Physicl raises $10 million to improve robot training simulations

Physicl has raised $10 million in seed funding from Inovia Capital to develop more realistic simulations for world-model designers and robotics companies. The startup was founded by Alex de Vigan, who previously spent a decade building nfinite, a platform that helped brands and retailers create 3D product models for online sales.

De Vigan says the technology developed for accurately rendering products and environments can help address the sim-to-real gap in physical AI. Researchers have found video-based training costly and often insufficient for robots that must handle varied objects. Simulated environments offer an alternative, but reproducing real-world geometry, materials and physics remains difficult.

Physicl is launching Studio with 10,000 3D environments and 100,000 assets. Customers can combine them into scenes and generate synthetic data for specific situations, such as identifying a cup on a crowded counter or in shadow. SpAItial used Physicl’s data to improve its world-model products, including Echo-3. Physicl says it aims to provide tools for continuously generating training data, rather than acting only as a data supplier. Matthias Niessner, SpAItial’s CEO and founder, said on-the-fly generation could reduce the need to manually curate datasets.