He Xiaopeng's five pushes for range extension have finally landed

Author | Wang Xiaojun

Editor | Chai Xuchen

In addition to opening up imagination and providing the outside world with some future-shocking robots, XPeng's current money-making business is still automobiles.

"After tonight, XPeng will no longer just be a pure electric vehicle company," said He Xiaopeng, Chairman and CEO of XPeng Motors, on November 6.

This statement was once interpreted by the outside world as a hint of disappointment. He Xiaopeng explained that in order to promote the establishment of the range extender project within the company, he had to "break through" five times before finally succeeding.

At last year's XPeng Technology Day, He Xiaopeng announced that XPeng officially entered the range extender track. A year later, the Kunpeng Super Range Extender's first model, the XPeng X9 Super Range Extender, has begun pre-sales, with the new car offering two versions: Max and Ultra, priced at 350,000 yuan and 370,000 yuan, respectively.

Specifically, the difference between the two versions of this car mainly lies in the number of chips. In terms of range, the X9 Super Range Extender is equipped with a 63.3kWh ultra-fast charging lithium iron phosphate battery and a 60L fuel tank, achieving a comprehensive range of 1,602 kilometers. For energy replenishment, this car still uses 5C+800V ultra-fast charging technology, allowing for a 70% charge in just 12 minutes.

To some extent, XPeng's turning point in range extenders has accelerated overseas. Currently, XPeng has established a network of 275 stores covering nearly 50 countries. In terms of sales, from January to September this year, its cumulative overseas delivery volume reached 29,706 units, a year-on-year increase of over 125%.

However, during his overseas inspections, He Xiaopeng genuinely felt the inconvenience of charging, which is a "bottleneck" pain point. It was precisely because he experienced the differences in energy replenishment methods in various places that XPeng decided to enter the range extender market.

He Xiaopeng further explained: "If a company wants to enter the global market, it needs a new energy replenishment method, which includes range-extended hybrid. Additionally, to achieve a range of over 1,000 kilometers, it is necessary to adopt a power form other than pure electric."

However, this car is not only aimed at the overseas market. He Xiaopeng found that even domestically, while pure electric vehicles have relatively convenient energy replenishment, XPeng still struggles to sell well in some areas due to a lack of adaptability to local conditions.

With the launch of the X9 Super Range Extender, XPeng has taken its first step in the dual-power strategy in the domestic market. In addition to the X9 range extender version, XPeng has planned at least five range extender models, covering sedans and SUVs, including the already filed P7+ and G6 range extender versions, which are expected to be launched in 2026 and 2027, respectively.

In the just-concluded October, XPeng's monthly delivery volume reached a new high, stabilizing at a level of 40,000 units. With the addition of range extenders, XPeng's sales are expected to reach new heights. He Xiaopeng revealed that XPeng's R&D expenses this year are nearly 10 billion yuan. This means that only if the automotive business performs well and contributes sufficient cash flow to XPeng can the entire ecosystem cycle effectively.

From pure electric to range extender, from automobiles to robots, XPeng is redefining its boundaries and stepping into new challenges. Next, XPeng's range extender will need to undergo market testing The following is a conversation with He Xiaopeng, Chairman and CEO of XPeng Motors, and Gu Jie, head of XPeng's powertrain, edited without affecting the original meaning:

Question: In a rather tragic way (cutting the robot on the spot) to prove yourself, especially seeing the robot walk past you off the stage, it looked very bittersweet. What were your feelings at that moment?

He Xiaopeng: For us, we want to create a very human-like robot because robots are quite a new thing.

As you can see, even today, people are still questioning whether new energy vehicles are safe enough and whether autonomous driving is safe enough, essentially proving a point. It was very difficult to persuade our robot team because they disagreed with this approach; they felt it was their child, their creation. However, from another perspective, we still wanted more friends, especially those not in this industry, to believe that this industry could eventually produce something different and that Chinese tech companies could create something unique.

We had never conducted such a test and didn't know how to do it. Today, I asked him (Gu Jie) if a pair of scissors could cut through it. The muscles are very firm. Most people opposed it at the time, but I still felt that if this approach could help accelerate the development of robots in China, it would be very good.

After the robot's performance at the Spring Festival Gala, it propelled investments in the entire robot industry nationwide. This is an inspiring event for the entire robot industry; however, there will definitely be many challenges in this process. But if we can instill more confidence and understanding in 1% of industry users and 99% of non-industry users, ultimately pushing the robot industry forward, I think that would be a victory.

Gu Jie: I actually found your question particularly touching. The power system of the robot, which is what our department is developing, has very few components displayed externally. The first generation of IRON shows many joint structures, while the second generation is very human-like. Of course, it was really painful to cut it open. It is also a demonstration; it is a real robot, and it displays a very elegant posture. The entire walking process is very graceful. I personally think it is a near-art piece. However, we need to prove ourselves, not just through technical self-affirmation but also to demonstrate to the audience that it is a real robot that can walk so well, showcasing our technical strength.

When XPeng told me, I felt somewhat helpless. But it still walked very well, even after cutting through the skin and muscles; it still walked elegantly, and I think that is enough.

Question: You mentioned when introducing the X9 range extender, “XPeng will no longer be a pure electric vehicle company.” I sensed a bit of resistance in that statement. Do you still have some reluctance towards making range extenders?

He Xiaopeng: On the contrary, I actually pushed for the range extender five times before it finally succeeded on the fifth attempt. I first advocated for XPeng to develop a range extender in 2019 because I always believed that the ultimate technology should meet customer needs. I feel that after pushing five times, we are finally about to launch it, and I have a sense of emotion Question: Will the XPeng X9 extended range version focus more on overseas markets? What is the core of sales growth? Should we already have a good judgment on the profitability target for the fourth quarter internally? Many products are scheduled for mass production next year, does that mean we need to set a profitability target for the entire next year?

He Xiaopeng: The XPeng Super Extended Range is not only for the global market but also for China. If you look at the distribution map of hybrid cars in different cities and regions in China, you will see that XPeng has been lacking in sales in many areas because we did not have cars suitable for local use. Therefore, I believe that whether in China or globally, the Super Extended Range has a very large usage scenario.

Regarding profitability, we are steadily moving towards achieving our fourth-quarter targets, but we cannot express a definitive stance. We have never provided clear guidance for next year. However, I can share that over the past two years, along with this year, XPeng has been guided by four words—steady progress for the long term.

As I rethink the future competition, future technology, and our positioning in this market, we need to rebuild many foundations, including technological foundations, product foundations, organizational foundations, and globalization foundations. Therefore, only by slowing down a bit and solidifying the foundations can we go further in the future, rather than pursuing faster growth in a particular year.

Question: Do you think there is a marginal effect in the large battery extended range technology route? Where is the balance? Should we continue to make the battery larger and the fuel tank smaller, or is our Super Extended Range 60 paired with 60 the optimal solution?

He Xiaopeng: First, when our Super Extended Range has a broader sales range by the end of this year or next year, we can gradually understand user behavior and infer their thought processes, because user behavior does not necessarily represent their true thinking.

Second, China is different from overseas. The large fuel tank we ultimately create allows users to add just a little fuel; 10 liters can accomplish a lot, and usually, it may not be necessary to fill it up. I have been to Xinjiang more than once, and when I was in Xinjiang, I particularly hoped my car could travel over 1,000 kilometers because refueling there takes a long time—about four hours to get to a gas station, and you often have to wait in long lines. So when I drive in Xinjiang, I generally prefer not to refuel and hope to drive for a very long time to save time. Therefore, I believe the demand in various scenarios is really different. I think a large fuel tank can provide users with the option to choose less or more. As for the large battery, I don't know, but I believe that the battery XPeng is currently using is already very suitable. Suitability is the most beautiful; it doesn't necessarily mean the largest.

Question: How is the boundary of the XPeng extended range power system defined? Will all models be equipped with the extended range system? How will the robotics and other businesses develop in the future? Will it continue through financing, or will the vehicle business strive to generate more cash flow to support R&D in other areas?

Gu Jie: From a technical perspective, each vehicle has the capability to develop both pure electric configurations and Super Extended Range configurations. Whether each vehicle undergoes such development depends on our judgment of the market and the models. The technical capability is there; it is just developed as needed He Xiaopeng: First, XPeng Motors' R&D expenses this year are nearly 10 billion, not including flying cars, which would make the expenses even higher. This ranks about 20th to 25th among private enterprises in China in terms of R&D investment. I personally believe that sufficient R&D intensity is necessary to create differentiated products.

Second, XPeng's R&D is very different from many car manufacturers; a lot of our R&D expenses are spent on software rather than just hardware. Hardware R&D expenses, such as mold changes during annual updates, are essentially gone with the next generation, while software has continuity. However, this year, large models have rendered previous training results almost useless. Software R&D is experimental science and can be very costly. However, from another perspective, software R&D should achieve platformization or modularization, which is standard in hardware but not in software.

From our ecosystem perspective, Gu Jie is responsible for automotive power, robot power, flying car power, and aircraft power; the autonomous driving team works on both automotive and robot autonomous driving, achieving strong modularization, such as the physical world simulation model, which was designed this way from the beginning. I believe that in about three years, the value of software in the physical AI world will truly emerge; currently, it is still in its infancy, and we have not yet seen the real effects.

I firmly believe that the endgame of this competition is not just hardware; hardware is the core foundation, but strong software is also needed, and not just hardware and software.

This year, we achieved profitability despite such high R&D expenses. I believe that next year and the year after, XPeng's revenue and sales scale will significantly increase compared to this year, and at that time, we will have the capacity to support more development.

Everyone should know that our significant R&D investment really started after 2022. Previously, we didn't have the money or the courage to invest so much in R&D. Although we have been doing this for a long time, the amount invested was very small, and we are only now starting to accelerate. I firmly believe in long-termism and the need to invest in R&D, even though it comes at a cost. I think companies need to do this. Many people ask whether XPeng can first do one thing well before building an ecosystem; this is an important judgment of mine. In fact, different thoughts on future strategies and each strategy will lead to different feedback.

Question: XPeng has always emphasized that the Kunpeng range extension technology benefits from the development accumulation in the pure electric field over the past decade. How much advantage does this technology reuse bring to the development of our X9 model? Additionally, how should this cost advantage be reflected in the final pricing of the product in terms of cost control and supply chain management?

Gu Jie: The technologies we use in range, energy consumption, batteries, and supercharging are the same as those in pure electric technology. This is thanks to what we call the super pure electric system, which is based on all our existing core key components, such as battery cells, drive systems, generator systems, thermal management systems, etc., all using the same platform as pure electric. Only by controlling the energy consumption of pure electric can we achieve equally excellent range performance in super range extension models, which is an extremely important foundation.

Additionally, the platformization of key components is also crucial. What we refer to as "one vehicle with dual energy" can help us expand production scale through platformization. At the same time, having a stable supply chain system can ensure the realization of economies of scale Question: Regarding the battery manufacturers' capacity being relatively tight in the fourth quarter of this year, everyone is scrambling for batteries, whether it's ternary lithium or lithium iron phosphate. XPeng has added extended-range models, and the demand for batteries has become very large. Does XPeng have any measures in place for battery production? The pure electric version of the X9 has already been produced in Indonesia; when will the extended-range version start production overseas?

He Xiaopeng: I can only say that in the past week or two, I have had drinks with all the bosses of our battery manufacturers.

As for Indonesia, the first issue to resolve is that the production efficiency of pure electric vehicles in Indonesia cannot meet our delivery demands there. Currently, our orders can only be produced in Indonesia for 11-12 months to meet one month's sales demand. We are already in a lot of pain communicating this, and we have not yet considered when to introduce the extended-range version there; we need to solve the pure electric issue first, which is the topic we are currently discussing.

Question: We have seen that the extended-range market has declined for three consecutive months, and the trend is very obvious. In this case, what is the overall sales expectation for the X9 with the launch of the extended-range model? What kind of volume can it reach? From your perspective, how much space is there for the extended-range market in the coming years? Where are the opportunities?

He Xiaopeng: I don't pay much attention to the market fluctuation data from the past few months. The decline you mentioned is a fact, but it may be a short-term fluctuation. Another possibility is that the entire market is looking forward to "super extended-range"—not just us, but other companies that have launched extended-range models may also introduce next-generation products. The market is like waves, with ups and downs.

Personally, I have a different feeling. Although there is no clear logical support, I believe that in the future, the likelihood of fuel vehicle users switching to super extended-range or extended-range will be greater. Currently, the power consumption of software in cars is still low; once the proportion increases, the power from fuel vehicles will not be able to support high power consumption demands—power consumption will be a major issue in the future, just like how powerful local AI is difficult to run on mobile phones today. The key is not the chip, but the insufficient power support.

Therefore, I firmly believe that super extended-range will form a clear transition from fuel vehicles. The proportion of fuel vehicles in China is still very high, and in some overseas countries, it is even higher, so this is a trend. This is also why we firmly believe that super extended-range will still account for a considerable proportion in the next 10 years, or even 20 years.

Question: There are industry views that believe there is an essential difference between assisted driving and autonomous driving, and that having a driver in an L4 level Robotaxi is meaningless. How will XPeng refute this view?

He Xiaopeng: I think this is a matter of choice and does not need to be refuted. It is a different perspective on the future, which could be right or wrong. We believe in our judgment—next year we will launch three Robotaxis, and we believe in the form of robots.

If autonomous driving (or automated assisted driving) can achieve a situation where a takeover is needed once a month, three months, or even six months, the demand for any robot will be very high. But if frequent takeovers are needed, then the value is indeed limited. If a significant technological change brings about a disruptive experience, it could create entirely new scenarios and demands, which is exactly what we are doing. There is no right or wrong, just different companies choosing different beautiful visions Question: Seeing the IRON robot elegantly walk on and off stage, while the slicing material of Hangzhou Yushu Robot shows it can breakdance and box. Would the average audience think that a robot that can breakdance and box is more advanced than one that can just walk? What are the current standards for evaluating the strength of embodied intelligence players?

He Xiaopeng: It depends on the scenario. XPeng aims to create high-level humanoid robots. I emphasized yesterday that during mass production, it needs to meet higher quality requirements, and the software should achieve a 1:1 balance with the hardware, rather than being hardware-centric. An important realization is that it must not only be fully self-developed but also integrate across domains to balance the brain, cerebellum, body balance, and posture coordination.

This is also why we abandoned the original quadruped or biped reinforcement learning path in 2022—this path could allow some joints to perform well but could not coordinate all 82 joints in the body. We chose a more difficult path, uncertain if we can reach the other shore, but we shared at the recent technology day because we see the possibility of success on multiple independent lines.

The challenge in the coming months is whether we can fully integrate hardware and software and achieve rapid internal OTA iterations after integration. This process may differ from other robot manufacturers.

Question: If future models like the P7 offer both range-extended and pure electric versions, they will face the reality of limited store space. How should the pure electric and range-extended versions of the same car be displayed? What are the criteria for evaluation? By region or dealer preference? In terms of sales strategy, do we hope the same sales team sells both energy forms simultaneously, or separately? What challenges will this bring to the sales system? What is the risk control rate for the components of the X9 pure electric and range-extended versions, and will it serve as a reference value for other models' range-extended versions?

He Xiaopeng: We implement a "one car, dual energy" strategy—same shape, same interior, different energy forms. Customers can choose between pure electric or range-extended, with a consistent perception. We hope customers choose based on the model rather than the energy form. Therefore, the sales team will not be divided by energy type but will sell all versions of the same model uniformly.

Gu Jie: The commonality rate of key components is very high, and many body parts also have a considerable commonality rate. As for specific data, I cannot provide a direct answer, but the most important thing is that we need to develop the one car, dual energy platform well, which is very beneficial for the entire manufacturing supply chain.

Question: Which do you think will become a new performance growth point for XPeng in the future, robots or flying cars? From a profitability perspective.

He Xiaopeng: Both flying cars and robots will have commercial value for us. Everyone will see that flying cars will definitely be mass-produced next year, and we will go all out for robots because the challenges for robots are even higher than for flying cars. I want to emphasize XPeng's three curves: the first is cars, the second is robots, and the third is that we will truly bring these two products to the global market. This is our expectation for the three curves, so I believe that robots may, in the next 10-20 years, become twice the size of the current global automotive market, representing a much larger market. We do not know when we will reach its turning point, just like when Apple launched the iPhone 4, we believed it would reach that point. We are slightly ahead in this race and will continue to push forward