Xpeng bets on humanoid robot Iron as part of expanding AI strategy

Xpeng places its humanoid robot Iron at the center of a broader strategy to move beyond electric vehicles, identifying humanoids, self-driving cars and flying cars as the three pillars of the next stage of development. In a podcast published on May 6, The Economist reported that founder He Xiaopeng has restructured the company as Xpeng Group and is ramping up investments in what he calls “physical AI”—a shift significant because it ties humanoid development directly to an established automaker rather than a standalone robotics startup.

Humanoid Iron became the public face of Xpeng

According to The Economist, at Xpeng's headquarters in Guangzhou, it wasn't electric cars or advanced flying cars that were getting the most attention, but humanoid robots. Iron, the latest model, was described as unusually human-like, with synthetic skin stretched over latticework “muscles” that felt soft to the touch. Videos showing the robot on the catwalk caused so much discussion online that some viewers doubted whether it was a person in a suit.

Along with Iron, the company showed earlier robot designs, including small four-legged machines called dogs and horses. Xpeng employees told the publication that they initially considered a four-legged architecture to be more stable, but in 2023, after the release of ChatGPT‑4, they switched to bipedal humanoids, believing that AI systems had reached a level that allowed them to operate on two legs.

This product evolution is now woven into a broader company overhaul. He said Xpeng has been focused on creating electric vehicles for the past ten years, and from 2025, the company sees its mainstay in three areas: cars, humanoid robots and flying cars. The renaming of Xpeng Motors to Xpeng Group is intended to reflect this expanded scope.

Why Xpeng connects cars and humanoids

He's argument in the interview is that Xpeng's auto business already contains much of the tech stack needed for robotics. He said he has long viewed cars as a type of robot, and Xpeng has spent years developing software-based cars, self-driving systems, AI chips and other research that could carry over into humanoid development.

That explains why the company is willing to continue spending in a market where many Chinese electric vehicle makers are under pressure. The Economist notes that Xpeng only turned profitable for the first time in the fourth quarter of last year, but still plans to invest 7 billion yuan in AI this year. He argues that many rivals are cutting back on R&D for short-term survival, while Xpeng aims to protect its technology leadership and avoid falling into the cheapest price levels.

The humanoid robot initiative was not initially accepted easily within the company. He said that when he bought the robotics company three years ago, all outside members of the board of directors opposed it. Now, he says, those same directors are asking when the robotics division might be spun off to attract outside capital, and he claims that Xpeng's robots are two to three generations ahead of most Chinese companies in the field - a claim given in an interview without independent verification.

Immediate jobs and long-term vision for care

He clearly shares long-term ambitions and immediate applications. While he says the ultimate goal is for a robot to be able to do household chores—laundry, cooking, and meal prep—he believes the first "physical AI" will be to replace simple, repetitive roles that people don't like. In his opinion, cleaning the house and caring for people still remains much more difficult tasks.

Part of Xpeng's thinking is shaped not only by engineers, but also by demographic realities. He said concerns about aging and caring for relatives — exacerbated by his father's illness — largely spurred his interest in humanoid subjects. This view also influences the design: it argues that a fully humanoid form with skin and "muscle" is more suitable for close interaction with humans than a machine that feels like steel armor.

Xpeng's bet coincides with China's overall greater support for robotics development. The podcast noted that "artificial robots" are mentioned in the country's latest five-year plan, that local governments are setting up development centers, and that the industry includes hundreds of companies. At the same time, analysts cited in the discussion warned of a possible bubble: so far, the largest buyer has been the state, and the commercial attractiveness of humanoids remains not fully proven.

What does Xpeng's bet on robots signal to the market?

The company is trying to differentiate itself at a time when advanced driver assistance systems (ADAS) are losing their uniqueness in the electric vehicle market. Xpeng has managed to create a reputation for strong ADAS - in an interview it was mentioned that Volkswagen bought a 5% stake in the company precisely because of its technology. The move into robotics is both an expansion plan and a response to increasing competition in the automotive industry.

According to the interviewee, awareness and acceptance of robots should increase after 2028, and he predicts that the robot market could eventually surpass the automotive market by several times. Morgan Stanley, quoted in the podcast, estimates the market for humanoid robots at a possible $5 trillion by 2050, although the discussion also emphasized that such projections are far from a proven need.

Xpeng's strategy reflects a broader trend in China: electric vehicle makers are moving into robotics and robotics companies are interested in cars. The next test will be the ability of companies like Xpeng to turn humanoid prototypes into repeatable work and sustainable revenue before the sector is again hit by the price wars and overcapacity that have already reshaped the electric vehicle industry.