Unitree Robotics, China’s largest humanoid robot manufacturer, filed for an initial public offering on the Shanghai Stock Exchange on March 20. It seeks to raise 4.2 billion yuan ($610 million). The Hangzhou-based company shipped 5,500 humanoid robots in 2025 alone, the most globally. It also posted an adjusted net profit of 600 million yuan on revenue of 1.71 billion yuan, a 335% surge from the previous year. If approved, the listing would be one of China’s biggest onshore tech IPOs in years, and a bellwether for the broader humanoid robotics sector. The numbers underscore a remarkable run. China now hosts over 100 humanoid robot companies, with Unitree and AgiBot together projected to account for roughly 80% of domestic shipments in 2026. Unitree’s robots have been performing martial arts sequences at the Spring Festival Gala for the third consecutive year, executing mid-air backflips and parkour routines that demonstrated clear mechanical advances over prior iterations. Its G1 humanoid recorded a running speed above 5 metres per second, a verified global best. EV makers BYD and Geely have deployed Unitree’s machines on production lines.
The industry has also managed to driven down the costs. In 2023, the average price of a Unitree humanoid was ~$85,000. By 2025, it had fallen to ~$25,000, a 72% drop over two years. Yet the company’s gross margin improved to nearly 60%, driven by vertical integration. While Elon Musk plans Optimus to eventually cost around $20,000, Unitree’s G1 has already brought it down to $16,000. But the cost story masks a deeper challenge: the use case. A breakdown of Unitree’s sales reinforces this challenge. In the first nine months of 2025, 73.6% of humanoid revenue came from research educational institutions, universities and labs purchasing platforms for experimentation. Another 17.4% came from commercial consumption uses, such as demonstrations, display environments and marketing spectacles. And only 9.01% came from industrial applications. Thus, over 70% of all humanoids sold were for research and educational purposes. The scientific research and education market is real, but scalability will remain a challenge. The industrial market, on the other hand, has the potential to offer scale. Unitree’s profit, as one Chinese tech publication noted, is less a victory for commercialisation than a victory for the dividends of the scientific research market - a niche where few competitors exist, budgets are sufficient, and products carry obvious differentiation. The calculus will change only when robots begin competing on productivity inside factories and warehouses. There’s a definite leap that the Chinese humanoid robot industry has taken. The challenge now is to find use cases to validate that leap.