GenRobot AI Raises $100M to Expand Physical AI Data

BEIJING -- China's GenRobot AI Inc., a startup providing robotics data, has secured $100 million of Series A+ capital to scale up its infrastructure to collect and process human activity data used to train robots.

The deal was announced on Nov. 2, stating that it had been finalized in the third quarter. The list of participants includes SAIC Motor, Baidu Inc., Shoucheng Holdings, Eastworlds (Malaysia), Galaxy General Robotics Co. Ltd. and several financial players.

The funds will finance increasing physical AI data infrastructure, developing datasets for world models, building robot evaluation systems, and entering overseas markets, the company said.

While companies creating robots themselves, GenRobot provides training data for developers of humanoid robots and robotic systems. In this case, the wearable device collects the information about people's movement and object manipulation when performing certain actions in home, factory, and logistics environments.

The company’s Gen DAS products provide video and people movement capture; the GenMatrix product of the company makes a pre-processing of the captured information. GenRobot claims that it has gathered over one million hours of data on thousands of skills and works with over 30 AI firms around the world.

Furthermore, the firm states that its quarterly provision of usable training data has grown from about 100,000 hours to 1 million hours. The firm predicts that its quarterly deliveries will hit 10 million hours by 2027, but the accuracy of these numbers has not been confirmed yet.

The firm was founded in May 2025 and has previously been funded by Ant Group, Didi and Momenta. It has not revealed its valuation after the recent funding round.