The wave of Chinese automakers crossing over into humanoid robotics is accelerating, with industry giants and new energy vehicle makers alike shifting from "four wheels" to "two legs" in a full-scale operational push.
BYD Company Limited has confirmed it will officially unveil a humanoid robot at its "Di Space" venue in August, marking the first time the company will publicly display a physical prototype as an official entity. Meanwhile, XPeng's humanoid robot, IRON, has commenced small-batch trial production at its Guangzhou factory, with mass production lines entering their final commissioning phase. Li Auto is planning two robot products—a two-wheeled and a two-legged version—with the two-wheeled robot set for public release within the year.
The flurry of activity from these three automakers signals that domestic car companies have moved from concept validation to a critical juncture in scaling their humanoid robot initiatives.
The appeal of this sector lies in both technological and commercial synergies. The technology overlap between smart vehicles and humanoid robots exceeds 70%, and their supply chains are highly shared. Furthermore, automobile factories themselves represent the ideal initial application scenarios for humanoid robots. Against the backdrop of sustained pressure on automotive industry profit margins, humanoid robots are increasingly seen by automakers as a core direction for cultivating a second growth curve.
BYD's August Debut: First Official Public Display of a Physical Prototype
The timeline for BYD Company Limited's entry into humanoid robotics can be traced back to 2022, when the company established an embodied intelligence research team, a move earlier than widely perceived.
In May of this year, BYD Executive Vice President Li Ke publicly addressed the progress for the first time in an interview for "BYD Shareholder Planet," outlining a vision for deploying robots in every store. She expressed a desire for the robots to handle standardized tasks such as multilingual vehicle explanations, car system demonstrations, and in-store customer engagement. In early June, online rumors circulated about BYD's self-developed humanoid robot, codenamed "Yao Shun Yu," and plans for an internal deployment of 20,000 units this year. BYD quickly refuted these rumors but simultaneously confirmed the fact of its self-developed humanoid robot.
On July 25, the "Di Space" WeChat public account under BYD released a poster featuring a humanoid robot, stating, "In early August, there's a new friend who wants to get to know you." On July 28, sources from BYD confirmed to reporters that the company plans to unveil the humanoid robot at "Di Space" in August. Industry insiders widely expect this debut to be the first public demonstration of a functional physical prototype, rather than a conceptual technology release.
Strategically, Li Ke stated that the core of robot competition lies in manufacturing capability, software capability, and hardware capability, and that capabilities related to automobiles share a common origin with those for robotics. She also revealed that BYD could offer an open platform, allowing for both in-house development and collaboration with other companies. As the global leader in new energy vehicle sales, BYD's deep expertise in core components like motors, electronic controls, and reducers overlaps significantly with the humanoid robot supply chain, forming the cornerstone of its confidence in entering this field.
XPeng's IRON Enters Mass Production Countdown, Li Auto Advances on Dual Fronts
Among new energy vehicle makers, XPeng is pursuing the most aggressive timeline. According to a July 24 report, XPeng's humanoid robot has officially begun small-batch trial production at its Guangzhou factory, with the mass production line simultaneously entering its final joint commissioning phase.
XPeng's robotics strategy was formally established during its 2025 Tech Day. He Xiaopeng announced then that XPeng Inc. would reposition itself as "a mobility explorer in the physical AI world, a global embodied intelligence company," and unveiled its first-generation humanoid robot, IRON. In June of this year, He Xiaopeng issued an internal letter announcing he would personally serve as CEO of the robotics business, elevating humanoid robots to the group's highest strategic priority.
According to plans, XPeng Inc. aims to achieve mass production of IRON by the end of 2026, with a monthly production capacity target of over a thousand units by year-end, initially deploying them commercially in its stores. It plans to enter nationwide physical stores in the first quarter of 2027, expand overseas in the second quarter, and target more households by 2028. The 178cm-tall IRON is equipped with a Turing AI chip (total computing power of 2250 TOPS) and a self-developed physical world large model. It is currently undertaking tasks such as sorting, handling, and quality inspection at the Guangzhou factory. Previously, XPeng Inc. also announced Baosteel as an ecological partner for the IRON project, with future plans to explore applications in complex industrial areas like inspection.
Li Auto is also accelerating its progress. Reports indicate that LI AUTO-W has planned two robot products: a two-wheeled robot primarily for factory settings, scheduled for public release within the year, and a two-legged robot. In terms of organizational structure, Li Auto reorganized its intelligentization department in February 2026, forming three teams for humanoid robots, software systems, and foundational models, with the humanoid robot project codenamed "Nexus." LI AUTO-W stated that it is simultaneously laying out its presence in both the first half of embodied intelligence—autonomous driving—and the second half—general-purpose humanoid robots. Its total R&D investment for 2026 is estimated at a substantial 12 billion yuan, with AI-related spending accounting for approximately 50%.
Automaker 'Robot-Making' Landscape: Full-Scale Rollout from New Forces to Traditional Giants
The intensive moves by BYD Company Limited, XPeng Inc., and LI AUTO-W are just a microcosm of a larger wave of automakers entering the field. According to public reports, nearly 20 mainstream automakers worldwide, including Tesla, BYD Company Limited, XPeng Inc., GAC, Chery, Changan, Xiaomi, LI AUTO-W, BMW, Mercedes-Benz, and Hyundai, have entered the humanoid robot sector through in-house development, investment, or incubation.
Among domestic traditional automakers, GAC Group released its fourth-generation humanoid robot, GoMate Mini, in February 2026. Changan Automobile invested 450 million yuan to establish "Changan Tianshu Intelligent Robot Technology Co., Ltd." Chery launched the "Mo Jia" robot brand. Additionally, SAIC Group's embodied intelligence robot, "Nengzai No.1," has entered mass production lines, Geely has invested through leading capital rounds, and Leapmotor has also ventured into this area.
On the international front, Tesla's third-generation Optimus started mass production in the second quarter of 2026, with Elon Musk confirming it will enter volume production at the Fremont factory in California in late July or August. The dedicated Optimus factory at the Texas Gigafactory has a final planned annual capacity of 1 million units. Hyundai Motor is advancing through its subsidiary Boston Dynamics, with its Atlas humanoid robot planned for deployment in Hyundai factories starting in 2028. BMW has chosen to purchase directly, deploying the Aeon humanoid robot, developed by Hexagon Robotics, at its Leipzig factory in Germany starting this summer.
Technological Convergence and Endogenous Scenarios: The Underlying Logic for Automakers
The rush of automakers to "make robots" is not simply cross-sector diversification; it follows a clear logic of industrial synergy.
From a technological standpoint, the overlap between smart vehicles and humanoid robots exceeds 70%. Core hardware for humanoid robots, such as joint actuation, energy supply, and perception control, is highly interchangeable with key components of smart vehicles. Perception hardware like lidar, cameras, and millimeter-wave radar can be almost seamlessly transferred. Algorithms for environment perception, path planning, and real-time decision-making are also highly shared. As He Xiaopeng noted, "70% of an automaker's technology reserves can be directly reused on robots." On the supply chain front, motors, gear systems, screws, bearings, batteries, and sensors are all critical components shared by electric vehicles and robots, differing only in precision requirements.
From a scenario perspective, car factories themselves are the most ideal initial deployment environments for humanoid robots—production lines are designed for humans, so robots can be directly adapted to existing systems without major modifications. Tasks like handling, assembly, quality inspection, and inspection are prime areas for early replacement by humanoid robots. Commercial scenarios like store sales and demonstrations also provide natural testing grounds. Furthermore, the high-quality training data generated from these internal scenarios can drive rapid product iteration, creating a technological feedback loop.
The commercial motivation is also clear. In 2025, the profit margin for China's automotive industry fell to 4.1%, the lowest since 2015. Squeezed by both stock competition and price wars, humanoid robots are seen as a vital direction for establishing a second growth curve. The humanoid robot industry, predicted by numerous investment banks to potentially create a market value of $10 trillion in the future, has become a new option for automakers. Morgan Stanley projects the humanoid robot market could reach $5 trillion by 2050.
Promising Future, Yet Challenges Remain
2026 is widely regarded by the industry as a critical validation period for humanoid robots moving "from 0 to 1." Global shipments of humanoid robots are expected to exceed 50,000 units this year and surpass 100,000 units in 2027.
However, challenges are equally significant. The requirements for actuator precision and dexterous hand tactile sensors in robots far exceed automotive-grade standards. The initial cost per unit is high, and there is still a considerable distance from large-scale commercialization. As noted in reports, the path from current "skill-showcasing" based on preset programs to achieving generality and widespread household adoption remains long. It requires not only sustained massive capital and talent investment but also breakthroughs in new technologies like solid-state batteries and flexible sensors, as well as the next "ChatGPT-level leap" in physical AI.
Market opportunities also require sober assessment. Robotic arms can already handle almost all large-scale production scenarios in manufacturing. The remaining tasks that cannot be replaced by automation are either extremely difficult to substitute or involve a limited number of replaceable workers. For automakers seeking to carve out new growth curves in the short term, they must find a balance between cost, performance, and substitution efficiency, while exploring sustainable, high-demand scenarios beyond their own factory floors.
Nonetheless, the automakers' advantage lies precisely in the fact that they do not need humanoid robots to be "perfect" before deploying them. In highly structured factory environments, robots capable of performing specific tasks already possess practical value. From this perspective, this industrial extension from "four wheels" to "two legs" is likely more than just a fleeting marketing spectacle.
Comments