yeah, we all know that. Also Japan, based on current inflation, everything in Japan cost is very attractive. Yeah, we thinking about other area like, Malaysia. Yeah, Malaysia cost, I think is, much lower than China, but Singapore is a little bit higher. Combined with these three region, I think the cost after some calculation, I think it's pretty much the same. Yeah, I don't expect the cost structure will increase.
I see. Good. That's it.
Mm-hmm.
Regarding the automotive business you mentioned earlier. Are we doing the HPC car chips or the companion chips inside the car?
Yes. The automotive area, we are focused on two areas. You see, you are right. Automotive HPC from Level 2 to Level 5, when we have a chance, we will try our best to win. That's on the edge side. Each car maker also have a plan to build out their data center. That's the server side. I think both of them, we have an opportunity to win.
Oh, okay. All right. Basically, we haven't really won the project, right? You stand a good chance to win those projects.
In fact, we already won a couple, and so many customers. Actually, a big-name car maker approaching us recently.
Oh, okay. All right. Sounds good. Okay. Congratulations. Thanks.
Thank you. Thank you.
Okay, next call will be the Gavan Mo, Evercore.
Thanks, Daniel. Thanks, Johnny. Congrats on the good results. I have a couple of questions. Maybe I can start. My first one would be on the total addressable market, right? So the 3A090 and 4A090 have obviously restricted a certain percentage of the market. China can't go beyond that certain threshold anymore. I'm just wondering, in terms of localization demand for chips below that threshold, do you see that as sizable? Is there just a lot of demand there as well?
Okay, Gavan, let me answer you this question. Actually, we don't have too much relations for the 4A090. 4A090 mainly defines the super computer system. Our job is just doing the IC turnkey service. The new class has more relationship with us is the 3A090. For 3A090, the most important spec in this class is the 600 Gb ps I/O speed. From this spec, you can see the main purpose of this new class is targeting GPU and AI GPU and maybe service accelerator. For CPU, usually don't need to have such high I/O speed.
Most of the CPU will not fall into this class. For GPU, to be honest, 3A090, the specs this class defines is not low. The benchmark is the NVIDIA A100, which is a pretty advanced chip. That's the reason why right now, we don't see a significant impact to our ongoing project.
Yeah, let me add to that. Daniel is right. We have a very limited impact on current project. Look, looking at the downside, I think when our GPU customers stepping into 5 nm or 3 nm, I think that will be easy to violate this 600 Gb requirement. So I think they have to have a few workarounds. I think continuously, future GPU project using most leading-edge technology, this kind of opportunity, I think will, our customer need to change their direction. I think that area will be slowed down. For CPU and other application, I think that will be no not much impact.
Sure. Just to clarify, what you're saying is that a large chunk of the market is being affected by these restrictions, at least going forward, right? 'Cause you can't grow in that space. But you're saying like, for example, the CPU market and even GPUs, you can repurpose it to be beyond, below the threshold. This chunk of the market is still quite sizable over the longer term. Is that how you view it?
Okay. Gavan, maybe Johnny can add some color.
Mm-hmm.
Later. That to us, as I mentioned, for the specs, right now, the spec is not low. Actually, the spec, as I mentioned, the NVIDIA A100 is a pretty advanced GPU. We don't see right now for this in the near term, we don't see it block a very big chunk of the market. In the future, maybe. According to the document BIS released, they will dynamically to adjust the specs in the 3A090.
Okay. Let me add more. From our customer's point of view, I think they are honestly a little bit nervous on the next generation. Thinking about the bright side, since NVIDIA cannot offer any advanced chip to China, and they already have a chip into production. They consider the lifecycle for their existing chip will be longer and volume will be bigger. For those kind of company, GPU has a spec near the limitation. I think most of them are our customer. In the near term, I don't expect too much revenue impact from this China CPU makers.
Sure. Great. Understand. Thanks, Johnny. One more follow-up question. As you switch your focus and strategy to focus more on non-China projects.
Yes.
Is this pivot away from China something very easy to execute? Because I already assume you, I mean, it's actually 70% revenue in other markets, right? It's very easy just to focus sales and marketing efforts on other countries. There won't be a lag time for you to sort of pick up marketing efforts in other countries, right? Because I would assume the demand there is very strong already. Is that the case?
Okay. When you look at the whole market, the design demand is, as I mentioned, very, very strong. The supply for the advanced technology node, the backend supply, the backend design supply is really, really limited. Maybe like GUC, or maybe one or two others can provide such kind of services in leading-edge technology nodes. We don't see taking projects from non-China region will be a problem, but it takes time. It is not that difficult to do.
Yeah, let me try to add to that. Yeah. Obviously, our U.S. revenue increased quite a bit recently and also all the opportunities. I think about this way, because unfortunately right now, the entire IC market is not so good right now. It's the market a little bit toward the recession portion. That's no good for our end customer. Thinking about the bright side, when the customer doing a cost down, the first thing they are thinking about is outsourcing. Yeah, to be honest, most of our customer and potential customer already issued the hiring freeze or even layoff. They need us much more than before. I consider they need us much more than before. In fact, I think the outsourcing to save the cost and we are the proven one. I think during this situation, we expect we have more opportunity, yeah, especially on the North American regions.
Okay. Let me move to the next one. Wilson Zhen, please.
Well, can you hear me?
Yes.
Well, okay. Thank you, management, for taking my questions. I'll ask some simple questions. Just, I think everybody was pretty concerned as by the end of second quarter, we see on your financials that China exposure account for 39% of total revenues. With the BIS American sanctions, share prices went down a lot. Then-
Mm.
Thank you for clarifying that. Going forward, let's say, we're concerned about the future, right? Do you see the China exposure at 26% currently to meaningfully go down in the foreseeable future next year? Let's say the first two quarters of 2023 perhaps, or is it going to stay the same? A second question is, we hear—we keep hearing from various sources that the company is in talks with some customers about autonomous driving projects related to that. If that's true, when do you see this revenue contribution to kick in next year, perhaps? And if that's true, maybe the NRE project would go on for several quarters, maybe? I don't know. When do you reasonably expect that to commence production, you know, if everything goes smoothly? Thank you.
Okay. Yeah. Daniel, let me go first, and you, maybe you can add on later.
Sure. Sure.
In terms of China business, yeah, like I mentioned before we taking some China business, especially the high compute related business, we need to be more conservative for sign our own related document, KYC related document, and working with the suppliers to do some DD before we take it. Yeah, after we're thinking about this, low risk and no risk, especially no military usage, those kind of stuff. Customer already also signed a related document back to us, then we will take. I try not to manage it for the total percentage, but I also mentioned about for every single account, we need to control then under certain percentage, like less than 10% or so.
Yeah, if it's any specific account, one account exceed this limitation, we can change the model back to COT or back to the royalty model. Total percentage, as long as we combine compared to the annual, I don't intentionally to further reduce the China concentration. Yeah, that's your first question. The second question is the autonomous driving contribution. Starting from Q4 and next year, I expect to have some NRE. But the actual mass production contribution based on current calculation, I'm thinking about by the end of or near the end of 2024, we expect to see a significant revenue, yeah, booming during that time. But prior to that, I think pretty much all the NRE. Yeah, as you know, the automotive are different compared to other application.
We need to engage much earlier. Other project, we're just doing purely back-end implementation, yeah, because most of IC makers in-house, they already have a team to do the front-end architecturally related design. For car maker, they need more support effort from us and also from our partner. That's why I think from the design start to production, it needs a longer period. Based on our calculation, the real contribution will be by the end of 2024.
Okay. Thank you, gentlemen.
Okay. Let me say this way. For the percentage part, I will say for next year, because the U.S. contribution will grow a lot, very significantly next year. The percentage from the revenue contribution percentage from China will go down. The absolute amount, the value, revenue, absolute amount, we don't expect it to go down because the demand there is still strong and the generation migration is still ongoing. I think that's the that is my best guess for the next year.
Mm-hmm. Thank you.
Okay.
Thanks again. Thank you.
Charlie, please.
Thanks for taking my follow-up questions. My question is more about the competition again. I mean, your U.S. peers like Marvell, Broadcom, I think they can pitch that they have a lower wafer price, right? Versus-
Mm-hmm.
Versus your GUC as a VCA, and they can support some front-end design. There could be some, you know, China constraints requirements, right? Because China's ecosystem just at a very young stage. How are you going to address those kind of competition from the U.S. competitors in those perspectives? Also, you know, speaking about that, the competition, whether you can really secure or sustain your largest U.S. customer's revenue into 2024 and 2025. Thank you.
Okay. Yeah. Let me try to answer that first. Yeah. First of all, if you specifically mentioned about Marvell, Broadcom and GUC, I think all of them are very respectable competitors. I think most each of us has some advantages, also disadvantages. Yeah. I don't want to mention too much. I think we are careful to do the competitor analysis. We know where our advantage is, and we know what's the disadvantage for other suppliers. Yeah, we have a good attacking point, try to penetrate in. Yeah. In fact, we are winning a few of them already. Yeah. The most important thing is the price from our supplier.
Yeah, I think that's before that's the biggest drawback from our side because the size compared to us to the U.S. provider are much less. When I talk to the major supplier, I think we can I can say I already get the consensus agreement. If a current business is supporting by us, when we try to approach their next generation and supplier will be fairly provide more competitive pricing to us in order to get a fair competition to the other competitors. Yeah, I'm looking forward to see that happen. I think I already get some agreement from the major supplier. Hopefully, that will happen. Yeah, local support, I think that's very important and also a neutral position.
I always mention about the neutral position, both us and the major supplier doesn't believe you're making a product still compete with your customer. I think some other, most of our competitor, I think they're making their own product. In the meantime, they also make an ASIC. I think little bit conflict of interest are there. I think that that's one of our advantage. We are always maintaining neutral position, never compete with any customer.
I see. Thank you. On this U.S. major accounts projects, do you have additional kind of a advantage that you can secure the project? I think investors are all very curious whether next generation projects is already secured. Can you give us some color?
Mm-hmm. Yeah. Daniel, do you wanna cover this?
Okay, sure. For the next generation, I won't say we will win. You never know. It hasn't had the result yet. Of course, when we do generation by generation project, we have certain advantage from the design perspective. Yeah, I can just say we try our best to win, and we have advantage to win.
Okay. Mm-hmm.
You never know.
Yeah.
Honestly.
Yeah, I know. The good news is, this customer keep prolonging their design. I think. When we talk to them, originally is thinking about the current design, maybe till the middle of 2024. Right now they have a plan to expand, to extend the production period. I can say, next generation, of course we all try our best to win, but since we are the incumbent, I think we are holding a good, biggest advantage to win the next generation. Yeah. Even though we can sacrifice a little bit about the current generation margin in order to secure the next generation. I think this is something we were thinking about.
Thank you. Fair enough. Yeah. Actually, that is going to my next question. Daniel, can you give us some kind of assumption for the future gross margin? Because probably you also need to be more flexible about the turnkey price, right, to secure next generation. Can you give us some guidance for 2023 gross margin level and also operating margin level? Thank you.
Okay. Firstly, I have to say, I'm lazy. I haven't done the number forecast for next year. I can offer you the best guess for next year because the production revenue will go up significantly. The gross margin will go down for sure because the gross margin for our production and the gross margin for our design, there is a relatively large gap between these two types of revenue. My best guess is the gross margin will be at about maybe high 20%, but that's just my best guess. I haven't done the number forecast right now for next year.
Mm-hmm.
Yeah. Overall, I think, Daniel is right. Gross margin will go down a bit. If we're thinking about the top line and net margin.
Yeah.
We have confidence to have significant growth.
Yeah.
Next year.
I have to emphasize here that I knew some investors are watching our gross margin. I'll offer you my view. For the production revenue, any gross margin better for us because we don't need extra human resources or let's say expense to process this revenue. Any gross margin from the production revenue will go directly down to the operating level. No matter how low the gross margin is for the production, it is a plus to our company. When you look back into our financial performance in the past six to eight or even 12 quarters, you may see our gross margin may go up and down. As long as the production revenue goes up, or let's say the whole revenue scale goes up, the operating margin goes up. I think that's my perspective to our business structure.
I see. Thank you. My very last question is about your 5 nm opportunities. Johnny, how many 5 nm projects you're going to see NRE contribution next year? And can you give us some hints about the end application? And, on a strategy wise, I know you guys just focus on a backend design place and routes, and there has been a great winning strategy. But do you kind of rethink this strategy? Meaning, for example, if a future automotive or China customers need some more front-end design supports, would you consider to increase that part of services as well? Thank you.
Okay. Yeah, to answer your question, 5 nm, we already win one. We already win couple. Not one. This morning we just had board meeting. I think our sales report, they are. We have a recent win. Yeah, 3 nm-5 nm design. So we have plenty 5 nm in our portfolio right now in the pipeline. To answer your question for the application, I think most of our applications, they are the same, HPC and AI related. Yeah. In Europe, in the U.S., in North America, in Europe, and also in China.
Hmm. Does it include the automotive HPC?
Oh, no. That particular one is not 5 nm.
Okay.
To answer your second question, yes. In order to win the automotive application, we need some front-end support. Other stuff, HPC related, if they want us to involve for the front-end support, to be honest, I plan to reject. Because myself and chairman always thinking about when you're making the chip, when you're making the system, you need to have a certain differentiation. You need to have put some capability. You cannot just 100% rely on partners to define your spec and architecture. Automotive is a totally different thing. The market already exists. They just are in a transition from a conventional one to the autonomous driving. They need many, many hands. In that area, I think we already have a few partners. Already locating a few partners. We are even thinking about to invest some of them. In order-
Mm.
To maintain our neutral position, I will not own the front-end team. But I try to control and drive a front-end partner, yeah, you know, help each other for us to win. We already have a couple of them. In fact, they already get involved during our pre-sales stage.
I see. That's a great answer. You guys are very diligent, not lazy. Keep going. That's all questions from my side. Thank you.
Thank you, Charlie.
Okay, next one is Qian Qiu from HSBC. You can unmute your speaker.
Oh, hi, Daniel. The first question was really going back to last year, when we had this Chinese VPN problem at that time. I just want you to update. I think at the time we were talking about three 7 nm projects were going on. What happened to them? Okay, I'll just ask this first, then I will follow up with this additional.
Okay. For Phytium, it is pretty simple right now. Everything is really clear right now that Phytium was put on the Entity List. Which means, for the new regulation, at the same time, BIS has a list of 28 companies. For Phytium, right now, their only choice for them is looking for license from BIS, and they will do so.
Mm-hmm. Okay.
Yeah. That's a full no. That's why, along with Phytium and the other 27 companies, they need to follow the tightest regulation applying for license. Yeah, just for your information, right now, fortunately, I think, before October 7th, we already ship out all the Phytium-related. There will be zero inventory in our hands. I think we will not suffer any AP/AR issue from this account, and we encourage this customer to apply for license to continuously produce their next generation.
It's good to know. My second question is. Thank you very much for clarification about this BIS impact on your revenue line. I just want to on the second order, right? Because in the past we did a lot of this paper test chip for Chinese companies using that specs. I think when you do a test chip or NRE project, you do have some expectations about in the future when they come in for the mass production, that number may be reflect in your future's revenue forecast. Will you be able to quantify that if that project is not going to go ahead for production anytime soon?
Mm-hmm.
Okay, for your question, I would suggest you can imagine if you were the customer and you are doing a project and the specs may fall into the class of 3A090, what will you do? You may scale down your spec. Right? From our perspective, we follow the regulation. When customer trying to do the high computing power project, we will ask them to do the ECCN classification first to make sure that their project is not subject to EAR.
Mm-hmm.
Okay. Let me add to that. For China customer, like I mentioned before, Daniel also mentioned, the only one customer, one product line of one customer, may vary the 3A090.
No, no. 3800.
Oh, sorry about that. Most other customers are not even close to that spec. Whatever the preparation, whatever the mask or reticle, we can continue to do the production. Yeah, you mentioned about the test chip. I think we rarely make any test chip for Chinese customers. I think most of our test chips, 3 nm, 5 nm, to prove IP, to prove some I/O speeds for Europe and also for North American customers. Yeah, I think almost 100% of the Chinese customers go for full mask directly.
My last question is about you were talking about the progress in the auto side, right? I remember in the past you were a bit hesitant going into the auto business. You just feel the project length is too long, and also we're taking up too much engineering resources, right? Now we decide to go in those concerns are still valid? And if they are still valid, what will be the impact on the margin?
Okay.
If auto become a bigger part of the business?
Okay. Let me try to answer your question. Before, we are a little bit hesitate to take automotive IC-related business. You are right. Because IC maker, if they try to compete with existing service provider like NVIDIA or AMD, we consider it's very difficult and also risk is much higher. We always mention about provide the ASIC service to chip provider. This business will not be sustained too long because they are making the chip. We helping them to the back end, we helping them to source. We are most of their cost of goods sold. Right now, system company and car maker approaching us, that will be different story. Yeah, that's a pure ASIC. We helping the system company, we helping the car maker to do the ASIC. That's our focus business.
Yeah, the ASIC by definition is helping the system company to do the chip. Recently, the reason we are open the door and start to do more investment, because so many car maker make the plan to do the ASIC. That's our business.
Okay, that's clear now. Yeah. Thank you very much. That's all for me. Yeah.
Thank you.
Okay, because of the time constraint, we take the desktop participants for the questions. Brian , please.
Hi, Daniel. Hi, Johnny. My question is really, well, we've been hearing reports that next year could be a $800 million year for you. I just wanted to first of all confirm that's the sort of number you're thinking of. Secondly, could you just break down a bit where, how do we get from where we are today to next year, and to the extent that we need some of the China business to grow to get there?
Okay. First thing first, we have never said $800 million. The Taiwan Stock Exchange does not allow company to give guidance, the revenue guidance to the outsiders. I can just say that, for this year, actually to us internally, it's kind of disappointing because of the insufficient substrate supply. For next year, we don't see the substrate supply as a bottleneck anymore. Because of that, we expect the production revenue to our U.S. customer to grow maybe three times, four times by comparing with the revenue contribution this year. That's the reason why I told the investors that next year the revenue growth will be quite significant. This project will be a major growth driver for our revenue next year. How high, how big the revenue can be, I would say, the number is really huge, I really don't want to mention it in the earnings conference meeting.
Thank you.
Mm-hmm. Thank you, Brian. I guess that's it. Thank you for your participation. Hope you can support us as always. Thank you.