--- format: "aidr-story-markdown/v1" id: "7c215b8d5f1890e1d3ecd603e862941d322da516d71275befd9d7386c1f0e99e" canonical_url: "https://aidr.today/7c215b8d?lang=en" title: "Nvidia Launches GR00T 2 Foundation Model for Humanoids and Grippers" lang: "en" requested_lang: "en" available_langs: ["en","vi"] translation_fallback: null fallback_fields: [] published_at: "2026-11-09T23:22:46.000Z" category: "Models" topics: ["nvidia","agent"] source_urls: ["https://huggingnews.com/ai/nvidia-launches-gr00t-2-foundation-model-for-humanoids-and-grippers-ba843165","https://marketbrief.now/ai/nvidia-launches-gr00t-2-foundation-model-for-humanoids-and-grippers-ba843165"] summary: "Nvidia launched GR00T 2, a robotics foundation model designed to be universal for diverse hardware including dexterous hands, parallel grippers, and humanoids. The model uses a new world-action model that predicts physical outcomes of movements to generate precise actions. Training utilized a cluster of 2,000 GPUs, including GB300 NVL72 and H100 chips, with 10,000 vCPUs for data processing." --- # Nvidia Launches GR00T 2 Foundation Model for Humanoids and Grippers > [Open the canonical story]() **Published:** 2026-11-09T23:22:46.000Z **Category:** Models **Topics:** nvidia, agent ## Summary Nvidia launched GR00T 2, a robotics foundation model designed to be universal for diverse hardware including dexterous hands, parallel grippers, and humanoids\. The model uses a new world\-action model that predicts physical outcomes of movements to generate precise actions\. Training utilized a cluster of 2,000 GPUs, including GB300 NVL72 and H100 chips, with 10,000 vCPUs for data processing\. ## Sources - [Story source]() - [Story source]()