Skip to content
World clockEU--:--UK--:--USA--:--CN--:--PLDEFRIT中文EN

portal about AI and technologyevents · analysis · interviews · technical background

Search
LIVE
›

Humanoid robots: China gets the factories ready and hands out the first diplomas

Changan wants to mass-produce humanoid robots by 2028 and to build a robot that builds robots. In Hangzhou, the robot school has handed out its first diplomas after practical trade tests.

TechnologyAnalysisRachel NwosuPublished: 25 September 20265 min readSources 3
Humanoid robots: China gets the factories ready and hands out the first diplomas

China treats humanoid robots as an industrial product to put on the line, not as a technology showcase. The clearest signal comes from the car group Changan. The company has set two goals: mass production of humanoid robots by 2028, and commercialisation of flying cars by 2030.

The strategy in stages

Changan's path is gradual, and it starts inside the cabin. The robots are first broken down into components inside the cars. From there they move into logistics, cleaning, energy storage, charging, inspection, security and firefighting. Executives sum up the formula in one line: build cars with robots, sell cars with robots, put robots inside cars. The stated goal is a factory where robots produce other robots. In the meantime, the company is validating a 20 Ah solid-state cell based on oxygen and polymers.

The problem of millimetres

Wang Xingxing, founder of Unitree, has pinpointed the technical knot. The bottleneck of embodied AI, he says, lies in the error of a few millimetres between what the model decides and what the robot does in the physical world. In his analysis, a ChatGPT moment for robotics will arrive when a machine can complete 80% of tasks in 80% of new scenarios. Today, efficiency and capability remain insufficient both for the factory and for the home. The way out lies in building more general models, able to transfer skills from one task to another.

Learning a trade

In Hangzhou the transition already has an organised form. The robot school has handed out its first diplomas, and there machines learn professional tasks through practical tests. The school collects training data in real contexts, and at the same time prepares the skills of the people who will have to manage fleets of machines.

What changes for manufacturing

For the European production system, and the Italian one in particular, the question is not replicating Chinese robots but integrating them. Flexible robotics becomes worthwhile when the cost per machine hour falls below that of the labour it replaces, and when reconfiguration times allow small batches, as specialised manufacturing requires. On this, the game is played more on software integration and on the operational safety of OT systems than on mechanics.

Comments 0

Sources

3
  1. 01IT之家 — Changan: robot umanoidi in produzione di massa nel 2028ZH
  2. 02IT之家 — Unitree, Wang Xingxing: il collo di bottiglia è l'errore di pochi millimetriZH
  3. 03ANSA — Da automi a lavoratori: in Cina i primi diplomi della scuola dei robotIT

All figures and quotations in this text come from the sources listed below.

Content prepared by the editorial team with AI assistance.

Rachel Nwosu

Rachel Nwosu

AI, models and technology

Rachel Nwosu covers AI, models and technology for FLASH24, working from public model documentation, benchmark releases and repository histories rather than press summaries, and she skips announcements that arrive without reproducible numbers. She checks training-data claims against dataset cards and reruns reported metrics where code is available. She spends much of her week interviewing researchers and engineers, tracking model launch calendars, and comparing vendor benchmarks with independent evaluations. Outside the desk she runs 3D printers, restores old computers, and tests how models learn from internet junk. She does not publish benchmark figures she cannot trace to a source.

Newsroom →

Comments

0
  1. No comments yet — be the first.

Write a comment

Comments are public. We do not publish abuse, spam or advertising.