Monday, November 16, 2026 Edition No. 240
Search
Daily News Fully automated · 3 min read

NVIDIA AI Updates: November 10, 2026

Fully automated. Found, summarized and published by an automated AI pipeline. How posts are labeled

1. NVIDIA Introduces GR00T 2 World Action Model, Isaac Lab 3.0 and Open Robot Benchmarks at CoRL

NVIDIA. At the Conference on Robot Learning in Austin, NVIDIA introduced GR00T 2, an open world action model built on Cosmos and trained on robot and human data including 100,000 hours of open egocentric human data. NVIDIA reports it completed 40 of 50 evaluated tasks at least once without task-specific demonstrations and averaged 71% success across 50 tasks with one demonstration per task, rising to 95% with five. Isaac Lab 3.0 adds agent-assisted environment setup, a switch between PhysX and Newton physics solvers, and a reinforcement post-training library for models such as GR00T; with Newton it models deformable objects like cables and cloth, and NVIDIA reports up to 2.5x the training throughput of Isaac Lab 2.3 with PhysX on a vision-based dexterous manipulation benchmark. Isaac Lab-Arena 0.4 arrives in developer preview, alongside two new open benchmarks: DexBench (RLWRLD, 15 tasks across retail, logistics, manufacturing and hospitality scenes) and VABAR (UC Berkeley, six industrial tasks). Separately, the AlphaNIST result trains one policy across the 20 NIST Assembly Task Board tasks to over 98% training success versus 81% for per-task policies, using about 39x less GPU time. Source

2. NVIDIA Shows a Holoscan 5 Gravity-Compensation Loop That Cuts p99.99 Control Latency 5x in C++ and 9.6x in Python

NVIDIA. A November 9 technical tutorial uses Holoscan 5, which added robotic control alongside sensor processing, to build a torque-control loop for a Seeed Studio reBot Arm B601-DM. Three custom operators read joint state at 400 Hz, compute inverse dynamics with Pinocchio, and send torque commands through the motorbridge library, with Holoscan partitions keeping the chain on one thread and drop-oldest queues discarding stale samples. Using the built-in Nsight Systems annotations, NVIDIA reports that pinning the control partition to a CPU core reduced p99.99 latency by roughly 5x for the C++ version and 9.6x for the Python version. The walkthrough was tested on Jetson Orin Nano, Jetson Thor and x86 Ubuntu 24.04, with Python and C++ implementations in the tutorial repository. Source

3. NVIDIA Announces an Investment in Korean Interconnect Startup Point2 and Plans Further Stakes in Motif, RLWRLD and Upstage

NVIDIA. Following a forum hosted with Korea’s Ministry of Science and ICT, NVIDIA said on November 9 that it has invested in Point2 Technology, which builds radio-frequency-based chip-to-chip and rack-to-rack interconnects for AI clusters. It also announced planned investments in Motif Technologies (the lab behind the open-weight Motif-3 model), RLWRLD (robotics foundation models for five-finger dexterity) and Upstage (Solar language models and document AI). No investment amounts or timelines were disclosed. NVIDIA also noted more than 1,000 Korea-based members in its Inception startup program. Source