Many foreigners are laughing: Why did the Chinese robot crash into a wall after finishing the 100-meter run? Because the sprint was fully autonomous, with no human intervention allowed. For humanoid robots today, running at high speed and stopping abruptly isn’t even the hardest challenge. When a robot runs at 8–10 meters per second, it’s already covering 8–10 meters every second. Assuming a camera operates at 30 frames per second, with each frame spaced about 33 milliseconds apart, the robot would have already moved forward 20–30 centimeters by the time the next frame is captured. Add to that visual recognition, algorithmic decision-making, controller commands, and the motor’s actual execution—by the time the robot “realizes it needs to brake,” it’s often already gone much farther. What’s even trickier is that humanoid robots can’t just slam on the brakes like a car. When running at high speed, the robot’s center of gravity is constantly leaning forward. The next step placement, leg swing, and torque of the hip and knee joints are all calculated based on continuing to run forward. Suddenly forcing it to stop is equivalent to instantly overturning the entire gait system: the front leg must catch the body’s inertia, the back leg needs to exert reverse force, and the upper body has to pull the center of gravity back within the support area. If any part of this process is even slightly delayed, the result isn’t just an ungraceful stop—it’s a full-on faceplant. So the simplest strategy for 100-meter robots right now is to maintain full speed before the finish line, focus all control capabilities on sustaining high-speed running, and then let a soft cushion absorb the remaining kinetic energy after crossing the line. In short, robots have already learned how to run fast, but they haven’t fully mastered how to safely stop at high speeds. That’s why it’s not that they don’t want to brake—it’s just that at this stage, forcing a stop is often more dangerous than crashing into a wall.