XPeng Inc.

Hello, ladies and gentlemen. Thank you for standing by for the second quarter 2025 earnings conference call for Xpeng Inc. At this time, all participants are in a listen-only mode. After 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. Alexei, Head of Investor Relations and Capital Markets of the company.

Thank you. Hello, everyone, and welcome to XPeng's second quarter 2025 earnings conference call. Our financial and operating results were issued by our Newswire services earlier today and available online. You can also view the earnings press release by visiting the IR section of our website at ir.xiaopeng.com. Participants on today's call from our management team will include our co-founder, chairman, and CEO, Mr. He Xiaopeng, Vice Chairman and President, Dr. Brian Gu, Vice President of Corporate Finance and VW Projects, Mr. Charles Zhang, Vice President of Finance Accounting, Mr. James Wu, 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 on 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 Securities Litigation Reform Act of 1995. Forward-looking statements involve inherent risks and certainties. As such, the company's results may be materially different from the views expressed today. Further information regarding these and other risks and uncertainties is included In the public filing of the company as filed with the U.S. Securities and Exchange Commission, the company does not assume any obligation to update forelooking statements except as required under applicable law. Please also note that EXPEN's earnings price release and this conference call include the disclosure of unordered GAAP financial measures as well as unaudited non-GAAP financial measures. XPeng's earnings press release contains a full reconciliation of the unaudited non-GAAP measures to the unaudited GAAP measures. I will now turn the call over to our co-founder, chairman, and CEO, Mr. He Xiaopeng. Please go ahead.

Hello, everyone. In 2025, the second quarter, the trend of small and medium-sized cars has a core business and financial indicators, including sales, In the second quarter, our delivery volume is 103,181 units, with a growth of 242%. Among them, the M03 Max version of Mora is on the market and on sale. In the middle of the whole series of orders, the ratio continues to exceed 80%. So the independent high-end supply value is near the general market. Not only that, in the face of the intense price competition and roll in the industry, Xiaopeng insists on long-term monitoring of news resources. At the same time, we continuously improve the capabilities of various infrastructure systems at a priority scale, including our ability to operate. Therefore, our interest rate has improved in the past eight consecutive seasons. In the second season, it increased by 3.8% and reached 14.3%. Hello everyone, in Q2 2025, we achieved a record high performance across our core business and financial metrics.

including deliveries, revenue, gross profit margin, and cash on hand. During this quarter, our deliveries reached 103,181 units, a 242% increase year over year. The Mona M03 Max immediately became a bestseller upon launch, accounting for over 80% of total Mona M03 sales and leading the way in bringing high-level ADAS technology to the mass market. Despite intense price competition, we remain focused on steady long-term sustainable growth, prioritizing scaling up whilst continuously strengthening the foundational systems. Our vehicle growth margin increased 3.8 percentage points quarter over quarter to 14.3%, marking the eighth consecutive quarter of improvement and raising the company's overall growth margin to 17.3%. net losses narrowing further. Free cash flow in Q2 exceeded RMB 2 billion, and total cash on hand at the end of quarter surpassed RMB 47.5 billion.

In July, Xiaotong G7 launched its first Turing AI chip, which we call Turing AI. In the past four weeks, it was the first to deliver or get the same level of sales, and it is very happy that the Osher version accounted for more than 50%. This means that we will start to lead the industry into the arrival of MAX and ULTRA, which are two more advanced automatic assist driving times. At the same time, in the fourth quarter of this year, Xiaopeng's Kunpeng super electric car will start production. This means that in 2025, we will fully complete the upgrade of the two new generation technology platforms, It includes the pure visual frame, AI model, Turing AI chip, and 51 super fast telecom and Kunpeng super electric series. I believe it will open the technical gap with competitors to ensure that in the next two years of large product cycle, we can achieve a better globalization, better supply chain management, and the ability to quickly improve production capacity.

In July, we launched the G7, our first model featuring in-house developed Turing AI SoC. Over the past four weeks, it has become the top among its competitors, with the ultra-trim making up over 50% of sales. The Max Plus Ultra experience marks a new phase in high-level ADAS in the industry. Our Kunpeng Super Electric models will begin mass production in Q4, leading us to fully upgrade to a generation technology platform with intelligence and the new vehicle dual energy system by 2025. This platform will include vision-based ADAS, advanced AI models, The Turing AI SoC finds the ultra-fast charging battery cells in the Kunpeng Super Electric system. These innovations will give us a generational lead, allowing for greater modularity, better supply chain management, and scalability during our product cycle over the next two years, boosting our sales growth.

我相信AI汽车的竞争在相当的一段时间内会是 a marathon that has no end. Therefore, we will continue to strengthen the overall infrastructure and focus on the four core pillars. The first is technology lead. We hope to build a full-scale technical war in the physical world of AI, including the full-scale speech in the vertical field and the cross-border fusion of hardware and software. In this field, we will work hard to make the design, AI, and quality of these three areas more solid. The second part of this basic ability is organizational capability. We hope that the team will have excellent and suitable talents to do the management, automation, and tools well. From the CEO, we will fully apply AI in the management, development, and cooperation of companies. to maximize the group welfare effect of the enterprise. Third, Xiaopeng will establish a stronger commercialization capability, to make good user products into good commercial products, and make the brand strong, know how to save money, know how to spend money, and make more money. Fourth, we will further improve the basic capabilities of globalization. Xiaopeng Globalization and many companies The race for AI-powered vehicles will continue indefinitely.

To succeed, I aim to strengthen the company's core capabilities in four key areas – First, focusing on technology leadership by developing AI capabilities in the physical world across all domains. This includes full-stack self-development within specific vertical areas and cross-domain integration. In vertical areas, we are dedicated to enhancing design, AI, and quality. Second, building organizational strengths by attracting and retaining top talent suited to each role. world-class management practices along with automated tools. Starting with a CEO, we will leverage AI to optimize the management, R&D, and collaboration, maximizing organizational compounding effect. The third focus is exciting in commercialization, shifting from creating user-centric products to developing commercially successful products and establishing a strong brand while reducing cost and increasing profitability. Lastly, our globalization approach is unique. Having integrated the global market needs into our product planning years ago, we attract international users with superior technology and quality to long-term growth with patience.

接下来,我们30万级的豪华自能轿跑全新的小鹏P7会在下周正式上市。 It represents an opening spirit for exploration. I believe this is also a performance of Xiaopeng's corporate values. PG does not follow the trend, because we believe that superior design and technological innovation itself can lead to the trend. Then PG will have a mobile style, extreme control and leading generation of AI technology. We hope PG will become a young dream car. Then PG represents a new goal for Xiaopeng's future design. Every new car launched by P7, we all pursue the technology first, and at the same time, we want to develop the product logic first. We will give more support and attention to the styling team. We hope that the future car, products, and the entire experience can bring you a different design and mood. We are also very happy that since the first show, we have seen the attention of users to P7. We expect that the sales of the new P7 will be the top three in the pure electric car market of 300,000 units. So with the delivery of the new P7, my goal is to start in September. The monthly delivery of Xiaobeng will be more than 40,000 units a month.

Next week, the all-new Exxon P7, a premium intelligent sports sedan in the RMB 300,000 price range, will be officially launched. The P7 reflects our innovative spirit and core values, showcasing the exceptional design and advanced technology can set market trends rather than following them. With its distinctive style, thrilling performance, and generation-ahead leading AI technology, the P7 is the ideal vehicle for younger consumers. It highlights EXPON's future design approach for all models following P7, combining top-tier technology with superior aesthetics. We are investing more strategically in our styling team to ensure future models deliver innovative, emotionally impactful designs. Since its debut, user interest has vastly exceeded expectations. with pre-sales orders surpassing all previous expo models during the same period. I anticipate the new P7 will rank among the top three best-selling pure electric sedans in the sub-300,000 RMB segment. With deliveries of the new P7, we aim for monthly sales to steadily surpass 40,000 units starting in September.

今年四季度我们还会 to launch Xiaopeng's first model, Kunpeng Super Electric Car, also known as X9. This iconic car will officially start the big product cycle of one car and two energy sources. Kunpeng Super Electric X9's pure battery life will exceed 450 kilometers, and its comprehensive life will exceed 1,500 kilometers. I believe this will be the champion of the same product. 那么之后我们还会发布多款超级电动车型的产品。 那么这些产品的目标都将是具备同类新能源产品中间最长的存电续航和最快的5C超级充电。 In Q4, we will introduce the X9 Kunpeng Super Electric Edition, our inaugural Kunpeng Super Electric vehicle, marking the beginning of our one-vehicle dual-energy era.

With over 450 kilometers of pure electric range and a combined range surpassing 1,500 kilometers, this model will lead in range within its category. Moving forward, we plan to launch several more super electric models, each offering top-tier pure electric range and 5C ultra-fast charging capabilities that exceed those of comparable NEVs on the market.

I believe there is a huge difference between Xiaopeng's car and other cars. It comes from our firm commitment to hardware and hardware in the core technology of the past 11 years. These ten years of persistence have allowed us to use AI driving to jointly develop AI chips and physical world models in a large-scale way, and to integrate them across the region, and to quickly deal with them through a vast amount of data and an automated sound system. In the long term, it will definitely be a super smart machine that will be built into the physical world. So from the launch of Xiaopeng G7 in the third quarter and the beginning of the upcoming PC, we will first start the generation of L3-level computing power. So Xiaopeng's ultra version will be equipped with three自然的图灵AI芯片. Effective computing power will reach 2250TB. So this is in the middle of the global mass-produced model, which is at the absolute top level. It is the latest flagship of the same line, with a speed of 3 to 10 times that of the previous model. The G7, P7, and subsequent models of the European version of Xiaopeng will be deployed in the VLA and VLM models. Based on the scale of the models deployed by Turing AI, we will increase the number of models to at least 1 billion parameters. What distinguishes us from automakers and software companies is our decade-long delegation to full-stack in-house development of core hardware and software technologies.

This commitment has enabled us to co-develop high-performance AI chips and foundational models for the physical world, achieving true cross-domain integration and speeding up iterations through extensive datasets and computing infrastructure, ultimately leading to AI superagents operating in the physical world. Beginning with the Xpeng G7 and P7 launches in Q3, We are leading the way in L3 level computing power. All ultra trims across our entire model lineup will feature the three in-house developed Turing AI SOCs, providing a total computing power of 2,250 TOPS. This positions us as the global leader among mass-produced vehicles with over three times the computing power of our latest flagship competitors. These ultra-trims also include in-vehicle VLA plus VLM models. Driven by Turing AI SoCs, the scale of our in-vehicle models will increase by an order of magnitude, reaching billions of parameters, with VLA models running at twice the frame rate of competitors, truly exemplifying smarter brain, agile control, and elevating the safety and user experience of ADAS.

So I hope that Xiaopeng's VLA and VLM will become the safest driver for users and the closest smart driver, so that the whole travel life will be more relaxed and fun. So for the G7 Ultra version of the car, we will start to promote the initial version of the VLA model in the near future, and in the next few months, through OTA, I will A bold goal for itself In the next 18 months, Xiaopeng based on Turing chip VLA and VLA model capabilities Will become more than the current industry mainstream program and auxiliary driver capabilities Increase by more than 10 times Then we will be safe In the whole scene and the overall user experience On the target of the L3 with a driver Realize a short-term lead Then I will share with you in advance here We support The L4 model will be launched in 2026, and will start operation and service in four parts of the region. We have seen that there are already a lot of cars in the market that support the L4 in the rear. However, there are not many cars that really have the power, software, hardware and cloud capability of the L4 in the front. Xiaofeng will be the first car in China. Our new generation of humanoid robots has also made a very remarkable breakthrough. I believe that based on Xiaopeng's Turing chip, V2A and V2M, we are moving forward quickly towards the mass production version of the 2JL4 capability in the robot field. We are working on the process preparation of our humanoid robots for the second half of 2026.

那么在之后小鹏的科技日,我还会跟大家介绍我们全新一代能行机程的相关情况。 I expect the Xpon's VLA plus VLM to become the safest driver and the most attentive smart assistant for users, simplifying and enhancing travel. We plan to roll out the initial VLA model to G7 Ultra owners soon, with rapid OTA updates over the coming month. Over 18 months, our Turing SOC-powered VLA is expected to outperform industry mainstream urban ADAS solutions by over 10 times. We aim to meet supervised L3 safety, coverage, and user experience, marking a generational lead. Our L4-capable vehicles are scheduled for mass production in 2026, with pilot robot taxi services launching in selected regions. Although aftermarket retrofits enable L4 in some vehicles, no mass-produced models currently have OEM integrated L4 hardware and software, a gap Xpeng aims to fill in China. In addition, our latest humanoid robots also have made promising advancements. With our Turing SoC, VLA, and VLM, we are quickly moving toward a mass-produced version featuring initial L4 capabilities in robotics, and we are actively preparing for mass production in the second half of 2026. During Xpeng AI Tech Day, I will unveil the new generation of robots to everyone.

关于VOM, 我们这次在行业内首次部署了一颗专门用于自能座舱, The AI chip of Turing AI is more than 12 times better than the chip of other flagships in the industry. Based on Turing AI's small-scale VLM chip, our goal is to become a smart steward on AI vehicles. VLM will not only cooperate with our VLA to bring a smarter service price, but it will also be a powerful car intelligence agent. We hope that in the future, it will not only be able to communicate in multiple languages, but also be able to have emotional exchanges and active services, which will further enhance and change the user experience of future AI vehicles. I believe that Xiaobeng's independent large-model chip plus VLM software will lead to a gradual update of the new generation of AI technology. In the future, it will be studied and approached by more industries. We have introduced the industry's first AI chip dedicated to foundation models for smart cabins.

providing over 12 times the effective computing power of the leading habit processors. Powered by the Turing AI SoC, our VLM model is designed to act as a smart assistant in AI-enabled vehicles. It not only collaborates with VLA to enhance assisted driving, but also functions as a robust in-vehicle intelligent agent. Looking ahead, it is expected to master multiple languages, communicate empathetically, and offer proactive services, greatly improving the user experience of future AI-driven vehicles. I believe our combined independent large model chips plus VLM solution will lead a paradigm shift in the next generation intelligent cockpit technology, encouraging more OEMs to adopt and reference it. Our AI technology demonstrates strong generalization abilities worldwide, not just in China. We welcome partnerships with top global players to explore opportunities for collaboration on our Turing AI SoC.

我们的海外业务在2025年上半年也继续保持高速增长并且成功打造了领先的品牌跟口碑。 Then in the first half of the year, Xiaopeng's delivery volume overseas exceeded 18,000 units, and the same growth exceeded 200%. We are in Norway, France, Singapore, Israel, etc. Ten countries are Chinese new market brands. It is also the highest selling Chinese car brand in the European market, the high-end pure electric market. Then in July this year, the first Xiaopeng X9 produced locally in Indonesia was delivered to the Indonesian car owners. This symbolizes that Xiaopeng has officially launched a global localization production strategy. By the second half of 2026, Xiaopeng will also fully complete the Kunpeng Super Electric and Ocean versions that are listed in the global overseas market. I believe this will greatly expand our model in the global market.

In the first half of 2025, our overseas business continued to show strong growth and build a top brand reputation. In the first half of the year, overseas deliveries exceeded 18,000 units, increasing over 200% year-over-year. Xpeng now ranks as the best-selling Chinese NEV startup brand in 10 markets, including Norway, France, Singapore, and Israel. and we lead in sales of mid- to high-end Chinese BEVs across Europe. In July, we delivered the first locally produced Xpeng X9 in Indonesia, marking a key milestone in our move toward global local manufacturing. By the second half of 2026, we aim to fully launch the entire Kunpeng super electric lineup in international markets, including auto trims, expanding our global total addressable market significantly.

随着G7和全新P7的交付跟放量, 我预期今年三季度的总交付量约为113,000至118,000辆, 同比上升142.8至153.6。 So the revenue is expected to be around RMB196 billion to RMB211 billion, which is the same as the rise of 94% to 107.9%. So from the fourth quarter, we will enter a big product cycle with one car and two people. At the same time, Turing AI's self-driving ability will gradually draw a clear contrast with the opponent. We are confident that not only will we achieve a scale advantage, but the company's management will also enter a new stage of profit and self-improvement. 那么在未来的三年,小鹏会在保持势能稳定向上的基础上加速扩大在中国和海外市场的规模跟市场份额。 我们希望用颠覆性的AI技术创新为全球的出行用户创造更大的价值。 As deliveries of G7 and the new P7 ramp up, we forecast Q3 deliveries to be 113,000

to 118,000 units, reflecting a year-over-year growth of 142.8% to 153.6%. Revenue is projected to reach between RMB 19.6 billion and 21 billion, representing an increase of 94% to 107.9% year-over-year. Starting in Q4, we will introduce the one-vehicle dual-energy strong product cycle, complemented by Turing AI-driven smart driving solutions, which will significantly strengthen our generational lead. We are confident in leading the market at scale, whilst advancing operational efficiency toward sustainable profitability. Over the next three years, our focus will be on expanding market share, both domestically and internationally. maintaining steady growth, and harnessing disruptive AI innovation to enhance value for users worldwide.

Thank you, everyone.

With that, I will now turn the call to our VP of Finance, Mr. James Wu, who will discuss our financial performance for the second quarter of 2025.

Thank you Xiaopeng. Now let me provide a brief overview of our financial results for the second quarter of 2025. I'll reference RMB only in my discussion today unless otherwise stated. Our total revenues were $18.27 billion for the second quarter of 2025, an increase of 125.3% year-over-year, and an increase of 15.6% quarter-over-quarter. Revenues from vehicle sales were $16.88 billion for the second quarter of 2025, an increase of 147.6% year-over-year and an increase of 17.5% quarter-over-quarter. The year-over-year and quarter-over-quarter increases were mainly attributable to higher vehicle deliveries. Revenues from services and others were $1.39 billion for the second quarter of 2025, representing an increase of 7.6% year-over-year and a decrease of 3.5% quarter over quarter. The year over year increase was mainly attributable to the increased revenues from parts and accessory sales in line with higher accumulated vehicle sales. The quarter over quarter decrease was mainly attributable to fluctuations in revenues from technical R&D services. Gross margin was 17.3% for the second quarter of 2025. compared with 14% for the same period of 2024 and 15.6% for the first quarter of 2025. Vehicle margin was 14.3% for the second quarter of 2025 compared with 6.4% for the same period of 2024 and 10.5% for the first quarter of 2025. The year-over-year and quarter-over-quarter increases were primarily attributable to the ongoing cost reduction and improvement in product mix of models. R&D expenses were $2.21 billion for the second quarter of 2025, representing an increase of 50.4% year-over-year and an increase of 11.4% quarter-over-quarter. The year-over-year and quarter-over-quarter increases were mainly due to higher expenses related to the development of new vehicle models and technologies as the company expanded its product portfolio to support future growth. SG&A expenses were $2.17 billion for the second quarter of 2025, representing an increase of 37.7% year-over-year and an increase of 11.4% quarter-over-quarter. The year-over-year and quarter-over-quarter increases were primarily attributable to the higher commissions paid to the franchise stores driven by higher sales volume. Moreover, the quarter-over-quarter increase was also due to the higher marketing and advertising expenses. As a result of the foregoing, loss from operations was 0.93 billion for the second quarter of 2025, compared with 1.61 billion year-over-year and 1.04 billion quarter-over-quarter. That loss was 0.48 billion for the second quarter of 2025, compared with 1.28 billion year-over-year and $0.66 billion quarter-over-quarter. As of June 30, 2025, our company had cash and cash equivalents, restricted cash, short-term investments, and time deposits in total of $47.57 billion. To be mindful of the length of our earnings call, I would encourage listeners to refer to our earnings press release for more details on our second quarter 2025 financial results. This concludes our prepared remarks. We'll now open the call to questions. Operator, please go ahead.

Thank you. If you wish to ask a question, please press star one on your telephone and wait for your name to be announced. If you wish to cancel your request, please press star two. If you're on a speakerphone, please pick up the handset to ask your question. For the benefit of all 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. Your first question comes from Tim Sal with Morgan Stanley. Please go ahead.

Hello, I'm Tim from Morgan Stanley. Thank you for answering my question. First of all, I'd like to congratulate Xiaopeng Motors on the high sales volume and the new technology. I have two questions. The first one is about product positioning. We've seen that the company has achieved significant results in the past two years, but at the same time, the average price of sales is still declining. So I'd like to ask the management to improve the brand positioning and create a more popular model with a price range of more than 200,000 yuan. When can we clearly improve the product's average price to speed up the company's profit? So my first question is about the brand position and the product pricing. Because Expo's average selling price has declined over the past two years, given the product mix. How are you going to reverse the trend and effectively upgrade the brand positioning and the sales mix of high-end models? And when are we going to see more meaningful ASP upgrades? and it can accelerate, that can help to accelerate expense-to-profit improvement. That's my first question.

对,谢谢啊。实际上我们小公司一直都在销售的价格区间在10到50万之间。实际上您的这个问题对我们内部来说分解成几个方向的问题。第一个是产品怎么布局。第二个是 how to use technology to have a better price. Another one is how to use emotions to have a better price. Another one is how to use brands to have a better price. These are the few lines that we are working on. For example, the P7 that we are about to release is worth more than 300,000 yuan. The X9 that we are working on today is worth more than 400,000 yuan. These are all higher than the average price of Xiaotong. In 2016 and 2017, . . . . . The original technology market, especially in the market of self-improvement value, we also added a new market target, which is the quality. So in another way, after the quality is done, the whole price, conversion, and marketing have changed a lot. Then the last one is the brand. Xiao Peng, in 2057 and 2027, we will see that the globalization of the brand will be faster than this year. So I believe that the changes in these four directions, in the future, you will see that our average price will be higher and higher.

Thank you very much for your question, Timothy, and the answer is as follows. As you can see, for our sales, it has always been in the range of 100,000 to 500,000 RMB. Internally, we are also looking at this, and we are approaching this from a few different directions. Number one is with respect to the different layout of our products. Number two is to leverage technology to increase premium. Number three, using emotions to increase premium. Number four, using the brand to increase premium. So if you look at the P7 that is soon to be launched, and this is in the price range of above 300,000, and soon to be unveiled, X9 will be in the range of 400,000 RMB. In 2026 to 2027, these will be our two major product sales cycles, and there will be multiple of different cars to be launched in the market, and their sales price will be above the $300,000 range at the moment. And number two, using technology. And technology has always been our strength, including ultra-trim versions as well as robotaxis, Starting from next year, you will see our investment being increased, and we believe that this will further drive up our premium. And number three, using technology. So we know that ADAS has always been our strength, apart from the strength of technology. And we have another strength, which is aesthetics. Becoming or making aesthetically more beautiful cars will also be able to drive the sales volumes. And finally, which is leveraging our brand. And I believe that in 2026 and 2027, these will be the two years of establishing more of a global brand of Xpeng and the speeding up or acceleration of our brand will pick up. So as a result, I do believe that you will see the ASP of our products will increase. So will the gross profit as well as the net profit. Thank you.

Thank you, Mr. Xiaofeng. The second question is about the long-term version of the technology we mentioned earlier, about the frame. We saw that many car companies in the opening year of this year have announced that their frame will be launched. Some car companies have also announced that they will be equipped with NVIDIA's Sol芯片 and VLA models. I'd like to ask the management team, what do we think of Xiaofeng's autonomous driving technology and some of the leading advantages compared to the industry? As President Xiaofeng mentioned, I'd like to ask, when will we be able to open the gap between the software experience and the twin-chip Ultra version? My second question is about smart driving, because we noticed that since the beginning of this year, several car makers have launched end-to-end smart driving solutions in new vehicles powered by Thor chips and VOA models. So how should we think about expense-to-technology advantage in smart driving against such a backdrop of competition? Specifically, could expense the touring chip An ultra-train is actively differentiated with a much better software experience versus our competitors. That's my second question.

Thank you. 实际上在当前中国各个车企或者TL1的智能辅助价值的第一梯队都是半斤八两。 有的通过算力,有的通过算力不足的就通过算力加上结构内打的组合。 So, in fact, as you all know, there is a scanning law in the big model. If you put these three models together and use better models, it will have a stronger effect. So, this is a theory in the past, but we, Xiaopeng, have already seen this effect in the test environment. So, in other words, the effective algorithm of OCHIAL has reached 2250 TOPS. And in the same industry, their flagship models are between 100 and 700 tops. So our new VIA model, I believe our data volume is the same as theirs. But the scale of the model may be ten times their operation rate may be twice their operation rate. So it will make the whole big brain smarter, the small brain more sensitive. So the question you just said about the Altair version, When will the experience gap be opened? I believe that in August of this year, we will start promoting the initial version of VLA in gray, and quickly deliver it in the following months. The goal of our initial version is to bring the experience of Qi and Max to the industry, which is the first version. But in the other versions of this year, especially at the end of the year, we hope to improve significantly. So when will there be a huge gap? I believe At that time, in the near future, we are looking forward to the next year's mass production. When our RoboTaxi starts to run, you will see that the car model of our Ultra is the same as the model on the L4 RoboTaxi. In addition to the presence of hardware, there is no cloud control. At this time, you will see the combination of Xiaopeng's Ultra and Turing's chip. Thank you for your second question.

And actually, if you look at the current players in the market, when it comes to software, and basically all the Tier 1 players, I would say that everyone is roughly the same, and for those either with computing power or the ones without sufficient computing power, they would have computing power plus LiDAR to make up or bridge the gap. And in a way, we know that with computing power plus model and plus data, these three factors will be able to provide with better results. And we know that this was an idea or a principle that people have in mind However, at XPOM, we have been able to validate this idea. And as you can see, our total computing power has already achieved 2,250 tops. In the meantime, if you look at our peers, they are still at 100 to 700 tops. And once we talk about this, when it comes to data as well as model scaling and in terms of our frame rate, which is running at two times, of our competitors and this again shows our smarter brain as well as a better control for motion. And in the meantime, for your next question about for ultra trim and when we would be able to ensure that there is a gap between us as in we leading the race ahead of our peers and for the ultra trim versions and we will start with our initial VLA launch and hopefully we will bring this to the same level as the MAX for now, and by end of this year, we would be able to see a significant improvement or difference between us and our peers. And if we're talking about huge differences between us and our peers, I would say that that time will come in next year, not long after this year, and roughly around when you will see the pilot runnings of our robotaxis, And with L4 as well as UltraTrim, they come from the same source of the same model. And basically, the only difference is that there is no cloud-end control. There is no hardware redundancy. So with the UltraTrim plus our Turing AI SoCs, and I do believe next year you will be able to see that our product will be leading ahead of our peers by 10 times or even more.

Thank you very much for sharing all the great details and exciting progress.

Thank you. My first question is related to your cooperation with Volkswagen. because recently you just announced to broaden your cooperation with Volkswagen's EEA and expand into more vehicle models, including ICEV. So could you give us more comment and guidance regarding the future revenue from the cooperation with Volkswagen?

Thank you, Ming. This is Charles here. So last week, we announced our fourth cooperation with Volkswagen. Basically, we're expanding our electrical and electronic architecture collaboration from the battery electrical EV in China into the ICE and PHEV of Volkswagen brand in China. And we believe that's a major expansion of our scope of the collaboration with our partner. And I think that through this collaboration, we're really creating a very significant strategic value for our partner, and I believe that Volkswagen is probably the only global auto company that possesses the technology that has one EE architecture platform across all the powertrain, through which you can see the value creation that we created by both parties. Obviously, I think from Xpeng's side, we're also benefiting from such collaboration. So, in terms of the second part of the question regarding the value of the potential creation from our partnership with Volkswagen, as you may be aware of, right now, since Q1 2024, we have started to recognize the IP licensing revenue from the G9 platform collaboration and later on the ERF-H collaboration. And we believe that in the past couple of quarters, the revenue from such collaboration has been relatively stable within a certain range. And for the second half this year, we see some growth potential in such revenue from the past two collaborations. So in terms of the newly entered expanded EEA collaboration, the revenue will start to be recognized once those Volkswagens, ICE, and the PHEV vehicles start SOP. So that's basically the third recurring revenue stream on top of the previous two revenue streams. I hope this clarifies your question, Ming.

Yes, Charles, thank you. So now I'm going to ask my second question. My second question, because there was a little crowd just now, also mentioned Robotech. I would like to know more about the company's future plans for the Robotech business. The first is that we have considered the 2C or 2B version of Mona. So, can we understand that in the future, if you work with this partner, you may have to design some differences with the current 2C products. In addition, we have already found a partner, so it can be mentioned that from 2026, there may be such a service. Our current car, which is a car from this point to the second half of the year, because it is equipped with this touring chip, so the chance is very strong. So the car that the second half of the year can be sold may be upgraded by OTA next year. So my second question is related to robot taxi business. I think the chairman already mentioned these are businesses. So in the future, will you launch the 2B version of vehicles to differentiate from your current product for the consumers. And also, for your current cars, you'll be able to provide level 4 functions next year when you provide an OTA service. Thank you.

I think there will be two types of cars in the future. One is a car that has L4 capability but can be driven. The other is a car that has L4 capability but can't be driven. Of course, I think this needs to be done in several years. At the same time, L4 needs to be allowed by the policy law. So in 2026, Xiao Pengle will have a car that starts at four points in some areas of China. So this is what we will do after we get the approval of the relevant policies. So, in another perspective, in the middle of the question you just mentioned, there are some points that are not particularly easy to answer, but I can explain some of them. I think the first one is that we have a big difference with other RoboTaxi companies. First of all, we are pre-installed vehicles. Second, we don't have regional restrictions. Because of our mapless model, We don't need to scan the data for each location to make a high-quality match. In addition, Xiaopeng's OSHA version, as I said, it doesn't have hardware, and it doesn't have cloud-based management services. But in terms of the overall software capability, we hope that next year, it will have the same model, but it won't be completely the same. For example, RoboTaxi may be We will be able to do it. But the car we sell will not be able to do it. So in terms of training, there are still certain differences. It's mainly about strengthening learning. So from another perspective, what you said about operation and commercialization, we will operate it ourselves when we test it. But once we pass it, we will look for Thank you for your question.

In the future, in my view, there will be two types of vehicles. The first type will be L4-enabled or L4-capable vehicles with people driving those. The second type will be L4-capable vehicles with no people driving it. And of course, this will take years to achieve. And we also know that with L4, this will need to obtain the permission and approval by different policies and regulations. In 2026, for XPOM, we will try out the L4 in different pilots, and of course, the precondition is that we are running these pilot schemes once after we have obtained the relevant approvals. In terms of the questions that you have asked, there are areas that I'm not really in the position to answer today, but what I can try to answer is that, number one, for expensive vehicles, and we are the front market where the mass-produced vehicles with the no map, non-HD map model. And that is to say we do not need LiDAR to scan the whole map, the whole city. And in the meantime, as I have mentioned earlier, we do not have the software redundancy, nor do we need the cloud to take over to manage this. And what we strive to achieve between our future products and Robotaxi is that there would be of models of the same sources. And however, there are differences. For instance, for our robo-taxis, and there will be functions such as when the passengers on the street wave their hands, the robo-taxis will stop. But for our own vehicles to consumers, obviously, there is no such functions. When there is a passenger on the street waving their hands, our own vehicles will not be stopping. So this is all about strengthened learning of the whole process. And we ourselves, obviously, will carry out our own trials and our own operations. And when all of these have passed, and we will be seeking out and speaking to partners and looking at how to expand this further.

Thank you, Charles. That's all my questions.

Your next question comes from Pingyu Wu with CityC. Please go ahead.

Hello, everyone. Thank you for the opportunity to ask this question. I'm Wu Pingyue, the Director of the Center. I have two questions here. The first is about our P7 project. Because it is also very congratulatory. In fact, after PT was launched, I also saw a lot of comments on social media. The attention and evaluation are very high. And its innovative design is also recognized by consumers with a sense of change. Earlier, the management team mentioned that our new P7's suspension is very strong. Can you ask the management team to share more about the suspension of this P7? And maybe we can look at some of the reasons behind this strong suspension. In addition, in terms of sales, is there a chance for the new P7 to see, for example, the P7 of the first generation, including the P7s released last year, become a more mobile model? Let me just make the question. First, congratulations on the successful launch of the pre-sales of the new P7. And it's a very highly original design. It's also appreciated by customers who are also leading the era. Could you please share more about the reason behind the P7 strong performance and how do we forecast itself? Is it going to be a high-volume model just like the first generation of P7 or our latest P7 Plus? Thank you.

Yes, thank you. Because P7 will be released next week, so... Sorry, I won't talk too much about the detailed data of P7. First of all, a few points. The first one is that the P7's attention is indeed far beyond what we imagined at the end of last year and the end of this year. So the small data broke down all the previous models, including MOLA, and the history records of the same period of P7. This is the first one. Then the second one, we saw that the P7's The number of blue cars in this small town is very high. And the number of young cars is the youngest among all our cars. It's a very interesting thing. And its industry is also very interesting. We still quite believe that PG has broken the circle this time. Because there are a lot of users who don't pay attention to the car owners of Xiaopeng. He is now paying attention to Xiaopeng. I think PG is not only good in appearance. I think Now, many of my friends are on the track or driving on the highway, and they see that the chassis is very solid. The chassis of our P7 is the same as the chassis of the new car that Xiaopeng and Dazhong collaborated on next year. So we think that the transfer rate of the users after the test drive will be better than what we expected. Because now all the internal users and media reviews are very good. So our goal is that P7 can achieve Thank you very much for your question

P7, yes, indeed. Next week, P7 will be officially launched into the market. Certain detailed data, I'm not going to be elaborating on those today. However, there are a few points that I can comment on. Number one is that the attention and interest that we have received in P7 has indeed far exceeded our expectation from end of last year to earlier this year. And in terms of the Pre-sales orders, as you have stated, it has exceeded all our historical data sales for different other series of cars, such as Mona, et cetera. And if you look at the purchasers and the users, consumers who have purchased P7, and it is also quite interesting, there is a very high percentage of male consumers. In addition, in terms of the age bracket of those who have placed an order before For P7, so far it is the youngest age bracket that we have, and their professions are also quite interesting. So this is what we have witnessed so far from the pre-sales orders. And for P7, it is not only that it is an aesthetically beautiful car, it is also a very strong, sturdy car, whether you are running on racetracks or running on highways. And what we have received the feedback from from the users as well as from the media reviews is that the chassis that we have worked together with Volkswagen is rather strong. And so we are getting very good positive feedback from the users, from the media. And our aim is to become the top three in the pure battery electric vehicles in the price range of $200,000 to $300,000. And of course, we don't know whether we can achieve a better ranking for that, but we will strive to do so.

Okay, thank you. And the second question is about Robotaxi. Because President He just shared some thoughts about RoboTaxi. In fact, I still want to ask, we are now actually going to re-reveal RoboTaxi's strategy for next year. Are we seeing some opportunities in this technology or even in some legal circumstances? Next year, for example, can RoboTaxi achieve a My second question is about the robot taxi. And as we renewed our announcement of our robot taxi strategy, and what opportunities such as technology or regulatory access do we see currently?

Thank you. Okay, thank you.

I think Xiaopeng is a whole car factory. It is very different from the software or service provided by pure RoboTaxi. We think that L4 cars are mainly sold to customers. And for a long time, the customer has to drive it himself. He has to be able to catch the wheel. There are legal problems, there are customers' needs. There are a series of reasons such as the need for technology to move forward. However, if a user can buy a car with such capabilities in the future, it can grow and the upper limit will be greatly improved. Indeed, we have seen the opportunity brought by the transformation of technology. I think the first reason is that we are very capable. Secondly, we believe that in the area, the possibility of RoboTaxi opening is increasing. But in the global scope, it is very difficult. But in the global scope, I think it is also increasing. In addition, our new technology, you will see, for example, now we are in the case of oil navigation, we start intelligent auxiliary driving. So can the new car's ability be done? Can intelligent auxiliary driving be done without navigation? Thank you for your question.

And as you can see, for XPeng, we are an OEM company, which makes us essentially different from from robotaxi software providers or robotaxi service providers. Therefore, of course, in order to sell this to the customers and for the customers to be able to drive such cars, there are issues that we need to resolve or we need to go about, for instance, for regulation issues and technology iterations, etc., But as you have stated, once we cross over those humps and once people are able to purchase L4 vehicles, in terms of the upper limit of the growth, that has greatly improved. And that is number one point. And number two, for robotaxis to operate within a certain region, I do think that there will be such possibilities here. However, to say that robot taxis will be rolled out on a global scale, I think it is difficult at this stage at least. But I would also say that globally speaking for robot taxis to be rolling out, it is also getting further expanded at the moment. And we know that in the past, people would need to rely on navigation to travel. use their cars and to drive the ADAS system. And in the future, would people be able to rely on ADAS without navigations? And would people be able to update their OTAs by simply relying on the same source codes? And those are questions and challenges that EXPON, we hope that we would be able to strive to resolve in the future.

Thank you so much.

Your next question comes from Tina Hu with Goldman Sachs. Please go ahead.

Thank you for accepting my question. My first question is about the interest rate of the whole car. Because we see that this second quarter is 3.8 points higher than the first quarter, which is actually much higher than the history of each quarter. Can you explain to us which aspects of these 3.8 points come from? For example, the change in the size of the car, the increase in the size, and the decrease in the cost of raw materials. Thanks, management, for taking my question. So the first question is regarding vehicle growth margin. We see that on a quarter-over-quarter basis, it's up by 3.8%, which is much bigger versus previous quarters. So could management help us break down the different factors contributing to this sequential improvement? Thanks.

Hi, Tina. This is James. Thanks for the question. You're right. I think you mentioned three aspects in terms of the product mix change, cost reduction, as well as scale. I think all three factors come into play. What I would say is the majority of the impact is still on the product mix side. So if you remember, the Mona M3 accounts for a pretty significant delivery percentage in the first quarter. And that percentage has reduced through the second quarter as we transitioned Mona into the new version, whereas our new G6 and G9 has increased its product mix in the second quarter of 2025. What we've seen is the new G6 and G9 has a very healthy gross margin thanks to the improvement technology as well as the specs that we offer to our customers that they like. In the meantime, we have achieved a higher level of platform carbonization through our engineering efforts. All of that have helped us to achieve healthier growth margin for our new model year changes. In the meantime, as you mentioned earlier, scale also come into play. We have delivered slightly higher volume in the second quarter as well. And last but not least, we have continued our effort through our supply chain optimization, which helped us to further optimize our material costs as well in the second quarter. Thank you.

感谢。 我下一个问题是关于咱们的这个费用端。 那首先我们这个研发费用是不是还是维持在全年85亿的一个指引? And then the second quarter, we also have a lot of marketing activities. But in fact, compared to the first quarter, it has increased by about 200 million. If we look at the rhythm of the new car release, the strength of the third quarter and the fourth quarter should be similar to the second quarter. So we can think that it is good to give channels. So my second question is regarding the operating expenses. So for R&D expense, is management still keeping the 8.5 billion R&D of annual guidance? And also in terms of the sales and marketing expense, Because we're seeing just around $200 million increase from 2Q to 1Q with the new model launch schedule. And if we look at 3Q and 4Q, it seems that the new model launch schedule is very similar to the intensity is very similar to second quarter. So should we expect excluding the channel fees, the sales and marketing expense should be quite similar to the second quarter? Thanks.

Got it. This is James again. So with regard to the expenses, I'll answer in two parts. From an R&D perspective, you're right. We have been expanding our investment in the R&D segment, particularly to try to basically bring the AI capabilities from the digital world to the physical world. As well as Xiaopeng mentioned earlier, we will start to launch one car with two dual energy platform in the fourth quarter of this year. So we have basically increased our engineering workforce and capabilities throughout this year as well as into the second half. In the meantime, our AI technology investment will help us to enhance our leadership in this area, which includes autonomous driving, robots, robotic investment as well as our enhancement of the engineering workforce. Also, the cloud-based computational power is something that we have continued to invest as well. So this is on the R&D side. With regard to the sales expenses, you're right, the second quarter has a slight increase versus the first quarter, partially also because of our higher deliveries in the second quarter. as we pay commissions to our franchise stores is part of that as well. We have product launches in Q2, as you mentioned, but remember we have more significant product launches in the third quarter. For example, the G7, which is a brand new SUV, as well as the new P7. These two launches, I would say, would be more significant. and probably would require higher marketing and advertising expenses in the third quarter. And going into end of third quarter and into fourth quarter, we'll be launching the Quimphome super electric platform as well in our X9 product, which we will also provide and deploy sufficient resource to make the successful launch as well. In that regard, we will continue to invest and make sure the marketing and advertising expenses are appropriate to support the product launches in the second half. Thank you.

Thank you very much, James.

Your next question comes from Xiaoli Li with Jefferies.

Please go ahead. Thank you. So my first question is about your product strategy. The next P7 has made a significant leap in the ethics design. Does this signal a shift in Xpeng's product strategy towards prioritizing design? Should we understand this not just as a short-term adjustment, but as a core component of the company's long-term strategic transformation?

Yes, this is Xiaofeng. The overall answer to your question is yes. Last year, when we were rethinking in the company, in the past of Xiaobo, technology is the most important, especially in the field of autonomous driving. In the field of autonomous driving, we invest a lot of money every year, but in fact, in many other fields, I think it is not enough. So in the past year, in the beginning of 2024, we decided that we wanted to The design is the same as that of the automatic driver. In the past 24 months, including for the next period of time, Xiaofeng's design will pursue no short version. It will increase the height of the mid-range version and have a number of long versions. Whether it's the automatic driver or the design, they are all the core of the long version that we hope to create. Therefore, we have given a very high weight to the design. and the entire product development process. We used to come first to estimate the project cost, positioning, and then make the shape. Now we come first to choose the shape, then to estimate the project, and then to position it. We have two buildings in Shanghai that are all designed by us. We have one building in Guangzhou. All three buildings are designed by us. We are building a larger design team around the world. Thank you.

Thank you very much for your question. And the overall answer to your question is yes. So last year, we were looking at this, and we were thinking that technology has always been our strength, and we have always focused on autonomous driving and invested a huge amount of money and resources in the area. However, we have lacked in the other areas, and therefore, in the beginning of 2024, we started to think that we should shift more focus onto design as well. And for the past 24 months, as well as the next whatever month to come, and in the near short-term future, I would say that the company's target is to ensure that our product does not have any shortcomings. We try and improve the areas that we are doing okay, and further adding strength to our existing areas that we already outperform others, which means that we are now currently giving more weightings and giving more attention to the styles and aesthetics of our products. In the past, when we look at styles or when we look at aesthetics and design of a car, It comes at last. We first start with the engineering. We first then look at the cost and the positioning of the car. And finally, it comes to the design and style and aesthetics of a car. But now, it is the reverse. Sometimes we look at the style and aesthetics of a car first. We then decide on the engineering, the cost, and the positioning of this vehicle. So, In Shanghai, currently, we have two buildings that are dedicated to work on the design and styles of our vehicle. In Guangzhou, we have one building, so that's three buildings in total. Globally, we are also building more different centers to help us with the designs. And overall, I would say that we are looking at providing our users with a better physical products, as well as a better overall user experience. And for the P7, and you can see that this is a process that we have been able to do so, and that is also the target that we strive to achieve.

Thank you. That's very helpful. 我第二个问题是关于这个反内卷政策。 我们看到反对转政策也是中国汽车行业近期重要的一个方向吧 那请问小鹏在经营层面做了哪些具体的调整来应对这一政策 那对于我们自身的这个业务会有哪些具体的影响 My second question is regarding the policy response As we noticed that China's anti-involution policy has become a key regulatory direction in the auto industry How is XPUN adjusting the business operations in response to this policy, and what specific impact will this have on your operations and market positioning?

Hi, this is Brian. Let me address this question. First of all, I think we noticed the recent policy announcements around anti-immolution measures. And I think this is actually a culmination of a number of discussions that aim to improve the Chinese economy, competitiveness, as well as healthiness. I think that's good for the long run. I think as an industry, we need to have healthy competition. At the same time, we need to have the ability to reinvest into new technology, reinvest into building better quality products, and this is actually the right direction. From our perspective, we actually always focused on achieving the most innovative technology and products for our customers through relentless innovation, relentless development. And I think this is actually, I would say, coincides with this new policy direction. We feel like by focusing on our target customers, by delivering innovation, the best quality, the best technology and products that actually has broader appeal is ultimately will be the right solution and will earn the right recognition both from customers as well as from our partners as well as from our regulators. So I think in that regard, we will continue to focus on full-stack innovation, continue to focus on building innovative, also broadly appealed appealing products to our customers, and at the same time also be able to maintain a disciplined and orderly operations. For example, you probably noticed that Xiaopeng Expand is actually one of the earliest OEMs to respond to the government's request of improving payment terms. We really have you know, making sure that our partners, our suppliers are all, you know, benefiting from this new trend. And we believe that ultimately will lead to a more healthier ecosystem and industry for our company.

Thank you. That's very helpful.

The next question comes from Ben Wang with Deutsche Bank. Please go ahead.

Thank you. My question is about the G7 SUV. The company seems to be facing some difficulty in the intelligent cockpit with the chips. When this issue can be resolved and what's the normalized monthly volume assumption the company can have?

Thank you.

Hey, Bing, just to clarify, there's no issue with G7. So we announced that we will deploy the specific touring SOC for infotainment in October, and then we will roll out the VLA and the VOM models for the G7 customers.

Okay, now the other question is about the 3G degree of horsepower. So the previous company said that it will only achieve double-digit car horsepower in 4G. So now the horsepower is over 2G. Can you say that 3G will achieve high double-digit car horsepower in advance? My second question is about vehicle gross margin guidance for the third quarter. Previously, the guidance is that only in the number four quarter, vehicle gross margin can go to high teams. Because the second quarter of vehicle gross margin has been higher than the market expectation, do we expect in the third quarter we already can achieve the high teams vehicle gross margin? Thank you.

Hi, Ben. This is James. So with regard to the vehicle margin, yes, we have improved significantly in Q2 and the reasons I have explained before. In the long term, we have always focused on taking into account the overall competitive landscape as well as improve our scale as our strategic priority. Overall, we'll take that as our overall guidance from a pricing as well as margin perspective. What I would emphasize is going into the second half as we launch the brand-new P7 as well as the X9 super electric version, I would focus on Q4, which we have communicated earlier, that we are aiming at profitability in the fourth quarter, and that target has not changed. And we have, as you remember, mentioned that we hope to achieve high-teens success profitability, overall profitability in the fourth quarter as well, and that target has not changed as well. Both of that will help us to solidify our target to achieve breakeven in the fourth quarter.

Basically, that's my answer to your question. Thanks. Thank you.

Thank you. That concludes our question and answer session. Now 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 contact Xpeng's investor relations through the contact information provided on our website or the Pearson Financial Communications.

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