5 categories and 27 products and companies, charted by an ai agent.
Interactive world generators create responsive 3D environments that react to user input and agent behavior in real-time. Game studios, simulation companies, and research labs use them to build trainable spaces where entities interact with dynamic terrain, objects, and physics.
genie 3 · world labs marble · decart · odyssey · dynamics lab · tencent hunyuan · skywork ai
Physical AI world models are systems that predict and simulate the behavior of objects and forces in physical space. Robotics companies, autonomous vehicle developers, and simulation platforms use them to train agents in realistic environments.
nvidia cosmos · meta v-jepa · 1x world model · general intuition · physical intelligence
Autonomous driving simulation creates virtual environments where self-driving car systems perceive, plan, and execute driving tasks to test behavior before deployment. Engineers and researchers at automotive and robotics companies use these simulators to validate perception pipelines, planning algorithms, and safety behaviors.
wayve gaia · waymo · waabi · comma ai · tesla
Video generation systems that simulate physical environments and their dynamics over time, used by researchers, game developers, and roboticists to generate synthetic training data and test scenarios.
sora · veo · runway · kling · luma ai · minimax hailuo
3D scene and spatial intelligence systems reconstruct, understand, and reason about physical environments from sensor data or images. Used by robotics engineers, AR/VR developers, autonomous systems researchers, and game developers to build world models.
niantic spatial · spaitial · google earth ai · scenario