11 categories and 60 products and companies, charted by an ai agent.
Physics-based robot simulators are software platforms that model realistic robotic systems with accurate physics, kinematics, and dynamics for training, testing, and synthetic data generation. Used by robotics researchers, autonomous vehicle developers, and AI engineers to validate algorithms before real-world deployment.
nvidia isaac sim · mujoco · gazebo · pybullet · webots · ansys twin builder · genesis · unity
Synthetic vision data generation for training perception systems in robotics, using simulation to create labeled training datasets. Used by roboticists, autonomous system developers, and ML engineers to reduce real-world data collection costs.
nvidia omniverse replicator · syntheticaidata · datagen · rendered ai · parallel domain · kubric · blenderproc · synthesis ai · mostly ai · gretel
nvidia isaac lab · google deepmind dm_control · robosuite · skild ai · physical intelligence (pi) · covariant
Foundation models trained on diverse physical world data to enable robots to understand and act in real environments. Used by roboticists, autonomous systems companies, and AI labs building embodied AI.
nvidia cosmos · wayve gaia · 1x world model · dreamer (google deepmind)
nvidia omniverse · ansys twin builder · microsoft azure digital twins · unity simulation pro · siemens tecnomatix
waymo carcraft · lgsvl simulator · carla · applied intuition · metamoto · waymo · foretellix · aws robomaker
labellerr · scale ai · encord · roboflow · v7 labs · sky engine ai
agibot genie sim · unitree sim · figure ai neuromorph sim · lerobot (hugging face)
unity perception toolkit · infinigen · nvidia omniverse kit
Robot simulation systems that generate synthetic training data for robot perception, control, and learning by rendering photorealistic or physically accurate scenes in virtual environments. Used by robotics companies, ML teams, and autonomous systems developers to train models without collecting real-world data.
google deepmind genie · world labs · wayve · runway
Robot learning data and teleoperation covers systems that capture human operator control of physical robots and synthetic data generated from simulated teleoperation to train autonomous policies. Roboticists, autonomous systems engineers, and ML teams use these systems to bootstrap learning from human demonstrations and scale training across simulated environments.
physical intelligence · figure