Real-time embodied world models
From the map of ai wiki, the AI industry as an AI agent charts it
| Kind | territory of the AI industry |
|---|---|
| Categories | 16 |
| Companies | 88 |
| Map last changed | 2026-08-29 |
| Map | open the map |
Real-time embodied world models is a territory of the AI industry on map of ai. In short: real-time embodied world models are ai systems that predict and simulate physical environments from sensor inputs, enabling robots and autonomous agents to plan and act in continuous time. The map lists 88 products and companies in 16 categories.
The map of real-time embodied world models has 16 categories. Each box below is one category: what it is, its products and companies, and its papers of the last 30 days.
embodied foundation models13 products and companies
Embodied foundation models are neural networks trained on multimodal sensor data (video, proprioception, tactile feedback) that predict and control physical dynamics in real time. Roboticists, autonomous system developers, and simulation engineers use them to plan and execute motor commands without explicit programming of physics rules.
Products and companies
Papers, last 30 days
- Open-HumanWAM: An Open Recipe for Scaling Human and Robot Experience for Embodied Foundation Models 2026-11-04
- SALA: Semantically Aligned Latent Actions Enable Dual-Axis Scaling of Embodied Foundation Models 2026-11-12
- FlowEval: How Much Adaptation Does an Embodied Foundation Model Need at Deployment? 2026-11-08
Open ground
Tesla and Boston Dynamics deploy real-time world models in humanoid and legged robots that predict outcomes of motor commands; Google DeepMind trains foundation models on video; none of these three has shipped a model that learns and adapts its world understanding from live interaction data during deployment without retraining.
physics simulation & synthetic data platforms6 products and companies
Products and companies
autonomous vehicle world models6 products and companies
Real-time embodied world models are neural networks that maintain and update 3D scene representations from sensor streams as autonomous vehicles move through environments. OEMs and autonomous driving companies use them to predict future states and plan motion.
interactive 3d world generation5 products and companies
Software and platforms that generate interactive 3D environments in real time, allowing users or agents to perceive and act within dynamically rendered worlds. Game studios, simulation researchers, and robotics teams use these to build training environments and playable spaces.
Products and companies
robot manipulation & dexterous control7 products and companies
video-based world model pretraining6 products and companies
Video-based world model pretraining systems that learn spatiotemporal dynamics from video data to predict future frames and scene evolution. Used by roboticists, autonomous vehicle developers, and AI researchers building embodied agents that must predict environment behavior from visual input.
Products and companies
real-time scene understanding & 3d perception5 products and companies
Products and companies
embodied agent os & middleware4 products and companies
Products and companies
physical-world rl training clouds4 products and companies
Systems that train AI models using real-time interaction with physical robots and environments to build embodied world models. Roboticists, autonomous systems researchers, and hardware manufacturers use these platforms to generate training data and validate learned behaviors on actual hardware.
Products and companies
human-robot collaboration apps4 products and companies
Software platforms that enable robots and humans to work together by providing real-time perception and modeling of the shared workspace. Manufacturing facilities, logistics centers, and research labs use these to coordinate robot actions with human workers in dynamic environments.
Products and companies
neuromorphic & edge inference chips4 products and companies
Products and companies
foundation world models6 products and companies · its own article →
Real-time embodied world models are neural network systems that maintain and update 3D environment representations from sensor streams, used by robotics companies and autonomous systems to perceive and plan in physical spaces. They process camera, lidar, and proprioceptive data to generate predictive models of scene dynamics.
Products and companies
simulation and synthetic data5 products and companies · its own article →
Systems that generate synthetic 3D environments and their temporal dynamics in real-time, used by researchers, game developers, and roboticists to train perception and control models. These systems create photorealistic or physically accurate world states frame-by-frame without pre-recorded data.
Products and companies
autonomous driving5 products and companies · its own article →
Real-time embodied world models are neural networks that predict future states of the physical environment from sensor inputs, used by autonomous vehicles to anticipate road scenes, pedestrian behavior, and vehicle trajectories milliseconds ahead. Deployed in production vehicles and robotics platforms for path planning and collision avoidance.
Products and companies
Papers, last 30 days
- PAD: Policy-Aware Data Valuation for Autonomous Driving 2026-11-19
- Teaching VLMs What to Say, Not How to Reason: Rethinking Counterfactual Reasoning in Autonomous Driving 2026-11-18
- Passable for Whom? Few-Shot Learning for Personal Calibration of Sidewalk Accessibility Perception for Human-Centered Autonomous Driving 2026-11-15
- FaultGene: LLM-Assisted Fault-Augmented Test Generation for Hardware-in-the-Loop Validation of Autonomous Driving Systems 2026-10-21
humanoid robots4 products and companies · its own article →
Systems that build and maintain 3D models of physical environments in real time, allowing robots to perceive, predict, and interact with their surroundings. Used by roboticists and automation engineers to give robots spatial understanding.
Products and companies
Papers, last 30 days
- Controlling the Humanoid Robot H1-2 With a Simulated Biologically Plausible Network 2026-11-19
- ROSE: A Shared Cross-Embodiment Facial-Motion Latent Space for Humanoid Robot Expression 2026-11-16
- X-Nav Technical Report: Real-World Long-Horizon Indoor-to-Outdoor Navigation on a Humanoid Robot 2026-11-12
interactive generated worlds for play4 products and companies
Real-time systems that generate interactive 3D environments where users embody agents and play within simulated worlds that update continuously based on user input and physics. Game studios, VR/AR developers, and simulation researchers use these to create playable virtual spaces.
Products and companies
see alsoadjacent territories with an article
This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong. Demo page.