China Humanoid Update October E01: Application in holiday

Uking Sun (0:00)

Hello, everyone, welcome back. This is our first update on China's humanoid robot sector for October. China's National Day holiday is a great opportunity to observe how humanoid robots are being deployed across the country. So my first question is, would you visit a theme park just because it has 300 humanoid robots in operation? AgiBot and Chimelong have launched the first large-scale deployment of its kind in the tourism sector, with robots working as guides, dancers, gaming partners and receptionists. Meanwhile, in Beijing, Galbot took advantage of the China Open to showcase its humanoid robot alongside world-famous tennis player Ekaterina Alexandrova. In the footage, the robot fell to the ground, got back up on its own and even returned a few balls. Pretty impressive, right? Over in Shanghai, at the opening ceremony of the World Skills Competition on September 22nd, 19 Unitree humanoid robots performed alongside 120 dancers in front of an audience of more than 10,000. The show was billed as the world's largest live broadcast performance featuring full-sized general-purpose humanoid robots.

Additionally, Leju Robot's Kuavo humanoid has taken on a new role as a guide at the Suzhou Bay Museum, greeting visitors, answering questions in real time, and leading guided tours through the exhibitions.

Turning to industrial robotics, DEEP Robotics demonstrated its quadruped robot working in a grain warehouse. As part of an automated inspection system, the robot can assist or replace human workers during routine safety checks by collecting data and identifying potential hazards. It can even verify whether workers are wearing safety helmets or smoking, and inspect fire extinguishers to make sure they're operational.

A new business model is emerging in China, crowdsourcing real-world data to train robots. Shanghai-based company Maniformer is renting out data collection equipment, allowing people to pick up tasks via an app and get paid for recording everyday activities at home like cooking, cleaning and doing laundry. Maniformer has built a marketplace for various types of data. In September, China officially recognized data collector as a new job category under embodied AI application technicians. Meanwhile, this collected data is increasingly being treated as a valuable commercial asset that can be registered, traded and even used as collateral.

Finally, I'd like to recommend an interview with Sergey Levine, co-founder of Physical Intelligence. It offers great insight into the current state of humanoid robotics. He shared several interesting takeaways, notably that robotics has entered a new phase where hardware is already good enough for many applications, meaning the main bottleneck now is developing intelligent decision-making systems, building a healthy robotics ecosystem and scaling the right solutions. All right, that's all for this episode. Thanks for watching. See you next time and welcome to New Zealand. If you ever need a private driver when visiting, feel free to check out my service at nowthis.nz.