Skip to main content

Tashi Zhihang (TARS Robotics)

Category: Robotics / Embodied AI

Shanghai-based embodied AI company building general-purpose humanoid robots and the AWE (AI World Engine) embodied foundation model for real industrial, logistics, and home tasks. Tashi Zhihang (TARS Robotics) was founded in 2025. The company is led by 陈亦伦 (Chen Yilun). Based in Shanghai, China. Team size: 101-500. Total funding raised: $697M. Latest round: Pre-Series A. Key investors include Hillhouse Ventures (高瓴创投), HongShan/Sequoia China (红杉中国), Meituan Strategic Investment (美团战投), Qiming Venture Partners (启明创投), Linear Capital (线性资本), Lanchi Ventures (蓝驰创投), Xiang He Capital (襄禾资本), Hongtai Fund (洪泰基金), CICC Capital (中金资本), Lenovo Capital (联想创投), TCL Capital, Hengxu Capital (恒旭资本).

Founded
2025
Headquarters
Shanghai, China
Team size
101-500
Total funding
$697M

Value proposition

Develops trustworthy physical-world AI and general-purpose robots via a "data-model-embodiment" trinity, using a human-centric data paradigm that digitizes skilled workers' real-world movements and tactile corrections into reproducible data assets to train its self-developed AWE embodied large model.

Products and solutions

AWE (AI World Engine) general embodied foundation model (AWE3.0/3.5/3.7, "Born as One" native multimodal architecture), T1 and A1 humanoid robots (A1 set a Guinness World Record for most sub-millimeter wire harnesses assembled by a robot in one hour), full-stack embodied AI tech stack spanning algorithm, embodiment, and soft-hard integrated mass production.

Unique value

One of the few companies simultaneously possessing embodied large-model capability, embodiment (hardware) R&D, and soft-hard integrated mass-production capability; pioneered a human-centric data paradigm and a single unretrained base model (AWE3.7) that generalizes across industrial, dynamic, and home tasks.

Target customer

Industrial manufacturers (automotive supply chains, complex wire-harness assembly), logistics operators, and home/service robotics applications.

Industries served

Industrial manufacturing, automotive supply chain, logistics, home/service robotics.

Technology advantage

Self-developed AWE (AI World Engine) architecture with native multimodal fusion of action, perception, geometry, and touch; end-to-end models, world models, and reinforcement learning; human-centric data collection that digitizes skilled-worker movements; million-hour-scale real human data; "Born as One" native embodied model design.

How they differentiate

Differentiates via a human-centric data paradigm (digitizing skilled workers' tactile corrections), a single unretrained base model (AWE3.7) that generalizes across industrial/dynamic/home tasks, and a "dream team" of ex-Huawei/ex-Baidu autonomous-driving leaders; holds China's largest angel and Pre-A funding records in embodied AI.

Main competitors

AgiBot (智元机器人), Unitree Robotics (宇树科技), UBTech Robotics (优必选), Galbot, Fourier Intelligence

Key partnerships

Tsinghua University AIR (talent cultivation partnership), Meituan (strategic investor & deployment partner), Shanghai municipal government (first 1,000-unit industrial embodied robot cluster deployment by 2027), Beijing Robotics Industry Development Investment Fund, Shanghai State Investment Guide.

Notable customers

Meituan (strategic investor & deployment partner), Shanghai municipal government industrial deployment, automotive supply-chain manufacturers.

Major milestones

Founded Feb 5, 2025, $120M angel round (Mar 2025, China record), $122M angel+ round (Jul 2025), AWE3.0 "world's first work-capable general embodied large model" launched at AWE 2026 (Mar 2026), A1 robot Guinness World Record (Mar 2026), $455M Pre-A round (Apr 2026, AWE3.7 demonstrated across industrial/dynamic/home tasks (Aug 2026), Shanghai 1,000-unit industrial robot cluster by 2027.

Growth metrics

Founded Feb 2025; raised ~$697M total in under 1.5 years; A1 robot set Guinness World Record for sub-millimeter wire-harness assembly; AWE model won WAIC SAIL award; planning first 1,000-unit industrial embodied robot cluster in Shanghai by 2027.

Market positioning

Positioned as a first-tier leader in China's embodied intelligence sector, holding the industry's highest single-round and Pre-A funding records and the highest post-investment valuation in the sector.

Geographic focus

China (Shanghai HQ), with plans for global competition in embodied intelligence.

About 陈亦伦 (Chen Yilun)

Ex-Huawei Autonomous Driving CTO & Chief Scientist (Intelligent Automotive Solutions BU); Ex-DJI Machine Vision Chief Engineer; Ex-Tsinghua University AIR (Institute for AI Industry Research) Chief Scientist for Intelligent Robotics. PhD in Electronic Engineering (Tsinghua University & University of Michigan).

Latest news about Tashi Zhihang (TARS Robotics)

More Robotics / Embodied AI companies

Official website: