Physicl raises $10M to make the digital world real enough for robots

Alex de Vigan, who previously built 3D rendering platform nfinite, has launched Physicl, a startup that has raised $10 million in seed funding from Inovia Capital. Physicl aims to help world model designers and robot builders train physical AI models using highly realistic simulated environments. Researchers have struggled with the high cost of collecting real-world robot training data and the limited returns from video-based training, while ensuring simulated worlds faithfully replicate real-world physics has proven difficult. de Vigan believes the technology pipeline his previous company built for converting unstructured real-world information into high-fidelity 3D representations with correct geometry, materials, and physics can now be repurposed for robotics. Physicl is releasing a product called Studio featuring 10,000 3D environments and 100,000 assets that lets users create scenes and generate synthetic training data, especially for difficult edge cases such as a robot struggling to identify a cup on a cluttered or shadowed countertop. The startup positions itself not merely as a data provider but as tooling for ongoing, automated synthetic data generation. World model builder SpAItial, led by CEO Matthias Niessner, worked with Physicl as an early customer to improve the fidelity of its products, and Niessner praised Physicl's ability to automatically generate training data on the fly based on a model's training needs. Physicl enters a crowded and still blurry category of startups supplying data and tooling to foundation model builders and robot operators.

(Source：TechCrunch)