NVIDIA Physical AI Dataset
NVIDIA'/%3e%3cdefs%3e%3cpattern%20id='pattern0_1240_293949'%20patternUnits='userSpaceOnUse'%20patternTransform='matrix(11.2%200%200%2011.2%203.93617%200)'%20preserveAspectRatio='none'%20viewBox='0%200%2016%2016'%20width='1'%20height='1'%3e%3cg%20id='pattern0_1240_293949_inner'%3e%3crect%20width='16'%20height='16'%20fill='%23f88376'/%3e%3crect%20x='6'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='6'%20y='8'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='4'%20y='2'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='4'%20y='10'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='2'%20y='4'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='2'%20y='12'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20y='6'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20y='14'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='14'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='14'%20y='8'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='12'%20y='2'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='12'%20y='10'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='10'%20y='4'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='10'%20y='12'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='8'%20y='6'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3crect%20x='8'%20y='14'%20width='2'%20height='2'%20fill='%23F7F6F6'/%3e%3c/g%3e%3c/pattern%3e%3c/defs%3e%3c/svg%3e)
NVIDIA's GR00T X-Embodiment Sim dataset, the most downloaded release in its Physical AI robotics collection, holds simulated trajectories used to post-train GR00T N1: bimanual Panda arms, the GR1 humanoid at tabletop tasks, single-arm kitchen manipulation and Unitree G1 loco-manipulation, published on the Hugging Face Hub.
Scoped to the GR00T X-Embodiment Sim release. NVIDIA publishes about twenty other Physical AI robotics datasets, several of them under a non-commercial license, and none of them is scored here.
Openness
5 high confidence- license
- cc-by-4.0(card metadata)
- access
- public(ungated on the Hub)
- dataset_card
- present
Downloadable from the Hub without a sign-up, under CC BY 4.0, with a card listing each subset. Other datasets in NVIDIA's Physical AI collection carry different terms.
- https://huggingface.co/api/datasets?author=nvidia&search=PhysicalAI-Robotics&sort=downloads&limit=50&expand[]=downloads&expand[]=cardData&expand[]=lastModified&expand[]=gated recorded 2026-09-26
Hub listing author=nvidia search=PhysicalAI-Robotics: 22 datasets; GR00T-X-Embodiment-Sim has the most downloads; several others, such as GR00T-Teleop-Sim, are cc-by-nc-4.0.
- https://huggingface.co/api/datasets/nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim?expand[]=downloads&expand[]=cardData&expand[]=lastModified&expand[]=gated recorded 2026-09-26
Hub API: gated false, license cc-by-4.0.
- https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim/raw/main/README.md recorded 2026-09-26
Card front matter: license: cc-by-4.0; card lists every subset with its trajectory count and download commands.
Adoption
5 medium confidenceHugging Face downloads of the GR00T X-Embodiment Sim repository. Loaders fetch its many subsets file by file, so one training run can count many times.
- https://huggingface.co/api/datasets/nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim?expand[]=downloads&expand[]=cardData&expand[]=lastModified&expand[]=gated recorded 2026-09-26
Hub API: 1,267,768 downloads in the trailing 30 days.
Capability
4 low confidenceThe collection spans humanoids, two-armed and single-arm robots, which few datasets here do, and at hundreds of thousands of trajectories it is large. It is simulated rather than recorded on real robots, and covers a handful of robot bodies against Open X-Embodiment's twenty-two.
- https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim/raw/main/README.md recorded 2026-09-26
Card: "Cross-embodied bimanual manipulation: 9k trajectories"; "Humanoid robot tabletop manipulation: 240k trajectories"; "Robot Arm Kitchen Manipulation: 72K trajectories"; "Unitree G1 Loco-Manipulation: 102 trajectories"; "used for post-training of GR00T N1".
Verified 2026-09-26