Perceptron, a startup founded by two former Meta Fundamental AI Research (FAIR) scientists, has launched Isaac 0.5, an open-weight vision model designed to help industrial robots navigate complex environments like warehouses and factory floors.
Co-founded by Armen Aghajanyan and Akshat Shrivastava, Perceptron's model aims to give machines the ability to perceive, reason, and act in physical settings. Unlike narrow models built for single repetitive tasks, Isaac 0.5 is general-purpose and flexible across different environments.
The model was trained on a million hours of general video data, plus ego video (first-person perspective recordings) and UMI video (repetitive human action recordings). Perceptron says it built petabyte-scale datasets spanning images, text, video, and robotic trajectories.
Isaac 0.5 is released as an open-weight model, meaning its parameters and training materials can be inspected by anyone. The startup previously raised $16 million from Bessemer Venture Partners and others, and is closing an additional funding round.
Target industries include manufacturing, logistics, warehousing, security, mobility, and media entertainment.
