Li Auto Inc.

Hello, ladies and gentlemen. Thank you for standing by for Lee Otto's third quarter 2025 earnings conference call. At this time, all participants are in listen-only mode. Today's conference call is being recorded. I will now turn the call over to your host, Ms. Janet Tseng, Investor Relations Director of Lee Otto. Please go ahead, Janet.

Thank you, Operator. Good evening and good morning, everyone. Welcome to Lee Otto's third quarter 2025 earnings conference call. The company's financial and operating results were published in a press release earlier today, and they were posted on the company's IR website. On today's call, we will have our Chairman and CEO, Mr. Xiang Li, and our CFO, Mr. Johnny Tia Li, to begin with prepared remarks. Our President, Mr. Donghui Ma, and CTO, Mr. Yan Xie, will join for the Q&A discussion. Before I continue, please be reminded 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 uncertainties. As such, the company's actual results may be materially different from the views expressed today. Further information regarding risks and uncertainties is included in certain company failings with the U.S. Securities and Exchange Commission of the Stock Exchange of Hong Kong Limited and the obligation to update any forward-looking statements except as required in the applicable law. Please also note that LIADO's earnings press release and this conference call include discussions of unaudited GAAP financial information as well as unaudited non-GAAP financial matters. please refer to Liado's disclosure documents on the IR section of our website, which contain a reconciliation of the unaudited non-GAAP matters to comparable GAAP matters. Our CEO will start his remarks in Chinese. There will be English translation after he finishes all his remarks. With that, I will now turn the call over to our CEO, Mr. Xiang Li. Please go ahead.

Hello, everyone. I'm Li Xiang. The third quarter of 2025 is also the first quarter of the second decade of Li Xiang. We have gone through the product cycle, public relations, public chain expansion, policy changes, and other problems. These factors have also had an impact on our delivery and operation. But today I would like to talk to you about our long-term thinking. The second decade of Li Xiang How should the three most important key choices be made? They are organization, product, and technology. The first key choice is organization. The choice we face is the management model of a startup company or the management model of a professional manager. In the past ten years, ideal cars have been the management model of a startup company for the first seven years. As the scale expands, it reaches the sales scale that I have not experienced before. In 2002, many people would suggest that we go to the management mode of professional managers. Because in history, the century-old car companies such as Benz and BMW and the technology giants such as Microsoft and Apple have achieved great success in this mode. Over the past three years, we have worked very hard to make ourselves the governance machine of professional managers. After the real experience of cooperation, we realized that entrepreneurship companies and professional managers are two completely different management systems. It has nothing to do with the process and organization structure, but the difference in management concept and elements, as well as the different stages and industry environment. The management of professional managers can be very successful, but it needs three elements. First, the industry and technical cycle are relatively stable. Second, the company's industry position is leading and stable. Third, the founders and the creative team Either there is no power to do it, or it is no longer there. If these conditions are met, the management model of the professional manager is a very good choice. Apple and Microsoft have also grown from thousands of dollars to tens of millions of dollars since the establishment of the professional manager. The conditions of the startup companies are quite the opposite. First, there is a huge change in the industry and technology cycle. Second, the industry pattern is not certain, and the enterprise is not a leader. Third, The founders and the startup team are still working hard every day, full of motivation. In today's era of artificial intelligence in all sectors, the environment we are in, our own characteristics, are more suitable for the startup industry. The core of the management of a startup company is four things. First, more conversations, deeper conversations, rather than more reports. In high-speed changes, A deep dialogue is a key element to improve awareness and bold decision-making. Second, it's not just about paying for the value of the user. It's about paying for the value of the user. It's not just about completing all kinds of tasks. Third, it's about continuing to improve efficiency. It's not about taking in more resources. Last year, we spent $10 to do something. This year, we'll spend $8 to do something. So that there is more resources to do short-term non-creative income long-term investment and ability construction The fourth is to realize the key problem, not to create a new trend Only added value, increased efficiency, and the key problem solved Can be in a highly competitive, highly variable environment to continue to meet the needs of users In the past three years, I have been working hard with the startup team to learn the management system of professional managers To force yourself to accept all kinds of changes But we have become It has become worse and worse. And Invita and Tesla are still managed in the way of a startup company today. If the strongest companies in the world can use the management method of a startup company, what reason do we have to give up our best way? I have 27 years of management experience in a startup company since 1998. And I have never worked in any big enterprise in the way of a professional manager. At the moment, we are facing a big environment of industry competition and technological change. I love cars, love products, love artificial intelligence. Work is my biggest hobby. Why not use my best ability and way to manage ideal cars? This is the first important choice for ideal cars in the second decade. Therefore, from the fourth quarter of this year, we firmly return to the management model of the startup company. To face the challenges of the new era and new technology. The choice of the organization is the basis of everything. In the past ten years, the more important question is what problem do we really want the user to solve? What is the product? Where does the technology go? This is the fundamental principle of universal convenience. First of all, about the product, we also have to make a choice. What are our products going to be in the next ten years? Is it an electric car? Is it a smart terminal? Or a giant robot? When the product stops in the electric car, The logic of competition will become a war of parameters More than 20 kilometers or less than 20 kilometers The car is more than two centimeters long or less than two centimeters If it's just an electric car competition It's a bigger space, more lines, cheaper prices By the way, copy those designs that have been successfully verified For example, ideal wine All other research and investment are a waste of research and development cost Stronger sensors Stronger models Stronger diamond power Stronger active suspension It's a waste of entire cost It can even generate strong computing power and active space, etc. The energy it generates will also bring negative effects to the virtual world. If we choose to do the smart terminal, we will pay more attention to the things in the screen. It can become a repeat of the smart phone function. In the past few years, the industry has made a lot of innovations on the smart terminal. Basically, it is to move the function of the smart phone to the car, put the APP of the phone into the car, and make the car a bigger prototype in the car. Even hope to be able to do coding in the car Do deeper research in the car User buying a car Is it for work or for a better life? Obviously have a better experience on mobile phones and computers More natural use Why do you have to put it in the car? These investments in the value of the user are very limited Even the company's selfishness We can also make the car into a giant smart product in the physics world To put it bluntly, it's a robot The cartoons and movies of Game of Thrones clearly tell us that the biggest division of robots is in two forms. One is like a person, and the other is like a car. Peter the Raccoon and the Riding Hood also clearly tell us that the car is one of the core forms of robots. How to turn a car into a robot? Give it the ability to sense the eyes and ears, give it the ability to model the brain and the brain, give it the ability to calculate the same as the heart organ, and transform the hardware into a stronger body. Let the car robot have the ability to drive, not only can it drive, but also can welcome you every day, help you drive, help you charge, open and close the car door for you, and provide you with convenience, experience and care. It can also play the role of parent assistant civil servant, provide you with the most convenient life service and space care in the car. Just like the same way the plane serves Just like when I was a kid Mom has been taking care of and caring for us all the time How to define a good giant robot From passive machines to automatic machines Then evolved into proactive robots In the next ten years The most valuable product of the giant intelligence Must be a car with automatic and proactive capabilities The competition of the product is also the automatic and proactive capabilities to what extent And these values will integrate into the life and experience of the high-end 有了就再也回不去了 电动车的选择不是不好 而是不够 智能终端的选择也不是不好 而是不够 只有选择巨声智能 这个最难的题 我们才可以真正改变用户的生活 提供自动和主动服务的 巨声智能机器人产品 恰好它就是变形钢的 汽车形态机器人 是赛车总动员 皮尤侠的汽车形态机器人 新时代给予汽车企业 and the best opportunity and the most difficult challenge for creators. Next is the choice of technology. More specifically, it is the technical choice of a complete AI system. Is the language intelligence in the digital world or the physical world's giant intelligence? This means that we need to build a completely different system. To build a giant intelligence, the most important thing is to build a set of AI systems that are different from the language intelligence. It includes the perception of the giant body, which is equivalent to the eyes and ears, the model of the giant body, which is equivalent to the brain, the operating system of the giant body, which is equivalent to the nerve, the power of the giant body, which is equivalent to the heart, the essence of the giant body, which is equivalent to the human body. At present, there is no third-party supplier that can fully supply this system, and even none of them can be supplied. The biggest feature of language intelligence is to pay attention to models and calculations. Bigger models and bigger calculations will bring stronger capabilities. And the giant intelligence must increase the perception of the physical world. Models must also be based on the understanding of the physical world. Accuracy is the first task, and the second is the countermeasure. The operating system needs to ensure the softest and most superior fusion, improve higher accuracy, and the overall system must be fast and accurate. The sense, model, operating system, and so on provided by the giant body intelligence also needs to be supplied at the end, not at the end. Finally, it is necessary to change the hardware to become the body of the giant body intelligence. For example, three-dimensional neural control and active AI reduction systems to improve the efficiency and accuracy of the execution of the body. If you use the giant body intelligence and the entire AI system, you will find that there are too many changes that can be changed and need to be changed. First of all, it is the perception based on the existing perception model and the algorithmic calculation. The effective perception distance of the current 3D-BEV-OCC-ZEN network 2D-VIP I am talking about is effective and not the most theoretical, only more than 100 meters, far beyond the human eye. If upgraded to 3D-VIP, the effective distance can be expanded by two to three times. Today, we can naturally solve more than 50% of the common problems of service driving. 3D VIT is not only limited to the area of automatic driving, but also the active interaction and service of the car and the owner, and the active interaction and service of family members in the car, which has become possible. This requires a lot of research and development breakthroughs in the sense model. There must also be M100, such a chip customized for G-Sense intelligence, and a powerful translation team to be able to achieve it. The second is the model. With 3D VIT, there is a real understanding of the physical world. The VLA model of VLA can better understand and understand the physical world. Human data can be used more efficiently for training. The data of the world model can be used better to improve training. There is the best computing platform, a 4B MOE model. The operating truth is only 10 Hz, and the implementation system is 60 Hz. If the real speed of the model engine can be twice or three times faster, some of the problems that supplement the driving performance and reflect the size of the problem can be effectively solved. This also requires the deepest modification and customization of the traditional GPU architecture and computing power, as well as the specialized operating system. The M100 chip is developed to solve the fundamental problems of high-tech. The last is the body. The fastest acceleration of human brakes is around 450 miles per second. At present, the total range of automatic driving is around 550 mm. Sitting in the driver's seat, the instinctive response is much slower, like an old man driving. The control system can increase the response speed of the entire chain to 350 mm. Don't underestimate the gap of 200 mm. This can probably increase the rate of failure by more than 50%. And it makes people feel that driving is better and safer than themselves. It is both the true security of reason and the sense of security. With such a demand, the whole control method will become completely different If you just pay attention to the improvement of the model size like language intelligence The model size is doubled The improvement of computing power is only 5% to 10% If you look at the overall size of the giant intelligence, each link solves the most critical problem in the whole system Then the next round of automatic driving performance improvement should be five to ten times The ability of GSM to provide automatic and active fast interaction and real-time service for users is more than ever. Over the past three years, we have been fully technical prepared for the GSM system, allowing our next generation of products to be full of confidence. The year-round of GSM robots has been officially launched from car robots, and the early income is just the beginning. The above three key choices have determined the foundation of our second decade of development. It is more difficult and more challenging than the last decade. We know that the future competition is not in a moment, but in the direction of long-term selection and continuous investment. With a firm financial foundation, we will remain focused on using our favorite entrepreneurial management model to build the leading giant smart products to ensure that ideal cars can travel through the cycle and use technology to become a long-term enterprise that will create value for users and society. I also hope that in the future, in this way, Now translation for Mr. Lee.