He Xiaopeng: The competition in the automotive market will be more brutal and bloody in 2026

Author | Wang Xiaojun

Editor | Chai Xuchen

Once considered to be in the "ICU," XPeng has made a stunning turnaround over the past year, with outsiders believing it has directly transitioned from "ICU" to "KTV."

Although in the eyes of XPeng's Chairman and CEO He Xiaopeng, it is still far from "KTV." However, the consensus within the industry and beyond is that "things have improved."

On November 21, XPeng officially welcomed its one millionth vehicle—the XPeng X9 Super Extended Range version rolled off the production line. It took XPeng 82 months to reach the first 500,000 vehicles; however, it only took 14 months to go from 500,000 to one million. XPeng has thus entered the million club.

The day before, the XPeng X9 Super Extended Range version was officially launched, priced starting at 309,800 yuan, significantly lower than the previous pre-sale price of 350,000 to 370,000 yuan. This also means that XPeng has offered a highly competitive price to enter the "one vehicle, dual energy" route.

At the same time, this vehicle also has certain competitiveness in terms of range and intelligence in the current market.

It is reported that this vehicle has a CLTC comprehensive range of 1,602 kilometers and a pure electric range of 452 kilometers.

In terms of intelligence, the Max version is equipped with a Turing chip with an effective computing power of 750 TOPS; the Ultra version is equipped with three Turing chips, with an effective computing power of 2,250 TOPS. The new car will be equipped with the second-generation VLA of the physical world large model, and according to the plan, in the first quarter of 2026, XPeng will first fully push the second-generation VLA for the Ultra model, and the MAX model will also be fully adapted in 2026.

This price was quickly recognized by the market. Although there were no specific orders, He Xiaopeng revealed on November 21 that the XPeng X9 Super Extended Range broke the historical daily order record within one hour of its launch. Orders from the northern region accounted for more than half, with the Ultra version accounting for 79%.

Prior to this, He Xiaopeng stated that he had pushed for the extended range internally five times, as XPeng's poor sales in certain areas were largely related to the inconvenience of local charging. Moreover, if XPeng wants to enter more global markets, the charging method cannot become a bottleneck.

The X9 is just the beginning. At the earnings conference, He Xiaopeng revealed that XPeng will have seven extended range models next year. Among them, three super extended range models will be launched first in the first quarter of 2026. This also means that next year will be a big year for XPeng's products.

In addition to the automotive business, XPeng's current narrative also includes embodied intelligence. In the past ten days, the newly unveiled humanoid robot IRON has been seen as a key to XPeng's reevaluation, and the capital market has also voted for XPeng with real money. In addition, XPeng's Robotaxi and flying car businesses have also made further progress.

With these supports, XPeng's market value once surpassed that of Li Auto, Geely, and other car companies.

However, He Xiaopeng stated, "I really don't like looking at stock prices; I think maintaining a calm mind is very important. Building cars is a marathon that never stops. At the same time, I believe that to truly become a global technology company, one must look further ahead. All the joys or pains today are just fleeting clouds, merely a process in between." Just like the fluctuations in market value, the above are all positive in the long run, but the market is always full of uncertainties.

He Xiaopeng himself believes that every car company is currently cautious, just as it was hard to imagine the changes in the Chinese auto market a year ago, it is also difficult to predict the uncertainties of the coming year. "The only thing I can be certain of is that the competition in the automotive market will be more brutal and bloody in 2026."

He Xiaopeng revealed that one "painful" issue for XPeng currently is that "the AI training costs for next year have already reached a very high number," and what to do afterward is something that needs to be considered.

However, the current sense of crisis is completely different from the previous difficulties. He Xiaopeng stated that XPeng needs to continue strengthening its capabilities in technology, business, globalization, and organizational systems. This is also the experience gained after XPeng faced challenges in 2022 and made "huge changes and adjustments" over the past three years.

From pure electric to extended range, from cars to robots, XPeng is redefining its boundaries at the milestone of one million vehicles. He Xiaopeng's vision is clear: to build XPeng into a "global embodied intelligence company."

But behind the currently imaginative strategic layout, how to balance investment and output, dreams and reality, remains a test that XPeng must face.

Below is a conversation with He Xiaopeng, Chairman and CEO of XPeng Motors, and Li Yifan, head of the XPeng Motors X9 product (edited without affecting the original meaning):

Q: There have been reports that XPeng's market value has surpassed Geely, and it has also surpassed Li Auto before. How do you feel about that?

He Xiaopeng: First, I really don't like looking at stock prices; I think maintaining a calm mindset is very important. Second, the automotive industry is a marathon, called "an unending marathon." I am more concerned about the medium to long term, how to avoid competing solely on price, quality, and performance, because these are the basics, but not everything. In Chinese internet companies, competition is not just about these factors; there are other elements and factors involved.

I used to work in the internet sector, and I have always believed that physics and AI will couple. In today's automotive field, physics accounts for a large proportion, so you don't see the value of AI as significant. However, in the internet world, software value is basically everything. In the world of robotics, physics and AI may each account for 50% of the value. Therefore, I am constantly thinking about how to truly combine physics and AI in the future; at that time, our efficiency may be better, our barriers higher, and our profits better.

From a certain perspective, I believe that to truly become a global technology company, one must look further. All the joys or pains today are just fleeting moments, merely part of the process.

Q: Regarding the revaluation of Chinese tech companies and the world's top tech companies—those trillion-dollar tech companies, when do you think this will happen?

He Xiaopeng: I personally feel that many times we shouldn't compare. From another angle, regarding who has pricing power in the capital market, this is often a matter of pricing power in the capital market. As a global technology company, we should firmly focus on doing our own things well If you do well, you might find that one day the valuation is very close, and you will be very happy; if it is still a bit far off, like a difference in exchange rates or something else, you can just keep a calm mind. There is really no need to be too calculative. Ultimately, if we do things well, others will respect us and like us.

Q: You just mentioned that you will enter more than 60 countries and regions next year. Which country will be the main market? And which model do you plan to focus on in the future? There will be many significant MPVs released next year. Why does XPeng choose to launch such an important first range-extended MPV at the end of the year instead of picking a better time for the release? The New Year is approaching, and many people want to pick up their cars before the holiday. How is your delivery going?

He Xiaopeng: Our main models overseas are the G6 and G9. By the end of next year, everyone will see multiple models overseas. The G6 and G9 are selling well mainly in Europe and some countries in the Middle East. Last month, we sold 5,000 cars overseas, and we calculate based on insurance registrations.

I have noticed a small pattern: just like in China, in large cities, at least 1,000 cars need to be sold, and in small counties, there should be at least 100 cars sold before brand recognition starts to develop. Without brand recognition, sales are difficult. In several countries like Northern Europe, Germany, and France, we went from a new car market share of 0.1% to 0.4% to quickly reaching 1% to 2% after a year, and in some countries, we have already reached 4% to 7%. This is very encouraging.

Q: Previously, you mentioned that the company might achieve quarterly profitability in the fourth quarter. What is your view on this issue? Additionally, from being in an "ICU" state over the past year to gradually recovering now, what do you think is the most important aspect of this process? As you mentioned, we need to take a long-term view of the company's development. In the second half of the competition in the new energy vehicle sector, what do you think is the most important point for XPeng to maintain its competitiveness?

He Xiaopeng: We just released our financial report. If I didn't dare to answer before the report, the probability of us being profitable is "five nines," but I can't say it's 100% yet.

In XPeng's system, we expect to excel in four capabilities: the first is technology, the second is business; the third is to build global capabilities well. The next five years will be a very good period for Chinese companies to go global. If we miss this opportunity, we might lose it. Of course, some Chinese companies can still create miracles, but we can't say we are fully there yet; the fourth is to build a strong system capability. After XPeng faced challenges in 2022, we made significant changes and adjustments in 2023, 2024, and 2025.

I believe that through our efforts, we have made some adjustments and improvements, but I am still very excited. I think the years 2026, 2027, and 2028 will be a new leap for XPeng.

In the third quarter, we spent about 2.4 billion on R&D, and it will be even higher in the fourth quarter. In fact, XPeng is a company with particularly high R&D efficiency. I believe that with such high quarterly R&D investment, in the next three years, everyone will see us entering a phase of AI-driven integration and innovation, producing more completely different hardware and software capabilities. This is what I am very much looking forward to. Based on that, we will strive to do better business globally and form better barriers. This is what we will work on in the second half of this year and in the coming years Question: At the previous XPeng Technology Day, the highly anthropomorphized design showcased by Iron was indeed impressive. I would like to understand what design considerations led XPeng to incorporate gender characteristics into Iron? Additionally, as robots approach high levels of bionics and even extreme anthropomorphism, how can the uncanny valley effect be effectively avoided? Does XPeng have systematic design, strategies, or technical judgments in this regard?

He Xiaopeng: There are manufacturers in the market that make non-humanoid robots, and there are those that make humanoid ones. I think both directions are valid because different scenarios require different types of robots.

XPeng chose a particularly anthropomorphized technical route for humanoid robots because we believe that different forms of robots have very different needs, such as data requirements, scenario needs, and user acceptance. After we developed a robot with muscles and skin, we made an unexpected discovery: previously, during the development and testing of robots, almost no one was willing to stand close to the robot, perceiving it as a machine. However, after we improved the muscles and skin, even girls dared to stand very close because they felt it resembled a partner in their lives or future work. This is a significant change we observed.

XPeng will soon enter the mass production phase of robots. We expect humanoid robots to take the lead in commercial scenarios, with the core premise being to ensure they possess a strong sense of affinity, so that children, adults, and the elderly are willing to approach them actively.

Just like what we are doing with VLM, next year we will have anthropomorphic multimodal voice (the Ultra version of XPeng X9 will also have VLM), which is a system derived from the same source as the robot, capable of speaking in different tones like a human. When you buy a robot, you can define its voice, and these are some of the things we are working on.

So from another perspective, we believe that versatile, anthropomorphic, and safe humanoid robots will become true partners of humanity in commercial, industrial, and home settings in the future. This is XPeng's choice; I believe other forms are also a correct path, but the market should flourish with diverse options, producing good products to serve everyone.

Question: There has been a serious shortage of lithium iron phosphate batteries this year. How much production capacity has been reserved for the XPeng X9, and if the XPeng X9 receives a surge in orders, how many can ultimately be delivered?

He Xiaopeng: Every company has its choices, and we have chosen platform-based battery cells. The battery cell is a platform-based cell, and there is no issue with the production capacity of the battery cells because all range-extended battery cells (including this year's X9 and next year's seven range-extended models) are all of the same type. Therefore, the production line for these battery cells is uniform and moves forward in the same way, so there is no pressure in this regard.

Question: The price announced at the press conference is quite explosive. The XPeng X9 Super Range Extender is positioned as a flagship product, starting at 309,800 yuan. What is the overall thinking behind this? Will it result in losses?

He Xiaopeng: This pricing was a very difficult decision for us internally; to some extent, what we did not anticipate while developing the range extender is that the entire hybrid market in China has been declining this year, while pure electric is on the rise XPeng launched the next generation of range extenders, with the core hope of bringing the next generation of range extenders to more users (those who originally drove gasoline cars and often travel long distances) and users in some countries/regions where charging is inconvenient. This is a new consideration for the XPeng X9 super range extender and is also the main reason for our final pricing.

Q: The pricing logic for range extenders is different from that of pure electric vehicles. So, compared to the pure electric version of the XPeng X9, is the cost of the range extender version higher or lower?

He Xiaopeng: The pure electric and range extender systems are different. Pure electric vehicles have larger batteries, so their prices are definitely higher than those of range extenders. This is also one reason why pure electric models have struggled against range extender models in the past. However, I believe that XPeng's entry into super range extenders through pure electric technology is mutually beneficial. This is the pricing for our first range extender vehicle, and next year we will have multiple "one vehicle, dual energy" cars. Everyone will see that under the stable support of pure electric systems, our range extender models will have a new pricing logic.

Q: The first question is to introduce some technologies behind the XPeng X9's tire blowout; the second question is how the internal division of the "one vehicle, dual energy" strategy is categorized, such as which cars are pure electric, which are pure electric + range extender, or if all are "one vehicle, dual energy"?

Li Yifan: The tire blowout stability technology we launched this time focuses on scenarios where users are easily exposed to safety threats on highways, mountain roads, etc., and where human intervention may further increase the danger. Therefore, we have achieved coordination at the technical level across the entire intelligent chassis domain, including the entire electric drive system, braking system, and steering system. During the entire cornering process, the vehicle's ESP system works stably, which is a fusion of multiple system integrations and an integration of the intelligent chassis domain.

The second key point is whether the signal acquisition is fast and accurate enough. In fact, we have deployed very high-precision sensors in the tires for tire pressure recognition and tire height recognition, with a very fast response speed. The signal source comes quickly, and processing is fast, making it easy to enhance safety AI through multi-domain integration. This is roughly the technical aspect.

He Xiaopeng: I cannot answer regarding product planning at this time. What I want to say is that "one vehicle, dual energy" represents that XPeng will have dual energy for suitable vehicles in the future, with the same appearance and type but different energy sources. However, some vehicles may not be suitable due to chassis or demand reasons, so they will be canceled based on demand, but most vehicles will be "one vehicle, dual energy."

Q: The XPeng X9 range extender version uses Max and Ultra chips, all of which are Turing. The XPeng X9 pure electric MAX version still uses dual Orin. Is there a timeline for switching the pure electric version to all Turing chips? Will XPeng say goodbye to NVIDIA after the X9 range extender and use its own Turing chips for the entire lineup?

He Xiaopeng: First, NVIDIA has always been a great partner for us, so in the future, everyone will still see that we will use many NVIDIA chips. Second, we will gradually have some Turing switches on different vehicles in the future, and we will see the specific rhythm when that vehicle is ready Question: XPeng has work points in Malaysia and Singapore. Can you please share XPeng's specific market expansion plans for these two countries? Including Thailand.

He Xiaopeng: We are making arrangements in Thailand and Malaysia this year, and we hope to have deeper layouts in more countries next year, not just for cars, but possibly also related to the supply chain and other areas. We are quite looking forward to more good Chinese cars entering (these two markets), and for the entire Asia to sell well.

Question: Now that it's November and we are about to welcome 2026, do you think there will be any disruptive changes in the car market next year? Do you have any predictions for the car market landscape next year?

He Xiaopeng: I think every car company is cautious. We insist on steady and long-term development. A year ago, it was hard to imagine the changes in the Chinese car market now, and it will be hard to predict a year later. I think there are many changes. Especially in intelligent driving, will we start to see the arrival of the (automated driving) iPhone 4 moment next year? The only thing I can be sure of is that the competition in the automotive market will be more brutal and bloody in 2026.

Supplement: Some car company CEOs have said that in the next five years, only a few leading companies may survive. A few years have passed, and although everyone is struggling, they are still alive.

He Xiaopeng: In the internet sector, due to network effects, it is easy to see who wins and who loses. The automotive industry does not have this, at least not currently.

The automotive market is a larger and more sustainable market, and it affects a large number of supply chains and ecological enterprises. So I have been telling our internal colleagues that this is a long marathon, or even a never-ending marathon.

I personally think that the (challenge nodes) will be graded and categorized. I believe that in the hardware era, the first, second, and third camps can find enough survival points; however, in the hardware + software era, some effects of the internet will be added, and at that time, the first and second camps will pull further apart and run further away.

For example, one painful thing we are facing now is that the AI training costs for next year have already reached a high number. What about the year after next? What about the year after that? Therefore, we will consider the logic of 3-5 years, how much revenue and profit are needed to support such a large investment, and what ways can I use to reduce investment while maintaining competitive advantages. These are all things that need to be thought about and researched in advance. So, because there are many changing factors, I still believe that the competition in the automotive market in 2026 will be more exciting, more brutal, and more bloody.

Question: We have already found that XPeng's G7, P7+, and the currently listed X9 all have super extended range versions. How will the differentiation between super extended range versions and pure electric versions be defined in the future?

He Xiaopeng: As mentioned earlier, pure electric, or "one car with dual energy," our logic is still based on the overall needs of new cars/platforms/customers. There is no one-size-fits-all approach, saying we will do everything or not do everything, or that certain (versions) must be done or not done. There is no such logic. Therefore, next year we will have multiple extended range vehicles appearing, and I think there is no need to pursue (such) logic and rules Question: In the user group of XPeng, there are three variations of how you are addressed. At first, they called you He Xiaopeng, now they call you General Xiaopeng, and I also see some people calling you Peng Shen. I think these three changes reflect the users' changing expectations for XPeng. What is your view on and response to this expectation? Additionally, I am curious, which of these three names do you personally like the most?

He Xiaopeng: I still prefer to be called "Xiaopeng"