General Trajectory

General Trajectory (GT) is developing a foundation model for physical intelligence, enabling autonomous operation of real-world machinery like robots, drones, and industrial equipment. The company uses a combination of generative AI, simulation, and real-world data to create models that can perceive, plan, and execute complex physical tasks. It is designed for robotics companies, industrial automation engineers, and researchers who need robust, generalizable control systems. GT differs from traditional robotic software by using large-scale pre-training on diverse physical tasks, allowing its models to adapt to new environments and hardware with minimal fine-tuning. The system integrates with existing hardware platforms through a unified API and simulation environment. The platform's core is a neural network trained on billions of trajectories spanning manipulation, locomotion, and navigation. GT claims to handle both structured settings like assembly lines and unstructured environments like warehouses or outdoor terrains. Development tools include a simulation suite, model zoo, and deployment SDKs. The company targets the gap between narrow industrial robots and general-purpose humanoids, offering a middle ground where a single AI can control multiple form factors. As of 2026, GT is in a closed beta, with pricing based on deployment scale.