Alchip Technologies Ltd

TWSE:3661

Earnings Call Transcript 2022-Q4

[Operator]

And this meeting will be English, but [indiscernible] all [indiscernible] in China inventory. So, you need Chinese slides or if you want to ask questions in Chinese, please go to please go to Mark compassionate to download the Chinese version of the slide. And you can also write down your questions through the room message function. We -- and or you can use the last 10 less hand function to restated we will take you more questions in the Q&A session. And this video on the content will upload to mask the cohesion that about 2 hours or 3 hours after this meeting. So let me introduce our CEO, Johnny Shen for the opening and the and his report to the investors.

[Johnny Shen, executive]

Good afternoon, ladies and gentlemen. This is Johnny Shen, President and CEO of Alchip Technologies. Thank you for joining our investor conference meeting. we appreciate the opportunity to share our 2022 Q4 results and also the whole year result and provide future business outlook for this year. In case some of you not go for me with Alchip technologies allow me to provide a brief company update.

Our company founded in 2003 and went public in Taiwan changed in 2014. The current cap count is about 600 people. The majority of our employee engineers for 3 quarters. Central company will being successfully paper more than 480 designs all in leading-edge technology. The revenue number, last year, we are achieving 460. Last year, we are concurrently [indiscernible] than 2s. 80% of our revenue coming from the HPC and AI area, we are one of the TSMC BCA members. Yes, we consider 2022 a reasonable year with a record-breaking number in all categories. Revenue at $460 million; operating income, $77.5 million, net income of $61.5 million contact EPS about TWD 25.7. Alchip growth bit compared to a year before. Theoretically, we couldn't have a much better result without the following invitations. -- capacity yes, from Q1 to Q3, capacity shortage from supplier, especially on the ABF substrate, limited our mass production shipments. -- geopolitical tension, [indiscernible] department, adding more senior and acquire new regulation to China as companies. It does slow down our China business quite a bit.

Fortunately, our customers' future demand remains very strong. Without a competitive delimitation this year, we are expecting tremendous revenue growth from now to 2025. Alchip continuously holds a great position in media engine technology business with 20 tape-outs and many design on last year. We have successfully diversified our business way from China to other regions. In year 2022, our North America revenue contribution already city. -- and majority of 5-nanometer design also coming from U.S. region.

In terms of China leading agent business, we are taking a more cautious approach, working with major foundry partner and IT partners closely. We are carefully reviewing customers' backlog end user and design spec before assessing the projects. We see believing and support China business as soon as they comply with the loan regulations. Another highlight for last year is our penetration for automotive business. The area we are focused on is ADAS L2-L4 type of ASIC application -- we truly believe the most high-end car maker will need to provide driving assistant or even FSD solution in the future. The outstanding car maker will develop their own ASIC solution to differentiate them to others. In addition to high-growth PCM AI areas, automotive application will be another key business driver for us starting from May 2024.

Furthermore, the automotive business has a lower chance to be affected by geopolitical factors. So major suppliers and partners are very positive and supportive for this application. Last but not least, I would like to emphasize Alchip's mutual position and diversified business condition again, similar to our foundry partner, El will never make a product to compete with any customer. We have a fully diversified and well-balanced business from all regions. In terms of headcount, we increased quite a bit recently. In addition to China, we also implemented a very aggressive retiring path to increase our engineering headcount in Japan, Taiwan and United States regions. We also implemented -- also plan to open additional offices in Southeast Asia. This effort will provide more sufficient and cost-effective solutions to meet our customers' requirements. Overall, we have confidence to say 2023 will be another outstanding and record-breaking year for Alchip. Thank you.

[Daniel Wang, executive]

Okay. The following session I will introduce the fourth quarter financial results and the breakdowns of Alchip and of course, our business outlook. For the fourth quarter last year, the total revenue, you may already know, we already published numbers already. So, the total revenue for forma last year is $148.3 million, which is 6.2% for the quarter and 66% year-on-year growth. And operating income, the operating income first quarter last year is $20 million, which is 7.5% quarter-on-quarter growth and 34.7% year-on-year growth. And the net income for [indiscernible] year is USD 16.4 million, translating into EPS of TWD 7.16 and for the yearly LL, the total revenue was 60.5% as just mentioned, which is 22.7% year-on-year. And operating income, 77.5%, which is 18.8% year-on-year growth -- and net income is 61.5% USD 61.5 million, which is 15.6% year-on-year growth.

And the full year ES is $25.7 -- for the revenue breakup breakdown by application. You may see, again, the HPC remains the majority competitor of our total revenue. For the fourth quarter last year, the HPC redated revenue accounts for 78% of our total revenue, while we niche market accounts for 7% networking unrelated auction accounted for 9% and 6% is for the consumer product. And for 2020, I'm sorry you may be using the last pilot. Well, tension fire. So, we will share the correct one. Sorry. My number is sake. It's kind of embarrassing. And for the fourth quarter last year for the fourth quarter last year HPC accounts for 85% of our total revenue. And while the mixed market accounted for 8% and the networking 3% and the consumer 3%.

And therefore, 2022, the whole year HPC accounted for 82% of our total revenue, while the others made up by consumer networking in the niche market. And for the imposes technology, I think we are still leaders about our industry. 71% of our total revenue came from 7-nanometer or best margin notes. The 16 accounted for 16% of our total revenue in last quarter and the other -- the others made up the rest. And for 2022, as a whole, the 7-nanometer or more of the best technology not related to revenue accounted for 68% part of the revenue by 16-nanometer 12-nanometer accounted for 20% of the total revenue last year.

For the regional breakdown, I knew that many of you are pretty concerned about our China exposure. So, in order to make it consistent, so we didn't separate the China into a single category. I will explain it later. So, over the last quarter 2022, Japan revenue from Japan accounted for 8% of our total revenue in Asia Pacific, which includes China and Taiwan accounted for 34% out of the revenue. And revenue from North America accounted for 45% of the revenue, while the others include the Middle East and Europe accounted for 13% of our total value last quarter. And for 2022, the whole year, the Japan accounted for 14% and Asia Pacific accounted for 38%.

North America accounted for 39%, while the others accounted 29%. Among the Asia Pacific, the total revenue exposure to China accounted for about 27% of our total revenue last year. And for our 2022 review, as mentioned, our sales, to be honestly lower than our expectation, given the supply shortage of mainly [indiscernible], so which cause us the shipment to our the AI human to our North American customers gradually impacted by this supply issue. But even so the company, -- as still managed to deliver record-breaking target bottom line and strong NI demand for both North America and the China region. Although the China IC industry was impacted by the geopolitical issue and of course, the EIS section list. But the demand -- the design demand from this region is still strong.

The NI revenue last year accounts for around 14% to 45%, our total revenue at investment was the production orders and the ASIC revenue. For the gross margin, I know many of you also pay a lot of attention to the gross margin. For 2022, the blended gross margin was 32.3%, which is a little bit lower comparing to 34.2% in 2021. And for the fourth quarter, the reason of our relatively low gross margin in the last quarter, the main reason because, first of all, we have pretty high percentage exposure to the production revenue in the last quarter. The production revenue last quarter exceed 60% of our total revenue. And for the NI in the fourth quarter last year, there are some big milestones, delayed or slipped to 2020 -- to this year 2023. And we don't see great risk for those milestones. It is just a normal business operation that we try to cooperate with our customers for the design for the design process.

For the operating expense, we are controlled. We are controlling it very well at $71.1 million for last year. And lit is slightly lower than the guidance, my previous guidance to the investors. So for this year, we still believe the operating expense on LG will be under control. For the business outlook, as I mentioned, HPC demand from North American market grows even harder. As currently is engaging in multiple AI-related projects from North America hyper-scalers and I would say the demand for AI application of U.S. customers came up increase for both NI and production.

We expect triple-digit growth for the ag revenue this year to North American customers. And the fiscal design tank demand for HPC gets stronger and stronger. The project pipeline is tanker, [ Averna ], -- in addition to the HPC and AI related area, we expect the automotive-related MI could be the seed for our -- another future worldlier. The concession migration kits are moving for North American market, as John mentioned, the majority of our financial meter or even 3-nanometer projects in our project pipeline coming from North coming from the North America region.

However, this migration, the process now operation is kind of slowing that in China because of the geopolitical reasons. And for this year, in addition to the numbers of P&L, another great task for us is to do the diversification, not only for the sales but also for our exit engineering resources. The company has managed to shave business focus to North America. Even we expect this year, we lean our China sales exposure will probably less than 20% of the total revenue. And the revenue to North American market will account for more than 50% this year.

And as I mentioned many times to the investors that we are massively diversifying our time engineering resource. We are expanding our Japan team and we are extending our time line team, and we are also -- we are planning -- we are doing the expansion in Southeast the Singapore and Malaysia. So, the company has already secured the engineering resource outside in Malaysia. And meanwhile, we will gradually build up our own team there. And I think those conclude our 2023 business outlook. So -- and the following is the Q&A session, please use the raise and button or write down your questions to the message function to us.

[Daniel Wang, executive]

Thank you. Charlie Morgan Stanley, please. So first of all, congrats again for your great results. And I think there should be some common questions from investors, right?

[Charlie Chan, analyst]

First of all, your next-generation AIS-related project with your big U.S. customer. Can you please update the timing for the decision? And what are the critical factors to win these projects. And if you win the project, when would be the timing for the 90 revenue and also the turnkey revenue. I hope that's clear for you.

[Daniel Wang, executive]

Okay. Charlie, to answer your question. For the business to the -- our current North America hyperscalers, the next generation of our 7-nanometer chip [indiscernible], vendor selection to most likely to happen in the late second quarter and to kick off the third quarter this year, that's the current schedule we receive at our customers. And we honestly actually is pretty confident that we should be ahead of our competitors for the last generation for the next generation is not -- and because this project is scheduled to kick off in the third quarter this year, most likely, we may not see the production to happen before 2026. So, because of that, the current 7-nanometer intrachip life cycle were extended to 2025 for sure. That's the current schedule for the most important project of Alchip.

[Johnny Shen, executive]

Let me add to a little bit. Yes, just like Daniel mentioned decision-making is keep delaying, I think, due to many reasons. As we all know, most many North American customers are doing a lot of cutting, so the budget car and also resource cutting. So, the advantage for us is like the current product will be colon -- and we have -- we consider we have a higher chance, more and more chance to do the well preparation and winning the next side. We are currently -- we've been working very closely to this specific customer, try to meet all their requirement. I think since we are incumbent for to generation already, we still believe the chance for us to win the next generation will be very high.

[Charlie Chan, analyst]

Okay. Got it. Yes. So, some investors are concerned that the delay of a decision may suggest some potential risk, but I think you explained that quite clearly. And next common question, I think could be the so-called the generative AI whether it is high not? I hope Eli can explain from your company's perspective, number one, whether there is any concrete projects for generative AI. And secondly, if there are any projects when and whether LC can really enjoy some project wins.

[Johnny Shen, executive]

Okay. Let me try to answer that -- to be straight, LT has a very high relationship to the AI application. A few years ago, when AI just both influence and training introduced to the market, we consider we are one of the winner, no matter the project from Japan, China or U.S. we win a lot. But unfortunately, after we're taping out the project, there is not too much production. To be honest, it's a little bit disappointed. But recently, the way we change quite a bit, we've been keep receiving more inquiry from our existing customers asking about the latest wafer price and potentially, they are planning to place a big order to us.

In addition to that, the -- they are talking about the new customer and also the existing customer talking about the next generation. The recent [indiscernible] become very popular. Most of our customers consider layer solutions are even better by any way, now we are waiting for the new application booming in the past, as long as any application has a more user require more compute power, Eventually, that will be the ACE opportunity. I think similar case has been proven over and over again. some as AI starts booming, I think Alchip for sure, is one of the winners. If you pay attention for even our slogan, we consider we are still part of the Ai.

[Charlie Chan, analyst]

Okay. Okay.

[Johnny Shen, executive]

One question for the message. Brian asked, there is some rumor that for the current generation 7-million to the North American hyperscalers may be at risk of being in-sourced. Can you please clear that? Is the rumor under planned Alchip should we already have signed the contract with the customer for the current generation. So, it is up to the company to decide how much to ship in the next few years since.

For your question, it is possible for the current generation 7-nanometer ACI, the contracts relate we were providing, and we have been providing the turnkey service to our customer. So, the business model is big for the shipment for the scheduled shipment, it is they are all caution by our customers and lip. So, I don't know why the Loomer comes out. But to me, it is impossible for the current generation that our customer to take back the production.

[Daniel Wang, executive]

And next question is from Jeffrey, I think. Jeffrey, please.

[Jeffrey Ohlweiler, analyst]

Can you give us any kind of numbers on your #1 client last year as a percent of sales and maybe this year forecast and maybe for #2 client as well?

[Johnny Shen, executive]

Okay. I would say for last year, the #1 customer of us accounted for about close to 30%, close to 30% of our total revenue. Back to you.

[Jeffrey Ohlweiler, analyst]

Any thoughts on this year? This year will be like...

[Johnny Shen, executive]

I would say, 40%, 50%? Maybe...

[Jeffrey Ohlweiler, analyst]

Okay. And how about your next biggest customer, roughly this year and last year.

[Johnny Shen, executive]

You mean the largest. Yes, the second largest IP lease last year should be around like 10%, high single digit. But for this year, I have to check digital numbers. I don't have it right now. I'll get back to you later.

[Daniel Wang, executive]

Yes. Gabe, I think the #1 customer are much higher than the #2. But between #2 to #5 is much closer. So, we don't have a clear #2 customer. Okay. And then last question before I let us jump in, can you talk a little bit more about the automobile related, whether it's you can give us more info on geography and type of scale maybe potentially?

[Jeffrey Ohlweiler, analyst]

Okay. For automotive-related projects, we already booked several projects not only on several projects almost all the major reasons like North America, China and Europe, indirectly or directly to...

[Johnny Shen, executive]

Thank you, Jeffrey. Charlie, please. Okay?

[Charlie Chan, analyst]

Sure. since or answering my previous question and also just wanted to follow up Jeffrey's question about automotive. I just wanted to get a sense about your strategy about the [indiscernible]. I know the company very, very focused on bake-in defense service, right? But I remember also mentioned that for automotive customers because it's kind of a new territory, right? So how do you work with customers to provide the frontiers?

[Johnny Shen, executive]

Okay. Yes. Yes, Charlie, you are right. I think for automotive, I think we can say it's totally different compared to other SoC design. Most of the carmaker it additional support from us, are sometimes not even from in, but even also including the software or even the EMS, PCB, we already built up many partnerships to front-end partners. Yet in fact, we are sometimes for a few particular one we are even invest on iron in service company. But within Alchip, I always considered the usual position and independent position is very important.

We were not doing architecture-related design for any customer, but we were using many partners because very difficult for me to believe I can design customer wants architect not leverage to the customer, too. So independent position will be very important, just like I mentioned, I never make a product to compete with my customer. So, working closely to the partner complete front-end solution. But front and, if you take into -- if you look into -- from and deeply, there will be a lot of category.

We don't touch architects, but we can help on the verification we can have on the DC generation or [indiscernible], those kind of work, we shall consider as independent. So yes, to answer your question, we already built up many partnerships in order to provide a more comprehensive solution to the car maker. And different car maker has a different requirement, for example, the key company from U.S., they pretty much have all the team member ready.

They need to design, and they need a service from us is, I think, is very straightforward. But for many car makers, especially in China, they never make a chip before. They need a lot of support from us. Yes. So, through the automotive so we can say we expand our service scope by a bit and touch some from and partner and from an IP and also on the software side and also on the PC side.

[Charlie Chan, analyst]

Got it. And given the very lenticular for crore industry, when do you think that can contribute meaningful revenue. Can you quantify, for example, real contribution percentage from automotive HPC in this year and maybe 3 years down the road.

[Daniel Wang, executive]

Okay. For this year and next year, we don't expect a very high percentage contribution to our total revenue because they are all MREs, although the 9 may be much bigger than the ordinary project, being our sales scale to reach our sales care could grow quite a bit this year. Now any given MI loan will contribute a lot of the percentage of our total revenue. It is a fact. So if you are talking about significant revenue contribution or let's say revenue driver, I would say, starting for 2025.

We estimate per chip, if the chip is used in the car, the achieved revenue contribution per project for automotive will be ranging from $50 million to $80 million kind of yearly kind of this kind of pitched contribution. You are talking about 2023, 2024, we expect combined with these 2 years per project, the NI would be about less 7-nanometer, the reticent to Volition.

[Johnny Shen, executive]

Yes. Just like Daniel mentioned about -- if we have the dependence -- revenue dependency on production, I think most likely it won't happen until late 2024, manning 2025. But Ana, I think that's a cost and comp situation. The from the automotive total NRE will be much higher than the conventional SoC design HPC. But automotive design period will be longer if we can take out on HPC design within a year, like 8 months also, but automotive, the store, I think, bigger combined front-end and back-end and many superiority. So the design cycle will be longer, at least 1.5 years for cheap.

[Charlie Chan, analyst]

Got it. So just to clarify, the USD 50 million to USD 80 million per project, you said NRE revenue or is the entire life cycle Yes. If the scope includes the fan IT Oh, I see.

[Johnny Shen, executive]

Okay. Yes, I got other questions, but I will let other ask back to the queue first. Okay. Jeffrey, please. Yes. I guess you also mentioned ABF shortages last year impacted your shipments. And also, earlier in the year, you also talked about the yield rates for ABF were also lower than expected. Has that changed over a few quarters now in terms of yield rates?

[Daniel Wang, executive]

Yes, the supply situation right now is much, much, much better than the same time last year. For this year, we will have 3 sees suppliers. And year-on-year is quite stable, not very high, but quite stable. And we have room to shifting or balancing the orders to different suppliers this year. So, the situation is totally different for the subject side. For the cobalt side, cars TSMC or [indiscernible] us a full commitment to this year's demand of ours. And the urine for the cohort right now is pretty high. It's pretty much is more than satisfied for that.

So for this year, I won't say there will be no other. For example, there may be minor collections, for example, in February, the revenue, the reason for the revenue posted month-on-month declines because our customers to revise this testing program the shipment mix in February will shift at the March or April. So the situation is nitens no longer the supply shortage issue for this product.

[Jeffrey Ohlweiler, analyst]

Okay. And maybe related to that, whether it's ABF substrate, Founder or anything else, any cost drops that are helping margins or...

[Daniel Wang, executive]

To be honest, this project is its cost plus margin pricing scheme. So the higher the cost, actually, we may benefit. Okay. And then the last question. Sorry. Right now, there is, we have this room for negotiating better pricing, but is all in progress, right?

[Jeffrey Ohlweiler, analyst]

Okay. And one more question for I jump off back in line. HBM, just had with memory. Can you talk a little bit about projects there and then how that's going...

[Daniel Wang, executive]

I'm not so sure about your question for the B. Actually, since we are focusing the PC area, the majority of our new project right now within financial meter, more even significant parts of our 7-nanometer projects are all using HBM.

Okay. Great. Thank you. Thank you. Chan, please.

[Charlie Chan, analyst]

Sure. Yes. So maybe I switch gear to the China business. First of all, I think people also care about the HPC business development, right, given ongoing on the restriction, right, but there's interest events. So whether your foundry supplier changed their stance or business strategy to engage with the Chinese customers, especially for the HPC side? And how that impacts your 2023 outlook?

[Daniel Wang, executive]

Okay. Yes. Like I mentioned before, every time when we interface with a new customer, we are working with the foundry partner very closely. There will be kind of SOP-related checking, check the company's background, their end user and also the application, make sure there's no valuation to the current regulation. Yes, once everything goes through, to be honest, I think the foundry partners still very supportive on the China business.

And also, we still believing a lot of business will provide a certain volume. So, I think to answer your question, yes, it will be more effort to do the corporation checking and also the post desire ECCN kind of applying yes. But once we go through this listing, I think the reasons itself is still very sexy. -- the percentage of NIE and gross margin from China customer is still higher than the United States, as long as they comply with all the regulation. I think we still are willing to reserve a certain resource to take this business.

[Charlie Chan, analyst]

Okay. So in general, because there is already some criteria, right, regarding the computing power for those GPU chip, right? So do you feel like does the Chinese consumer becomes more passive, right? Because if there is such kind of restriction, right, they would never compete with the foreign suppliers or opposes, right? That becomes even much more aggressive for the HPC desire.

[Johnny Shen, executive]

Okay. Can let me answer your question. I would say the BIS new regulation restriction on the tax, actually, now for China IC makers, they won't try to risk their future to go beyond the limitation. Otherwise, the TSMC will support the project. So that's the reality. But if you are thinking about competition, I would say the China localizing market, there are some marketplace be launched to the government purchase. I would say it's quite special.

The government purchase market in China is relatively much bigger than the other countries in the world. And it is -- for this market, of course, you can overstate not a fair 5. They will choose the local suppliers first, measure the performance will be -- I guess that's the current reality. Those IC makers will still have some room for them to survive. I think there are no fixed restriction for that on IOP and total compute power and so does their competitors. Yes, so on China still has a market. I think there will be a few winners.

[Charlie Chan, analyst]

Okay. Okay. Got you. But I think last year, you also mentioned some opportunities in a more kind of consumer type, right, like a speaker or ISP for smartphones, supposedly, it was a big growth drivers, right? But what's the status for this the consumer China projects?

[Johnny Shen, executive]

Okay. This project is still ongoing, but the reason why it is delayed is because actually, you -- some of you may know that this project is a company which is a joint venture between 2 different companies, latus wishing ownership during the past days. So right now, the onesies switching to one of the owners. So the project will resume and we are used for some major near devices.

[Daniel Wang, executive]

Yes. Just for your nonproject already paper, we expect to get the city [indiscernible] soon. Hopefully, we are working with Japan. The are joint venture company to do the order in the future. Yes. To answer your question, we still have a high expectation on that area because everything is proven and also the customer already have a certain market share.

[Charlie Chan, analyst]

Okay. Great. Yes. So lastly, from my really just the financial guidance. So maybe, Daniel, can you give us some guidance about top line gross margin, especially OpEx, right? You're increasing the resource -- can you some...

[Daniel Wang, executive]

Actually, by regulation, I'm not allowed to do the numerous guidance, but I would say this way, I knew that some research reports has a revenue forecast for us this year for this year, ranging 700-million to $800,000 million where we were grade even more aggressively. To us, we will say $800 million to us this year is not a very difficult target to achieve. That's for the revenue on. And our gross margin, yes, the increasing revenue exposure to production, we address that our gross margin, but we still believe we can deliver high 20s percentage gross margin this year, even we -- the production revenue grew significantly this year. And therefore, operating for the operating expense, OpEx. Last year, as I mentioned, I guided to you last year, it's about 70% to 73% and the actual number turns out to -- for this year, we are targeting $85 million, $84 million, $85 million for the whole year. And I think we have really high confidence to achieve this target.

[Charlie Chan, analyst]

Okay. That's super helpful. Yes. So actually, when I look at this number, especially OpEx, right? 2 questions came to me. First of all, how come you spending overseas operation, but the cost increase just very minimal. And maybe second question to Johnny, right? If you look at your business model, you only spent like less than USD 100 million and can play such an important role, right, for the future as a custom chip design. So why there are many much bigger companies than Alchip, right? They can spend USD 100 million easily. Why can duplicate this business model nominator internally or externally?

[Johnny Shen, executive]

Because we are simple efficient. Okay. Anyway, that's our guess would be -- to be honest, I think LT is a pure independent company. We don't have any so-called rich bothersome other. We're doing everything on our own. So control the cost to provide the more efficient service to our customer. I think that's our company is building that, I think, for the past 20 years. OpEx is a key for service company. Yes, there's many ways to control the OpEx. I don't need to mention too much, I think, put the right people on the right position and try to utilize more resource on their -- from the junior engineer and increased the senior junior ratio and balance from all region and make a customer satisfaction.

I think those kind of things you can control OpEx and also put the incentive on the future do not control the salary and cost and metal expectation on the stock option, those kind of stuff. I think we'll be improving quite well overall, even though our OpEx is low, but if you're thinking about the turnover rate compared to most of competitors, we are much lower, the executive level, almost 0 turnover and management level, I think the turnover is very legal Hopefully, that answered your question.

[Daniel Wang, executive]

Okay. And Chad, for your other question, the expansion of our human resource this year is kind of flat we are committed to expand our overseas design resource for sure. But in China, we may do our resource plan very conservative. That's the reason for the OpEx. And...

[Charlie Chan, analyst]

Yes. So maybe just to follow-up what TSMC is providing to investors -- if you compare, I don't know, right, the engineered cost overseas versus in China, what was the cost gap, right? So, we can do some calculation by our own.

[Johnny Shen, executive]

Right. Yes, the area we try to build out the resources from our existing subsidiary, Japan and Taiwan. And also we mentioned about the Southeast Asia, Malaysia or Vietnam. Yes, honestly, this full region compared to China, the costs are even lower. It's not like our foundry partners try to get the resources from North America. I think that's a probably different story. China is not -- honestly, China is not a cheap base. Even we diversify that China is a resource try to open more subsidiaries in other regions. I think the overall cost will not increase significantly. There might be some onetime costs when we build out a cloud server and also the office. But eventually, I think the costs are even more stable.

[Charlie Chan, analyst]

Yes. Yes. So, this question comes from one of the investors before, right? The question is you get us some subsidies from China governments because big you had a very big operation in China, right? But now if you are expanding outside of China, can you still enjoy those non-attach parties or government subsidies from China?

[Daniel Wang, executive]

Yes. To answer your question, yes, we do receive some subsidy from China, but honestly, it's not so significant mind we still continuously receiving this support from the local government. It doesn't change. It's not -- the support is not related to the headcount. Yes, related to the business in for, especially for the leading-edge technology.

[Charlie Chan, analyst]

I see, I see. Okay. That's all my questions. Those are very helpful.

[Daniel Wang, executive]

Thank you. Okay. Investors adds from message that given your biggest North American customer also does business directly with TSMC. So our chief point of view, how does these customers decide what projects to outsource to the [indiscernible] vendors? Is it cost [indiscernible] or something else?

I'll answer the question first, and maybe Johnny can add [indiscernible] one. First of all, yes, our biggest North American customer this direct business with TSMC. But as I mentioned many times to investors, I would say, is Altair customer is kind of a partnership relationship. Our customer has its own in-house reengineering resource and how to allocate their internal resources efficiently, we believe by partnering with the outport by our customers is the right direction. And the things has been proven. It is a good choice. And as for you are asking about is it the cost of other IP or vote or something else. I think it is -- it is a consideration and decision for many, many reasons, not only cost or IP or modern, I would say, a have a lot of time. The -- many, many things, it's not decided by some specific reasons for our customers must outsource or in-sourced.

Okay. And for another question from message. Could you please give us a sense of how gross margin will trend this year as well as how much OpEx will grow trends?

I think I just gave the guidance to you that we are still targeting high 20s for the gross margin this year. and questions from investors, can you please also talk about the other potential opportunities after the other North American type skills and what are the applications, DPU, CPU or AI accelerators. Yes. In additional to our current #1 customer, we have -- we are approaching to many others, all of them are Tier 1 hyperscaler kind of a customer and the application, they are pursuing, I think, quite similar, is AI and HPC related. That in fact, for a few of them, we're already stepping to the last stage, the diversified the concentration from a single customer is our -- one of our goals. Starting from next year, I hope there will be more company will contribute more significant revenues in addition to our Tier 1 customer.

[Johnny Shen, executive]

Okay. I'll say it this way. We are engaging in discussions, we cannot disclose the details on the age negotiations or discussions at. I can tell you that we are engaging in project discussions with all what was all the major hyperscalers in the North American market.

[Unknown Analyst]

[indiscernible], please. Just want to check some details. Can you just mention about hyperscale, the second generation product. You just said you will kick off in the third quarter and contributing to revenue in 2026. Is that right?

[Johnny Shen, executive]

Not -- I would say that's the schedule. I'm not saying -- at kick off the project, the vendor select is not deciding yet.

[Unknown Analyst]

So will it contribute to our revenue in next year?

[Johnny Shen, executive]

We don't know. The revenue But as I mentioned, the vendor selection, which means, we are not winning the project yet. So, you are asking about if there is revenue contribution, I cannot answer your discussion. Yes.

[Unknown Analyst]

Okay. And so about automotive customer and what's the contribution in the next year? Is there any turnkey or a contribution in the next year?

[Johnny Shen, executive]

Is, we already have the revenues in automotive-related projects already.

[Unknown Analyst]

Okay. Sorry about that. I just mentioned about the United States, the Tier 1 customer, will you contribute to any revenue contribution in the next year?

[Daniel Wang, executive]

Tier what, I really don't know who are you referring to?

[Unknown Analyst]

Okay. The EV customer.

[Johnny Shen, executive]

We have multiple EV customers, and we all consider they are Tier 1 papers.

Questions for message. In the beginning, Jon mentioned, there will be FSD-related ship for auto customers is [indiscernible], specific client or it is just a general function for various auto clients I may say this way, the FSU is considered basic. It's not a general chip be used by different automates. As at least for now, we are -- our FSD-related projects is almost all C type of business, right?

[Daniel Wang, executive]

Yes, what I mentioned is because the solution we are providing either is -- just we are driving assistant and by some of the customer, I think, is presuming L3 to L4 is full self-drive this kind of function.

[Johnny Shen, executive]

Okay. It seems we already mentioned, we already answered all the questions. Is there any questions you can raise -- you can use the recent or the message [indiscernible].

Okay. If you thank you for today's participate and see you next time.

[Daniel Wang, executive]

Thank you very much. Thank you for your support. Thank you very much.