On December 25, 2025, Global Times reported that Beijing will host the second Humanoid Robot Half-Marathon on April 19, 2026, following the successful inaugural race in 2025 where humanoid machines competed alongside humans in a real half-marathon setting. This is more than a quirky race—it showcases rapid advances in robotics, autonomy, and embodied AI, hinting at a future where AI Employees and Non-Human Workers are tested in real-world environments.
In April 2025, the first global robot half-marathon took place in Beijing, drawing entries from cutting-edge humanoid systems like Tien Kung Ultra and models from Unitree Robotics and others. This year’s event will push capabilities further, featuring both autonomous and remote-controlled robots on the same course as human runners—though separated for safety. Participating teams compete not just for trophies but for R&D subsidies, incubation support, and industrial orders, bridging innovation with commercialization.
Officials organizing the 2026 race are raising the bar with stricter standards for autonomy, endurance, dynamic stability, and perception—core traits for real-world service robots. New awards like “Best Endurance” and “Best Gait” aim to accelerate breakthroughs that matter for practical applications: industrial inspection, logistics, and emergency response. Alongside the marathon, a robot development community and training camp will support teams with shared bodies, computing resources, and tools, reducing barriers for innovators.
This event reflects broader global trends in how robots and Voice AI Agents are transitioning from research labs to frontline roles. While robots racing might seem niche, the technology tested—autonomous navigation, decision-making in complex environments, extended operation time—is the same foundation for Non-Human Workers across industries. These AI-driven systems can complement human work, automate repetitive or dangerous tasks, and ultimately reshape work dynamics in manufacturing, logistics, and beyond.
Key Highlights:
Reference: