BYD to Unveil in August, XPeng IRON Enters Trial Production — Automakers Accelerate Humanoid Robot Push
0xBroomberg
BYD confirmed it will debut a humanoid robot prototype in August; XPeng's IRON is already in trial production in Guangzhou with mass output targeted for year-end. Three Chinese automakers are betting on embodied intelligence — backed by a 70%+ technology overlap between smart cars and humanoid robots.
What is BYD about to reveal?
BYD confirmed to China Securities Journal on July 28 that it will officially unveil a humanoid robot at its "D-Space" venue in August — the company's first public showing of a physical prototype.
Earlier online rumors — a codename "Yao-Shun-Yu" and a plan to deploy 20,000 units internally this year — have been denied, but BYD confirmed the in-house robotics program itself.
Executive VP Li Ke said the competitive edge lies in manufacturing, software, and hardware. BYD positions itself as an open platform — developing in-house while collaborating with outside partners.
How far along is XPeng's IRON?
IRON has entered small-batch trial production at XPeng's Guangzhou plant. The mass-production line is in final commissioning, targeting volume output by end of 2026 at 1,000+ units per month.
Rollout timeline: late 2026 in retail stores → Q1 2027 nationwide → Q2 2027 overseas → 2028 into households.
IRON stands 178 cm tall, runs on a Turing AI chip (2,250 TOPS total) and a proprietary physics-world foundation model. It already handles sorting, carrying, and quality inspection on the factory floor.
This means → He Xiaopeng personally took on the CEO role for the robotics unit, elevating humanoid robots to XPeng's top strategic priority — this is not a side experiment, it is a core bet.
What is Li Auto doing?
Li Auto is developing two form factors: a wheeled robot for factory use, set for public debut this year, and a bipedal robot codenamed "Nexus."
The company restructured its intelligent-systems division in February 2026 into three teams: humanoid robotics, software, and foundation models.
Full-year 2026 R&D spending is projected at RMB 12 billion, with AI-related work accounting for roughly 50%.
In plain terms = Li Auto treats autonomous driving as the first half of embodied intelligence and general-purpose humanoid robots as the second half — it is playing both halves at the same time.
Why can automakers build humanoid robots?
Smart vehicles and humanoid robots share over 70% of their core technology — joint actuators, power systems, and perception-control hardware are largely interchangeable.
Lidar, cameras, and millimeter-wave radar — the sensor suite on a car — can transfer to a robot almost seamlessly. Path-planning and environment-perception algorithms are equally shared.
On the supply-chain side, motors, gears, lead screws, bearings, batteries, and sensors are key components common to both EVs and robots.
In plain terms = cars and humanoid robots run on the same "muscles and nerves." Automakers are not crossing into a foreign industry — they are growing down their existing technology tree.
Why are automakers' own factories the ideal proving ground?
Automotive production lines were designed for humans. Robots can slot directly into the existing setup — carrying, assembling, inspecting, and patrolling are all tasks ripe for early substitution.
This means → the training data generated on-site feeds rapid product iteration. Automakers are both the builders and the first customers of their own robots.
The commercial incentive is clear: China's auto-industry profit margin fell to 4.1% in 2025, the lowest since 2015. Humanoid robots are seen as a path to a second growth curve.
Where does the broader industry stand, and what is still missing?
Beyond BYD, XPeng, and Li Auto, GAC, Changan, Chery, SAIC, Geely, and Leapmotor have also entered the field. Tesla's third-generation Optimus began mass production in Q2 2026; Musk confirmed output at the Fremont plant by late July or August, with the Texas Gigafactory's dedicated Optimus line ultimately targeting 1 million units per year.
The industry views 2026 as the critical "zero-to-one" validation period. Global shipments are expected to pass 50,000 units this year and exceed 100,000 in 2027. Morgan Stanley projects the market could reach $5 trillion by 2050.
Challenges remain: actuator precision and dexterous-hand tactile sensors — components in robot fingers that detect force and texture — must meet standards far above automotive grade. Early per-unit costs are high, and large-scale commercialization is still some distance away.
This reflects a pivotal test ahead: whether BYD's August unveiling can demonstrate a prototype with real practical value will be the first public validation checkpoint in this wave of automakers building humanoid robots.
Content is for reference only, not financial advice.