Yushu Technology is making money, while some new forces are still operating at a loss.

On March 20, the humanoid robotics company Unitree Technology disclosed its IPO prospectus for the STAR Market. On the same day, XPeng Motors released its 2025 financial results. Although the two sets of performance records appear to belong to different tracks, they form a striking contrast due to a high degree of homology at the underlying technology level—smart cars are viewed as “wheeled robots,” while humanoid robots are “intelligent cars that walk on two legs.” Both rely on shared three core systems: perception, decision-making, and execution.

Data show that Unitree Technology expects to generate a non-GAAP net profit of over RMB 600 million in 2025. On the same day, XPeng Motors released its financial report showing that while its profit turned positive in the fourth quarter of last year, its full-year net loss still reached RMB 1.14 billion. NIO’s previously released financial report also indicates that its full-year net loss was RMB 14.943 billion.

In terms of gross margin, XPeng Motors and NIO posted full-year gross margins of 18.9% and 13.6%, respectively, while their automotive gross margins were 12.8% and 14.6%, respectively. In the same period, Unitree Technology’s gross margin reached 60.27%.

Although the underlying technology is highly homogeneous, there is a huge disparity in profitability. One of the reasons lies in the different stages of investment. XPeng Motors’ R&D expenditures in 2025 were RMB 9.49 billion, accounting for 12.4% of revenue; NIO’s R&D expenditures were RMB 10.605 billion, accounting for 18.9% of revenue. Unitree Technology’s R&D expenses for the first nine months of 2025 were RMB 90.21 million. After reaching a peak of 31.39% of revenue in 2023, the figure fell to 17.84% in 2024, and further declined to 7.73% in January–September 2025.

This divergence between investment and profitability reflects two distinctly different development stages across the two tracks. Cars are a “heavy-asset long-distance race,” requiring investments on the scale of tens of billions of RMB to secure future market share. Robots are a “light-asset positioning play.” With Unitree not yet facing a large-scale arms race, it has entered the stage of technology monetization first.

Despite the difference in development stages, as embodied intelligence rises and leading companies pivot, the two tracks are accelerating toward convergence.

In January 2026, Tesla CEO Elon Musk announced that the company would retrofit its California factory into an Optimus humanoid robot production line, and expects that the future value of the robot business will exceed that of the automotive business. The related market valuation could reach $2.5 trillion by 2050.

In China, new energy vehicle companies’ attitudes toward robot businesses are showing polarization.

In its financial report, XPeng Motors explicitly mentioned its robot business layout, and founder He Xiaopeng proposed a goal of “scaling and mass-producing high-level humanoid robots.”

Li Xiang, chairman of Ideal Automobile, said at an internal all-employee meeting in January this year that the ultimate form of cars is robots, and Ideal wants to make cars the best carrier for embodied intelligence.

Li Bin, founder of NIO, believes that the capability models of intelligent electric vehicles and robots are highly consistent at the levels of technology, supply chain, and manufacturing. NIO’s current strategic focus remains on its core automotive business, and it plans to enter the robot market when it has crossed the threshold for large-scale commercialization.

Zhu Jiangming, chairman of Leapmotor, said that Leapmotor’s application of technologies such as embodied intelligence and industrial robots prioritizes practicality and benefit orientation, and will not blindly follow the trend.

(This article comes from Yicai)