Once and for all: Are robots from China ready to replace us?

On Tuesday this week, the humanoid robot Tiangong Ultra crossed the finish line of a 100-meter dash in Beijing in just 8.86 seconds, thereby "shattering" the human world record set by Usain Bolt in 2009, which stands at 9.58 seconds. In the 100-meter heat last year, a previous model of the Tiangong Ultra crossed the finish line in 21.5 seconds. Will another record be broken next year?

However, the images from the track also clearly illustrated the other side of Chinese robotics. Robots fell during the run, one of them fell apart, another caught fire after crossing the finish line, and in some of the heats, the robots struggled to brake after the finish line. This gap, between almost superhuman physical ability and the difficulty of reliably performing daily actions, is currently at the center of the race to develop humanoid robots.

The humanoid robot games held in Beijing attracted more than 2,000 robots that competed in 51 disciplines. This year, a more significant dimension was added to the Olympics: alongside running, jumping, boxing, and soccer, 21 of the heats were dedicated to scenarios simulating work in the real world. The robots were required, among other things, to connect cables, tighten screws, fill shelves, tidy up a hotel room, prepare food, and perform tasks in an industrial environment.

For a humanoid robot, in many ways, these are more difficult tests than a sprint. This is perhaps the "real Olympics" of the humanoid robot: not to beat a human in a race, but to independently and reliably perform the tasks for which it would be worthwhile to employ it in a factory or warehouse.

But the Olympics were only the most photogenic part of a busy week. At the same time, the World Robot Conference was held in Beijing, where more than 300 companies presented thousands of new systems. The message: the Chinese industry wants to move from the stage of tricks and capability demonstrations to the stage where robots perform work of economic value.

From the stadium to the factory floor

This transition is already starting to happen, albeit on a limited scale. Chinese companies are operating humanoids in pilots in automotive, electronics, and battery factories, in tasks such as component transport, sorting, quality checks, and loading and unloading lines.

"The demonstrations we see today from China testify to a significant leap in the capabilities of the robots, but the transition from an impressive demonstration to widespread deployment in an industrial environment is a challenge of a completely different kind," says Amnon Avi Guy, manager of Robotize from Mitvach. "The moment a robot enters a factory and works alongside humans, safety, reliability, maintenance, efficiency, and of course, return on investment enter the equation."

The numbers are also starting to become significant. According to some reports, more than 40,000 humanoid robots were supplied in China in the first half of 2026, and the country is responsible for about 97% of global shipments. And yet, Unitree CEO Wang Xingxing admitted that the industry has not yet reached its "ChatGPT moment" — a situation where you can put a robot in a new environment, tell it what to do, and it will know how to do it without specific programming. By his estimate, this turning point may arrive within 2-10 years. In other words: serial production has already begun; serial work has not yet.

"China has a significant advantage here, because it is already the world's largest manufacturing hub today and has a huge infrastructure of factories, supply chains, and engineering power where these technologies can be tested and implemented," says Avi Guy. "It seems that China is currently putting a lot of its chips on robotics and automation to maintain its relative advantage in manufacturing in the coming decades as well."

Big money is coming in

This week also illustrated that Chinese robotics is no longer just an engineering story. Unitree, the company most identified with the Chinese humanoid boom, went public on the Shanghai Stock Exchange and its stock jumped more than fivefold on the first day of trading. At the same time, XPeng announced that its robotics division had raised more than $900 million, at a valuation of more than $6.3 billion, in the largest private funding round to date in the humanoid field in China. AiMOGA, the robotics division of automaker Chery, has already delivered more than 3,000 robots and is planning its own IPO.

The entry of automakers is not accidental. A humanoid robot relies on some of the same infrastructure built for the electric car: motors, batteries, power systems, sensors, electronics, and suppliers. From China's perspective, the humanoid may be the next product on which the advantage already built in the supply chain can be leveraged.

The managed economy enters the lab

Here lies perhaps the most important difference between China and the Western model. In the United States, a company needs to convince investors that it has a product and then customers that it provides a return on investment. In China, the state tries to create the demand as well.

Central government and local authorities are setting up training centers for robots, subsidizing research, and opening sites for pilots. The robots generate data there, which is used to train the models and improve them for the next round of deployment.

The model is reminiscent of the way China built the electric vehicle industry: create a market large enough for dozens of companies to compete, prices to drop, and the supply chain to mature. However, some of the "demand" today comes from universities, research centers, and government programs, and not from customers who have decided that it is profitable for them to replace a human worker. Therefore, the number of robots sold is not the same as the number of "robotic workers" operating in the market.

And what will happen to the workers?

Here too, the Chinese framing is different from what is common in the West. While in the United States robots and AI immediately raise questions about job loss, Beijing presents them mainly as a tool for increasing productivity, dealing with an aging population and a shrinking workforce, and replacing dangerous and grueling tasks.

There is an inherent tension in this. A state ruled by a communist party, which puts the worker at the center of its ethos, is also one of the most aggressive in the world in promoting technologies that can replace human labor. As long as robots fill jobs that are difficult to recruit workers for, it is relatively easy to resolve the contradiction. If they start replacing workers on a large scale, the discussion will be more complex.

And that is still not where the industry is. The big week of Chinese robotics showed that a robot can already beat Usain Bolt. The next test will be less photogenic and much more important for the economy: can that same robot arrive at a factory, identify a part it has never seen, pick it up, screw it in the right place, and work for an entire shift, and then return the next day and do it again. In this race, China has not yet crossed the finish line. But it is already building the entire track.