World Humanoid Robot Games 2025

World Humanoid Robot Games 2025

Beijing, August 17, 2025

The 100-meter sprint event at the 2025 World Humanoid Robot Games welcomed an intense “Man vs. Machine” showdown. Unitree Technology (宇树科技) and the Beijing Humanoid Robot Innovation Center (co-established by local authorities and the Intelligent Humanoid Robot Innovation Center) each fielded two teams to compete in the finals. Ultimately, TIAN GONG Ultra clinched the championship with a time of 21.50 seconds, while Unitree Technology’s Lingyu team and Gaofei Technology’s team secured second and third place, respectively.

Article content

The TIAN GONG team, adhering to a philosophy of discarding remote controllers, enabled their humanoid robots to rely solely on their own vision and lane-tracking technology used in autonomous driving. This innovation allowed for fully autonomous sprinting. Under competition rules, TIAN GONG’s recorded time benefited from a 0.8 coefficient reduction. Although Unitree’s robots demonstrated superior raw speed in the 100-meter sprint, their reliance on remote operation meant that their total score placed them behind TIAN GONG’s autonomously sprinting robots, resulting in a runner-up finish.

Article content

After the race, Wang Xingxing of Unitree Technology commented that autonomous navigation is technically straightforward. However, because Unitree is focused on pushing the limits of robotic sprinting performance, their current entries relied on remote navigation. For future competitions, Unitree will certainly include robots equipped with autonomous navigation capabilities.

On the path toward the evolution of embodied intelligence, “abandoning remote controllers” is far from a trivial step; it signifies a leap toward robots possessing their own “brains.” The essence of remote operation lies in allowing robots to temporarily rely on human cognition. Operators substitute for robots in tasks such as environmental perception, path planning, and assisted dynamic balance control. In contrast, under fully autonomous mode, robots must independently complete the cycle of perception, decision-making, and execution, enabling them to autonomously respond to unexpected scenarios.


Article content


To view or add a comment, sign in

Explore content categories