Physicl

The page offers path-traced video with depth, normals, and instance IDs for custom scene variations based on user descriptions.

Rendered indoor scenes for robot perception and navigation training. Depth in metres, surface normals, instance IDs and camera pose on every frame.

Rendered scenes for world model training and evaluation. Run the same room again with one thing changed, with depth, normals and camera pose on every frame.

The page describes AI Habitat’s public scene datasets, which include 211 homes, and Physicl’s custom, physics-accurate Habitat datasets.

Writing from Physicl on synthetic data, ground truth, world models and robot policies, and what physical AI models need to learn from.

Physicl creates physics-accurate, USD-compatible assets for Isaac Sim, addressing inaccuracies in existing assets.

Physicl is a synthetic data platform for physical AI. With Studio, teams generate images and videos with exact ground truth on every frame.

Talk to the Physicl team about generating training data for robot policies, world models and perception models with Studio.

Physicl is the synthetic data platform for physical AI. Describe a scene, vary it millions of ways, and get exact ground truth on every frame.

Physicl, Inc. outlines its privacy policy for personal data collection, use, sharing, protection, and deletion in Physicl Studio.

Legal information for Physicl, Inc.: registered entity, address and contact details.

Physicl secures $10 million in seed funding to develop a platform generating targeted training data for physical AI. The company focuses on addressing edge-case challenges in physical AI models, aiming to improve their accuracy and reliability. Based in Paris, Physicl targets advancements in AI-driven physical systems through specialized data solutions.

Nov 2026The SaaS NewsNews article

Physicl secures $10 million in seed funding from Inovia Capital and US Venture Partners to develop synthetic training data for physical AI models. The company focuses on generating realistic datasets to advance AI applications in robotics and physical systems. This funding supports its mission to bridge gaps in AI training for real-world physical interactions.

Nov 2026SiliconReportNews article

Physicl secures $10M in seed funding from Inovia Capital to develop a data infrastructure for physical AI systems. The company addresses limitations in robotics and vision models by enabling teams to generate synthetic data—images and videos—matching specific edge cases, such as unusual lighting or cluttered environments. Physicl Studio allows researchers to input model errors and receive tailored datasets to improve accuracy.

Nov 2026SignalbaseNews article

Physicl secures $10 million in funding to adapt its digital product-rendering platform for robotics applications. Founder Alex de Vigan repurposes technology originally designed for e-commerce visuals to enhance robotic interaction with virtual environments. The platform aims to bridge digital and physical worlds by creating realistic simulations for robotic tasks.

Nov 2026TechCrunchNews article