China's Robotics Revolution: Inside 11 Companies Racing to Build Humanoid Robots
Key Takeaways
- ▸Over 140 Chinese companies are developing humanoid robots, with some already achieving mass production (AgiBot's 5,000-unit milestone) and demonstrating increasingly complex motor skills
- ▸Deep learning and AI are the core technology enabling robots to learn dexterity from data patterns rather than explicit rules, similar to how LLMs like ChatGPT learn language
- ▸China has committed £100 billion in government funding for robotics and other strategic technologies, with major cities adding their own investments—signaling unprecedented state-level commitment to automation
Summary
A comprehensive on-the-ground investigation reveals that China is rapidly advancing toward mass-produced humanoid robots, with over 140 companies competing to build autonomous machines capable of performing factory labor. The article follows visits to 11 robotics firms across five Chinese cities, including AgiBot (which has already produced 5,000 humanoid robots) and Guchi Robotics, a Shanghai-based automation specialist focused on car factory final assembly. The technology relies heavily on deep learning and AI—the same mathematical frameworks powering large language models like ChatGPT—applied to physical robotics to enable machines to develop human-like dexterity without explicit programming.
China's government is backing this ambition with unprecedented resources: a £100 billion strategic technology fund announced in 2025 covers robotics alongside quantum computing and clean energy. The visible showcase of progress came during the lunar new year festival gala, where robots demonstrated increasingly sophisticated capabilities—from synchronized cheerleading routines last year to cartwheels and parkour this year. While fully autonomous humanoid robots capable of handling 80% of remaining factory assembly tasks remain technically challenging, the scale of investment and the speed of incremental progress suggest the sci-fi vision of large-scale robotic labor replacement is moving from fantasy toward feasibility within the coming years.
- The ultimate target is automating 'final assembly' in manufacturing (currently 80% non-automated), which would displace hundreds of millions of workers globally, making this a transformative labor and economic issue
Editorial Opinion
China's robotics push represents one of the most consequential technology races underway today, blending deep learning advances with manufacturing automation at an unprecedented scale and pace. The combination of state funding, dozens of competing companies, and rapid visible progress (from choreography routines to parkour in a single year) suggests humanoid robots capable of general factory work are no longer decades away. However, the article wisely frames this not merely as a technological achievement but as an economic and labor question with profound implications—this is a race not just to build robots, but to reshape who does work and how.
