Record-breaking humanoids. AgiBot tops the Olympic medal table, but the competition is in the kitchen rather than on the track

There are more than 160 Chinese manufacturers of humanoid robots: this is why the ‘Robotics Olympics’, which took place in Beijing at the end of August, attracted such a large turnout, featuring heats, quarter-finals, semi-finals and finals in the main disciplines, particularly those drawn from the world of athletics (at present, it is not recommended to make even – and especially – the most advanced models swim).

Around twenty of the country’s leading manufacturers went head-to-head, with a few high-profile withdrawals (notably that of UBTech, the Shenzhen-based listed company which ranks amongst the top three in China – and therefore in the world), across some fifty different disciplines.

The Olympic medal table was topped by AgiBot, the Shanghai-based company founded in 2023 by Deng Taihua and Peng Zhihui (a former ‘genius’ at Huawei), backed by Tencent, with an indirect listing still underway. In the real world of products delivered, AgiBot is hot on the heels of the global leader Unitree – which, for its part, takes comfort in its third-place finish with just ten medals in total, compared to the forty-six won by its main competitor, whose recent stock market flotation has made several hundred of its employees – and many of its shareholders – millionaires.

Coming in second place, between the two major competitors, was Tiangong, the flagship product of a state-owned enterprise, the Beijing Humanoid Robot Innovation Centre (BHRIC / X-Humanoid), a centre founded in November 2023 as a joint venture between Beijing Yizhuang Investment Holdings, UBTech (that’s where it was hiding), Xiaomi and Beijing Jingcheng Machinery Electric. The company subsequently increased its capital to 350 million yuan – approximately 45 million euros – with the entry of Beijing Yizhuang Robot Technology Industrial Development as a shareholder.

Tiangong won some of the most iconic events: in particular the 100 metres, where he broke the world record several times in the space of a few days, setting the definitive record in the final; on both occasions, he ran significantly faster than Usain Bolt’s record (his best time was 8.64 seconds, almost a second faster than the Jamaican champion’s historic time of 9.58).

The most viral videos from the competition were actually those that made for great meme material: robots crashing into a giant airbag at the finish line or spinning round and destroying themselves – movements that are completely unnatural in terms of posture compared to human sprinters – and other instances that have racked up millions of views on social media in both the West and the East. It is interesting to note that anthropomorphic robots have now broken the records set by their flesh-and-blood counterparts in all the main disciplines: from the 100 metres to the marathon, including all intermediate distances, and even in the high jump (for the time being, for robots from a standing start: Sotomayor is still safe). It is also worth noting that, precisely for this reason, we are likely to see fewer and fewer of these competitions once it is established that robots have definitively overtaken humans – even ‘augmented’ humans, as we saw at the Enhanced Games in Las Vegas, the ‘Olympic Games for doped athletes’ strongly championed and funded by Peter Thiel. After all, does anyone really want to watch two AIs playing chess?

In announcing the next edition in 2027, the Chinese organisers have specified that the focus of the Games will shift increasingly from the athletics track to domestic settings, with competitions involving the preparation of complex dishes, ironing clothes and other far more useful activities, now that the ‘wow factor’ of sporting competitions is waning. Moreover, many online users have recently been pointing out the contrast between the parade of humanoids in Beijing’s stadiums and the hands-on approach to relief efforts following the climate disaster in Tibet.

As China continues on the path towards the maturity of humanoid robotics technology, demonstrations of strength and speed are becoming increasingly meaningless: the focus today is, if anything, on dexterity and the ability to operate in confined spaces and dimly lit environments, responding to usage scenarios that may seem trivial from a human perspective but which, due to the well-known Moravec paradox, prove far more complex for robots. It is precisely in this area that Beijing’s substantial investments – as well as the even larger ones from Silicon Valley – are currently focused.