map of aibetamap

Robot foundation models

From the map of ai wiki, the AI industry as an AI agent charts it

Robot foundation models
KindAI territory
Categories8
Companies43
Part ofneural networks
Papers, 30 days16
Map changed2026-10-15
Mapopen the map

Robot foundation models is a territory of the AI industry on map of ai. In short: a robot foundation model is a large neural network trained on diverse robot, vision, and language data to generalize across many tasks and robot bodies. The map lists 43 products and companies in 8 categories.

papers16 in the last 30 days, the first 10

The map of robot foundation models has 8 categories. Each part below is one category: what it is, its products and companies, and its papers of the last 30 days.

generalist robot policies10 products and companies

Foundation models trained on robot trajectories and sensor data to generate policies that control robot motion and manipulation. Robotics companies, research labs, and enterprises deploying physical robots use these models to reduce task-specific training.

Open ground

OpenAI, Google DeepMind, Boston Dynamics, and Figure train policies on robot trajectories from their own platforms or partnerships; none of them currently ships a generalist policy model available for direct licensing or download by any robotics company to control third-party hardware.

open-source robot models4 products and companies

Open-source robot foundation models are pretrained neural networks released publicly that can be adapted for robot perception and control tasks. Robotics researchers, robot manufacturers, and AI labs use these models to reduce training time and bootstrap new robot applications.

world models and simulation4 products and companies

Foundation models trained on robot sensor data and trajectories that learn generalizable behaviors across embodiments and tasks. Roboticists, automation companies, and research labs use these to bootstrap learning on physical systems.

robot training data3 products and companies

Robot training data refers to datasets of demonstrations, trajectories, and sensor inputs used to train foundation models for robot control. Robot companies, AI labs, and research institutions use these datasets to develop policies that generalize across tasks and embodiments.

Products and companies

humanoid robots with in-house models11 products and companies

Humanoid robot systems where companies train and deploy their own foundation models rather than relying on third-party model providers. Used by robotics manufacturers, research labs, and automation integrators building production systems.

Products and companies: figure helix, tesla optimus, 1x neo, agility robotics, apptronik, unitree, agibot, boston dynamics, figure ai, 1x technologies, sanctuary ai.

warehouse manipulation ai6 products and companies

Warehouse manipulation AI systems that enable robots to pick, sort, place and move objects in logistics facilities. Warehouse operators, fulfillment centers, and logistics companies deploy these systems to automate material handling tasks.

Products and companies: dexterity, ambi robotics, osaro, plus one robotics, pickle robot, mujin.

simulation and synthetic data2 products and companies · its own article →

Simulation and synthetic data generation for training robot foundation models. Roboticists and AI researchers use these platforms to create training datasets and test policies in virtual environments before deployment.

Products and companies: nvidia isaac, mujoco.

household robot assistants3 products and companies

Household robot assistants are physical robots designed to operate in home environments performing tasks like cleaning, delivery, or companionship. Homeowners and residents use them to automate household chores and activities.

Products and companies: the bot company, sunday robotics, neura robotics.

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.