Summary of Key Points
Xiaopeng's humanoid robot business, Pengxing, has raised $900 million in a single round of financing (600 million from external investors, 200 million from Xiaopeng itself, and 100 million from senior executives), pushing its valuation to 43 billion yuan, setting a new record for a single-round financing in the embodied intelligence industry. Investors such as IDG, Tencent, and Alibaba are betting on Xiaopeng's autonomous driving technology being applicable to robots, its full-stack self-development capabilities, and the potential for a "car manufacturer to cross into the robotics sector." However, behind these impressive figures lie several concerns: the team has undergone three reorganizations in three years, with key personnel leaving one after another; the company's technical capabilities (hardware, algorithms, production volume) fall short of its valuation; the synergy between autonomous driving and robotics is uncertain; and more importantly, Xiaopeng's main automotive business is facing a tough period, urgently needing to rely on robotics to create a new narrative to sustain its growth.
I. Why Are Investors Willing to Pay 43 Billion? They Bet on "Reusability" and "Imagination"
The investment is not without reason; investors are focusing on three main aspects:
1. Autonomous Driving Technology as a Shortcut: Xiaopeng has been in the autonomous driving field for 12 years and has developed its own chips (Turing), physical AI models, and data infrastructure. Investors believe these technologies can be directly applied to robots, allowing for environment perception and real-time decision-making without starting from scratch.
2. Full-Stack Self-Development + Supply Chain Advantages: IDG notes that Xiaopeng handles everything from chips to hardware and can leverage the scale of the automotive supply chain, potentially enabling faster mass production.
3. Support from Tencent and Alibaba: These companies have previously invested in Zhiyuan and Yushu and are now backing Xiaopeng, indicating they see it as the most promising candidate among car manufacturers entering the robotics sector. Both automotive and robotics involve "physical AI," and Xiaopeng has a solid foundation in this area.
In short, investors are betting that Xiaopeng can transform its autonomous driving strengths into new advantages for robotics.
II. The Turbulent Team: Core Personnel Leaving One After Another
Xiaopeng's robotics team has undergone three major reorganizations since 2020, with many key figures departing:
- First Time: In 2023, founder Zhao Tongyang (who later left to start his own robotics company) left due to differing visions—Zhao favored rapid experimentation, while He Xiaopeng focused on mass production through the automotive supply chain. Zhao took with him a group of key members.
- Second Time: In May 2026, a robot demonstration resulted in an accident, leading He Xiaopeng to restructure the team. In June, product director Shi Xiaoxin resigned after more than four years, and in July, business leader Mi Liangchuan (with 15 years of experience at NVIDIA) also left. Even the AI architect was poached by OpenAI.
- Third Time: He Xiaopeng took on the role of CEO for robotics personally and established nine new departments, but it's uncertain whether the new team can bridge the technical gaps left by previous departures.
The consequences of these team disruptions are clear: inconsistent technical approaches, talent loss, and a disruption in progress. With money coming in but key personnel gone, this represents a significant risk.
III. Is the 43-Billion Yuan Valuation Matched by the Actual Capabilities? It May Be "Overinflated"
Xiaopeng's robotics capabilities are still far behind leading players:
1. Lack of Hardware Advantages: Although IRON claims to have 76 degrees of freedom and self-developed chips, its competitors like Yushu have 95% of their core hardware self-developed, and their motor costs are half lower than imported ones. Zhiyuan has already produced 15,000 units, while Xiaopeng is still in the small-scale trial production phase, focusing on tasks that traditional robots can perform.
2. Meagre Production Volume: Zhiyuan aims to produce over 5,000 units in 2025, and Yushu's bipedal robots have sold 5,500 units; Xiaopeng doesn't even disclose specific production figures.
3. Outdated Algorithm Approach: Xiaopeng is betting on the VLA architecture (visual + language + action), but the industry has shifted to a "brain + cerebellum" layered architecture (higher-level decision-making + lower-level control) in 2026. VLA struggles with complex tasks and lacks sufficient data (only 5% of the required million hours of real-world interaction data has been collected).
In summary, the valuation may be at the forefront, but the actual capabilities might be at a secondary or even tertiary level, a typical case of overpromise.
IV. Can Autonomous Driving and Robotics Really Be Integrated? It's Challenging!
He Xiaopeng often talks about the integration of autonomous driving and robotics, but the reality is more complex:
1. Incompatible Data: Autonomous driving data comes from road scenarios (driving, traffic lights), while robots need data from factory or home environments (screwing, moving objects). It's like asking a ten-year-old driver to learn factory tasks from scratch—the data is fundamentally different.
2. Contradictory Methodologies: Autonomous driving emphasizes safety, leading to conservative algorithms, while robotics requires trial and error and a higher tolerance for errors. These two cultures are difficult to reconcile within one company.
3. Insufficient Resources: Xiaopeng's main automotive business is struggling, with a 13.4% loss in the second quarter and a 12.8% decline in sales, achieving only 37% of its annual targets. It also has to develop robotics, flying cars, and Robotaxi, leaving it short of both funds and personnel.
Therefore, the integration of autonomous driving and robotics remains uncertain.
V. The "New Narrative" Forced by a Challenging Main Business
Xiaopeng is pushing robotics so urgently because its automotive business has lost investors' interest:
- The new energy vehicle market is highly competitive, with price wars and thin profits (gross margin dropped from 14.3% to 12.1% in the second quarter).
- Failing to meet sales targets means the stock price cannot rise.
Robotics is Xiaopeng's last hope to attract investors again by presenting a story about the future of "physical AI." He Xiaopeng is personally in charge of robotics not because he wants to, but because there's no other choice; its main business is failing, and a new narrative is essential to survive.
Final Summary
Xiaopeng's 43-billion yuan valuation for its robotics business is a bold gamble: betting on technology reuse, team reconstruction, and the ability to turn its main business into a profitable robotics venture. However, the chances of success are uncertain due to team instability, lagging capabilities, and the burden of its struggling main business. Investors are investing in the potential for the future, not the current reality. Only time will tell whether this gamble pays off.