Humanoid Robot Completes Half-Marathon in Beijing, Setting Record for All-Bot Race
A humanoid robot developed by Honor completed a half-marathon in Beijing this month, setting a record for an all-robot race. The event showcased advances in China's physical AI sector and marked a shift from lab demonstrations to public performance benchmarks.

A humanoid robot built by Chinese technology company Honor completed a half-marathon in Beijing this month, setting a record in what organizers called the first all-robot race of its kind.
The robot finished the 13.1-mile course, demonstrating endurance, navigation, and the ability to handle varied terrain over an extended period. The event drew attention from robotics researchers and technology investors who have been tracking China's rapid progress in physical AI.
Physical AI refers to AI systems that operate in the real world through robotic bodies, as opposed to software-only applications. China has invested heavily in the sector, with government support for robotics research and a growing number of private companies developing humanoid robots for industrial and consumer use.
The Beijing half-marathon follows a series of public demonstrations by Chinese robotics companies over the past year. Analysts say the events are designed to show that Chinese robots can perform reliably outside controlled laboratory settings.
Samsung also revealed a new AI robot concept this month called Project Luna, a round-screen assistant with a swiveling head. The device is designed as a home companion, capable of answering questions, playing media, and responding to voice commands. Samsung said it is exploring how AI assistants can move beyond phones and laptops into ambient home devices.
Coco Robotics announced a partnership with BlindSquare this month to use delivery robots as mobile sensing platforms. The robots will collect data about sidewalk conditions and provide real-time spoken alerts to blind users about hazards like construction, uneven pavement, and blocked paths.
Google's Gemini Robotics-ER 1.6 model, also released this month, allows robots to reason about their physical surroundings with greater precision, including reading gauges and interpreting visual information in industrial settings.


