ShanghaiTech Team Secures Multi-Million Dollar Funding for Embodied World Model Infrastructure
InstAdapt, a company founded by a team from ShanghaiTech University, has completed a multi-million US dollar seed round of financing. The funding will be used to build physical AI data infrastructure and develop embodied world models.
The company's goal is to address the core bottleneck in embodied intelligence: how robots can quickly acquire new experiences instead of collecting data and training from scratch.
Dr. Shi Ye, founder of InstAdapt, has been researching diffusion models, generative reinforcement learning, and embodied intelligence.
InstAdapt's four-layer Physical AI infrastructure includes robot data acquisition and governance, 3D physical asset generation and scene reconstruction, unified world action model and robot strategy training, and generative reinforcement learning and embodied recursive self-improvement.
The company has built a complete technical pipeline covering data production, physical asset generation, model training, and robot deployment. InstAdapt's relevant technical solution was invited to share at NVIDIA GTC 2026.