Technology
Former Meta scientists launch industrial vision AI startup Perceptron
The company released its open-weight general-purpose model, Isaac 0.5, after raising $21 million in venture funding.
The short version
- Former Meta research scientists founded Perceptron to build frontier vision models for industrial and physical automation.
- The startup released Isaac 0.5, an open-weight, general-purpose model designed to help robots perceive, reason, and navigate physical environments like factory floors and warehouses.
- Perceptron recently secured $21 million in funding led by Bessemer Venture Partners to target sectors including manufacturing, logistics, security, and mobility.
Key facts
- Perceptron was founded in November 2024 by former Meta research scientists Armen Aghajanyan and Akshat Shrivastava.[TechCrunch]
- Perceptron raised $21 million in a funding round led by Bessemer Venture Partners.[TechCrunch]
- The company launched Isaac 0.5, an open-weight vision model designed to assist robots with navigation, spatial reasoning, and operational tasks in industrial settings.[TechCrunch]
- Isaac 0.5 was trained on a million hours of general video alongside ego video, UMI video, and internally built petabyte-scale datasets.[TechCrunch]
- Perceptron plans to market its technology to sectors including logistics, manufacturing, security, mobility, and media.[TechCrunch]
What remains uncertain
- Perceptron has not publicly disclosed the specific sources used to gather its training video data.[TechCrunch]