The Second World Humanoid Robot Games began with an event held at the Beijing National Speed Skating Oval on the evening of Saturday, August 22, 2026. A total of 2,056 humanoid robots and more than 600 teams from 16 countries took part in the 5-day event, which showcased the limits of autonomous technologies in the physical world. The robots competed in athletics and combat sports, as well as on the tennis court, which requires complex coordination.
Autonomous Humanoid Robots Take to the Court for the First Time
Humanoid robots developed by Beijing-based robotics company Galbot staged history’s first human-robot tennis demonstration as part of the event. The autonomous system stood out in singles and mixed doubles matches with its racket control and footwork tracking. Within seconds, the robot tracked the incoming ball, calculated where it would land, activated its body and arm motors simultaneously, and hit forehand and backhand shots.
During the exhibition match, the robot lost its balance and fell backward while tracking the ball, then rolled over and stood up again on its own before resuming the game from where it had left off. Its ability to adapt to constantly changing dynamics in the physical environment demonstrated the success of reinforcement learning algorithms and advanced joint actuators.
A thousand 301 Matches Across 51 Different Disciplines
At the large-scale event, where 1,301 matches were held across a total of 51 different disciplines, performances were showcased in a wide range of areas, from athletics and table tennis to kickboxing, industrial manufacturing, and hotel and logistics scenarios. The event, previously announced as a competition featuring 2 thousand machines, also included disciplines such as table tennis and combat this year, both of which require rapid decision-making.
In track races, the Chinese humanoid robot TienKung Ultra ran 100 meters in 9.39 seconds, surpassing the human record of 9.58 seconds held by Usain Bolt.
The record of TienKung completing 100 meters in 9.39 seconds clearly demonstrated the engineering level the machines have reached in the areas of leg design and actuator torque balance.
Technical Challenges and Limitations in the Field
Although their autonomous mobility capabilities have advanced, critical obstacles faced by the humanoid robots throughout the event also drew attention:
Maintaining balance control during high-speed dynamic movements and the risk of sudden falls
Limited continuous operating time due to restricted battery capacity
The need for networks with higher bandwidth for data processing and real-time perception
The response time of computer vision systems in real-time object tracking
Nevertheless, the autonomous performance demonstrated in a dynamic sport such as tennis is regarded as an important milestone in artificial intelligence’s transition from digital simulation environments to complex tasks in the physical world.