At the 2026 World Humanoid Robot Games in Beijing, a robot called Tiangong Ultra ran the 100-meter dash in 8.64 seconds. Usain Bolt’s world record is 9.58 seconds. A machine was able to beat the fastest human who ever lived by almost a full second.
Other robots broke jumping records. One played table tennis against an Olympian. Then the organizers handed them hammers and asked them to drive nails into a corkboard. The results were negative, and even the robots being puppeted by humans wearing motion-capture gloves could not do it.
Sprinting is the same motion repeatedly. Nothing changes halfway through. The track does not move. Engineers can run that motion in simulation thousands of times until it is perfect and electric motors are just stronger than muscle.
Bolt reached his physical limit because he is a human. Robots have not reached theirs because their limitations are battery output and material strength, both of which improve every year.
Hammering is a completely different kind of problem. Every swing changes something. The nail shifts, the board gives a little, the angle drifts and a hand corrects all of that before the person consciously notices. The average human hand has 27 bones and thousands of touch receptors constantly sending updates. Pack all of those aspects into the fist with motors, wiring and sensors inside the robot and make the software read all of it fast enough to react mid-swing. Touch sensors also wear down over time, which is a problem when the job requires hitting something repeatedly.
Researchers have known about this for decades. Tasks that are hard for humans are often easy to automate. Tasks that a toddler can do without thinking can be nearly impossible for robots.
The gloves did not fix the problem either, because knowing where a hand is positioned is not the same as feeling what it is touching. The operators were simply working with their hands asleep.
The scoring told the real story. Organizers gave more points for autonomous runs than remote-controlled ones, and that choice says exactly what the industry wants. A robot that needs a trained person in a mocap suit is not saving anyone labor. It is an extremely expensive puppet and it still requires a human for each machine. The whole point is to eliminate the operator.
Humanoid robots will likely show up in everyday life in stages. Warehouses and factories will probably come first, where the same objects, lighting and motion all shift. Houses and job sites are messy and unpredictable, and that is exactly where good hands matter most.
