Hello, ladies and gentlemen. Thank you for standing by for Hesai Group's Second Quarter 2024 Earnings Conference Call. [Operator Instructions] Please note that today's conference call is being recorded. I will now 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. Thank you for joining Hesai Group's second quarter 2024 earnings conference call. Our earnings release is now available on our IR website at investor.hesaitech.com, as well as via Newswire Services. Today, you will hear from our CEO, Dr. David Li, who will provide an overview of our recent updates and address our financial results before we open the call for questions. Before we continue, I refer you to the safe harbor statement in our earnings press release, which applies to this call as we'll 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. David Li. David, please go ahead.

Thank you, Yuanting, and thank you, everyone, for joining our call today. Let's begin with an overview of the LiDAR market trends and some recent business highlights before moving on to our second quarter financials and operational results. We are thrilled to announce that Hesai has been recognized as the number one automotive LiDAR company by market share for the third consecutive year according to the latest LiDAR for Automotive reports from Yole Intelligence, a well-renowned European independent research firm. 2023 was a stellar year for Hesai marked by record-breaking revenues and shipments, strategic design wins and expanded product lineup and new partnerships. Our expertise in both passenger cars and robotaxis enabled us to capture 37% of the global LiDAR market and an impressive 74% share of the global robotaxi LiDAR market in 2023, demonstrating our clear leadership of the global automotive LiDAR market. Yole's recognition inspires us to aim even higher as we continue to deliver the most advanced and powerful LiDAR products, driving further growth and innovation in the dynamic automotive industry. Next, a brief update on autonomous mobility business. At present, one of the most exciting developments in China is the thriving robotaxi market, which is scaling and commercializing at an impressive pace. We're delighted to witness this momentum and proud of Hesai key role in promoting this trend. Baidu's Apollo Go and autonomous driving travel service platform has recently launched operations in Wuhan, one of China's largest cities. This launch has garnered significant market interest and numerous orders highlighting the clear development path and the tremendous potential of the robotaxi business. According to media reports, Apollo Go has provided over 6 million robotaxi rides in 11 Chinese cities since 2019, we've been serving as Baidu Apollo Go exclusive supplier for their vehicles main perception LiDAR. With…

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

I have two questions. The first one, we noticed that the Financial Times reported that the U.S. Defense Department has decided to remove Hesai from the backlist. Could you please elaborate a bit more on the current development, the progress and of course, the implication to your project wins and operation in overseas because we think this is a very important milestone. So it is great, if you can share more details. That's my first question.

Thank you, Tim, and thank you for the question. We have consistently maintained that our inclusion on the 1250H list was a mistake. Our products are strictly for commercial and surveillance use, and we have no connection to any military bodies. We're not a vendor to any military bodies either. As we discussed in the previous earnings calls, being on the DoD means strictly that the U.S. Department of Defense, aka the Pentagon, cannot buy our products starting mid-2026. But to be honest, as far as we know, we've never sold anything to DoD or any military. And we definitely don't plan in the future. Or in simple terms, we've never generated any revenue from DoD or any other military. And again, we don't expect to. However, being on the list did seriously impact our reputation. And as you can see, the stock price, it also impeded some of the business opportunities over the past few months since we've been included. And again, to be honest, it actually made things very tricky for some of the customers because they're worried that we might have ties to the Chinese military, as that's what the list meant. Unfortunately, it's hard to disapprove until DoD officially takes us off the list. We will not be able to speak to the actual results until it's official, but by any chance, if we are removed with our leading technology and the best-in-class financial strength, we're very optimistic that we're locking more global deals. And this is the information we would like to share regarding the current status of the DoD. Unfortunately, I will not be able to speak to the status of the Financial Times article because that's speculation of the case, and we definitely need to wait for the official ones. But I think it's more important that people understand our position, understand the fact and also understand that we have indeed by this inclusion and if by any chance, we are removed, we are extremely optimistic that a lot of the progress that was indeed and apart will be removed and hopefully, on an accelerated path. Does this answer your question, Tim?

Yes. And my second question is currently on investor focuses a lot on the robotaxi. And David, you also shared quite a lot of updates during the presentation. But as you mentioned, I think Hesai chosen by Baidu Apollo Go for its next generation robotaxi a sole suppliers across 11 cities. But if there's any further details, how should we think about the market opportunities? Because you also mentioned in addition to Baidu Apollo, the top-notch robotaxi operators or makers also want to do the business with us. You mentioned like a 4 light up the car. Would that kind of business also boost the value content with like a potential assembly or software being integrated? In the meantime, when do you think we're going to see the inflection point with more meaningful revenue uptick, it's going to be next year or the year after, or we are going to see some progress later this year? So that's my second question.

Thank you, Tim. I will answer the first part. I think for split of a second, I missed your question on the software side and which I might have to ask you to -- ask again. But let me first take on the question of the business development of the robotaxi. I think it's fair to say that we are the biggest robotaxi LiDAR supplier in the world by any measures. And in China, we are the exclusive the main perception LiDAR supplier for all of the top 5 robotaxi supplier. And most people believe that we have a product that's way superior than the peers. And including Baidu Apollo Go, right, and based on the Yole report for 2023, we have a 74% market share of the global LiDAR market for robotaxi as we, I think, have been communicated. And historically, those people use mechanical spinning LiDAR. But now as you see, as we are scaling up, a lot of people are adopting the ADAS AT128, again, including Baidu Apollo, which is a very interesting transition as we see because AT128 is a new sensor that we've shipped hundreds of thousands of and it's a very powerful, much more affordable sensor. And I think this actually helps the customers to achieve a better balance between price and performance. At this point, to me, the biggest inflection is more commercial than technological because technological enhancement has always been continued -- continuously advancing. Well, this year is the first year, most people are talking about commercialization in the sense that you need to look at the total cost of equipment, you need to look at amortization, and in the end, you need to come to a business case in which if your sensors are still worth tens of thousands of dollars each, as it used to, it makes it tricky for people to do that. And that's why we are working very closely with robotaxi people to do that. And by the way, it's one of the example. And they're not only one, by the way. Many other robotaxi have also adopted a similar trend. So as a result, what we see is that we start to receive significantly larger LiDAR orders for robotaxis for the years to come and some of them have already been locked in. And as we expect this market to grow and we expect the customers to move from a smaller scale, high-priced mechanical LiDARs with the larger scale ADAS LiDAR. And also international robotaxi players are still focusing on the performance. So they are still focused on the mechanical LiDAR, which is more high priced and a more -- much higher margin business for us. And so this is on the business part. And again, I remember you had a question on the software. Can you repeat that so that I can clear the answer?

Sure, sure. So basically, I just want to tell that if there would be any differences between our business with the robotaxi maker and the traditional car maker. So in robotaxi's case because we are going to provide like a multiple LiDARs. So just wondering if there will be any potential upside to the battery content. So in addition to LiDAR, Hesai also provide like assembly services or together with some additional software to facilitate the rollout of the robotaxi? And in short, will there be any value content upside in addition to the LiDAR you sell?

Okay. So your question is to make it simple, can we make more money from them by not only something the hardware, but also selling the software, right? So the quick answer is, I believe it's a different business model in a sense that we do provide a lot of tools. We're trying to stay away from the data per se, but we do provide tools for installation and calibration, stitching a different Point Cloud into one panorama image. And those are the tools we have. But instead of targeting them separately, we think the business model is that we provide them with the software as part of the hardware package, and that way allows us to have a better margin and a better -- a more competitive product. So in other words, that's already been embedded and that is one of our advantages of being able to cater to this market with a better offering.

Yes. Let me clarify on that. I think Hesai's key focus is exclusively on supplying LiDAR hardware but not the software. I think we are very, very strict about we never have any data security or privacy risk that can be posted because we never transmit data wirelessly. We do not start even a second of the Point Cloud, and we only supply the LiDAR hardware, not the software. I think that's key.

I'll give you another example that I consider as analogous. If you think about, for example, Apple, right, they sell you cell phone, right? Technically, they could charge you for the iOS system or if you look at your MacBook, they could charge you for the macOS system, they never do that. They just sell you a more expensive hardware, because most people don't have the habit of buying a separate set of software for the hardware, even though they go together. I think the software is served as a higher barrier and a better user experience for the hardware product. I think that's kind of the philosophy we're following.

Got it. Super clear. Thanks for sharing all the details.

Your next question comes from Tina Hou from Goldman Sachs.

So I have two questions. The first one is regarding the robotaxi LiDAR. So you report autonomous mobility LiDAR shipment separately. But now I understand that would include the AT128 LiDAR as well, right? So just wondering like how much of that [5,700] is AT and then the rest is Pandar. Could you give us like a breakdown? And then also for the AT128 that you sell to the robotaxi customers, is it the same pricing and margin as you sell to the EV OEMs?

So that is a good question. So as we've said, currently, within the Chinese market, we are seeing a trend where our Chinese robotaxi players when they would like to commercialize their business, they're switching from the transitional mechanical spin LiDARs to adopt the ADAS LiDAR. And as a result, we have received significantly larger LiDAR orders for use in robotaxis for the years to come. That means that the big order will be splitted into several years, and we will record the ADAS LiDARs sell to these type of Chinese robotaxi players in the coming years. So holistically, I think on the robotaxi side in China, we can expect that they are moving from the smaller scale, higher-priced mechanical LiDAR to larger-scale ADAS LiDARs, and that will boost our revenue and gross profit in the long run as the robotaxi business grow in China. And with regard to your second question, are we having the higher price? Are we having the higher margins for the ADAS LiDAR we sell to the robotaxi players in China? The answer is yes. Because the amount they are buying from us is not as significant as the passenger car OEMs they're buying for us. So basically, they will enjoy, of course, a relatively better price. And of course, we are selling the ADAS LiDARs to them with the same cost, so the margin will be relatively better as well. So that's why I'm saying that will boost our revenue and gross profit for the long form.

Yes. So the other angle to look at it is that to us, because they're buying the identical product. And in the end, the pricing is strictly tied to volume, right? So if you kind of look at the ADAS volume, usually, we're in the range of hundreds of thousands of units a year. Well, clearly, robotaxi isn't there. So that's why they have to pay for a higher price and add the volume in and there. But having said that, the long-term frame agreements we have with different robotaxi makers is such that by the time they reach similar level -- range level as the ADAS product, the price to them will be also go to the ADAS product. So -- and it is the nature of such a business.

Yes. And financially, I think it's an equation, right? It's about the price you have multiplied by the volume you have, then multiplied by the gross margin you have. And so based on this equation, I think in the long run, having the ADAS LiDAR shipped to our larger amounts of robotaxi players that will help their -- to help them commercialize their business. And in the end, the equation for us, I think will be beneficial for us financially as well.

That's very clear. It makes a lot of sense. The second question is regarding your gross margin because the second quarter gross margin has improved quite a lot from first quarter. So could you give us a breakdown in terms of the factor for the improvement? How much is from product mix, how much is from like scale benefit? And also, as David mentioned, the ADAS, ADAS LiDAR gross margin has also significantly improved. So just wondering like how much difference is there between the ADAS gross margin and autonomous gross margin now? And maybe too many questions on gross margin as well. So as you guided for 3Q and 4Q gross margin at 40%. So wondering why is it lower than our second quarter margin? And lastly, how should we think about gross margin in 2025?

Thank you. Yes, there are many, many questions in this. So I will try to give you the top level of thinking and maybe we can have a more interactive discussion based on this. The first is despite the downstream adjustments to accommodate the postponed SOP time line for certain clients, our ADAS ASP will remain relatively stable within the year. The pricing was negotiated at the time of the contract on sign typically fixed for the year. The mechanical LiDAR for each specific product will also remain stable because most people recognize us being the by far the best mechanical LiDAR on the market. And then you're right that we had a 45% gross margin blended in Q2, highlighting our financial strength and definitely the resilience of the business despite some of the shorter-term headwinds. I think one thing I want to point out is that our quarterly gross profit total is more than RMB 200 million, that's roughly 3x that of the next the highest publicly listed competitor on the global side. And I think a few reasons. One is our cost management effort have paid off because we upgraded AT128 featuring a more integrated design, along with improvement of our in-house ASICs. I think that's one of the biggest effects there. And the other one is definitely we're benefiting from the economies of scale as we ramp up the shipment. And also, we were able to take advantage of that to drive down materials and the manufacturing cost. And there's also -- I have to say there's also a one-off high-margin service fee aka NRE from one of the global leading OEM in the second quarter, which is very helpful on the gross margin side because it's the service revenue. So these are the status of it. And looking ahead, I think we initially guided a blended gross margin of the higher end of 30% to 35% range for the full year. We now expect the margin for Q3 and Q4 to be closer to 40%. To be honest, we are not aware of any other player in the global LiDAR industry that can match our financial strength at the level of the revenue and the number of units we deliver, which is already at scale. And with effective expense controls and optimize operational efficiencies, our OpEx is expected to grow 10% to 15% compared to the full year of 2023. And these initiatives bring us closer to achieving profitability in the fourth quarter of the year. We are also optimistic about approaching non-GAAP breakeven for the second half of 2024. And these are what we expect for the future and also the current status of the year.

So in short, our gross margin will be very robust. We delivered 45% in 2Q, and we are expecting Q3 and Q4 to be closer to the 40%. You mentioned why there will be a Q-o-Q slightly decline from 45% to 40%, it is because we recorded a one-off high-margin services revenue during the second quarter, as David mentioned. However, on the other side, if you look at the cost management capability we have, if you look at the economies of scale benefits we are enjoying, the 40% level of the gross margin is probably a very leading position in the industry. And if you look at the gross profit absolute value, we have recorded more than RMB 200 million in 2Q alone. I think we can -- as of now, we are not aware of anyone else in the industry actually have this kind of financial strength. And that is helping us to garner a lot of deals from other passenger car OEMs as well because the best-in-class financial strength, they have more confidence in us to partner with us in the following years to come.

Does that help you understand the gross margin fluctuations?

Yes, very helpful.

Your next question comes from Jessie Lo at Bank of America Securities.

My first question, you sort of just answered just now because I was wondering that we have already reiterated our breakeven for the fourth quarter of the year. And then you just mentioned that we are taking the non-GAAP breakeven in second half of 2025, just making sure that I have heard it all correct.

Excuse me, Jessie, would you mind repeat your question? We missed the first part.

We answered that already. Is there a new question?

Is that about the breakeven time line? I think we just answered the question when Tina bring that question...

Yes, yes, sure, sure. And then I would like to ask about, as [indiscernible] mentioned that they will not use LiDAR going forward. Of course, they are not a meaningful clients to us as of now. But then we were just wondering that would it be the chance that other LiDAR maker will try to follow what they are doing on the technology side because what we see is very diverse. For us, we are seeing a lot of orders coming in and a lot of slight inches, a lot of sampling with that client. But on the other hand side, we are still also seeing like someone decided not to use. And then I guess they have their own reason as well. So how do we see this to develop in the future?

So may I clarify your question? If you are asking about the domestic project pipeline?

I am asking about [indiscernible], they are not using -- they decide not to use the LiDAR on the future models. And then with other OEMs [indiscernible] in the future trying to follow or copy what they're doing right now.

So let me make sure I answer your questions.

So you're asking about the question about Tesla's vision-only solution and expand switching from LiDAR solution to non-LiDAR, am I correct?

Yes, yes. And then, of course, it does not impact us. But then what is other OEMs trying to copy and/or like what they're trying to do right now?

Yes. Sure.

So the Tesla question is the same question as it is before, right? The one thing I think people should pay closer attention is that there has been a publication by the Wall Street Journal, and they've been lifting out a lot of the publicly available sources on the different technological challenges on the safety side Tesla have. And definitely, they believe some of them could have been avoided with a LiDAR. I think that, even though it's public data, it was very interesting for people to collect them in such a way and to see, it's a much more intuitive image for us to understand LiDARs as the seatbelt or the airbags. And the other fact I want to point out is that, in my mind, Tesla flash a lot of the carmakers today, they're shipping Level 2 or Level 2+ sensor systems in which ultimately, it's not human, right? It's not on the machine. Well, clearly, there have been a lot of efforts globally for sure and also in China, that the industry is moving towards Level 3. The nature of Level 3 is you need redundant measures to achieve that level of functional safety function, function safety meaning that you need the redundancy to have different ways to measure the same results or same objects, and that requires a LiDAR. And at least that's what we see of almost every other global OEM and the vast majority of the OEMs in China when they develop Level 3 system. So the message is that, a, look at Tesla's recent the accidents; b, for Level 3 globally, everybody else agrees you need a LiDAR. Look at China, Level 3, everybody agrees you need LiDAR. For Level 2, some people think you could save that money, but even with that, we're making that affordable enough that most people are widely adopting that and a lot of the players decided to make it standard configuration moving forward. And so this is really the status of the industry.

Yes, sure. And then second, I still have one more question, which is we previously guided third quarter shipment should be 150,000, fourth quarter 200,000 and full year at 500,000. What that have an update?? And if you could sort of shed some light on the key client breakdown as well, for example, [indiscernible] versus the past?

Yes, sure. So I can take that question. So for the second half of ‘24, we are now expecting to 300,000 to 350,000 shipments in the second half of the year. And this is due to downstream adjustments by certain ADAS clients where some of their models have been pushed out to early next year. However, on the other side, we are seeing 2 trends. The first trend is that our ATX has garnered significant interest among the domestic OEMs. And as we stated in the earnings is that we have already secured 7 OEMs for our ATX product already. And the second trend is that we have noted that in China, especially recently, there has been a growing trend where the domestic OEMs are also moving toward Level 3 standards. So they are requesting for even better performance of the ADAS LiDAR. So like we said in the earnings that one of the leading EV maker has already signed the agreement with us to adopt our next-generation ultra high-performance ADAS LiDAR for all of their new models scheduled to SOP in 2025. And we believe that this will open a gate for others to follow as well for these ultra higher performance LiDAR. So as we offer both ultimate performance and also ultimate value-to-cost products, we are seeing some clients have decided to switch from our competitors to us. So that’s why we are – we will be anticipating very strong and solid order pipeline in 2025 and 2026. And in 2024, as we stated because the pushouts for certain models from the clients, we are guiding 300,000 to 350,000 for the second half of the year.

Your next question comes from Zhang Yu at Huatai Securities.

Yuanting and David, can you hear me?

Yes, please go ahead.

Okay. We see the ATX about go on the market. My first question is, what's the difference between the ATX and the AT128 for the [indiscernible]? And what's the impact for the gross margin? [Foreign Language]

That's a good question. Thank you, [Zhang Yu]. So our ATX is a high value-to-cost, long-range ADAS LiDARs and comparing with AT128, I think the key point is that it's having a very reasonable price, of course. And unlike similar products for our competitors that sacrifice performance for their lower-priced ADAS LiDARs, our ATX offers both performance upgrades and cost efficiencies and that's why we are securing a bigger number of partnerships from the OEMs. So since its launch, we have seen that the ATX has secured several design wins with 7 major OEMs by the second quarter, as we said. And the SOP timing for ATX, we expect that it will begin in the first quarter of next year with significantly larger volumes expected in 2026 as well. And we anticipate that it will penetrate into a broader range of vehicle pricing segments with small models featuring as a standard configuration and further driving our flywheel of scale and cost optimization. And because of this, I think the gross margin for the ATX product will be rather reasonable, even though we didn't specifically disclose for the gross margins will have.

Okay. And my second question is, when you see other competitors that they have received orders for the robotaxi LiDAR. Could you tell us more about the products there for the robo? Do you have done robotics order?

You're talking about the robotics...

Yes. [Foreign Language]

So the question is, do we have a specific product for robotics industry. And depending on the definition of robotics, we always have non-ADAS sensors, for example, the QT series, the XC series, they were not for ADAS product. So we always have that. And having said that, we do recognize that there have been a lot more embodied AI and robotics applications who would benefit from an even more miniaturized product with a lower price point. And so we are looking into that. And we do have a lot of interest and some of the orders already at a reasonably large scale from the non-ADAS industry. But we don't have plans to disclose their names yet, but we have been aggregating them into the generic robotics sector or no-ADAS sector.

Yes. I think industrial robotics has always been part of our AM business. And remember that we are the global leader in AM capturing most of the market worldwide and all thanks to the best performance of our mechanical spinning LiDARs. For example, if you look at the recent news, in July, we just signed a deal with Westwell to speed up the global expansion in AD for logistics. And our LiDARs can, of course, be used in many kinds of robotics applications, namely logistics, mining trucks, last-mile delivery, robotics, electronics, even human robotics in the future potentially. So we see huge growth opportunities ahead, I would say, and also stay tuned for our IAA event in Germany in September this year. Okay?

Your next question comes from [Jacky John] from CICC.

Management team, I'm [Jacky John] from CICC. So congratulations on this quarter results. I have 2 questions to ask. The first one is regarding to the Level 3. Just to mention that Level 3 requires a certain level of redundancy. So I want to go to more specific regarding the conditions. So are there any scenarios or conditions that can only be met with LiDAR so that to make a LiDAR necessity for Level 3?

Yes. That is a very good question in the sense that if you think about Level 3 and the functional safety requires that, first, you need to have multiple measures of the same market in case one sales, right? That's the nature of such a system. And then directly to your question, when we see the best use of LiDAR being the one that stepped up when cameras failed, there are 2 main scenarios. And those are not hypothetical cases. Those are the data we see from the hundreds of thousands sensors, which have shipped and when they're operating on the road, we get those feedback. One is they call that general objects and for AEB functions, if you have an object like a truck laid down on sideways on the highway or an unidentified object, lost cargo, that type of object, your newer network training would require your system to have seeing such an object in the past to be able to recognize that. That's one. The other is generally a nice -- again, I encourage you to look into the Wall Street Journal data set, which I think they include more than 200 cases. And if you look at those cases, a lot of them are generally in the change of lighting conditions, especially in darkness or in or out of the tunnels in which your computer vision is struggling to see anything until it's too late, and that is where LiDAR could reliably and stably and precisely identify the best timing to break or not. So those are the 2 scenarios we consistently see across different even Level 2++ plus applications. But again, I said for Level 2++, you can always blame the human and that's not big deal if machine fails because humans are supposed to be the backup plan. But to Level 3, you're giving humans freedom to not to pay attention, and that is a big safety risk.

And we also talked about the introduction of regulations mandating higher safety standards, for example, the Level 3 policy rolling out from China and also the AEV speed requirements rolling out from NASA. I think with these kinds of safety standards rolling out, the content per vehicle financially will be higher as well. That means that LiDAR -- the number of LIDARs per car can potentially increase that will be beneficial to us financially. I hope that answers your question, Jackie.

Yes, that's very clear. So for my second part, I have a few -- many questions regarding to the ATX product. The first one is how is the progress of ATX to be integrated with headlines. I guess there should be some -- there'll probably be some technical issues with that. And the second question is regarding how do you guys achieve a lower cost for ATX other than the economic scale? Do we employ more maybe the Chinese supply chain, et cetera?

So the -- I'll do the second question first. The way we achieve a lower cost for ATX are a few things. The first is that the first -- next level of economies of scale as we expect millions of units to be shipped over the lifetime of such a product, and we already see that. And second is the further level of integration and ASICs because it's using our Gen 4 semiconductor, well, you remember the AT was Gen 2. And so that is a major improvement on the level of integration. Integration in semiconductor industry means cost savings because this is less area of part of the wafer to build such a chip. Number three is the product definition fit in the sense that AT series was more generic because that's the first time we define an ADAS forward-looking sensors. By the time of ATX, we had hundreds of thousands of sensors of experience knowing where to focus and where we can relax the specs a little bit to achieve a relatively similar higher performance. So we are able to fine-tune such a sensor to fully utilize where it's needed best and then cut costs as a part that is not necessary. So with the combination of those 3 factors, we are able to achieve a superior cost-to-performance ratio.

And with regard to your first question about integration, we signed a partnership together with Marelli several months ago to integrate our ATX LiDARs into the headlines. And remember that Marelli is, of course, one of the most famous Tier 1 supplier for a global OEM. So that’s about the integration. I think ATX can be applied to many placements on the cars, and we will be seeing large adoption for the ATX product starting in 2025 and even bigger volume in 2026.

That concludes our question-and-answer session. I'd like to turn the call back over to the company for closing remarks.

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

Thank you, everyone.

Thank you.

This concludes today's conference call. You may now disconnect your lines. Thank you very much.