Honor's humanoid robots break three world records

Honor's self-developed humanoid robots, "Lightning" and "Yuanqizai," delivered a stunning performance at the 2026 World Humanoid Robot Games in Beijing, breaking three human world records in fully autonomous racing events.

In the 100-meter final on Wednesday night, Lightning clocked an extraordinary 8.83 seconds. That also broke human world records in the 400-meter and 1,500-meter events.

Honor said what set the competition apart was its fully autonomous format — robots had to rely entirely on their own sensors and on-device computing power for perception, decision-making and execution, with no remote control intervention. This demands excellence in power output, motion control, system coordination and overall stability.

Lightning stood out not just for its speed but for its stability — it was the only robot that remained firmly standing after the sprint finish. High-speed deceleration triggers fluctuations in sensor data, requiring millisecond-level computation of balance angles and velocity, rapid motor mode switching for impact absorption, and real-time foot placement algorithms for dynamic center-of-gravity recovery, Honor added.

According to the company, Honor's robotics achievements build on its consumer electronics expertise. The company adapted its Qinghai Lake battery technology for robotic energy systems, developed proprietary integrated joint modules delivering 400 Nm of peak torque, and optimized its phone and PC thermal solutions for high-load robotic operation. The competition served as an extreme testing ground, with insights flowing back to improve consumer devices.

Honor plans to bring technologies validated through these extreme scenarios to its upcoming Magic9 series, scheduled for release on September 28.

Honor said humanoid robotics represents a key direction in Honor's Alpha strategy as it expands into the physical AI era. For Honor, robot competitions are vehicles for technological iteration, not endpoints. The company aims to continue advancing AI from the screen into the physical world, pushing the boundaries of AI-powered devices.