This Battle, "Weilai Xiaoli" Cannot Afford to Lose

"Heavy traffic towards Shanghai, please slow down and stop using assisted driving."

The just-concluded National Day holiday, on the last day of the return journey, I captured a slogan on the Shanghai-Chongqing Expressway, and suddenly felt a mix of emotions.

In 2025, what was expected to be the inaugural year of widespread smart driving adoption was instead doused with a cold splash of reality due to various internal and external challenges, almost reverting back to square one. In the minds of many ordinary consumers, the newfound sense of novelty and trust that had been painstakingly built up quickly crumbled.

In my view, for intelligent driving to mature and for everyone to truly be willing to accept and use it with peace of mind, there is undoubtedly a long way to go. The current situation resembles the darkness before dawn, requiring all players to continuously refine, cultivate, and iterate on themselves.

Certainly, during this process, there is one undeniable point: as competition becomes increasingly fierce, in order to stay in the game in the Chinese car market, no one can afford to lose the battle of intelligent driving, which must be regarded as a top priority for a company's development.

After all, as the battle for electrification in the first half comes to a close, the second half's struggle for intelligence is gradually unfolding. Performance in this dimension, whether good or bad, may ultimately determine the final ranking positions.

Similarly, those "Wei Xiaoli" who have long been under the industry's magnifying glass must feel even more deeply. Recently, whether by coincidence or intentional design, the intelligent driving teams of the three new car-making forces have all experienced significant personnel changes.

The adjustments and shifts in strategy fundamentally reflect a sudden change in underlying technological approaches. At the same time, they also highlight the anxiety, pressure, and determination of "Weixiaoli" to seek leadership in the intelligent driving dimension.

01The old era has come to an end.

Messages are more interesting when read together.

Yesterday morning, there were reports that Li Liyun will no longer serve as the head of Xiaopeng's autonomous driving center, and Liu Xianming, the head of the world base model, will take over the position.

In response, the official statement from this emerging car manufacturer was, "We thank Liyun for leading the XPeng intelligent driving team since 2023 and for the significant contributions made to the company."

The resume of the "successor" Liu Xianming shows that he graduated with a Ph.D. in Electrical and Computer Engineering from the University of Illinois at Urbana-Champaign in 2016.

Previously worked at Facebook (now Meta) and Cruise, an autonomous vehicle company under General Motors, engaging in cutting-edge research in the fields of machine learning and computer vision.

In March 2024, officially joined XPeng as the head of the AI team, and made a public appearance in June of the same year as the head of the World Base Model.

It is worth noting that before this round of personnel changes, he was already reporting directly to He Xiaopeng.

Coincidentally, also this week, it was reported that Bai Yuli, the head of NIO's artificial intelligence platform, and Ma Ningning, the head of the world model, have recently resigned. According to public information, Huang Xin, the head of intelligent driving products, has also left the company.

In response, the new automotive power company officially stated: "In order to strengthen the absorption of the latest technologies in general artificial intelligence and to efficiently, swiftly, and with high quality deliver the intelligent driving experience, we have proactively adjusted the organizational structure of our intelligent driving department."

The organizational restructuring will be more conducive to fully sprinting towards the development and delivery of the World Model 2.0. According to NIO's commitment, this iterative version will be gradually launched on multiple platforms from the end of this year to the first quarter of next year.

Even more rare is that against the backdrop of constant personnel upheaval, Ren Shaoqing, who holds the position of "Intelligent Driving No. 1," has stepped into the spotlight for the first time to accept an interview, discussing his various reflections along the way. From the outside, this may also be seen as a roundabout response to the multiple rumors of departures this year.

In contrast, Ideal's intelligent driving team underwent a "disruptive" restructuring as early as September.

Before the adjustment, this new force in car manufacturing had three core R&D teams, each led by a senior director. Among them, Jia Peng was responsible for the development of the VLA and world models; Xia Zhongpu was in charge of the end-to-end model development; and Wang Jiajia was responsible for mass production R&D.

However, these three leaders all submitted their resignations this year and subsequently chose to start their own businesses.

After the adjustment, the three core R&D teams were divided into 11 secondary departments, all reporting to Lang Xianpeng, Senior Vice President of Autonomous Driving R&D.

The official explanation for this is, "The organizational restructuring aims to achieve technological breakthroughs and product implementation with greater speed and higher quality, allowing the autonomous driving organization to continuously evolve into an AI organization."

In summary, this is the basic situation of the recent actions of "Weilai, Xiaopeng, and Li Auto."

What have you perceived from the continuous wielding of the knife? From my personal perspective, the subtitle of this section is the answer. All three are bridging the somewhat closed gap that was previously established for "functional implementation," swiftly adjusting to a more suitable "AI large model" flat development structure.

The latter, in the intelligent driving sector, has apparently been elevated to an unprecedented level of strategic importance.

02A new war has begun.

"The first half of intelligent driving is about the ability to implement engineering and who has stronger execution; the second half of intelligent driving is about imagination, about who has a more powerful underlying architecture and who has a broader technological vision."

While browsing Weibo by chance, I came across this perspective. I deeply agree with it, and it precisely explains the fundamental reason behind the series of major personnel changes at "NIO, Xpeng, and Li Auto": "Without decisive action and seeking change, one will fall into the quagmire of stagnation."

The three new power car manufacturers deeply understand that "the old era has ended, and a new war with AI large models as the backdrop has begun." In the following sections, we are not inclined to delve into the somewhat obscure technical dimensions. Instead, we would rather discuss the real situations they are facing.

First, focus on Xpeng.

When it comes to this new power in car manufacturing, "intelligent driving" is perhaps the key label that has been firmly attached to it, and also its golden signature for selling cars. In terms of overall experience, it even had a period of distinct lead.

In the past two years, with the downfall of "oligarch" Huawei and the aggressive push by many leading suppliers, the so-called first-mover advantage has gradually been leveled out.

It is no exaggeration to say that Xpeng is facing fierce attacks from both front and back. If it fails to capture the consumers' mindset in this regard, the impact will be devastating.

Secondly, looking towards NIO.

Without exaggeration or bias, in terms of product design, energy replenishment experience, and comprehensive services, this new car-making force has performed remarkably and can be regarded as one of the exemplars of high-end Chinese new energy brands.

However, when it comes to its intelligent driving, it's hard to say everything in one word. Taking several models on the NT2 platform as an example, the most criticized aspect is that, despite being equipped with four Orin chips for top-notch hardware support, the experience has never been on par. Now, switching to the NT3 platform, there still hasn't been a significant transformation.

By the end of this year or the beginning of next year, the promised World Model 2.0 version's level of advancement remains largely uncertain. However, it is certain that if there is a prolonged lag in this dimension, it will inevitably impact sales.

Regarding intelligent driving, NIO urgently needs to clarify its position.

Moreover, it is the ideal.

Compared to Xpeng and NIO, the progress of this new electric vehicle startup in the field of intelligent driving is evident. Whether in terms of user experience or marketing, its efforts are very aggressive.

Under the relentless pursuit of progress, it's inevitable that the foundation becomes unstable. In just the past few years, its autonomous driving technology has ventured into three different tracks: initially relying on high-precision maps with a "rule-based" approach, then learning from Tesla's "end-to-end" method, and now fully betting on the "VLA visual language model."

Such frequent changes are undoubtedly a huge test for the team's stability and technical accumulation.

Furthermore, looking at the specific segment market, Li Auto is currently also the most aggressively targeted by AITO. From its perspective, only continuous iterations in intelligent driving can offer a long-term competitive edge against Huawei. According to the current trend, this new force in car manufacturing has already placed its future entirely on the VLA card.

The pressure I bear is much heavier than I imagined.

Of course, at the level of intelligent driving, "Weilai, Xiaopeng, and Li Auto" face many challenges that are actually similar. All three need to solidify their labels, need to justify themselves, and need to bravely move forward under the pressure from Huawei and leading suppliers.

Anyway, the same thing goes, "In this battle, no one can afford to lose."

As they enter the final stages and attempt to stay at the table, the strength in this dimension will play a decisive role, especially for new forces in car manufacturing. The recent major personnel changes at "Wei Xiaoli" are essentially a typical reflection of the intensifying arms race.

As the article approaches its conclusion, I would like to emphasize: "At first glance, the enthusiasm for intelligent driving in China's auto market in 2025 has shown a slight decline. However, all major manufacturers should understand that there is often a period of unusual calm before the big battle begins. If you become complacent and slow down during this time, you are likely to be the first to be defeated when the full-scale attack occurs."

I have always felt that electrification determines the lower limit of a car company, while intelligence determines the upper limit. Smart driving is the jewel in the crown.

Truly dazzling things will never have their brilliance obscured by temporary clouds. Believe that one day, they will go from being usable to being user-friendly, making everyone love to use them, and will become the winning factor in the fierce competition among car companies.

After the large-scale adoption of L2, L3 and L4 will soon make a brilliant debut.