Key Takeaways:
XPeng has completed nomination agreements with suppliers of robot joint modules, actuators and precision structural parts and is now scheduling production lines, the clearest signal yet that the Guangzhou automaker intends to build humanoid robots at manufacturing scale rather than in a lab. The company disclosed the status at its first robot supply-chain partner conference, where component makers from those three categories gathered in one room for the first time.
"Nomination is the point where a program stops being a prototype and starts consuming capital," said a person familiar with the program, who asked not to be identified because the agreements are private. "Scheduling lines means tooling orders are being placed."
XPeng did not disclose production volumes, unit pricing, delivery dates or the identity of the nominated suppliers. Those figures are not yet disclosed, and the absence of them caps the near-term earnings impact: a nomination agreement commits a supplier to capacity, but revenue only appears when parts ship against a purchase order.
The program's public reference point is XPeng's Iron robot, which the company has said runs the same autonomous-driving perception stack and in-cabin AI as its vehicles, giving it 720-degree visual coverage and voice interaction. That reuse is the core of the investment case. XPeng already pays for LiDAR, cameras, inertial measurement units, neural processing units and high-torque electric motors at automotive volumes, and a humanoid platform amortizes that engineering spend across a second product line.
The scale of the prize is what makes the timeline worth tracking. China's humanoid robotics market reached nearly 2.8 billion yuan ($380 million) in 2024 and is projected to grow into a 100 billion yuan market by 2030, according to the China Academy of Information and Communications Technology. Total financing in the country's embodied-intelligence sector exceeded 10 billion yuan ($1.4 billion) in the first half of 2025 alone, CCTV reported.
XPeng is not alone in making the jump. BYD self-produces about 80 percent of core robot components including harmonic reducers and torque sensors, a vertical integration that has cut costs 30 to 40 percent, and plans to deploy 2,000 robots on its own production lines this year. GAC's GoMate robot uses a force-control system adapted from its EV drive platform, which the company says improved single-leg balance response time by 50 percent and pulled force-control development nine months ahead of the industry average. NIO and Geely have opened factories to Unitree and other robotics developers as testbeds.
The difference for XPeng is that it is buying rather than building the joint modules and actuators, which puts the near-term revenue in the hands of its suppliers and leaves XPeng's own margin story dependent on software and systems integration. That is a faster path to a production line and a thinner one to gross profit per unit.
For investors, the trade splits in two directions. XPEV carries the optionality: the stock is valued primarily on vehicle deliveries and gross margin, and a robot program that reaches volume would add a second multiple to the story without a second factory. The parts makers carry the nearer-term earnings leverage, because a nomination converts into tooling and pre-production orders before it converts into robot revenue. Both legs of that trade depend on the same undisclosed variable — how many units XPeng intends to build, and when.
The next checkpoint is the production schedule itself. XPeng has not given a date for first line output, and until it does, the supplier conference remains a signal about intent rather than a number about volume.
This article is for informational purposes only and does not constitute investment advice.
