XPeng Inc.

hello ladies and gentlemen and thank you for standing by for the third quarter 2021 earnings conference call for expeng inc at this time all participants are in listen only mode after the management's remarks there will be a question and answer session today's conference call is being recorded i will now turn the call over to your host mr alexey head of investor relations of the company please go ahead alex

Thank you. Hello, everyone, and welcome to EXPOM's third quarter 2021 earnings conference call. Our financial and operating results were issued by our newswire services earlier today and are available online. You can also view the earnings press release by visiting the IR section of our website at ir.shop.com. Participants on today's call from our management will include co-founder, chairman, and CEO, Mr. He Xiaofeng, vice chairman and president, Dr. Brian Gu, vice president of finance, Mr. Dennis Lu, vice president of corporate finance and investments, Mr. Charles Zhang, and myself. Management will begin with prepared remarks, and the call will conclude with a Q&A session. A webcast replay of this conference call will be available in the IR section of our website. Before we continue, please note that today's discussion will contain forward-looking statements made under the safe harbor provisions of the U.S. Private Security Solidification Reform Act of 1995. Forward-looking statements involve inherent risks and uncertainties As such, the company's results may be materially different from the views expressed today. Further information regarding these and other risks and certainties is included in the relevant public findings of the company, as filed with the U.S. Securities and Exchange Commission. The company does not assume any obligation to update any forward-looking statements except as required under applicable law. Please also note that Expense Earnings Press Release and this conference call include the disclosure of unaudited GAAP financial measures as well as unaudited non-GAAP financial measures. Expense Earnings Press Release contains a reconciliation of the unaudited non-GAAP financial measures to the unaudited GAAP financial measures. We will now turn the call over to our co-founder, chairman, and CEO, Mr. He Xiaopeng. Please go ahead.

Hello, everyone. In the third quarter of this year, the number of small and medium-sized vehicles has increased to 25,666, which is 199% of the total growth. In the previous nine months, the total number of vehicles has exceeded 56,400, which is twice the total number of vehicles in 2020. In September and October, the number of vehicles has exceeded 10,000 in two consecutive months. Our goal is to hit 150,000 vehicles in one month. Then, with the rapid growth of the yield, the scale effect and the ratio of P7 yield reached 99% and 77%. Then, in the third quarter, our interest rate reached 14.4%, and the ratio increased by 2.5%.

In the third quarter of 2021, Xpeng achieved another record with deliveries reaching 25,666 units, representing a year-on-year increase of 199%. Our total deliveries for the first three months of 2021 exceeded 56,400 units, more than double last year's full-year delivery. Furthermore, our monthly deliveries surpass 10,000 units consecutively in both September and October and will strive to reach the target of 15,000 units of delivery per month. as rapid growth in deliveries continued to drive our economies of scale, and P7's mix in our deliveries grew substantially to around 77%. Our growth margin in the third quarter increased to 14.4%, up 250 basis points from last quarter.

Xiaofeng will insist on the technical route of all-war software and core hardware self-designed, and continue to expand its smart advantage.

Our unwavering commitment to in-house developed full-spec autonomous driving software and core hardware underpins Xpeng's ability to continuously strengthen our technology leadership.

On the 1024 Science Day, we displayed the X-Panel 3.5 X-Panel 3.5 GT. This is the development version. The road performance of this version is very good in the center of Guangzhou. From this, we can see the technological innovation and expansion of self-support driving. It is experiencing the acceleration of users faster than we originally imagined. And Xiaofeng is accelerating the technological innovation and the acceleration process.

At our annual Tech Day event, we showcase a P5 equipped with our city and GP development version, which was able to navigate through the complex driving scenarios in downtown Guangzhou. This highlights technological breakthroughs and rising popularity of advanced driver assistance systems are transforming users' mobility experience at an unprecedented speed. As our mission, FPeng is accelerating technology innovation and disruption in this field.

We believe that the competitiveness of the smart driver is due to the comprehensive capabilities of hardware, software, and data, including security. With the ultimate comprehensive cost and global-wide delivery efficiency, we have accumulated data and formative avoidance capabilities in a large number of Chinese driver scenarios. We are increasingly realizing that this will become Xiaopeng's important competitive advantage in the smart driver industry, and the gap between it and other companies. In the past three seasons, the X-Panel 3.0 software penetration rate of P7 has been close to 20%, and the software revenue is growing slowly. As of September 30, we have delivered more than 50,000 P7s in total, and more than 11,000 P7s have been delivered to X-Panel 3.0. In the three seasons of 2021, we have reached more than 60% of the maximum usage rate of NGP, and the penetration rate of NGP is also more than 60%.

In our view, our competitive advantages on driver assistance systems stem from our capabilities, including passenger safety to vertically integrate hardware, software, and massive data set in-house, overall cost efficiency, and rapid product iterations globally. Our ability to accumulate corner cases from complex and real-life driving scenarios across China is XPILOT's core competitive edge and will continue to strengthen our technology leadership over peers. The attach rate of XPILOT 3.0 software reached close to 20% in P7s in the third quarter, and revenue from software increased quarter over quarter. Among the more than 50,000 P7s that we have delivered as of September 30th, over 11,000 units were equipped with XPILOT 3.0. During the third quarter of 2021, the average monthly utilization rate of our NGP for highway and the highway NGP mileage penetration rate each exceeded 60%. And the NGP assisted our customers in driving approximately 5.51 million kilometers or 3.42 million miles on highways.

X3 launched X-Pilot 3.0 in the second half of 2022 will begin to explore the Robotech C business.

Our strategic goal in the near term is to improve the robustness and safety of our autonomous driving algorithms through generalization. I believe our ability to mass-produce turnkey solutions, including vehicles and software, to potential mobility service operators will bring a revolutionary mobility experience to users and create significant business value.

In addition to the soft hardware of the self-support value, we also insist on self-development at the core hardware level. For example, in China, we will first produce 800V high-pressure coal-fired coal platforms and 480,000W superchargers, which will charge for 5 minutes and carry 200 kilograms. In our fourth vehicle, which is the G9, our pre-controlled machine and electrical structure are also fully developed. The self-development of the core hardware layer will maximize the advantages of our full-time resource software, reduce costs and strengthen delivery, thereby strengthening our leading position in technology and cost.

XPeng not only develops full stack software in-house, but also design and integrate core hardware in-house to achieve rapid iteration of technology. For instance, we'll be one of the first in China to mass produce 800 volt high voltage silicon carbide cut platform and 480 kilowatt high voltage supercharging piles, which will allow 200 kilometers of driving range through only five minutes of charging. The domain controller and the electric and electronic architecture equipped in our fourth EV model, G9, will also develop in-house. We believe our ability to design this core hardware in-house will compound our competitive advantage in software, which will further cement our technology leadership. 我们将继续发展产品的快速迭代,并且扩大目标市场的用户渗透率。

In the future, we will accelerate the launch of products with a price range of 150,000 to 400,000 yuan. In the third quarter of 2022, our goal is to deliver our fourth car, which is the G9. The G9 hardware platform can support X-Panel 4.0 intelligent auxiliary driving. We believe that the launch of X-Panel 4.0 will truly change the customer mentality, and promote the current new engine car era to enter the real We are accelerating our product roll-out in order to increase our product penetration in our target market with price ranges between 150,000 to 400,000 RMB.

In the third quarter of 2022, we'll begin the deliveries of our fourth EV model, the G9, which supports XPILOT 4.0. In our view, X-PILOT 4.0 is going to fundamentally transform mobility experiences into the next stage. Our mission is to make smart EV accessible to a broader user group in the largest passenger vehicle market segment, delivering revolutionary smart EV experience across our target market segment.

Our third mass-produced P5 will be launched this year. In the past few decades, this price range has been unchanging. We believe that the P5 of Xiaopeng will subvert this traditional pattern and accelerate the spread of intelligent technology. Although we are facing a shortage of chips in a period of time, we can see that the demand for P5 orders has clearly exceeded our expectations. Currently, the delivery of P5 must be done after the Spring Festival holiday.

Officially launched on September the 15th, our third mass-produced model, the P5, began deliveries in October as planned. The launch of P5 marks for the first time that an industry-leading advanced driver assistance system and smart cockpit system in China can be deployed to smart EV models with price at around 200,000 RMB. We believe that P5 has achieved a purchase cost parity with its equivalent internal vehicles. I believe P5 will disrupt the family sedan market, which has been dominated by traditional ICE models that have remained largely unchanged in terms of technology over the past decade. Despite the recent challenges from semiconductor shortages, the orders for P5 have exceeded our expectation and the delivery of P5 has been scheduled until after Chinese New Year already.

Then last Friday, our fourth car, which is the self-propelled SUV Xiaopeng G9, was officially launched in Guangzhou at the Guangzhou Motor Show. Xiaopeng G9 will become the first mass-produced car of our X-Panel 4.0 self-propelled driving system. At the same time, G9 will be equipped with 800V high-pressure carbon-coated platform and Xiaopeng's XEA 3.0 electric engine structure. And I would like to emphasize that G9 is the first model developed by Xiaopeng for the first time in the world.

Last Friday, we unveiled G9, our flagship SUV, at the 2021 Guangzhou International Automobile Exhibition. The G9 will be our first mass-produced smart EV that supports XPILOT 4.0. Meanwhile, the G9 will be equipped with an 800-voltage high-voltage silicon carbide platform and XPANG's new proprietary XEEA 3.0 electric and electronic architecture that adopts a domain controller. deeply integrating hardware, software, and communication architecture to achieve powerful performance and high flexibility in OTA upgrade. Also, I would like to highlight that G9 is our first model to be conceived and developed from the ground up for the international as well as the Chinese market.

Next, I will introduce some of our progress in the overall strategic layout.

Next, I would like to walk you through our strategic initiatives.

I believe that in the coming years, we will see more breakthroughs in the self-driving, electric vehicles, and automotive technologies.

Accelerating our research and development for new models and core technology innovations, we continue to build out our R&D team. As of the end of third quarter 2021, XPeng's R&D staff comprised of over 4,000 employees, more than two-fold our staffing at the end of 2020. Over the next couple of years, you'll continue to see our technology breakthrough in uncharted areas.

The sales and service network of Xiaobong is expanding rapidly. As of September 32, 271 sales stores are covering 95 cities, 167 of which are headquarters stores and 104 are authorized stores. By the end of this year, we plan to have more than 350 sales stores. In the third quarter, the sales of our single-storied goods will increase. This reflects the overall ability and efficiency of our sales system.

Our sales and services network sustains rapid expansion. As of September 30th, Xpeng's physical sales network consisted of 271 sales centers across 95 cities in China. Of these sales stores, 167 were directly operated by us and 104 were franchise stores. We expect the number of sales stores to reach more than 350 by the end of 2021. Meanwhile, the average monthly store sales volume witnessed a sequential increase in the third quarter, reflecting enhancements to our store productivity and efficiency.

In terms of the network layout of the charging facilities, we will continue to work quickly, so the coverage of the charging network will be more comprehensive. As of September 30, we have 439 brand charging stations covering 121 cities. The total number of excellent free charging stations has reached 1,648, covering 221 cities. In the industry, we first went through the high-speed charging route of Jinghu and Jingang-2. Through 30 charging stations, an average of 170 kilometers per hour is achieved. There is a Xiaopeng charging station. By the end of 2021, we expect that there will be more than 600 small and medium-sized shopping malls, and the charging layout will be further expanded beyond the first and second-tier levels. We hope to complete the signing and construction of small and medium-sized shopping malls in all sub-states in China this year, and cover most of the sub-states and high-speed roads in China on the basis of perfecting the key areas in 2022 and 2023.

We also continue to rapidly build out our supercharging network. As of September 30, the number of X-Plane-branded supercharging stations increased to 439 across 121 cities, and the total number of free supercharging stations reached 1,648, covering 221 cities. We are the industry pioneer deploying supercharging capabilities across the entire Beijing-Shanghai and Beijing-Hong Kong-Macao expressways. The 30 XPeng-branded supercharging stations deployed alongside the expressways allow XPeng customers to access our supercharging facilities every 170 kilometers on average. We expect XPeng-branded supercharging stations will to be more than 600 by the end of the year, accelerating our supercharging network expansion across lower tier cities. We expect to accomplish the construction of infrastructure supporting XPeng's supercharging network spanning all prefecture-level cities in China this year. Furthermore, we will build on our well-established charging facilities in main areas to expand our network to cover a majority of country-level cities and highways in 2022 and 2023.

In terms of overseas market growth, Xiaofeng P7 has officially launched in Norway on October 25. We will continue to actively deploy in Norway and other European markets, including Sweden, Denmark and the Netherlands, and continue to improve our local sales and delivery service system.

On the overseas business front, Xpeng P7 was launched in Norway on October 25th. We will explore opportunities to expand in Norway and other European markets, such as Sweden, Denmark, and the Netherlands, and improve our sales, delivery, and service network continuously.

In terms of production and manufacturing, Xiaopeng's first factory, Zaoqing Factory, produces 20 hours a day. China China China China China China

With respect to manufacturing capabilities, our first plant, Zhaoqin Plant, is currently running on double shift, with production lines working up to 20 hours per day and production is ramping up rapidly. Total defined production capacity of Zhaoqing plant and the plants being built in Guangzhou and Wuhan will be more than 400,000 units per year. By adopting a double shift production at these plants, annual production can reach up to 600,000 units, laying a strong foundation for our high growth in the next two years.

So in terms of the layout of the smart transportation ecosystem, on October 19, Xiaopeng Qizhou's ecosystem enterprise, Xiaopeng Huitian, announced that it had completed more than $500 million of A-wheel farming. This is the largest unit of farmland acquired in the industry in Asia's low-altitude air-to-air aircraft field so far.

On October the 19th, Xpeng's urban and mobility affiliate, HT Aero, announced that it has entered into a definitive agreement with a consortium of investors to raise over $500 million for its Series 8 capital funding. This marks the largest single-trench fund raising to date in Asia's low-altitude flying vehicle sector.

Xiaofeng Car Self-Defense Force has set the highest standards in ESG, using green low-carbon smart transportation to help China achieve its carbon footprint and goals. On October 15, Xiaofeng released its first ESG report, and received the MSCI ESG level AA level for two consecutive years, which is the highest level of global car level. Then in the future, we expect to continue to use technology to carry out full-scale research in the nuclear field to achieve better results through technological innovation.

Xpeng is always committed to upholding the highest standards in ESG practices and is driving contributions to China's carbon neutrality goal by advancing our green, eco-friendly, smart mobility offerings. On October 15, Xpeng released its inaugural ESG report, and notably, for two consecutive years, Xpeng received an AA rating from MSCI ESG Research, topping its global auto peers. In the near future, we expect to make greater contributions to carbon neutrality through technological innovations utilizing our in-house developed full-stack technologies.

Finally, in the fourth quarter of this year, we will work hard to overcome the challenges of chip shortcomings and the pandemic in some areas. We expect that in the fourth quarter of 2021, the total number of deliveries will be 34,500 to 36,500. The revenue is expected to be between 7.1 billion and 7.5 billion yuan.

In conclusion, we'll make every endeavor to overcome the challenges stemming from chip shortages and the COVID-19 pandemic, which is still prevalent in some parts of the world. In the fourth quarter of 2021, we expect our smart EV deliveries to be approximately 34,500 to 36,500 units, and our total revenue to be approximately 7.1 to 7.5 billion RMB.

Thank you, everyone.

With that, I'll now turn the call over to our VP of Finance, Mr. Dennis Liu, to discuss our financial performance for the third quarter of 2021.

Thank you, Mr. He, and hello, everyone. Now, I would like to provide a brief overview of our financial results for the third quarter of 2021. I will only reference the IMB in my discussion today unless otherwise stated. Our total revenues were $5.7 billion for the third quarter of 2021, an increase of 187% year-over-year, and an increase of 52% quarter-over-quarter. Revenues from vehicle sales were $5.5 billion for the third quarter of 2021, an increase of 188% year-over-year, and an increase of 52% from last quarter. The revenue increase was mainly attributable to higher vehicle deliveries, especially of the P7, resulting from channel expansion and brand equity improvements. Revenues from service and others were $0.3 billion for the third quarter of 2021, an increase of 182% year-over-year, and an increase of 47% quarter-over-quarter, mainly attributed to more income from service, parts, and accessory sales, in other words, higher accumulated vehicle sales. Growth margin was 14.4% for the third quarter of 2021 compared with 4.6% for the same period of 2020 and 11.9% last quarter. Maple margin reached 13.6% for the third quarter of 2021 compared with 3.2% for the same period of 2020 and 11% last quarter. The vehicle margin improvement was primarily attributable to better product mix and manufacturing efficiency enhancement driven by economies of scale. R&D expenses were $1.3 billion for the third quarter of 2021, an increase of 99% year-over-year, and an increase of 46% quarter-over-quarter. Many due to, one, increasing employee compensation as a result of expanded research and development staff, and two, higher expense relating to the development of the G9, the P5, and related software technology to support future growth. SG&A expense was $1.5 billion for the third quarter of 2021, an increase of 28% year-over-year, and an increase of 49% quarter-over-quarter, mainly due to, one, higher marketing, promotional, and advertising expenses to support vehicle sales, and two, expansion in our sales network and associate personnel costs and commission for different store sales. Our income was $0.2 billion for the third quarter of 2021, including government subsidies of approximately $0.3 billion, partially offset by the location and disposal costs, about $0.1 billion related to the high-mark rent. Lost farm operations was $1.8 billion for the third quarter of 2021, compared with $1.7 billion for the same period of 2020 and $1.4 billion last quarter. Discluding share-based compensation expense, the non-GAAP loss per operation was $1.7 billion for the third quarter of 2021 compared with $0.2 billion for the same period of 2020 and $1.3 billion last quarter. Net loss was $1.6 billion for the third quarter compared with $1.1 billion for the same period a year ago and $1.2 billion last quarter. Non-GAAP net loss was $1.5 billion for the third quarter of 2021, compared with $0.9 billion for the same period of 2020, and $1.1 billion last quarter. In July, we completed dual primary listing in Hong Kong. As of September 30, 2021, our company had cash, cash equivalent, restricted cash, short-term and long-term investment, and deposit in total $45 billion. To be mindful of length of our earnings call, I will encourage listeners to refer to our earnings press release for further details. This concludes our prepared remarks. I will now open the call to questions. Operator, please go ahead.

Thank you. Again, to ask a question, please press star 1 on your telephone keypad. For the benefit of all the participants on today's call, If you wish to ask your question to management in Chinese, please immediately repeat your question in English. For the sake of clarity and order, please ask one question at a time. Management will respond, and then feel free to follow up with your next question. Our first question comes from Tim Hsiao from Morgan Stanley. Your line is open.

Hi, Mr. Hurd, Brian, Dennis, and Tim. Thanks for taking my questions, and congratulations on the great results. I have two questions. The first question is just one to quickly follow up about the rubber taxi plans we just mentioned during the presentations. Could you share more details in terms of the scale, business motive, business model, and more? Is it more like a demonstration of the capability of X pilot 4.0? And are we going to have any different or facelift model for the rubber taxi fleet operation next year, or simply leverage P5 for the fleet operation in Guangzhou? And are we going to do that in other places or other province? So that's my first question. 首先可不可以请管理层再多分享一下 刚才简报中提到Robert Taxi的这个计划 目前针对计划的规模 背后的动机 盈利模式以及覆盖的市场 是不是可以再多讲解一下 The main question is to show the X8's 4.0 full-screen AI driving point-to-point function. Will there be any modified models or directly use P5 as our operating vehicle? This is my first question.

OK, thank you. We are still in the middle of the internal communication. I can only share with you some very initial framework. We hope to use the car we are now planning to mass production, from X3.3.5 to 4.0, and even more OTAs in the future. We are doing local testing to test the ability of the software, hardware, and data used in this program for a long time. After this training, we will improve our future 4.0 and even 5.0 capabilities. Thanks.

Thanks for your first question. Now, in regards to RoboTaxi and its development and plan, we are still in the process of doing internal discussion. So I can only share with you some of the basic and initial frameworks and some thoughts. So basically, we are going to adopt the current models that we have, and we are going to use Expeller 3.5 to 4.0 or even further future OTA versions on these leads. And our motives or target is to test and train our closed-loop capability of integrating software, hardware, and also our data capability in urban roads so that we can serve the future development of XPilot 4.0, 5.0, or even further more advanced versions of our XPilot system. software upgrades. And so we are also not looking into doing mass services in the future. We are looking to partner with the current operators so that we can together contribute to the development of this industry.

Great. My second question is about P5. I think we already touched on the topic a little bit during the presentation, but we have more updates on P5's order intake and backlog at the moment. Probably just some rough quantitative guidance would be great, and roughly how many percentage of the consumers are willing to take the cards first without radar. In the meantime, We have a rough idea about the order mix of 550P and 600P, i.e. the two models with LiDARs integrated. Thank you. 管理層想跟進一下有關於P5的訂單, 是不是可以稍微量化一下目前我們整個在手訂單, 新的訂單的狀況跟在手訂單的量。 另外就是就這次的新車交付, How many percent of the car owners are willing to pick up the car first and then install the radar? In addition, can you follow up on how many percent of the car owners are willing to choose the two highest-end vehicles with laser radars?

Hi, Tim. It's Brian. Let me just first of all respond. We never give guidance or forecasts on our specific orders and backlogs. So that has been our policy since our listing. But just to give you some color on P5, we actually saw very robust demand, as Mr. He stated in the opening speech. We actually saw the backlog now is thanks to post the Chinese New Year. So an average in some models up to four months of backlog wait time so that's also very in a robust demand in terms of the mix the only thing i can tell you just a few facts one is that uh more than 50 percent of the orders that we receive are for software subscribing models which incorporates xpilot 3.0 or 3.5 hardware architect And also, regarding to your question, what's the percentage of orders that actually choose to take delivery and have future installation of the radar, actually more than 80% of people actually have chosen that option. So we actually see very strong favorable sort of reception to our P5.

Thank you, Brian. Thank you, Mr. Ho. And congratulations, again, on the great resale. Thank you very much.

Our next question comes from Jeff Chong from Citi Airlines Open.

Hello, everyone. I'm Jeff. I have three questions. I'll speak Chinese first and then translate into English. First, should the service plan of RoboTaxi be half as cheap as Baidu and Poli? This is our starting point. If so, I'd like to ask, is RoboTaxi's daily travel ten times as much as a private car? If so, does this mean that in terms of the data that we accumulate, in terms of the software AI learning process, it will be faster than our competitors? That's why we want to take the RoboTaxi route. I want to clarify this first. The second question is, I would like to ask, how do we see the tick rate of our autopilot in the first half of next year? Because we know that P5, there are five cars, there are three cars, all of them can have this service. But P7, there are only three cars, only one car has this. And then next year, in the first half of next year, our city's autopilot 3.5 can also be OTA, FOTA can upgrade. I would like to ask, how do we see the tick rate of this software in the next year? And the last question is about the very good four-week index. But the market has already predicted that there should be 15,000 units in December. If the whole four-week index is like this, it means that there will be only 10,000 units in November. That means that there will be no growth in October. So I'm thinking, is it because I think that November's 10,000 units are too conservative, or is it because Okay, so let me translate that into, so question one is about the robo taxi. So can we quantify our solution on robo taxi? How much cheaper than Baidu and Tony and other competitors? And secondly, it's about our autopilot penetration rate into the first half next year. Given that our autopilot 3.5 urban version OTA will be activated soon. And also for P5, the 3 out of 5 model will offer this function. And this is much higher than the P7, one model out of three. And my last question is about the November and December monthly sales run rate guidance. Thank you.

Thank you. Let me answer the first question. I believe that we are using the current small and medium-sized vehicles to support and serve. So I believe that it will be lower than most other prices from a hardware point of view. So from our previous data, A car like this can run nearly 80 to 10000 kilometers in a month. We believe that a lot of data will be useful for us in the field of sensor, map, and color case data. This is what we have always wanted to do. At the same time, we want to prove that in addition to software and algorithm, hardware and reliability, How can we improve safety and reduce costs in the future? This is what we are thinking about in the Logotype system.

Thank you. Let me take the first question. Now, in regards to RoboTaxi, basically, we're going to adopt all of our mass-produced models. So, technically speaking, their costs would be a lot cheaper to our competitors. And on average, every month, we expect to see 8,000 to 10,000 kilometers of driving mileage that we can accumulate from our fleet. And so the purpose for using those kind of fleets at Robotaxi is to gain further ability and capability of getting more knowledge in the maps and also accumulated corner cases in our database so that we can further explore and enhance our overall capability of integrating the software, data capability, hardware, and also to build up a higher level of safety and reduce costs further. Thank you.

And, Tim, this is Brian. Let me address your next two questions. First of all, on the question of P5 software subscription, again, I don't want to offer any guidance and forecast, but I want to point to a few facts. First of all, right now we're seeing more than 50% of the orders for P5 are for the software subscription-ready hardware models. So definitely you can see the mix is higher than the P7 that we actually see in terms of applications. the software-capable models. Secondly is that the delivery of the software subscribing models also are subject to some supply chain bottleneck. As you saw earlier, we actually need to install some of the radars post the delivery. So as a result, we also offered software for free for some of these customers. So I think in the short term, you will see the impact due to a number of these factors. But in the long run, we're confident that T5 will see higher software substitution rates for both Xpilot 3.0 or Xpilot 3.5 software. As you know, expired 3.5 software is priced higher than expired 3.0. Your third question on the guidance, again, I don't want to give additional color specific month delivery, but I want to reiterate two things. One is that we are still striving for hitting a monthly peak delivery of 15,000 per month in the next two months. So that's our goal, and I think we have a good chance of hitting that second thing i want to mention is that the supply chain constraint is still very severe um so the visibility to uh the chip shortage as well as some other uh constraining factors uh is it makes the forecast is very short uh and not very reliable in the coming months or so so that's why i think we provided the guidance with the best knowledge that we see today. But I think, you know, this is where we see and also how we feel about the monthly delivery peak will arrive.

Our next question comes from Ming Li from Bank of America. Your line is open.

Hello, I'm Ming. I have two simple questions for you. The first question is, we noticed that in the third season of G3, there is a production switch to our own factory. I would like to know if the production switch to our own factory is different from the original working mode. If not, Because this quarter, we also noticed that the sales ratio of PT has increased a lot. G3, in comparison, is actually similar to the previous quarter. When will we put the new G3 facelift or the new generation out in the future? And then improve the design power of our G3. This is my first question. The second question is about, can you ask Guanlin Chen to briefly talk about X-Palette 3.5, My first question is regarding the G3. Because in the third quarter, you start to make G3 in your own plant. And how do you see the margin improvement after you make the car in your own plant? And also, in the future, when do you expect to launch the facelift or new generation of G3 to improve? the margin of the product. And my second question is, what are the different functions between X-PILOT 3.5 and 4.0? Thank you, Juanita.

Yeah, I mean, this is Dennis. Let me address the first question. We actually, we moved our G3 to our Zaoqing on-prem in August to commence with the FASD model, the GSUN-I. As you rightly pointed, we actually, we can see the scales improvement. So we see the, you know, better labor and overhead efficiency compared with in old days, we have the high amount to do the country manufacturing. So we are seeing the margin improvement on Qishan I already, which we have launched and we are setting the vehicle already. That's number one. And number two.

Go ahead. As for the functional differences between XP Pilot 3.5 and XP Pilot 4.0, we will not discuss them in detail today. After next year, we will discuss XP Pilot 4.0 in detail. They will get support. But we hope that X-Panel 4.0 can be supported in China, in Europe, and in more cities in more roads. And X-Panel 4.0's entire electronic appliances, the signing room, and the perception and reasoning platforms are actually quite advanced. We believe that these are two sets of self-supporting platforms with both inheritance and bigger vision.

Thank you. For your second question, actually, we're not going to disclose any more details of the difference between XPilot 3.5 and XPilot 4.0 because we're going to launch and talk more about XPilot 4.0 next year. But right now, I can give you a general idea about the differences. Basically, for XPilot 3.5, we are able to cover the urban NGP function that supports quite a number of cities and some urban roads. But that is a comparatively smaller number compared to XPilot 4.0 because we expect XPilot 4.0 to be able to support the majority number of roads and city in not just China but also Europe as well and on top of that xpilot 4.0 is also equipped with a higher level of sensing capability electric and electronic architecture and also a more stronger sort of algorithm calculation computing power as well thank you

Our next question comes from Ben Wang from Credit Suisse. Your line is open.

Thank you. I got two questions. Number one is about the margin improving. Can you break down the different factors about the gross margin increase? For example, how much came efficiency and how much came from the product mix. Thank you. That's number one. Number two is more a technology question, because you actually will migrate to the 800-voltage system. Does that mean for the current supercharger that would need to have a dramatic rebuild? What's the robbery cost from this rebuild from kind of maybe 350-voltage to 800? Thank you.

Can you repeat the question in Chinese? Okay.

The first question is about the change in the interest rate. Could you please elaborate on the three reasons for the increase in the interest rate? For example, how much comes from the increase in the operating rate of the software? How much comes from the improvement of the combination of products? How much comes from the cost savings brought by the scale efficiency? This is the first question. The second question is that the G9 will soon enter the 800-volume cycle. Thank you.

Hey, Ben, this is Dennis. Let me address your first question. Regarding the quarter over quarter margin improvement, majority of improvement came from the mixed improvement. You know, in the quarter two, our total P7 accounted for about 66% of our sales in quarter two. But in third quarter, because of the G3 migration, we actually saw more P7. P7 accounted for about 77% of our total sales in quarter three. So that's the major reason for the marginal improvement. And the other improvement would be for the labor and overhead efficiency due to the scales, the economy improvement. because now we are producing all the vehicles in our Zhaoxing plant, so we see the kind of efficiency improvement. However, we also have some cost increase to offset partially the good news. For example, the raw material increase and also the chip shortage, that offset partially of the good news I mentioned briefly. So I hope I answered your first question.

Okay, let me simply answer the second question. From the G9 to the next generation, we will gradually introduce a 800V charging system. The current charging station is 180,000W. Next year, we will upgrade to 360,000W. A little bit later, we will reach 480,000W. So, I believe that the construction of the high-speed railway station is a land-based method. Only the new station or the new station will be used to upgrade from 360 to 480 km. We believe that there will be no obvious waste in this system. We will directly use the latest high-speed railway station system through continuous construction stations.

Let me take the second question. Now for G9 and other future models, we are going to adopt the 800-volt high-voltage charging. And so for our current charging piles, we are going to upgrade them to, I mean, in the future to 360 kilowatts and also in the future, 480 kilowatts. And so as we lay out our deployment of supercharging stations, we are going, I mean, the newer ones, we're going to take the high-voltage charging facilities forward. such as 380 kilowatts and 480 kilowatts for sure. So there won't be any waste because we will gradually roll out our plan and network building for a construction for the charging facilities.

Yes, because each charging station consists of different piles of different voltage, and so we are going to be able to diversify the kind of piles in each single station. Yes.

Our next question comes from Nick Lai from JPMorgan.

Nick, we lost you. You still there? Nick?

His line got disconnected. Our next question will be from Edison Yu from Deutsche Bank. Your line is open.

Great. Thank you for taking the question. I actually just have one. Could you maybe go over long-term what kind of opportunity you see in Europe? Obviously, don't expect anything too concrete, but what kind of potential volume or what kind of percentage of market you can maybe target or expect to see there in the longer term?

So, hi, Edison. It's Brian. First of all, I mean, to answer your question, I want to emphasize that our current plan in Europe is mainly focused on building our presence capabilities and also make sure we have the right network and brand awareness built up. So we are not emphasizing on delivery numbers in the near term. And in the long run, I would say that multi-year, we actually see ourselves as a global company with around half of the volume ultimately will come from outside of China and half the volume inside China. That's our goal. Again, it's a multi-year goal, but that's what we're striving for.

Thank you.

Our next question comes from Paul Gong from UBS. Your line is open.

Hi. Thanks for taking my question. I have two questions. The first one is still regarding the autonomous driving. My understanding is from the Xpilot 3.0 to 3.5, the hardware-wise, you cannot migrate to the newer version. because of the different configuration of the hardware. How about from 3.5 to 4.0, we will support upgrades around the hardware for the previous buyers of the system. 我有兩個問題,第一個問題是關於這個 ADAS 的可升級性。 就現在從 3.0 到 3.5 的話,因為就是說硬件上面不兼容, 這硬件上面沒辦法升級。 從 3.5 到 4.0 的話,這裡會有就是說硬件的可升級嗎?

OK. OK. It's just an improvement, but we can lay the foundation for the future of real L4. So we have a lot of, especially about the previous ACC, LCC, and the current one, including APA, we have all rewritten. So there are a lot of different things in this system.

Actually, the hardware cannot be migrated to XPilot 4.0 because of the fundamental upgrade in computing power and sensing capability. We also upgrade the whole electronic and electric system as well. So basically, these two XPilot systems are built on top of two very different infrastructure. Also, we are laying the foundation for the future development of L4 level of autonomous driving. So you will see fundamental changes in terms of the ACC, LCC, and also AP8 programs.

Okay. My second question is regarding the Norway operation. I think G3 has been available there for over one year, and P7 was available since last month. So far, how have you been feeling the learning process in Norway? Like, what lessons you have learned? And what has been preventing you selling more vehicles over there, given the cost advantage? And some of the smart features of the Expo cars are pretty competitive in that market. 第二个问题的话是关于这个挪威的运行, 我记得就是说G3已经是在那边有一年多了, 然后P7的话也都从上个月开始了。 In terms of China's cost advantage and the level of smartness of small cars, what factors have restricted you from selling more cars? In our internationalization strategy, our initial goal was to

From 2020 to 2022, we will lay the foundation for internationalization in three years. This foundation includes hardware development, security and data protection, including software development, China China China This means that between 2023 and 2025, all the new technologies will support the safety and data standards of China, Europe, and even more countries. After that, we will have greater capabilities to deliver and ensure that the entire logic system of the policy is correct. This is our entire thinking strategy.

Our international strategy determines that our target between 2020 to 2022 is to make a good foundation for our international expansion in the following aspects, which are the software and hardware R&D development, our safety and data protection, and also our team and organization structure revamp. So basically, in Norway and other international market, you would see us doing more restructuring and exploration in terms of our products, technology, and also our organization. So sales right now is not our priority in the international market. But in the future, from 2020 on to 2025, we are going to conceived and developed models that are suitable for not just the Chinese market, but for the international market, in the sense that we are going to comply with all the safety regulations and environmental regulations to the highest standards in the European market, as well as the global market in general. So by that time, I think we'll be more prepared to hit the south target or further expand into the overseas market.

Thank you very much. Hi, operator. We are ready for closing remarks.

Thank you. All right. Please go ahead.

Hey, operator. Nick, you still have one more question left for JP Morgan, Nick Lyon.

OK. Our next question, again, is from Nick Lyon. Your line's open.

OK. Can you hear me? Hello?

Yes.

Okay. Let me ask questions very briefly. I know I'm running out of time. Maybe two questions quickly. One is on the profit margin outlook, and the second is on G9 sales outlook. Let me ask quickly in English and then translate in Chinese. On the profit margin, indeed, great result in third quarter. As we move to first half next year, how should we think about the possibility considering positive and potential industry headwind? On the positive side, P5 should contribute to the initial volume, and P5 share large part of component with G3, so we should have efficiency gain. But at the same time, we have input material price hike, And also P5 carry lower profit margin, that could be margin diluted as we sell more P5 into our first half next year. And that's the first question on margin. And the second on G9 sell, considering G9 addressable market, how should we think about the monthly run rate when G9 is fully ramped up? Is it fair to compare G9 with comparable peers like Neos ES8 or maybe Li Auto's product? It's a very good performance in the third quarter. In the fourth quarter and the first half of the year, from a positive point of view, P5 and G3 share a lot of components. There should be a lot of efficiency gain. But at the same time, P5's horsepower is relatively different from P7. When we sell more G3 and P5, the overall horsepower will have a little bit of a lackluster effect. Thank you. These are two simple questions.

Hey, Nick, this is Dennis. Let me address the first question regarding the margin. As I mentioned previously, our quarter three margin improvement compared with quarter two is primarily due to the product mix improvement more P7. You're right. Going to the first quarter, we will deliver P5, and then the G3i volume will be better than the old model. So the mix will have the impact. I mean, the P7 percentage in terms of total sale will reduce. But having said that, we actually, in our future product, including the G9 and the future model, We will have better margin. For example, bigger car, we have seen better margin for the future models. We will see the near-term impact due to the mix change. But going to next year, we actually will see more good news from the better product mix, including the new product introduction.

Yeah. So, Nick, it's Brian. Again, no forecast on G9, unfortunately, but I just wanted to tell you that we are very confident that G9 will be a flagship product that, you know, will deliver, I think, results similar to our top-selling models for a number of reasons. First of all, the G9 is positioned and the size more similar to ES6 than ES8. So the addressable market is actually much bigger. Secondly is that upon the launch of G9 at the time, it will offer superior capability in a number of areas in terms of autonomous vehicles. highly capable autonomous driving capabilities, as well as fast charging capabilities and leading electric architecture components for much faster upgrades, etc. So we think it will be probably one of the most a technologically superior product coupled with, you know, the attractive design and also by then a leading, you know, infrastructure of Exxon network. So we're very actually confident that it will achieve, you know, strong debut. And also to be note that G9 is designed for the global market. So that product only will be sold in China. It will also be designed and sold in international markets and with europe specifically in mind so i think we think that this is going to be a very important product for expansion forward thank you there is no further question this time you may continue

Thank you once again for joining us today. If you have further questions, please feel free to contact EXCOMM's Investor Relations through the contact information provided on our website or the Pearson Group Investor Relations.

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