news

If it doesn’t win sales, what else can Xpeng Motors win?

2024-08-03

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina


Image source: Xpeng Motors

On August 1, each company released its sales results as scheduled. Xpeng's sales, as always, remained above 10,000 units. Compared with Ideal's over 50,000 units and Weilai's over 20,000 units, such sales results are still lower than market expectations.

At the same time, compared with Xiaopeng's intelligent driving, its sales performance has been criticized. In response to this, in the communication after the Xiaopeng Motors' innovative technology conference, Xiaopeng Motors founder He Xiaopeng admitted that there are three main reasons: AI intelligent driving is not good enough; pure electric vehicles have been under great pressure in the past two years; Xiaopeng Motors' products focus on the 200,000-300,000 yuan range, where the competition is the most intense.

Over the past year or so, Xiaopeng Motors has carried out a series of reforms and adjustments. In He Xiaopeng's view, running a business requires a long-term perspective. Xiaopeng Motors' recent series of actions are all sharpening the knife, which does not delay the chopping of wood. Soon everyone will see obvious changes.

He Xiaopeng's confidence in the future may come from two aspects. One is Xiaopeng's advantage in intelligent driving; the other is the cooperation between Xiaopeng and Volkswagen.

So for Xiaopeng, what else can it win if not sales? In fact, the answer is self-evident. Among Wei, Xiaopeng and Li Auto, Xiaopeng’s label is “intelligent driving”.

Intelligent driving has greater potential than new energy

When asked, "Why did you persuade Lei Jun to build a car?"

He Xiaopeng admitted that he once told Lei Jun that the ultimate goal of automobiles is intelligent driving. "From another perspective, we are happier to have more friendly competitors."

He Xiaopeng firmly believes that "AI intelligent driving is a very important moat for car companies and provides valuable data. I believe that in three or four years, rather than five years, everyone will agree that intelligent driving will be a huge change."

He Xiaopeng told Titanium Media App that people's doubts about intelligent driving are very similar to their attitudes towards new energy vehicles five years ago. In 2019, many people dare not say what the future of new energy will be.

He originally predicted that sales of gasoline vehicles would decline rapidly in 2025, but he did not expect that the development of new energy vehicles would be faster than he expected, and initial results have been seen this year.

According to the latest data from the China Passenger Car Association, from July 1 to 28, 2024, the retail sales of new energy vehicles will be 722,000 units, a year-on-year increase of 31% in July last year; but the cumulative retail sales of new energy passenger vehicles this year have reached 4.835 million units, a year-on-year increase of 33%. At the same time, in June this year, the penetration rate of new energy vehicles in the domestic market exceeded 50% for the first time, reaching 50.72%, and it is expected to continue to break historical records in July.

However, in his opinion, intelligent driving has a greater potential than new energy, because the situation of new energy in China is different from that in the world, and AI intelligent driving is the real way to overtake. And the key node of the development of intelligent driving will surely come soon, and it is still at the starting point of change.

This is also why Volkswagen has continued to increase its cooperation with Xiaopeng since last year. On the one hand, it has seen the importance of intelligent driving in future new cars; on the other hand, it has seen Xiaopeng's capabilities in intelligent driving.

He Xiaopeng revealed at the press conference that the cooperation between Xiaopeng and Volkswagen continues to deepen, and the first model based on the jointly developed electronic and electrical architecture is expected to be put into mass production within 24 months.

Looking back on the cooperation between the two parties, the first was equity investment. On July 27, 2023, Volkswagen Group signed a strategic minority equity investment in Xpeng Motors.

The second time was joint procurement. On February 29, 2024, Xiaopeng Motors and Volkswagen Group signed a joint development agreement on platform and software strategic technology cooperation.

The third time is the EEA electronic and electrical architecture technology strategic cooperation. On April 17, 2024, Xiaopeng Motors and Volkswagen Group signed the EEA electronic and electrical architecture technology strategic cooperation framework agreement.

He Xiaopeng said that Xiaopeng has benefited a lot from the cooperation with Volkswagen. The two sides have slightly different views on quality. He shared a story that when Xiaopeng Motors conducts vehicle model testing internally, it will use multiple cars to conduct a million-kilometer test. But when Volkswagen conducts the test, it will only select two cars. "When you drive each of the two cars for 500,000 kilometers or 1 million kilometers, you will see a lot of things that you haven't seen before," said He Xiaopeng.

At the same time, the cooperation between the two parties in the supply chain also enabled Xiaopeng to reduce costs. He Xiaopeng believes that jumping out of the original logic is a win-win situation.

At present, Xiaopeng has officially entered the Singapore market. He Xiaopeng admitted that Xiaopeng Motors is also focusing on the overseas expansion of intelligent driving. To achieve the globalization of Xiaopeng intelligent driving, multiple conditions must be met. First, the hardware conditions must be met, second, it must comply with policies and regulations, and third, it must meet the needs of local customers. The needs of users in different countries vary greatly, and personalized preparation is essential.

End-to-end big model accelerates the development of intelligent driving

Although the domestic intelligent driving industry is also very popular, and all companies are also competing for end-to-end, when talking about intelligent driving, we must talk about end-to-end. However, in He Xiaopeng's view, "the intelligent driving environment in the United States is significantly different from that in China. The American industry is more focused on how to improve end-to-end intelligent driving solutions. However, when I returned to China, I saw that my competitors were still busy making sales rankings and only thought about how to make money, rather than technological innovation and breakthroughs."

At the Xiaopeng Innovation Technology Conference, He Xiaopeng proposed that Xiaopeng Motors has officially entered the stage of "easy to use across the country" and even "easy to use globally", and proposed three "unlimited" requirements for achieving "easy to use": no restrictions on cities, no restrictions on routes, no restrictions on road conditions, and this standard will continue to improve.

He Xiaopeng pointed out that the marketing concept of "can be driven throughout the country" has become rampant. While many competitors are still barely able to drive throughout the country, or are only able to drive throughout the country on designated roads on the whitelist, Xiaopeng Motors has entered the "next level" and is more focused on the "practicality" of intelligent driving.

It is reported that Xiaopeng Motors has invested 3.5 billion in AI training, including computing power and manpower, which is difficult for most companies to do. Intelligence has a network effect. The stronger it is in big AI, the more data it obtains, the more money it needs to invest, and the more cars it needs to sell to support this system. Now almost all supply chains cannot handle this cycle.

"With the support of 'end-to-end', XNGP has stronger learning capabilities and can handle more complex road conditions, which was difficult to achieve in the era of rules. With the increase of data learning, XNGP's capabilities will continue to improve, and intelligent driving that does not use large models will be eliminated." He Xiaopeng said.

He Xiaopeng's goal for the intelligent driving team is to use the current high-end model (Max version) to upgrade its intelligent driving capabilities to the level of the foreign Waymo autonomous driving company through OTA by the second half of next year.

"This means that a person who drives an average of 70 kilometers a day only needs to take over once or twice a week. This experience will be disruptive and completely different from the current driving experience. I think this is not far away and can be achieved within 18 months or even less," said He Xiaopeng.

Xiaopeng Motors completed the mass production of its first end-to-end large-scale model in May. He Xiaopeng revealed that with the support of end-to-end, Xiaopeng's intelligent driving evolution is accelerating unprecedentedly, and the improvement of large-scale version capabilities has been gradually reduced from 13 months. By the end of this year, Xiaopeng will be able to achieve a true "door-to-door" experience, such as being able to automatically pass through ETC and open up internal roads in the park.

As for the end-to-end data problem, He Xiaopeng believes thatThis is too technical." He suggested that there are two things you should never believe: First, don't believe someone who says he saw a model on the Internet, or someone made a model, and that he can do autonomous driving; second, don't believe someone who says he has a lot of cars, so he has a lot of data.

Regarding whether Xpeng Motors' intelligent driving technology will enter the Robotaxi industry in the future, He Xiaopeng admitted that Xpeng Motors will get involved in Robotaxi, but is not willing to take over the operation. He hopes that Robotaxi will not only serve China in the future, but also the world.

In the past month or so, He Xiaopeng has met with the CEO of Didi and the CEO of Uber. In his opinion, they have done a good job in the field of autonomous driving, and there will be many such operators in the world in the future. Xiaopeng Motors hopes to cooperate with them.

Making cars is difficult, and operating them is even more difficult. It’s not just ‘extra difficult’, it’s ‘really difficult’. We should have more partners to work together to do this well.” He Xiaopeng sighed.

He Xiaopeng believes that the real difficulty lies in being able to drive well. It is really not easy to make it easy to use and drive well.

"Is it considered unmanned driving if it takes over once every 200 kilometers? What about once every 20,000 kilometers?" He Xiaopeng asked in return.

In his opinion, this may be a very good autonomous driving, but it may still not be safe enough. There is still a long way to go before the technology is perfect, but under the previous rules, this road will not be finished in ten years, and full-scenario autonomous driving is impossible. He believes that end-to-end is the future, and not only L2 can be developed, but also L4 in the future.

He Xiaopeng believes that end-to-end and past rules have their own advantages, at least for now. However, if end-to-end develops two or three versions, the speed will be faster. To analyze from a plain perspective, it is very difficult to globalize past rules because the global environment is too complex, so end-to-end and map-free are the only way to globalize.(This article was first published on Titanium Media App, author: Zhang Min)