Humanoid robot half marathon postponed due to weather

Humanoid robots take the stage during the opening ceremony of the 2025 Zhongguancun Forum (ZGC Forum) Annual Conference, held in Beijing, China, on March 27, 2025. (Photo: Xinhua)

Severe weather has forced the postponement of Beijing’s much-anticipated Yizhuang Half Marathon, which was supposed to make history by showcasing humanoid robots running alongside human athletes.

Authorities issued an orange gale warning, predicting strong gusts and a precipitous drop in temperature, which caused the race, which was originally scheduled for April 13, to be delayed to April 19.

Given the robots’ vulnerability to wind-related balancing problems, organizers were worried that the harsh weather may jeopardize both human safety and robot performance.

This unique event marks the world’s first official half marathon to include humanoid robots as participants.

The 21.0975-kilometer race will unfold in Yizhuang, an innovation hub in southeast Beijing that hosts over 100 robotics firms and is backed by a 10 billion yuan (approx. $1.36 billion) government fund aimed at accelerating China’s dominance in artificial intelligence and robotics.

With over 12,000 human runners and more than 20 robotics companies expected to compete, the marathon serves as both a tech showcase and a sporting spectacle.

Leading the charge is the humanoid robot Tien Kung, developed by the National and Local Co-built Embodied AI Robotics Innovation Center.

A dedicated support team and referees will shadow the robots to quickly intervene in case of technical difficulties, ensuring smooth operations throughout the event.

Participating robots must meet specific criteria: they must resemble humans in both form and movement, finish the race within 3.5 hours, and maintain autonomy throughout—no outside control, battery swaps allowed only at designated points.

Ultimately, the event is more than a tech gimmick—it’s a real-world benchmark of robotic capability and an ambitious demonstration of how artificial intelligence is transitioning from controlled labs to open environments.