Watch: Moment robot attempts to beat Usain Bolt’s 100m record

A landmark moment for humanoid robotics development unfolded recently at China’s World Humanoid Robot Games, where a bipedal android came achingly close to beating Usain Bolt’s iconic 100-meter world record, according to Honor, the leading Chinese smartphone manufacturer that backed the project. The unprecedented attempt marks one of the most high-profile milestones in modern humanoid robot mobility, showcasing how far engineering teams have pushed the boundaries of robotic speed and balance over the past decade.

Footage captured at the competition shows the humanoid robot powering through the 100-meter straight track, with its synchronized leg movements and dynamic balance control allowing it to hit a top speed that came closer to Bolt’s long-standing record than any other humanoid robot in history. While official independent verification of the robot’s final time is still pending, Honor’s announcement has already sent ripples of excitement through the global robotics and tech communities, highlighting the rapid progress in artificial intelligence, actuator design, and locomotion control for humanoid systems.

The World Humanoid Robot Games, hosted annually in China, brings together leading research institutions and tech firms to test the latest capabilities of humanoid robots across a range of human-centric tasks, from running and jumping to obstacle navigation and complex manipulation. This 100-meter race challenge was designed specifically to push developers to prioritize speed and stability, two of the most persistent technical hurdles for bipedal robots that aim to match human mobility.

Industry analysts note that advances in humanoid running speed do not just serve as a publicity stunt—they have direct practical implications for the future deployment of humanoid robots in real-world scenarios, including emergency response, industrial logistics, and personal assistance, where the ability to navigate varied terrain quickly is a critical requirement.