Introduction: Auto companies are encountering each other on another track.

The automotive industry is undergoing a somewhat familiar transformation.

Previously, real estate companies transitioned into computing power, data centers, and AI; now, it's the turn of automotive companies.

Xpeng is developing robots and secured over $900 million in financing in August; SAIC has directly integrated humanoid robots into its battery production lines; GAC has established an independent robot company; Changan has scheduled robot mass production for 2028; Chery has already delivered over 2,000 robots.

According to incomplete statistics, as of September 2026, 14 mainstream domestic auto companies have clearly laid out plans for humanoid robots, ranging from Xpeng, Li Auto, and NIO to Changan, SAIC, Chery, and further to GAC, Geely, BAIC, and FAW.

Auto companies are encountering each other on another track.

The question arises: Why are automotive companies suddenly eager to 'create humans'?

The answer is not complex: Making money from cars has become increasingly difficult.

Data from the National Bureau of Statistics shows that from January to July 2026, the operating income of the automotive manufacturing industry above a certain scale was 6.08 trillion yuan, a year-on-year increase of 2.7%, but the total profit was only 216.24 billion yuan, a year-on-year decrease of 20.4%, with an operating income profit margin of approximately 3.6%. During the same period, the profits of industrial enterprises above a certain scale nationwide increased by 17.6%.

Data from the Ministry of Industry and Information Technology indicates that from January to July, national car sales reached 17.602 million units, a year-on-year decrease of 3.7%. Although new energy vehicles continue to grow, the traditional automotive industry as a whole has entered a phase of Stock competition (stock competition).

The People's Daily also directly pointed out this change in September 2026: In the financial reports of automotive-related companies in the first half of the year, some complete vehicle companies experienced 'increased revenue but decreased profits.' Meanwhile, the profit performance of upstream sectors such as batteries, chips, and intelligent driving solutions showed significant differences.

Cars are increasingly resembling a typical heavy-asset, low-profit business—sales remain, but profits vanish.

Consequently, various auto companies are to happen to coincide (in unison) turning their attention to robots.

01 The Robot Dilemma of Auto Companies

At first glance, auto companies venturing into robotics somewhat resembles real estate companies suddenly delving into computing power.

One sells houses, the other manufactures cars, yet both are suddenly talking about AI and robots, seemingly searching for a 'second growth curve.'

However, the relationship between robots and cars is much closer than that between real estate and computing power.

What intelligent vehicles essentially do is: perceive the world, understand the world, make decisions, and execute actions.

Humanoid robots follow the same pattern.

Cars use cameras, LiDAR, and millimeter-wave radar to perceive the environment, employ chips and models to judge what lies ahead, and then decide to accelerate, brake, or steer.

Robots simply have a different body.

They use cameras and sensors to perceive their surroundings, employ AI to understand the environment, and then decide to raise their hands, grasp, walk, or avoid obstacles.

Therefore, in March 2026, when People's Daily Online reported on auto companies flocking into humanoid robots, it directly summarized this relationship as follows: Auto companies' Layout (layout) in humanoid robots inherently involves significant technology collaboration.

This is also why auto companies dare to charge ahead.

They already possess some of the most challenging elements for robot companies to acquire: supply chain, manufacturing capabilities, intelligent driving algorithms, computing power, real-world scenarios, and even data, as in the case of Xpeng.

Unlike many companies, Xpeng does not merely feed car data to its robots but directly integrates its automotive AI system with its robots.

The company had previously clearly stated that robots and intelligent driving research and development have synergistic effects and planned to have robots use the same physical AI foundation as cars. In August 2026, Xpeng's robot business completed its first round of financing exceeding $900 million, with a post-investment valuation exceeding $6.3 billion; in September, it officially launched its robot production line.

This is no longer just a case of an 'auto company casually researching robots' but treating robots as a genuinely new business.

However, there is an easily overlooked aspect in the market—car data does not equal robot data.

This is both the greatest advantage and the biggest misconception for auto companies venturing into robots.

Driving data primarily addresses road perception, driving decisions, and vehicle control.

What robots truly lack is another type of data: How do I reach out for a cup? How much force should I apply after touching it? How do I move it after grasping? What should I do if I encounter an obstacle during movement? What if the object slips?

These are all 'operational data.'

In April, People's Daily Online cited relevant research reports stating that by 2025, China had over 140 humanoid robot companies, with shipments reaching 14,400 units, accounting for 84.7% globally; however, the industry is still in the stage of transitioning from technical verification to industrialization.

Therefore, the truly advantageous auto companies are not those with 'many cars' but those capable of integrating automotive data, robot data, models, and real-world scenarios. Among them, SAIC is a fascinating example.

In March 2026, 'Nengzai No.1' officially entered the Buick Zhijing E7 battery mass production line, undertaking tasks such as cell grasping and feeding. SAIC claimed this to be the first batch of humanoid robot applications in China's automotive industry to truly enter mass production lines.

Having a robotics company purchase dozens of robots for training is entirely different from an auto company directly placing robots into a car factory that produces daily.

The former is conducting tests, while the latter is creating a data closed loop.

02 The Fork in the Road for Auto Companies

It is the real-world business scenarios brought by factory production lines that complete the crucial leap from prototype testing to data closed loop.

However, even with the unique advantage of a ready landing ground, it does not mean that all auto companies entering the field can smoothly achieve the commercialization of humanoid robots.

The supply chain, manufacturing, intelligent driving algorithms, and road data accumulated in the automotive sector merely lay the foundation. To truly succeed in robotics, another set of capabilities, such as body control, dexterous operation, and vertical scenario adaptation, must be acquired. Different auto companies hold different cards, leading to diverging development paths.

The first category is companies like Xpeng, which focus on 'software + AI + robots.'

Xpeng's advantage does not lie in its car scale but in its early push towards becoming an AI company.

In the first quarter of 2026, Xpeng had already defined itself as a physical AI company covering intelligent vehicles, autonomous driving, flying cars, humanoid robots, and AI chips; its robot mass production base covers approximately 110,000 square meters, with plans to achieve mass production by the end of 2026 and enter stores and commercial scenarios in 2027.

By September, the robot production line was officially launched.

In other words, Xpeng has completed the first round of the 'model—robot—scenario—manufacturing' closed loop.

The second category consists of traditional automotive giants like SAIC, GAC, and Changan.

Their greatest advantage lies in their factories and industrial organization capabilities.

SAIC has already integrated robots into its mass production lines while making industrial investments in robot bodies and core components.

GAC has gone a step further. In February 2026, GAC officially established Hui Lun Technology, transforming its robot business from an internal R&D project into an independent industrialization entity; in August, Hui Lun Technology secured financing exceeding 100 million yuan to continue promoting the large-scale commercialization of humanoid robot Complete machine (complete machines) and core components.

Changan has explicitly stated its goal of achieving mass production of humanoid robots by 2028 and is Layout (planning) scenarios in factories, stores, and homes.

Their advantages may not be as glamorous but are highly practical, including factories, supply chains, workers, customers, and after-sales systems.

Once robots are truly deployed on a large scale, these elements will become crucial.

Besides the three categories mentioned above, Chery deserves special mention because it does not merely bet on 'general-purpose humanoid robots' but focuses on the commercialization of robot products.

In April 2026, Chery's Moja robot completed a 1,000-unit order signing and a 100-unit Centralized delivery (centralized delivery); by July, Moja robots had accumulated over 2,000 deliveries globally, with products and services covering more than 60 countries and regions.

These 2,000 units cannot be simply understood as 2,000 humanoid robots because Moja's product matrix includes humanoid robots, intelligent police robots, medical guide robots, robot dogs, and other categories.

However, this achievement remains significant as it proves one thing: Robots do not necessarily have to wait until 'general-purpose humanoid' technology matures to generate revenue. Starting with vertical scenarios such as security, medical guidance, industry, and services is also a viable path.

This actually poses two different questions for auto companies: Do they want to become a 'robot company,' or do they merely want to establish a 'robot business'?

These two paths are vastly different.

03 Being Able to Build Cars Does Not Equate to Being Able to Build Robots

Someone might ask: Can any auto company build robots?

The answer is clearly no.

One of the most common illusions in the robotics industry is the belief that because one can build cars, one can also build robots.

This does not hold true.

Cars can rely on platformization, standardized components, and a mature supply chain to distribute complexity.

The biggest challenge for humanoid robots currently lies in the 'last meter'—hands, feet, joints, dynamic balance, dexterous operation, and model generalization.

This is why individuals within the automotive industry themselves are highly cautious.

The Dongfeng R&D team has publicly pointed out that the 'brain' of humanoid robots is still rapidly evolving, with technology routes not yet truly converged; simultaneously, the industry lacks massive amounts of high-quality real-world scenario data, and issues regarding safety, reliability, as well as the trade-offs between overall machine cost and work efficiency, still need to be resolved.

This implies that several types of auto companies may be at risk.

One type consists of companies with only concepts but no robot R&D systems. Launching a robot, shooting a video, and participating in a robot conference do not equate to having a robot business.

Another type comprises companies with only manufacturing capabilities but no AI capabilities. Such companies can 'assemble' robots but may not be able to make them 'learn.'

Yet another type includes companies with only automotive data but no operational data. Millions of cars can teach autonomous driving models how to drive but cannot automatically teach robots how to screw in bolts.

Therefore, the robotics industry may not necessarily eliminate companies without automotive manufacturing capabilities in the future. Instead, it could be those companies that equate 'automotive supply chain advantages' with 'general-purpose robot capabilities.'

Entering the field is merely the first step; the real challenge lies in the next decade.

Transferring automotive AI capabilities, supply chains, and factories into training grounds, and then selling robots.

If any link in this chain fails, the second growth curve could turn into a second R&D black hole.

Xpeng is proving whether it can become a 'robot + car' company; SAIC is demonstrating whether industrial capital, automotive factories, and robot companies can form a closed loop; GAC is proving whether robots can evolve from a group laboratory into an independent company; Chery is taking a more direct approach, seeking commercialization through products and scenarios first.

As for a large number of auto companies such as Changan, Geely, NIO, Li Auto, and Dongfeng, the real test has already been laid out on the table.

This race will not ultimately reward the auto company that 'releases robots the earliest,' nor necessarily the one whose robots 'most resemble humans.' Instead, it may reward a different capability: Who can make robots truly go to work.

In the automotive industry's first few decades, the competition revolved around engines, later batteries, then chips, software, and autonomous driving. Today, auto companies may be vying for a new industrial entry point—combining the robot's body, the AI's brain, and the automotive manufacturing system.

Therefore, this time, auto companies venturing into robotics is not entirely the same as real estate companies transitioning into computing power.

Real estate companies transitioning into computing power aim to leave their old world behind; automotive companies transitioning into robotics somewhat resemble continuing forward along their old world.

Only by reaching the end will they discover that cars can teach robots how to be manufactured but may not necessarily teach them how to 'live,' and this is the true watershed after auto companies collectively rush into robotics.