The world’s first half-marathon for humanoid robots in Beijing offered an entertaining yet sobering look at the current state of robotics.
While 21 ambitious bipedal machines lined up at the starting line, only 6 managed to cross the finish line – with X-Humanoid’s Tien Kung Ultra robot winning at 2 hours and 40 minutes, nearly three times slower than the human winner’s 1 hour and 2 minutes.
Noetix Robotics’ N2 robot claimed second place with a time of 3 hours 37 minutes, while DroidUp’s X02 finished third at 4 hours 50 minutes.
Robots Running on Fumes
The race quickly turned into what could be described as a mechanical comedy of errors, starting when an unofficial Unitree G2 robot fell over at the starting line to much amusement amongst onlookers.
A small robot from Beijing Polytechnic University dramatically overheated and went up in smoke, while others needed water spraying to keep them cool.
As Tang Jian, CTO of X-Humanoid, admitted, even the champion fell once during the marathon due to battery failure and required three battery swaps.
The race demonstrated that, whilst AI technology has enabled some impressive applications of humanoid robots, when it comes to keeping up with real people, robots are still very much in the slow lane.
Video source: NBC News
Brilliant Brains, Brittle Bodies
While the physical performance of the various robots may have been underwhelming, the artificial intelligence powering them is cutting-edge technology.
These machines aren’t remote-controlled or pre-programmed toys; they’re complex systems with sophisticated control algorithms, and just completing the course would have been unthinkable just a few years ago.
X-Humanoid spent “seven to eight months preparing for this marathon by doing a lot of tests on the robot and fine-tuning its algorithm,” according to Tang. The company reduced their robot’s weight, improved its heat dissipation system, and added cushioning functions to minimize wear on its sole and joint modules.
As one of the developers noted, the marathon aimed to test robots’ “cerebellums” (ability to control the body) rather than their “cerebrums” (ability to make decisions).
This distinction highlights an important reality: while AI software has made remarkable strides, translating that intelligence into graceful, humanlike physical movement remains challenging, and more importantly, may ultimately limit their real-world applications.
The Commercial Future of Humanoid Robots
Despite their current limitations, companies remain optimistic about humanoid robots’ commercial potential.
Tang Jian outlined a three-phase commercialization plan:
The commercialization of humanoid robots will have three phases. In the first phase, robots will be used in general industries; then, they will be used in specific commercial environments to finish tasks such as packaging and scanning codes; finally, they will be deployed to homes to play the role of nanny and housekeeper.
This timeline might seem ambitious given the robotic stumbling displayed in Beijing. As Hu Chenxu, founder and chief scientist of Noetix, stated, “Robots need a high adaptability to go from the laboratory to the marathon track.”
Nevertheless, China is pushing ahead in the humanoid robot race.
The Gaogong Robot Industry Research Institute projects that Chinese companies may sell about 7,300 humanoid robots globally in 2025, out of a worldwide total of 12,400 units worth 6.34 billion yuan.
Conclusion: Not Quite Ready to Take On Humans
The half-marathon showed that while AI continues to make impressive strides, the robots themselves are still finding their feet, quite literally.
Battery limitations, overheating issues, and basic stability problems all need solving before humanoid robots can truly hit their stride.
As these machines slowly jog toward commercial viability, one thing is clear: humans don’t need to worry about being outrun by robots anytime soon. But with companies investing heavily in both the brains and bodies of these machines, it’s only a matter of time before they pick up the pace.
For now, though, these robots aren’t exactly running circles around us – they’re more likely to be running out of battery.
