In Beijing a humanoid robot runs the 100 metres in 9.32 seconds, according to CCTV
The measurement comes from a test session linked to the World Humanoid Robot Games, which began on August 22. The figure originates from the Chinese state broadcaster and has no independent verification; the conditions are not comparable to an athletics race.
A new story compared to this morning’s edition, and one that belongs to an area the newspaper follows for technical rather than sporting reasons: the dynamic control of bipedal locomotion.
The humanoid robot Lightning, developed by Chinese phone manufacturer Honor, covered 100 metres in 9.32 seconds, reaching a top speed of 14.5 metres per second. The measurement comes from a demonstration session linked to the second edition of the World Humanoid Robot Games, which began in Beijing on August 22, 2026 (ANSA).
The origin of the figure
A clarification is needed here, because this is what distinguishes verified news from reported news. The measurement comes from the Chinese state broadcaster CCTV: this is how it is attributed by Swissinfo, Open and ANSA itself, while Il Sole 24 Ore picks it up from the Reuters agency. International coverage therefore traces back to the same communication: the timing does not appear to have been verified by a third party, and should be treated as a figure declared by the event’s organisation and disseminated by state television.
Why the comparison with athletics does not hold
News reports pair the time with the human world record in the 100 metres, 9.58 seconds set by Usain Bolt at the 2009 World Championships in Berlin. The comparison is suggestive and technically improper, for three reasons worth stating.
The first is procedural: an athletics record exists because there is a set of rules establishing its conditions — starting blocks, electronic timing, wind measurement, ratification. The available material indicates nothing similar for the Beijing trial, which is described as a test event.
The second is mechanical. A humanoid reaching 14.5 metres per second does so with an actuation system, a balance and a ground-impact management entirely different from those of the human body: the comparable variable is not the time, but stability of gait at high frequencies.
The third is stated by the manufacturers themselves. For the Games, researchers lengthened the robot’s legs by 10 centimetres, bringing them to 1.05 metres. This is a modification that increases stride length — a parameter that can be adjusted at will in a machine, but not in an athlete — and it alone explains why the comparison belongs to communication rather than measurement.
What the figure actually says
The interesting technical information is the progression. In April 2026 the same robot had won the Beijing half marathon, 21 kilometres in 50 minutes and 26 seconds: an endurance task, dominated by energy consumption and heat dissipation. Moving in four months from an endurance exercise to a maximum-speed trial means working on opposite constraints: in the first case, efficiency per kilometre matters; in the second, instantaneous power and the ability not to lose balance.
On one point the available data offer no answers: nothing indicates whether the trial was repeatable, how many runs were carried out, or what the fall rate was. These are the pieces of information that distinguish a demonstration from an established engineering result, and their absence is part of the story.
The second edition of the World Humanoid Robot Games began in Beijing on August 22: it will be the schedule of official trials, with published rankings and rules, that provides figures comparable across different machines.
Sources: ANSA, Swissinfo, Il Sole 24 Ore, Open. Primary source of the measurement: CCTV (Chinese state broadcaster).
← Archive · Front page · Past editorials · Report an error · Original article (in Italian)