Xiaopeng Motors (NYSE: XPEV) today released its financial report for the first quarter of 2024: total revenue was 6.55 billion yuan, up 62.3% year over year and down 49.8% month on month. Net loss of 1.37 billion yuan, compared with net loss of 2.34 billion yuan for the same period last year, and net loss of 1.35 billion yuan for the previous quarter. Excluding US GAAP, the adjusted net loss was $1.41 billion, compared with a net loss of $2.21 billion for the same period last year and a net loss of $1.77 billion for the previous quarter.

After the financial report was released, Xiaopeng Motor Co-founder, Chairman and CEO He Xiaopeng, Vice Chairman and CEO Gu Hongdi, Vice Chairman of Investment Charles Zhang, and Vice President of Finance and Accounting Wu Jiaming attended the subsequent earnings conference call to interpret the key points of the financial report and answer questions from analysts.

Here are the highlights of the analyst Q&A session of this conference call:

Morgan Stanley analyst Tim Hsiao: The first question I would like to ask management is, can management quantify the scale of this portion of revenue for platform software technology generated by cooperation with Volkswagen in the first quarter? Also, is the contribution expected to gradually increase over the next few quarters? Is the electronic and electrical architecture agreement revenue announced by the company last month an opportunity to be included in this year's financial report?

Charles Zhang: Platform software cooperation revenue from Volkswagen has been included in service and other revenue for the first quarter. This revenue is recurring. I believe that every quarter in the future, platform software technology service revenue will be included in financial reports. In the first quarter, the scale of this revenue was hundreds of millions of yuan. We believe that in the following quarters, this revenue will increase quarterly. Obviously, given the nature of platform and software revenue, this has generated very high profit margins for the company. I believe that with the support of the automobile sales business and platform software technology service revenue, our gross margin at the company level can continue to be at the level of 11%-16%. As Xiaopeng pointed out, we have created a very unique business model in the automotive industry.

In addition, we will continue to reduce costs through technology and supply chain to further increase the gross margin of our automotive business. Regarding the electronic architecture revenue you mentioned, we expect this portion of revenue to be included in financial reports starting in the second half of this year.

Tim Hsiao: The second question is about the company's autonomous driving business. Earlier, we also saw Ho Xiaopeng mention that in the third quarter of this year, the smart driving capability of autonomous assisted driving will be raised to a level that can be driven on every road in the country, and then achieve the level of L4 smart driving next year. I wonder if Xiaopeng Motor needs a corresponding large-scale increase in the deployment of R&D personnel and resource investment during this process? How long does management expect it to take to transform Xiaopeng's technological leadership into growth in new car sales or technology monetization?

He Xiaopeng: First, to correct the statement, next year we expect to be able to achieve quasi-L4 capabilities, not L4, because achieving L4 autonomous driving requires hardware, and simultaneous policies and regulations. This is the first aspect. The big model will actually have an impact on the autonomous driving development of all companies today, which means that the people who originally wrote the rules will now have to make changes, and eventually the model will generate the rules, so there has been no significant increase in overall personnel. Instead, we have freed up more personnel to do things related to global business.

Now it's actually difficult to judge the launch of the big model. The impact on acceleration and the changes caused by the experience and acceleration were easier to make clear judgments when people used to write programs. For example, I had more people, then I had more time, so I could do it. However, models are based on data capabilities, based on data scale, training capabilities, and model capabilities. I originally thought that in the latter half of 2025, autonomous driving capabilities would be an extremely strong demand. After the big model is launched, this period may be brought forward to next year.

Bank of America Merrill Lynch analyst Ming Hsun Lee: The first question is about some of the later models. Does the company have more detailed instructions? What is the approximate pace of supply of Mona models to companies and delivery to consumers in the third quarter? Also, management previously mentioned that at least one new car should be launched every season next year. Can you ask management to quantify how many of these cars are Xiaopeng brands, and then how many are Mona brands?

He Xiaopeng: First, I would like to explain that for all of our models, our main customers are individual consumers, and we will not consider launching certain models on a large scale in the corporate customer sector, including the Mona. Therefore, when Mona was initially designed, it was designed with the highest appearance value. Under the premise of high appearance, we don't expect it to have huge penetration in the corporate market. After the third quarter of this year, we will launch different cars in different quarters. Some quarters will have more, while others may be a combination of only a few modified cars. We will continue to have related models from the third quarter of 2024 to 2026, and it is not convenient to share specific product planning maps with the public here. However, we have just shared this year's two models with you in the financial report.

Mona will be a series by Xiaopeng, and next month we will share with the market some of the company's thoughts on the brand. Today, let's repeat what we just said. Mona has a very good face value. We expect good profits. We expect good profits. Install different levels of intelligent driving in Mola models, including the highest level of intelligent driving functions, and set the price of models equipped with smart driving software from the current price range of 200,000 or more to less than 200,000.

Ming Hsun Lee: Regarding the company's overseas market business, the Xiaopeng brand officially entered the Hong Kong market last week. I know that recently the company has also entered many different countries. Could we ask Ho Xiaopeng and Gu Hongdi to share our overseas market layout this year? Including the tens of thousands of units expected by management, is the company's progress in entering overseas markets in line with our original expectations? Is there a chance to get ahead of the curve? In overseas markets, He Xiaopeng just mentioned testing of high-speed autonomous navigation and assisted driving. Will the company invest a relatively high amount of capital in this technology?

Gu Hongdi: First, with regard to our overseas expansion plans, we have accelerated the pace of international business development this year. Our goal is to promote Xiaopeng's industry-leading technology and the best tram models in more than 20 countries. The specific number of markets that can be achieved is also obviously related to our ability to provide these markets with models suitable for local needs. For example, currently in the European market, we have started selling the G9 and P7 models, and will launch the left-hand driven G6 this month. In other markets, such as Southeast Asia, or Hong Kong, China, as you mentioned, we will launch the right-hand drive G6, which is expected to be delivered within a few months. Additionally, we are planning to launch the right-hand drive X9 by the end of this year or early next year. Our international business expansion is progressing according to plan. The company's previous sales outlook and total delivery volume of more than 10% this quarter are still valid.

Regarding intelligent driving technology related to Xiaopeng Autonomous Driving Assistance (XNGP), we are actively testing it in various overseas markets, and we are also cooperating with regulators to ensure that it complies with local regulations and standards. Currently, our R&D work is mainly focused on improving the performance and user experience of artificial intelligence chauffeur functions. Our goal is to expand the coverage of Xiaopeng's autonomous driving assistance system to 50 cities by the end of this year, and further expand it to more cities next year. In addition, we are continuously optimizing the performance and functions of Xiaopeng's autonomous driving assistance system to improve its applicability and stability in different scenarios.

Deutsche Bank analyst Ben Wang: I have two questions here. First, can you ask management to talk about our four-quarter products? The management just mentioned that sales in the fourth quarter will have a very high growth compared to last year. In the fourth quarter of last year, Xiaopeng's sales volume should have been around 20,000 units. Can we assume that sales in the fourth quarter of this year will increase to 30,000 units? If it were possible to achieve this level of sales, how much would existing products, Mona brand models, and the F57 large sedan each contribute? How to forecast the gross margin level for the fourth quarter? Can it reach 15%?

Gu Hongdi: First, in the fourth quarter of this year, full delivery of our Mona products will begin, and delivery of our B-Class models will also begin. These two products will make a significant contribution to monthly deliveries in the fourth quarter, but their delivery situation is at different stages due to differences in delivery start times. We're very excited about the growth prospects for these two products, as they have the potential to generate significant sales growth. Assuming that all orders can be delivered for both models, we expect a significant increase in monthly deliveries compared to the same period last year, which is our expectation. However, I won't give specific predictions and breakdowns.

Also, as you've just heard, these two products are just the beginning of our superproduct cycle, which could last 18 to 24 months. At the beginning of next year, we will also launch other products and facelift models, which will be available over the next few quarters. We want to maintain a steady growth trend rather than intermittent sharp growth. Therefore, we want to achieve a steady and orderly increase in monthly sales volume rather than a sudden surge like when other models were launched. I think this growth trend is likely to continue until next year, and we will show you the details in our financial reports for the next few quarters.

On the profit side, I would like to talk about the Mona model first. Although it is only an A-class car, we still expect it to contribute a relatively healthy positive profit margin, and sales will also be quite good. For the B-Class, we think its profit margin may be higher than our current models other than the X9. Both cars will make a significant contribution to the company's overall gross profit, and at the same time, as Charles mentioned, the contribution from Volkswagen will continue to be a stable recurring revenue. Additionally, we anticipate additional revenue from other partnerships. Taking all of the above into account, and our view of the potential gross margin level of the new product, we will maintain a gross margin level of 11%-16% by the end of this year.

Ben Wang: Second question, I would like to ask that many of our rivals have launched extended-range models, which has led to a significant increase in sales. Has Xiaopeng considered launching extended-range models for some time to come?

He Xiaopeng: We are generally unwilling to answer clear plans for future products because we are still in the process of planning. What I can answer today is that in the field of flying cars, we will use range-enhancing solutions. In a flying car, the part of the car has an extended range, while the flight system side is pure electric. This is very clear. If we have plans for related extended-range products, we will share them with you again.

I'd also like to add some of our thoughts on range-extension technology here. The first is that customers actually have real needs for hybrids. There are many car owners now that in the past year or more, and over the next year, hybrid technology, including range extension, is a product that has huge demand for use. However, we also investigated many users. After accepting the hybrid, they discovered that I only needed to add oil a few times a year, and usually needed to be charged frequently, so they began to accept more pure electric products. Therefore, we will accelerate the construction of ultra-fast charging and even S5 ultra-fast charging devices this year. Recently, we have been building very fast. We want to speed up the user's charging experience. I believe there is a high probability that many hybrid car owners will be willing to switch to pure electricity in the next car. This is the first demand we have seen.

Second, from a technical point of view, the transition from extended range to pure electric range is much more difficult than from pure electric to extended range. We believe that the current generation of extended-range vehicles have actually seen many challenges, such as the many challenges encountered in globalization. For example, Chinese cars mainly drive in cities, while some overseas countries account for a significant proportion of cities and highways, and extended-range vehicles still have many problems with high-speed experience. Looking at it another way, for Xiaopeng, what is the next generation of range extension technology like? We will think more deeply and hope to share our views with you again in the future.

UBS Securities Analyst Paul Gong: I have two questions. The first one is about the big end-to-end model. Currently, there is some feedback from the industry that there will be uncertainties when using the end-to-end model, because the decisions made based on its calculated results may not be accurate, so consumers are unsure to use it, and there is also feedback that as the model gets bigger, its updates will slow down. What kind of thoughts on the management side?

He Xiaopeng: Yes, when we first started with the end-to-end model, we agreed with you from a certain point of view, and we all had very big concerns. But as technology progressed, we saw tremendous value in an end-to-end model. We all know there are uncertainties on all big models, but the most important thing is how you build your controller, like your braking system and black box safety system. In this regard, I think Xiaopeng Motor has done a lot of things related to safety control and safety games in autonomous driving in the past, and there are many technologies that can continue to be used in the end-to-end model to increase our control over these capabilities. This is number one. Second, we need to build a huge simulation testing system, and it takes a lot of effort to do this thing.

Regarding the second question, I don't really understand it, so I can try to answer it. In fact, the big model of autonomous driving is not entirely consistent with the big language model. You can understand the logic of the entire big model based on the Transformer technology architecture and write it yourself instead of using an open source big model. When you have the ability to use big models yourself, you actually don't need such big data changes, such as the 30 B or larger B models as the industry says. There is enough data in the appropriate model for pre-training, so in the future, we also hope to add post-training capabilities to maximize the value of the suitable model. In fact, compared to using people to write autonomous driving programs, the ability to write large models has greatly improved.

Paul Gong: Second question, as far as existing sales channels are concerned, they are often better locations, and the area is relatively small. As Xiaopeng Motors launches more and more models, including the Mona brand, do we need to expand the size of our stores? Even at the cost of sacrificing the location of the store.

He Xiaopeng: We have been optimizing channels. For example, we are now sinking channels to cover more low-tier cities. In the past, Xiaopeng mainly expanded sales business in Tier 1 and 2 cities. Second, as I just mentioned, we will definitely increase the proportion of general stores, and 4S stores can have better support for service and multiple models. Third, our new store in this shopping mall still won't have a large area, or will use a relatively modest area. There is no need to put all models in. The goal is to achieve very good drainage and conversion.

We began channel optimization in the fourth quarter of last year. Looking at the first and second quarters of this year, the progress and speed of recovery is quite good. At the same time, we expect the number of Xiaopeng sales stores to increase to 600 in the third quarter of this year, preparing for a wider range of sales.

Goldman Sachs Analyst Tina Hou: I have two questions. The first is our Mona brand. On the consumer side, what are the competitive models in the market? One aspect of our design's appearance is likely to be characteristic, and what other aspects can compete with other competitors? Has the company conducted a survey of consumers in this price segment? Compared to consumers with more than 200,000 models, will their demand for intelligence be slightly lower? Is the Mona brand expected to sell on the corporate customer side?

He Xiaopeng: Actually, we should have discussed this question at the Mona press conference. Mona's car buyers need good looks to meet their car buying needs. If the price is lower than other cars, we used to have a model called 10,000 US dollars, which means that for every 10,000 US dollar reduction in price, the sales volume of a certain model will probably increase by 1 to 3 times. Among the more than 200,000 models, we have all seen this trend, and it is also the main competition segment for new car builders. If we can reduce some quotas on this basis, the specific amount will be told to everyone at our press conference. I believe the scale of demand for it is very large; it will probably be at least two to four times the current scale.

So in the past situation, why can't anyone make a car that is both beautiful, has good space, and intelligent driving in the price range of 100,000 to 200,000? It's because he can't do it. Xiaopeng has been developing this area with our partners for four years. With such a large investment in R&D, I feel that preparations can finally begin until now. This is our overall thought. Currently, there are some cars like this. They sell a lot, between 20,000 and 50,000 units a month, but Xiaopeng's products are still very different in positioning from these models. Xiaopeng's products meet consumers' needs for beauty and technology, which did not exist in the past.

We've also been discussing the need for intelligence for a long time and thinking about it for a long time. I think there are two aspects. First, we will provide different intelligent hierarchical versions to satisfy some users. They may not need particularly strong intelligent technology. However, here, I can cite an example from when I was still starting a business in the mobile internet industry. High-end mobile phone voice users are often not high-end users of data traffic; conversely, users who make fewer phone calls are high-end users of data traffic. So according to our data, I think that among the 100,000 to 200,000 yuan model-level consumers, more people may need high-grade intelligent value, but they think it's too expensive. Once he can afford such a high-tech product, its appeal is huge. In this regard, next year everyone should see changes in data.

Gu Hongdi: Let me answer questions about enterprise users. In this price range, products that are successful on the consumer side are also suitable for enterprise users. Therefore, I believe that the success of this product, even if it is not aimed at corporate users in the market segment, will still be widely used in related travel scenarios. Furthermore, our collaboration with DiDi will also drive the mass appearance of Mona in multi-use scenarios.

Tina Hou: For business users, do we expect to achieve 100,000 sales in the first 13 months, then 100,000 sales in the next 12 months? Is that a reasonable expectation?

Gu Hongdi: Our cooperation agreement with DiDi is still in effect. If they can reach the agreed number of purchases, we will still provide relevant car purchase discounts.

Tina Hou: So sales volume still mainly depends on DiDi's side, right?

Gu Hongdi: No, if you read the public agreement we published, you'll find that if Mona products are used in the DiDi system or its mobility ecosystem, and they reach a certain amount of use, they will receive additional rewards. This has nothing to do with their own buying behavior.

Tina Hou: The second question is about our software services revenue. Since this quarter, technology revenue has generated high profit margins, which has contributed well to the increase in the company's overall profit level. Since our agreement with Volkswagen is not exclusive, I would like to ask management whether Xiaopeng is actively exploring cooperation with other partners under a similar model?

Charles Zhang: Our strategic cooperation with the public has actually created great strategic value for both of us. For example, we cooperate on technology and also on the supply chain. In the process, we also discovered opportunities for closer cooperation between the two sides in other fields. I believe that as strategic partners, the two sides can do a lot.

Also, from the perspective of cooperation, our cooperation is not exclusive; we are also open to seeking strategic cooperation opportunities with other participants. However, I think what we value most is how we can bring value to the partnership and create value for each other, not just collaboration on supplier related aspects. (end)