Studio's private beta lets physical AI teams build labeled training images from adjustable 3D scenes, with Physicl's $10 million seed round supporting hiring and continued product development.
Inovia Capital led the seed financing, while U.S. Venture Partners joined as an investor. Physicl said the money will cover additional hires and ongoing platform work.
A selected setting or text description gives users a starting scene. They can change the lighting, camera angle and object placement, with rendered images carrying labels for depth, surface normals, segmentation or camera pose. Physicl founder and CEO Alex de Vigan said, “Models fail on edge cases.”
A base of 10,000 ground-truth environments and an inventory topping 100,000 assets underpin billions of scenes the platform can generate.
An Oct. 19 release from CNTXT AI described training facilities for robots and a data collection network spanning more than 300 use cases; EconoTimes previously covered its data expansion plans.
Studio outputs labeled images; it does not simulate physical processes over time.
