A visit to humanoid robot marathon contender L7

Añadido: 2026-05-05

[00:00:01]This is L7, a fierce competitor in the 2026 Beijing E-Town Half Marathon for humanoid robots. Before the race, L7 is putting in the hard miles, training at full intensity, and pushing for peak performance.

[00:00:18]>> [music] >> We made several major upgrades for this competition. First, the course is more than 20 km, so battery changes are critical. We built a hot-swap battery system, which means L7 can switch batteries without powering down, saving precious time. Second, we improved cooling in the power modules for long-distance, high-speed running. Our motors can generate up to 400 N-m of torque, so keeping them cool is key to staying fast and stable. We also upgraded autonomous navigation and obstacle avoidance, adding LiDAR sensors and a high-precision positioning system, so L7 can run the course on its own.

[00:01:05]Lap after lap, L7 keeps going while engineers track every movement, turn, and data point.

[00:01:13]They're testing endurance, balance, >> [music] >> navigation, and speed. Each run helps refine algorithms and push performance a little further.

[00:01:22]After a few laps around the stadium, its arms and legs starting to warm up a little bit, but not too much. And it can still stand upright without any trouble.

[00:01:31]This is now the second edition of the robot half marathon, and I'm stunned by the progress from last year.

[00:01:38]Back then, many robots needed technicians running behind them with remote controls. Some could barely keep a steady pace. But this year, they're so quick that you may need an electric cart to keep up.

[00:01:52]And that leap forward comes from breakthroughs in both hardware [music] and AI.

[00:01:58]Robot development depends on progress across multiple areas: structural design, motor systems, and algorithms.

[00:02:06]Only when all three improve together can we achieve the kind of progress seen this year. For a marathon, robots need stronger bodies, more durable motors, and smarter algorithms to manage speed [music] and control over long distances.

[00:02:19]That combination has driven rapid progress in recent years.

[00:02:23]As these robotic runners continue running the track, the final results may matter [music] less than what the competition represents.

[00:02:31]Because this is more than a competition, but a live stress test of speed, [music] stability, endurance, and control. And what is learned here could soon be used in factories, logistics centers, and daily life. Every step [music] those robots take on the course could mark a giant leap forward for embodied intelligence. Chen Shaode, CGTN, Beijing.