📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $NVIDIA(NVDA.US) NVIDIA Unveils GR00T 2, Isaac Lab 3.0 and New Robotics AI Benchmarks at CoRL👉 𝗞𝗲𝘆…

📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $NVIDIA(NVDA.US) NVIDIA Unveils GR00T 2, Isaac Lab 3.0 and New Robotics AI Benchmarks at CoRL

👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:

➤ 𝗡𝗩𝗜𝗗𝗜𝗔 introduces 𝗚𝗥𝟬𝟬𝗧 𝟮, 𝗜𝘀𝗮𝗮𝗰 𝗟𝗮𝗯 𝟯.𝟬, and open robotics benchmarks.

➤ 𝗚𝗥𝟬𝟬𝗧 𝟮 trained using 𝟭𝟬𝟬,𝟬𝟬𝟬 hours of human egocentric data.

➤ Model completed 𝟰𝟬 𝗼𝗳 𝟱𝟬 tasks without task-specific demonstrations.

➤ Achieved 𝟳𝟭% success with one demonstration, rising to 𝟵𝟱% with five.

➤ 𝗜𝘀𝗮𝗮𝗰 𝗟𝗮𝗯 𝟯.𝟬 delivers up to 𝟮.𝟱𝘅 higher simulation training throughput.

➤ 𝗔𝗹𝗽𝗵𝗮𝗡𝗜𝗦𝗧 achieved over 𝟵𝟴% training success using 𝟯𝟵𝘅 less GPU time.

➤ New 𝗗𝗲𝘅𝗕𝗲𝗻𝗰𝗵 and 𝗩𝗔𝗕𝗔𝗥 benchmarks evaluate industrial robot dexterity.

➤ 𝗙𝗼𝘅𝗰𝗼𝗻𝗻, 𝗦𝗶𝗲𝗺𝗲𝗻𝘀, and 𝗡𝗲𝘅𝘂𝗻𝗶 are testing GR00T 2 applications.

➤ 𝗨𝗖 𝗕𝗲𝗿𝗸𝗲𝗹𝗲𝘆 introduces 𝗚𝗿𝗮𝗽𝗵-𝗮𝘀-𝗣𝗼𝗹𝗶𝗰𝘆 for agent-controlled robotic workflows.

➤ NVIDIA research papers named among 𝘀𝗶𝘅 𝗖𝗼𝗥𝗟 𝗕𝗲𝘀𝘁 𝗣𝗮𝗽𝗲𝗿 finalists.

👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:

➤ Accelerates 𝗽𝗵𝘆𝘀𝗶𝗰𝗮𝗹 𝗔𝗜 adoption across manufacturing, logistics, and industrial automation.

➤ Higher simulation throughput could reduce 𝗿𝗼𝗯𝗼𝘁 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗰𝗼𝘀𝘁𝘀 and development time.

➤ Stronger manipulation capabilities could expand robotics beyond basic movement into 𝗽𝗿𝗲𝗰𝗶𝘀𝗶𝗼𝗻 𝗮𝘀𝘀𝗲𝗺𝗯𝗹𝘆.

➤ Open models and benchmarks strengthen NVIDIA's 𝗿𝗼𝗯𝗼𝘁𝗶𝗰𝘀 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗲𝗰𝗼𝘀𝘆𝘀𝘁𝗲𝗺.

➤ Broader robotics adoption could create long-term demand for 𝗡𝗩𝗜𝗗𝗜𝗔 𝗔𝗜 𝗰𝗼𝗺𝗽𝘂𝘁𝗲.

👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁:

𝗞𝗲𝗻 𝗚𝗼𝗹𝗱𝗯𝗲𝗿𝗴, Professor of Robotics at UC Berkeley and Lead on VABAR and GaP Projects:

“To me, agentic robotics is the most exciting technological development since the internet in 1995.”