Hesai Group (HSAI) Q1 2023 Earnings Report, Transcript and Summary

Hello, ladies and gentlemen. Thank you for standing by for Hesai Group First Quarter 2023 Earnings Conference Call. [Operator Instructions] Please note that today's conference call is being recorded. I would now like to turn the call over to our first speaker today, Yuanting Shi, the company's Investor Relations Director. Please go ahead.

Thank you, operator. Hello, everyone, and thank you for joining Hesai Group's First Quarter 2023 Earnings Conference Call. Our earnings release is now available on our IR website, investor.hesaitech.com as well as via newswire services. Today, you will hear from our CEO, Dr. David Li, who will start the call with an overview of our recent updates. Next, our Global CFO, Mr. Louis Hsieh, will address our financial results before we open the call for questions. Before we continue, I refer you to our safe harbor statement in our earnings press release which applies to this call as we will make forward-looking statements. Please also note that the company will discuss non-GAAP measures today, which are more thoroughly explained and reconciled to the most comparable measures reported on the GAAP in our earnings release and SEC filings. With that, I'm pleased to turn over the call to our CEO, Dr. Dave Li. David, please go ahead.

Thank you, Yuanting, and thank you, everyone, for joining our call today. Let's start with the big picture. We're pleased to deliver the strongest quarterly financial performance in our history. In the first quarter, our net revenues were up 73% year-over-year to a new record high, and our industry-leading gross margins increased quarter-over-quarter to 38%, up from 30% last quarter. Expense-wise, excluding share-based compensation, R&D remains steady and our product lines and organizational structure are approaching maturity. As a result, and most excitingly, we achieved non-GAAP profitability and positive operating cash flow in the first quarter, demonstrating our tremendous future profitability prospects. Mobilization is more encouraging than the recognition from our customers. Having delivered more than 135,000 LiDAR units to date, Hesai has become the most commercially successful LiDAR manufacturer in terms of shipments in both Robotaxi and ADAS market. We stand out from others for two reasons. First, our proven automotive-grade product quality built upon our methodology of having in-house manufacturing as a crucial part of the R&D to enable rapid iteration of our product. And second, in addition to taking advantage of economics of scale, we have committed to continued innovation on our proprietary ASIC platform, allowing more affordable solutions via highly integrated components. Since 2017, we've gone through 4 generations of our proprietary ASIC development with contributions from more than 150 dedicated semiconductor staff. We're proud to be uniquely positioned to offer LiDAR solutions with better combination of high performance, high-quality and affordability. Now I'd like to share some exciting first quarter business updates. First, we're thrilled to have the privilege of collaborating with 11 pioneering OEMs in the industry, a multiyear ADAS contract, including Li Auto, a leading Chinese EV maker, as well as another top-selling EV maker in China. We're also working with a leading consumer electronics maker in China on their planned EV debut and 2 of the largest automakers in China, including Taiwan and many other leading auto OEMs. In the first quarter, we continued to gain momentum in signing more contracts in the domestic market. We were selected by Li Auto for its new fully battery electric vehicle platform and by Jidu, a joint venture between Baidu and Geely Group, for its new EV model. We also initiated a partnership with a new customer, SERES, a leading EV maker in China, whose new car models will be shipped with our AT128 LiDAR. As the China EV market is growing exponentially, we'll be shipping to 6 OEMs by end of 2023 and 11 by end of 2024. Hesai has a global vision since its inception. With our proven ability to ship large volumes of high-quality ADAS sensors, we already advanced in 8 RFI and RFQs with 5 leading global OEMs who are expecting a new round of sourcing decisions to materialize in the next coming 12 months. In the autonomous mobility market, we continue to lead as the best-in-class player. We recently signed the largest Robotaxi contract in our company's history. According to an independent third-party report, we are the primary LiDAR provider for 12 out of the world's top 15 autonomous driving companies with nearly 60% global market share. Next, we're dedicated to in-house manufacturing as we see this as the only way to meet automotive-grade standards at this early stage of development in ADAS market. During the first quarter, we completed the first successful production sample from our Hertz center, a dedicated manufacturing center in Hangzhou, which takes automation to the next level. Featuring over 90% automation, Hertz produces 1 LiDAR every 45 seconds. With an optimized design, the AT series produced at Hertz would improve our ADAS gross margin as it reaches the economy of scale in manufacturing and utilization. Additionally, our advanced R&D and intelligent manufacturing center in Shanghai, Maxwell, is a true innovation center for new products to be designed, tested, calibrated under one roof before they are handed over to mass production. Its construction is scheduled to be completed in Q3 of this year. Leveraging the development of our latest ASIC Gen 4.0, we're able to continue releasing state-of-the-art products that meet the ever-increasing needs of both performance and design aesthetics while keeping them affordable. ET25, a groundbreaking in-cabinet LiDAR, featuring an extremely thin profile of only 25 millimeter in height with extended range and enhanced resolution, was recently released at the Shanghai Auto Show. ET25's revolutionary concept is attracting interest from OEMs globally, who are demanding extremely miniaturized LiDAR behind the windshield to achieve both an aesthetically pleasing exterior vehicle design and optimal air dynamics. To conquer the great challenges of putting LiDARs behind the windshield, we team up with global automotive glass leader, Fuyao Group, to create a specialized windshield to enhance the laser transmittance efficiency by up to 3x. Finally, before I hand it over to Louis for a deeper look at the financial numbers, I wanted to congratulate the whole team from operations to product innovation on an exceptional quarter. With non-GAAP profitability and positive operating cash flow in the first quarter, which showcased our potential for continued growth momentum, we remain confident in our ability to seize development opportunities in both ADAS and the AM market, bringing a higher level of safety to the future of intelligence driving system. Now I'll turn over the call to Louis to share more details on our financial performance and outlook. Louis, please go ahead.

Thank you, David, and hello, everyone. We further demonstrated our commercial success with a stellar set of financial and operational numbers in the first quarter. Be mindful of the length of our earnings call today. I encourage listeners to refer to the first quarter earnings press release for further details. We achieved record net revenues of RMB 429.9 million, USD 62.6 million, up 73% year-over-year and 5.1% quarter-over-quarter. Product revenues grew even faster, increasing 77.7% year-over-year to RMB 424.1 million, USD 61.8 million, during the quarter. Gross margin was 37.8% for the first quarter, compared with 50.9% for the same period of 2022 and 30% for the fourth quarter of 2022 as our revenue continues to shift from higher-margin autonomous mobility products to higher volume ADAS sales. Although we were pleased with this quarter's gross margin, we do, however, anticipate a lower gross margin for Q2 and Q3 due to a larger -- large volume but low margin order from a leading Robotaxi OEM. We anticipate gross margins will rebound in Q4 as the Hertz center production line kicks in with anticipated higher gross margins with updated ADAS products shipping and volume. At this early stage, we are willing to sacrifice some gross margin in the short term in exchange for market share gains in the long term in ADAS and to some degree in the autonomous mobility as well. The ADAS and AM LiDAR markets are at a nascent stage of development. And at this critical point, we are prioritizing market share expansion, but we are still targeting blended gross margins to stabilize in between 30% and 35% in the long term. First quarter deliveries totaled close to 35,000 units, an increase of over 400% year-over-year, but a slight decline from Q4 due to seasonal factors. We strategically pulled forward some volume in December to mitigate production constraints and manufacturing slowdown in Q1 due to the Chinese New Year holiday. AT deliveries accelerated in March and April, and we anticipate a record second quarter of ADAS and AM deliveries to reach almost 50,000 units, and increase of 40% quarter-over-quarter, and a tenfold volume increase year-over-year, which would represent a record high quarterly LiDAR volume for Hesai. In the first quarter, we achieved non-GAAP net income of RMB 1.6 million, USD 0.2 million. Excluding share-based compensation expense of RMB 120.5 million, USD 17.5 million, this increase in share-based compensation expense was mostly attributable to the recognition of one-off expense of RMB 90.9 million, USD 13.2 million, upon the completion of our successful IPO in the quarter in relation to the stock option granted with an IPO trigger condition. Before discussing Hesai's 2023 outlook, we would like to reiterate the statements in our April 17, 2023, press release responding to Ouster's IP infringement lawsuits against Hesai. We believe Ouster's claims are deeply flawed and lack merit. Hesai's independently developed LiDAR technology is the result of years of investment in research, development and engineering. Hesai disputes Ouster's allegations of patent infringement and we'll vigorously defend ourselves against such allegations. Since the lawsuits are pending, upon the advice of counsel, we will not comment further or take any questions on this matter. Now moving to the financial outlook. For the second quarter of 2023, the company expects our net revenue to be between RMB 410 million, USD 59.7 million, and RMB 430 million, USD 62.6 million, a year-over-year increase of approximately 94.3% to 103.8%. This represents an expected doubling in Q2 2023 revenues year-over-year and an expected tenfold increase year-over-year in LiDAR shipments to almost 50,000 units, demonstrating our continued optimization in operational execution and strong market demand for our products. It should be noted, however, that last year's Q2 was negatively impacted by the COVID lockdown in Shanghai. The above outlook is based on the current market conditions and reflects the company's preliminary estimates of market and operational conditions and customer demand, which are also subject to change. In closing, it should be clear to you, investors, by now that Hesai is the undisputed global leader in the ADAS and autonomous mobility LiDAR market. In fact, we continue to easily outpace the competition, easily delivering higher annual and quarterly revenues in LiDAR shipments than our 6 other U.S.-listed LiDAR peers combined. This concludes our prepared remarks for today. Operator, we are now ready to take questions.

[Operator Instructions] Your first question today comes from Tim from Morgan Stanley.

Congratulations on the strong first quarter results. So I have two questions. The first question is about our gross margin guidance because during the presentation, I think Louis mentioned second quarter, third quarter margin would be lower despite a record volume sales due to the product mix. So can we have the rough breakdown between the comps driving the mechanical LiDAR and the ADAS [indiscernible] in the first quarter, in the second quarter? And separately, when are we going to see the inflection point of more meaningful gross margin improvement, as Louis mentioned, back to like 35% or above? So that's my first question.

Thank you, Tim. I'll take this one. For the gross margin for Q1 on AM versus ADAS, we are no longer disclosing the individual breakdown for competitive reasons. You can imagine that our customers see the same information and use it against us in negotiations. So we are now disclosing a blended gross margin for our company. And as we noted earlier, the 37.8%, up 180 basis points from the prior quarter. So we will no longer be breaking down each individual product line for gross margin. For the trend, Q2, as you know, we have disclosed and Q3 will be slightly lower because of a large -- the largest order in our history, for Robotaxi LiDAR. And so that one, we conceded a lower price to get the volume. The total contract is approximately USD 200 million. So it's a huge contract. So it's worth it to us to do that. We do believe the inflection point, on your last question, relates to probably Q4 when the Hertz facility is up and running and producing 80 units in volume. At that point, we expect the gross margin to rebound. Our total gross margin guidance for the year remains at 25% to 30%, and our long-term gross margin target still remains at 30% to 35%.

My second question is about the new product. So could the management team elaborate a bit more about Hesai's new product, ET25, the in-cabin LiDAR, regarding the project wins, potential shipment [ rounding ] contribution this year, and how does the margin of this product look like? And separately without the change to their car body design with the roof line, would ET25 be -- actually be easier for OEM to install this product in their cars or by any chance to accelerate the process of LiDAR adoption? So that's my second question.

Thank you, Tim. The two questions, right? One you wanted to know -- by the way, this is CEO, David Li. So you wanted to know the expected gross margin compared to ET, right? So we don't have this number yet because we don't have signed contracts to disclose on the number -- on the selling price yet. What I could comment is that I think any LiDAR or any, in general, hardware product, if you are unique, you are irreplaceable for some of the technical features that competitors don't have, usually, a product like this enjoys a better margin. We tend to think ET is a unique product in a sense that it solves quite a few extremely technically challenging aspects of LiDAR. One is that ET stands for extremely thin. It's of 25 millimeter height, which makes it possible to install behind the windshield in the cabin. If you didn't have that, if you're very thick, then the problem is that it will be too bulky and it will take up a good part of your windshield as the entire system. That's impossible. Second is that if you install something behind your windshield, you can imagine the windshield doesn't really help with long-distance detection, it actually hurts it because it absorbs the light on the way out and on the way back in. So it does that twice. So you need to have a very special process, especially highly integrated process with both the OEM and especially the glass makers to do that. We disclosed that our partner is Fuyao Group, one of the biggest glass -- auto glass maker in the world. And so it takes a lot of time to develop processes like this to make sure the seamless integration. For those reasons, and also, of course, on top of it, you need to be very quiet, the NVH needs to be very good. For all those reasons, ET is a very unique product that make it more unique, and we expect them to have higher margins if we do this right with the right partner. And then your second question is how hard it is to integrate it with the OEM, is that right?

Yes, correct.

Okay. So yes, it's going to be a little more challenging than AT. AT is extremely standardized, right? Everyone pretty much gets the same hardware. For ET, the platform is the same, but there is integration work. And that's why we are starting to engage with quite a few domestic and global OEMs in starting to try to integrate that into their vehicles before they give us official nomination because it's well understood that it is an interactive process with a customer. I actually consider this is a good thing because this makes it further unique as a product, especially if you use our ET, it would be more challenging for you to -- for somebody else to come in and try to replace us just because they have the performance. Integration is also very challenging.

Your next question comes from Bin from Crédit Suisse.

I also got a question about the gross margin because of guide full year would be 25% to 30%. If I want to be coupled by quarter, I have an easy calculation show that in second and third quarter may be it's 15% gross margin and in the last quarter, maybe 30% gross margin can achieve full year 25% to 30% guidance. So can you correct me if my calculation is wrong about the margin, especially in the #4 quarter, because all eyes on the margin recovery in the #4 quarter? Can you give me the #4 quarter margin data?

Thank you, Wang Bin. It's Louis. As I said earlier, we won't give you quarter-by-quarter analysis on the gross margin. But based on your calculation, it makes sense what you said. So 37.8% for Q1. Q2 will be slightly lower as we already indicated. Q3 will be lower as well, and Q4 will rebound. And our target is 25% to 30% for the year. It's still intact.

So Wang Bin, I wanted to give you a little more insight on what's happening as we transition. So we talk about that, it's going to be a V shape on gross margin this year. And Q2, Q3 are relatively lower compared to Q1. And Q4 will be better on all the products and also on the blended. So let's look into -- look at them in two factors separately, right? So Robotaxi, we had the biggest order in our history, but that order is a big order that's spanning over a couple of years. So you can imagine for a big order like that, it's a negotiated price of all the lifetime, but the truth is that for this year, as we're starting to ship in small volume, the margin is low. That's why Q2, Q3 have some impact, limited impact but some impact, on the margins. But starting next year, we will have upgraded products on the Pandar series that will bring the Robotaxi margin to a reasonable level for Robotaxi, which is historically higher than ADAS. Similar phenomenon happened for our ADAS gross margin. The fact that we were starting in small volume last year and started to ramp up this year. And then as part of the negotiation with the customer, we have slightly lower ASP compared to last year. And we also try to address this by upgrading our technology on the semiconductor side on the product. And also, as we ramp in volume, we will be further able to take advantage of the economy of scale. That's why we expect by Q4, the ADAS margins go back to a healthy number. That's why we're taking this V shape, but we already see, on the technical side and on the economic side, that this will happen.

The final point there, Wang Bin, is that the Hertz center will be open in Q4, and that has lower manufacturing costs.

And my second question about volume. Can you provide a full year volume guidance again, especially in the second quarter because it seems like the ASP will decline because your revenue guidance is similar as the first quarter, but one has been 40% increase. Can I assume that second quarter ADAS will have a much higher growth than 40% Q-on-Q, but on AM, Pandar will be stable? Is that the right assumption for the mix for the second quarter in volume?

Yes, certainly, ADAS is growing in Q2 because in Q1, as we mentioned, Chinese New Year holiday, the deliveries were pulled forward into December of last year. I think the year-over-year guidance stays -- is about 240,000 to 260,000 total units, of which 220,000 to 230,000 are still expected to be ADAS. So there's no change in the annual guidance, which means that Q3 and Q4, obviously, will be a higher ramp in ADAS. And remember, by Q3, late Q3, 6 OEMs will be shipping. And by Q4, we have 11 OEMs for next year. Many of them will be taking some significant delivery numbers in Q4 as they ramp up for their model release in 2024. So Q4, we expect the largest volume for ADAS in Q4, which has traditionally been the case for us. I know it's only in 2 years, but we expect a larger volume of ADAS in Q4. And then with the Hertz center up, the margin should go up.

How about the second quarter mix?

Second quarter mix, well, I think the -- we're targeting over 40,000 ADAS units and then the remainder will be the autonomous mobility, Pandar QT and XT, so it should be slightly over 40,000 units for ADAS is what we're targeting. It's not done yet. There's still a lot of work to do. So we said almost 50,000 is our target.

Your next question comes from Olivia from Goldman Sachs.

There are two questions from my side. The first question is about the pricing cuts in the EV industry. Has the general pricing cut in the first quarter been passed to the upstream LiDAR yet? And also, what's the expectation of the EV price in China in the next few quarters? If there is another round of pricing cut, how would that impact the price and the margin for Hesai's ADAS LiDAR products?

Thank you, Olivia. This is David. We anticipated limited impact for a few reasons. Well, the first reason is that the top of price first, right? So the nature of those multiyear contracts is that we negotiated price at the earliest date of this contract. Essentially, they're locked in. And they have a projection on the volume. But yes, of course, it's depending on their actual volume shipment. But when there is a price cut on the customer side, the contract we have with them on the nomination still holds. And then, of course, more importantly, it's the volume. We're very fortunate that our biggest SOP customer, Li Auto, is going very strong. And that's one of the bigger reasons that we still are very optimistic on the guidance we gave, especially given the fact that earlier this year, when we had the original guidance, we already knew there was going to be a price war, and usually OEMs knew that way before the rest of the market. And that was also part of negotiation we had prior to this. So we're actually now seeing that our customers, especially the top lines, are going very strong. And then if we put money aside, if we purely talk about technology, if you think about LiDAR, it's not your leather seat because it provides a very strong active function on the safety side. Most of the OEMs are using LiDARs as one of their top selling points to help them sell LiDAR -- sell vehicles as opposed to be just purely a cost. So that's why a lot of people, even for the cheaper model, they're still considering using LiDAR to help them differentiate from the rest of the peers who offer a slightly inferior product just because they didn't have LiDAR-equipped driving systems. So that's why we're also seeing a lot of interest, even though some of the prices have gone down, LiDAR hasn't been the considerations that they turn off. And then the last point I want to make again is that I feel like if you treat LiDAR as a pure hardware, it's kind of unfair because traditionally, when you buy a car, the value of the car depreciates over time as a pure hardware. But an intelligent driving ADAS system with LiDAR, the moment you buy it, the value actually goes up over time as they keep developing more software on it and they push to the customer. So essentially, after buying a car, in 6 months or 12 months, your car could be more valuable to you because it provides more functions via the software. One of the interesting example we started to see on the market is, for example, for Li Auto L9 with our LiDAR, right, and their AEB function with LiDAR became much strong, and there was a verifiable evidence that this became a much safer feature. And then L9 also scored 5 stars in the C-NCAP crash safety test. I think it would be fair to say that LiDAR was a critical part of this. And that is all the value that LiDAR is providing to the customer, and that's also the reason we believe the market will start to adopt that actually in an increasing manner.

That's very clear. My second question is about the 4D millimeter wave radar. What's your view on that? Would that be a supplement or the substitute to LiDAR?

Yes. Great question. Yes, we do believe they're complementary at this stage. So -- well, it's kind of a bigger question because we probably won't be able to comment without mentioning one of the companies who hasn't used LiDAR, Tesla, right? So at least from my opinion, it's great that Tesla is considering using 4D radars because for me, from a technical standpoint, this is a strong message that Tesla agrees that Vision's only system isn't sufficient. At least, they used to say that all I need is a camera because humans don't need other than our eyes to drive. But now they're voting to incorporate more sensors because it makes perfect sense for technology to complement each other for sensor fusion to function in a much safer way than 1 sensor alone as they see different parts of the world, right? So if we agree with that, I think the natural question is that we do agree you need multi-sensor modality, multi-sensor fusion. There's radar, there's LiDAR and there's a camera. How do people make the decision in the end? I feel like in the end, now it's not a technical decision. More sensors is always better. It's a commercial decision how affordable we can make it. Today, 4D radar isn't cheap. 4D radar is still a fraction of LiDAR, but not by a huge factor, probably 2, 3, roughly. And we're seeing the LiDAR being more affordable. If you see the history of LiDARs, they're being much more affordable, and this trend is going, right? And then the last thing is you want to compare the performance. 4D radar, due to its nature of the wavelength that is roughly 1,000x longer than the LiDAR, it sees limited set of objects. For example, it will be much more challenging for 4D radar to see stuff like lost cargo. It will be hard to see a tire because the connectivity of those objects are identical. Even for humans, it's much more difficult for radar if you compare to LiDARs. So it doesn't cover the full set. And lastly, but very importantly, the resolution. Resolution for 4D radar is very limited. Used to be either worse for 3D Radar. Today, 4D radar is about roughly 1% of the resolution of LiDAR, like 100x of difference. A good analogy would be when the resolution is so low is that you have short-sighted eyes without glass, and you're driving like that because you only have 1% of the resolution. Is it better than not having your eyes? Yes, I would agree, but that might not be very good argument if you want to replace LiDAR. So then in the end, if you look at it, really, LiDAR is much better in performance, becoming cheaper and cheaper over time, and this is a verified sensor to be the critical partner of the driving system, and everyone agrees that sensor fusion is critical. That's why we're still very optimistic, actually more excited that more companies are looking into the fact that vision only isn't going to be sufficient.

Your next question comes from Paul from UBS.

I have only one question. It is regarding the innovation in terms of the direction and the pace. So when you are seeking today and you imagine the LiDAR products next 2 years later, what would be the key areas to improve? Is that going to be resolution? Is that going to be range? Is that going to be durability? Is that going to be the cost structure or just the size and the shape of the LiDAR? So the reason I'm asking this questions is, if the innovation pace is too slow, of course, that would limit the adoption of LiDAR. But if it's too quick, for anyone who buy Li Auto L9 today, driving for the next 10 years, for the first 2 years, he is happy, the state-of-art LiDAR. But for the remaining 8 years, Hesai high-end expensive large SUV would [indiscernible] backwards LiDAR compared to the latest products by then in the market. So how should we think about this balance in terms of the innovation, in terms of both the pace as well as a direction? The products, let's say, 2 years later, if it's better than today, where it would be better?

Thank you, Paul. This is David. I totally agree that you don't want to be too fast and you definitely don't want to be too slow in pushing forward the pace of the innovation. So let's divide that into a few technical aspects, right? I will talk about what we think we've done and what we think remains as a potential and that path moving forward. First is distance right? And so today, it's roughly 200 meters and some of the competing technology has 250 meters. And we're also moving in that direction. I think it's roughly meeting the expectation. Second is form factor. Form factor is interesting because it's not about how big it is, it's about how you install that. So I think we made a major progress in being able to install that -- make it so small and install that behind the windshield. I feel like that part, the market is extremely excited about. And we definitely see a lot of potential for doing so. And so that part is done. And then I want to talk about what hasn't been fully done and remains as the opportunity and the challenges. One is resolution. Let me give an apple-to-apple example. So if you think about the typical good camera for cars today, it's 8 mega pixels per frame. 8 mega pixels, right? So LiDAR, what we have today is only 1.5 per second, which means that it's 1/10 of it per frame, which is 150,000 per frame. 150,000 versus 8 million. That's like a 50-some times room of improvement for LiDAR resolution to meet camera, which is a natural thing people would want. So we have 50 more times to innovate. Of course, our next generation will significantly shorten this -- close this gap, but that's one of the opportunities we need -- we were looking at. And the other one is price. I won't be able to give you specific numbers, but we definitely see, for some of the lower-end cars, they probably don't want to pay the close to $1,000 level price, and they want some cheaper version that will have limited function, but still LiDAR is like your invisible airbag. You definitely wanted it for added safety, so we're also seeing that. And then Hesai, as a company, is better determined to be able to take advantage of our in-house manufacturing and our strong semiconductor platform to be able to drive down the cost. If anybody, I feel like we have the best chance to do that. So that remains a challenge and the biggest opportunity moving forward. Those are the things that are on my mind to move forward.

Your next question comes from [ Tony Chen ] from Huatai Securities.

As we saw the intense EV competition in China and some [ e-specs ] trend from the product, many functions will be dropped in their low-end model and become optional in their high-end product. This is one trend. And as Tesla, you just mentioned, they choose to use pure Vision and they use only camera to be their platform for autonomous driving. Did you see China EV brand will follow this trend from your perspective? Will these 2 trends slow down the adoption rate of the LiDAR from your perspective? That's my first question.

Sure. Thank you. Let me make it very clear. I think what I was -- the reason I mentioned Tesla is a friend, not an enemy in a sense that they're actually acknowledging that camera alone isn't sufficient, right? Because they're introducing 4D radar. So again, my argument is that Tesla used to claim very confidently that Vision only is sufficient. And now by introducing 4D radar, at least we know they now agreed that multi-modality sensor fusion is going to be very, very helpful for increased safety. So that's what I see. That's the reason I'm more excited for this because I feel like if LiDARs, we keep going in the direction, make it better and cheaper, there's a very good chance that people would agree that this should be part of their system. So -- and then back to your question on the China EV. So I think I also talked about it already. We are seeing, even for the cheaper models, they want this to be a differentiator and then -- and for small EVs, LiDAR is one of the iconic [indiscernible]. And yes, today, for very low-end models, it's relatively expensive. We're also making the effort to hopefully to drive down the cost. So -- but we're already seeing wide range of adoptions this year and by next year, we'll be shipping with 11 OEMs. Some of them are on the more affordable models, but they still use it and some of them even use that as a standard configuration for all the cars they ship because they see that as a strong safety features that they cannot live without.

Okay. That's great. And my second question, could you give us more color about the R&D progress in ASIC solution in your LiDAR, you just mentioned in opening remarks? And what's your strategy? And how important this technology will be to win competitive advantages?

Sorry. Can you say it again, the question?

Yes. Can you give us more color about the R&D progress in ASIC solution [indiscernible]?

ASIC, so we're at the -- in the development of the fourth generation. So each generation is a major step forward in more highly integrated components to further reduce the cost while having exponentially higher performance. When we talk about performance, we're talking about a few things. One is the range. Today, we're about 180- to 200-meter range. And then our next generation is going to go way beyond that. And the reason we're able to do that, we made a lot of progress on the component, the laser and the receiver side, to be able to drive that part of the efficiency up in the way, for lack of analogies, like your cellphone camera, right? 20 years ago, if you take a photo at night, you barely see anything. Now it's super clear and it sees more things than your eyes, and we're driving our components in that direction to be able to detect further, right? ASIC also allowed us to handle way more data as we send out the light pulses and receive them because we used to -- our current AT is 1.5 million points. Next generation will be many times more than that. The reason for that is our ASIC handles way more LiDAR pulses, and we can just multiply a big factor on where we do many points per second wise. So resolution-wise, it's a huge leap of faith. And also, if you are familiar with Moore's Law, we're also seeing that trend in the sense that for every unit, the ASP isn't declining as very fast, but the performance per point is declining a lot because we're having many x of increase on the resolution. So this is what ASIC is allowing us to do. And on top of it, it gives us a better security chain management ability because we used to source a lot from different vendors for some of the semiconductors. Now we just own the design and just ask the foundries to do that for us, and we're using very mature process node. So it's very much safer for us to work with foundries, not having to manage a global supply chain system.

Your next question comes from Joel from Nomura.

I'm asking for a -- when management mentioned about the annual cost cut from OEM side, I want to -- can we share generally the situation here in terms of the annual cost cut level? If we cannot talk about the detailed numbers, can we say that it's low single digit or maybe high single digits? Can you just talk about that? And also in terms of the OEM attitude toward LiDAR, we're talking about the low-end economy because also a lower-end car brand also thinking about the LiDAR adoption. But in terms of the current situation, I would say because of the price competition in OEM market, do we see any further pressures or OEMs be more cautious in cost efficiency in this year and maybe impact the LiDAR market in the short term? Do you see any impact or any concern from OEM side?

On your first question, Joel, this is Louis. Regarding the price decline, we would estimate about 5% to 10% a year -- year-over-year decline in the ASP side. But as David mentioned earlier, our contract was signed for the whole year. So they're kind of locked in sometimes multiple years where the price is locked in. As far as the OEM adoption of lower-priced LiDAR in the competition, like I said, it's a long uphill process because they're more integrated with our LiDAR into the new model. It takes time to change. There is severe price competition, and we face that the last 2 or 3 years. And that's why the price has come down. A lot of companies used to pick just based on price, and then they regretted it since because the perform is quite poor or the quality. So I think is that we're hoping that and expecting that price competition abates compared to the performance as the best LiDAR providers come out and able to show that they can produce auto-grade LiDAR at affordable prices. So I think the price competition is beginning to wane a little bit. It's more about your product and a price that's affordable. You had a second question? Sorry, I didn't catch that correctly.

I still have one follow-up. So in terms of the -- because of the current market situation, so in terms of the short-term target, I mean the task for the companies, so what's the key? It's about the cost down or it's about the new technology which in terms of short term, I mean maybe within 2023, I guess that actually the cost down will be the first task?

Short term is not as big a factor for us because the prices were locked in last year or at this point, and those models are coming out. So -- and most of them are actually going to be delivered in 2024. So other than Li Auto, Lotus, we view 4 or 5 players in our universe, the ones that we sell to, there aren't a lot of new companies releasing cars this year for the short term. So -- and then they are gearing up for 2024. And because LiDAR is a competitive advantage differentiator in the market, as competition increases, you want to have it. So in the short term, to answer your question, is probably very, very little impact since not that many models are shipping yet. Longer term, we'll have to see.

Your next question comes from Olivia from Haitong International Securities.

Could you provide an update on your new factories in Shanghai and Hangzhou? If I'm not wrong, I remember you said both will be put in operation by Q3. And I also wonder if your factories are compatible for ET25 or other new products beyond?

So let me try to clarify a few phases, right? So moving forward, we have 2 major factories. One is called Maxwell, we talk a lot about. It's an R&D and manufacturing center in a sense that the main goal for Maxwell is to be able to fast iterate product development and the manufacturing under 1 roof. So that's why the name is Maxwell. It's the name of the scientist who came up with important equation, right? And then the other factory's name is Hertz. And Hertz is like a cycle time, how many units produced per second, that type of nature. You can see it's for mass production only. So for Hertz factory, it's extremely highly automated with 90% plus automation and the cycle time from -- used to be more than 1 minute, now Hertz is 45 seconds per LiDAR. And also because Hertz is used for high automation, and in Hangzhou, it's a relatively cheaper place, and it's actually the labor cost and manufacturing cost is also lower. So having said that, with the two functions, there's another important concept I want to make it clear. So building a factory or retrofitting a factory is one thing, populating a factory with all its full production capabilities is another. And those are completely different things. And sometimes people are confused by those. Building a factory like Maxwell, it's like buying a big land. So when you do that, you wanted to have a lot of space, right? You can buy a couple of acres of land that technically can host hundreds of people to live, like you do at a hotel, right? But we're not stupid. What we do is actually, we populate the production line based on actual needs projected down the road. We can build it very quickly. Within 6 -- less than 6 months, we can populate those production lines. And we definitely don't want to build more than we will be able to ship. So that's why either for Hertz factory, each line. Today, we have is about 300,000 to 500,000 units per line. And currently, we only have 1 line because we already see this will be -- and together with Maxwell will be able to meet our need. It's like you have a big land, you buy it because the land itself is cheap, but then you started to build those houses as you grow on the people, but you don't want to build all those houses on the land yet because it only takes 6 months to do that, there's really no point of doing that. So that's why moving forward, we will only slowly expand Hertz and some part of the Maxwell to keep it very, very lean because we know with the visibility we have, it takes less than 6 months to do so. And even for each line, it isn't extremely expensive, right? Louis, do you want to comment on the cost down the line?

Yes. I think the -- yes, I mean, David makes a very good point is the thing that is confusing the investors is we're not building $200 million, $300 million plant where the CapEx will eat up our cash. We've already guided CapEx this year and probably next year will be about USD 50 million and that's all for mostly plant equipment. So it's about $10 million to $15 million per line for the 18 fully -- 90% automated line and much less for the anticipated lines for the autonomous mobility because that requires more manual labor, so even much less of that. So don't get confused that we're building these plants that cost $100 million that are eating up our cash, that's not the case. So that's why even today, we have $457 million of cash at the end of the quarter. That's much higher than any of our competitors. So we're not spending a lot of money on CapEx other than for equipment and for the assembly line.

I hope I answered all the other quick questions. Do you think those lines can be used from 1 product to another? Well, that's a no. Every time you need to -- for a different product, you need to redesign the full automation line to do that. So that's also another reason we have the new AT product building, Hertz factory.

Okay.

Right. And then, sure, even though the production lines are redesigned, some of the equipment can be reused as we are ramping down on some of the legacy products. And then because of the automation process are shared in great similarity, the majority of those equipment can be used, but the line needs to be redesigned.

Your next question comes from Jesse from Bank of America Securities.

First of all, I know that you mentioned ET25 hasn't received some order yet, so we cannot disclose on ASP and margin. But wanted to understand from an OEM standpoint. So when they need to purchase an upgraded windshield to go with this ET25, how does this cost going in the future for them? So it's more expensive compared to the traditional solution or could be cheaper for them?

So first of all, we are in the different processes for some -- a lot of the sourcing decisions. I think in the script, we explained, we were in 8 RFI and RFQ processes for 5 major global OEMs. So I wouldn't say that we don't have progress, it's that we don't have the final nomination we could announce. That's the status. And then your next question is about when they make decisions, do you expect them to be cheaper or more expensive. So the biggest reason that this will be cheaper is that it's a product that's later -- a few years later than AT series, right? AT series is we started to ship in 2022, right? ET is a few years later. And it will be able to leverage our new development on the semiconductor side. And as we move forward, even though the performance is going up, and we always try to have a better integration on the semiconductor side to control the total [ bond ] cost.

She's asking for the total cost for the windshield [indiscernible]...

So you mean like how much more money will be added on the glass side?

Something like that.

Okay. Well, that part is specific to OEMs. I don't expect that to be a major part of it because for the special processes [indiscernible] a large production space is just one of the relatively simple process people do. It's no significant price increase.

It just looks so much better. So [indiscernible] going to be a big factor, right, as OEMs make decisions.

And of course, OEMs always want things that are better, more beautiful and the cheaper, and that's reasonable.

Yes. Totally makes sense. So another question would be another peer also state that the software that goes with the LiDAR would be the key to win the long-term process. So what's our view on this?

I agree. I think today, a lot of customers are already seeing that they may or may not be the most experienced user for the LiDAR. That's why we provide the reference software system to the customer, especially that we know how to use them. We already have a package that will help them jump start. We don't fully take over their development because most of the customers, they prefer to own a process, but we would provide the reference software to help them from -- anything from perception to tracing the target, to calibration and to some of the algorithms in calibrating and essentially helping them to get things up and running. And that part is very important because without that, it's a piece of some hardware that no one can take advantage of. Did I answer your question?

Unfortunately, that does conclude our time for questions today. I'd now like to turn the call back over to the company for any closing remarks.

Thank you once again for joining us today. If you have any further questions, please feel free to contact our IR team. This concludes today's call, and we look forward to speaking to you again next quarter. Thank you, and goodbye.

This concludes today's conference call. You may now disconnect your line.