--- format: "aidr-story-markdown/v1" id: "37ade9e0dc04966b9de69087a0d20de9ea7eb504083a7f1779de51cb5329c3e3" canonical_url: "https://aidr.today/37ade9e0?lang=en" title: "Nvidia Releases Training Data and Hardware Designs for New GR00T 2 Robotics Model" lang: "en" requested_lang: "en" available_langs: ["en","vi"] translation_fallback: null fallback_fields: [] published_at: "2026-11-10T16:23:28.000Z" category: "Open Source" topics: ["nvidia","agent","open-source","data-engineering"] source_urls: ["https://huggingnews.com/ai/nvidia-releases-training-data-and-hardware-designs-for-new-gr00t-2-robot-609f5355","https://marketbrief.now/ai/nvidia-releases-training-data-and-hardware-designs-for-new-gr00t-2-robot-609f5355"] summary: "Nvidia released 100,000 hours of robotics training data and development tools for its GR00T 2 foundation model. Nvidia made its hardware designs, agentic framework, and simulation and annotation engines available to the public. The GR00T 2 system runs entirely on a robot's onboard compute for tasks involving more than 20 steps and lasting over five minutes." --- # Nvidia Releases Training Data and Hardware Designs for New GR00T 2 Robotics Model > [Open the canonical story]() **Published:** 2026-11-10T16:23:28.000Z **Category:** Open Source **Topics:** nvidia, agent, open\-source, data\-engineering ## Summary Nvidia released 100,000 hours of robotics training data and development tools for its GR00T 2 foundation model\. Nvidia made its hardware designs, agentic framework, and simulation and annotation engines available to the public\. The GR00T 2 system runs entirely on a robot's onboard compute for tasks involving more than 20 steps and lasting over five minutes\. ## Sources - [Story source]() - [Story source]()