Embodied AI Landscape

Figure AI is an American robotics company developing general-purpose humanoid robots powered by artificial intelligence.

1X Technologies (formerly Halodi Robotics) is a Norwegian robotics company developing humanoid robots for consumer and industrial use, backed by OpenAI.

Unitree Robotics is a Chinese robotics company and the world's largest producer of humanoid robots by volume, known for the G1 and H1 humanoids at disruptive price points.

Physical Intelligence (π) is an American AI and robotics company developing general-purpose robot foundation models, including the π0 vision-language-action model for cross-embodiment robot control.

Boston Dynamics is an American engineering and robotics design company known for highly dynamic mobile robots including Atlas, Spot, and BigDog. Owned by Hyundai Motor Group.

Apptronik is an American robotics company developing the Apollo humanoid robot for industrial logistics and manufacturing, with backing from Google and Mercedes-Benz.

Figure 02 is Figure AI's second-generation humanoid robot, built for commercial deployment in warehouse, logistics, and manufacturing environments.

NEO is 1X Technologies' soft-bodied, tendon-driven consumer humanoid robot designed for household chores with built-in LLM integration.

Digit is Agility Robotics' bipedal humanoid designed for warehouse logistics, with purpose-built capabilities for navigating human-scale environments.

Apollo is Apptronik's full-size humanoid robot built for industrial logistics and manufacturing, featuring 34 degrees of freedom and modular design.

Atlas is Boston Dynamics' advanced humanoid robot, now in its fully electric third generation, designed for industrial material handling with 56 DoF and exceptional mobility.

G1 is Unitree Robotics' entry-level humanoid robot notable for its aggressive $16,000 price point, targeting research and education applications.

Helix is Figure AI's proprietary VLA model for generalist humanoid control with zero-shot manipulation and multi-robot collaboration.

pi0 (pi-zero) is Physical Intelligence's generalist VLA robot foundation model for zero-shot dexterous manipulation across 8 robot types.

OpenVLA is a pioneering open-source 7B VLA model combining a pretrained VLM with action de-tokenization for zero-shot robot manipulation.

RT-2 is a Google DeepMind VLA model transferring web-scale knowledge to robotic control via VLM fine-tuning.

Octo is an open-source generalist robot policy — a transformer-based diffusion policy trained on 800k robot trajectories from Open X-Embodiment, available in 10M/27M/93M parameter variants.

Open X-Embodiment is a Google DeepMind-led collaborative dataset spanning 22 robot embodiments, 34 institutions, and over 500 skills — the foundational dataset for training generalist VLA models.

DROID is a large-scale in-the-wild robot manipulation dataset with 87k demonstrations across 564 scenes, serving as a key benchmark for real-world VLA generalization.

BridgeData V2 is a large-scale robot manipulation dataset from UC Berkeley RAIL Lab with 60,000+ WidowX trajectories across diverse scenes for training generalist robot policies.

NVIDIA Isaac GR00T is an open reference platform and development project for general-purpose humanoid robotics, combining open foundation models (GR00T N1 series), data pipelines, simulation frameworks, CUDA-X accelerated runtime libraries, and NVIDIA Jetson Thor for real-time robot inference and control.

LeRobot is an open-source robotics library from Hugging Face providing state-of-the-art policies, standardized datasets, and hardware interfaces for democratizing robot learning.

Isaac Lab is an open-source GPU-accelerated robot learning framework from NVIDIA on Isaac Sim, with 30+ environments and 16+ robots for RL, IL, and motion planning.

AI Habitat is a high-performance 3D simulation platform from Meta FAIR for training embodied AI agents in photorealistic environments.

Genesis is an open-source, GPU-accelerated universal robotics simulation platform combining multi-physics engine, photorealistic renderer, and generative data engine.

MuJoCo (Multi-Joint dynamics with Contact) is a general-purpose physics simulator by Google DeepMind with 13,900+ GitHub stars, widely used for robotics and embodied AI research.