China's TARS raises record US$120 million in angel round for embodied AI

by Jiang Yiman

TARS, a Chinese start-up focused on embodied artificial intelligence (AI), has raised US$120 million in angel funding, setting a record for China’s emerging embodied AI sector.

Announced on Tuesday, the round was co-led by Lanchi Ventures and Qiming Venture Partners, with backing from Linear Capital, Hengxu Capital, Hongtai Fund, Lenovo Capital, Xiang He Capital, and Hillhouse Venture Capital. The funds will go toward product development, model training, and application expansion.

Founded on February 5, TARS positions itself as a full-stack embodied AI company with capabilities in foundation models, hardware R&D, and mass production of integrated AI products.

Its founding team includes former executives and scientists from Huawei, Baidu, and DJI. CEO Chen Yilun was previously chief scientist for intelligent robotics at Tsinghua University’s AI institute, CTO of autonomous driving at Huawei, and machine vision chief engineer at DJI.

Chairman Li Zhenyu, formerly president of Baidu’s Intelligent Driving Group, led the Apollo open platform and autonomous ride-hailing service “Luobo Kuaipao.”

Chief scientist Ding Wenchao, a Fudan University robotics researcher and Huawei “Genius Youth” alum, helped build Huawei’s end-to-end decision networks and Fudan’s first humanoid robot, while chief architect Chen Tongqing headed Huawei’s ADS Navigation Department.

According to Tianyancha, TARS raised its registered capital from 1 million yuan to 1.36 million yuan on March 6. Investors from the latest round are now listed as shareholders.

Embodied AI — which integrates large AI models with robotics — is seen as a key future tech. McKinsey estimates the global market could reach tens of trillions of yuan by 2030.

The sector received a boost this year after “embodied intelligence” appeared for the first time in China’s annual government work report, alongside quantum tech, 6G, and bio-manufacturing.

Still, challenges remain in supply chains, datasets, and scaling real-world use. As the space evolves, only time and market validation will tell who delivers on the promise of embodied AI.