Physicl has secured $10 million in seed funding led by Inovia Capital to help world model designers and robot builders solve training challenges with realistic simulations. Founded by Alex de Vigan, the startup repurposes a tech stack originally built for e-commerce 3D modeling to faithfully render physical environments and aid physical AI development.
The company aims to address difficulties researchers face when training robots on video data or expensive remote operations by providing high-fidelity digital simulations. Physicl is launching a product called Studio, which features 10,000 3D environments and 100,000 assets that allow users to generate synthetic training data for specific edge cases. Early customers like world model builder SpAItial have collaborated with Physicl to improve product fidelity.
This funding will help Physicl provide ongoing synthetic data generation tooling for foundation model builders and robot operators. By transforming unstructured real-time information into accurate 3D representations, the platform allows developers to automatically generate diverse training permutations on the fly.
Why it matters: Realistic physics simulations can accelerate physical AI development by reducing the time and expense required to train robots for complex real-world tasks.