Tashi Zhihang AWE 3.0 Raises $455M, Setting New Embodied AI Funding Record
Tashi Zhihang (它石智航) has announced a groundbreaking $455 million Pre-A round funding for its AWE 3.0 native embodied AI model, marking the largest single funding round in China’s embodied intelligence sector.
Technical Breakthrough: AWE 3.0 Native Embodied Model
AWE 3.0 represents the world’s first native embodied foundation model, achieving natural alignment across three core modalities: visual perception, language understanding, and motion control. Unlike traditional robots that can only see or hear in isolation, AWE 3.0 genuinely transforms perception into coordinated physical action.
The model’s built-in Failure Recovery capability allows robots to predict multiple possible future outcomes within latent space and adjust decisions in real-time, significantly improving precision and adaptability in industrial operations.
Industrial Application: Wire Harness Assembly Success
The practical results speak for themselves in wire harness assembly tasks:
- Sub-millimeter precision for flexible operations
- Over 100 assemblies completed within one hour
- Guinness World Record achieved, breaking traditional robot limitations in delicate assembly
Data Ecosystem: 100,000+ Hours of Operation Data
The company has accumulated over 100,000 hours of high-quality operational data through its SenseHub suite, combining data collection gloves with first-person cameras for direct human-centric data capture at real work sites.
A Spark Plan targets building a 100 million-hour industrial-grade data sharing ecosystem to enable cross-scene generalization.
Investment Significance
Hillhouse Capital and Sequoia China co-led this round, signaling investors shift from hardware-focused to brain-centric embodied AI companies, highlighting the market value of technological generalization.