LIVE PROTOCOL
EET--:--:-- edition--.--.--

China targets 10,000 humanoid robots in new national program

China targets 10,000 humanoid robots in new national program

China has launched a nationwide program to fast-track humanoid robots and embodied artificial intelligence into industry, calling for the deployment of more than 10,000 units by the end of the year.

China has launched a nationwide program to fast-track the rollout of humanoid robots and embodied artificial intelligence into industry. The initiative, unveiled on Tuesday, sets ambitious targets for humanoid robots and related key products. It signals a coordinated national push to move the technology out of the laboratory and toward real industrial use across the country.

At the center of the plan is a concrete deployment goal. Authorities are calling for more than 100 high-value applications to be created, intended to support the deployment of over 10,000 units by the end of the year. The targets tie the technology's progress to measurable use in the field rather than to one-off demonstrations, giving manufacturers a clear benchmark to work toward.

The program builds on a market that is already expanding quickly. Official data shows that China shipped around 17,000 humanoid robots in 2025, produced by more than 140 companies. That figure points to a crowded and competitive field of manufacturers now working on the same class of machines, and helps explain why the authorities want to channel the effort.

According to those involved, the emphasis has shifted away from raw hardware performance. The stated goal is to teach robots to understand the physical world, to perceive their surroundings, make decisions autonomously and learn new tasks on the job. In industry terms, this effort is described as building a physical world model, a step beyond simply improving a machine's mechanical abilities.

Engineers acknowledge that significant hurdles remain before that vision is realized. A robot's brain still needs stronger generalization capabilities and the ability to carry out tasks with precision, while developers are also weighing battery endurance, accuracy and safety when machines complete real work. The understanding of the physical world, they say, is still being built gradually from digital sensing and computing, and the data and environments involved need to become more dynamic rather than purely static.

Supporters of the approach argue that a concentrated national effort can speed up progress. Although six months is described as a short window, the push is expected to help the industry converge more quickly on workable engineering solutions. It is also intended to pull companies along the wider industrial chain toward the same goal, so that gains are shared across suppliers and developers rather than confined to a handful of firms.

Loading article...