Salem Robotics Builds AI-Classical Hybrid Software for Hazardous Industrial Inspection Robots

Salem Robotics, a YC S26 startup founded by former Los Alamos National Laboratory roboticists, has developed software that gives existing mobile robots the intelligence to perform complex inspections in hazardous industrial environments. The platform targets facilities such as nuclear plants and oil and gas sites, enabling robots to carry out physically interactive tasks like contamination smear tests and leak detection inspections. A core technical challenge the company addresses is generating constrained manipulation plans in real time based on geometry the robot actually observes, rather than requiring manually authored trajectories for each surface or component. Salem's approach combines classical robotics methods — valued for predictability and explicit constraints — with AI for semantic understanding of unfamiliar scenes. The founders bring roughly 15 years of combined nuclear industry experience and prior robotics research from UT Austin to the venture.