Physicl Raises $10 Million for Physical AI Training Data Platform

Physicl raised $10 million in seed funding to develop synthetic training data for physical AI models, with its new Studio tool focused on visual cases such as objects obscured by clutter and changes in lighting.

Inovia Capital led the round, and U.S. Venture Partners participated. The company plans to use the proceeds to expand its team and develop its platform.

Studio lets users select or describe a 3D scene, then adjust elements such as lighting, camera viewpoint and object placement. Its library contains 10,000 ground-truth environments and more than 100,000 assets, with billions of possible scene variations. The product is in private beta.

Studio renders scenes into image frames with labels that can include depth, surface normals, segmentation and camera pose. It does not simulate physics over time, making the tool a way to produce labeled visual data rather than a physics simulation.

The launch adds a synthetic-data approach to the broader push to expand physical AI training data. “Models fail on edge cases,” Physicl founder and CEO Alex de Vigan said.