Mark RadarMARK RADAR
中 EN
Event File AI NVIDIA Edge Computing

NVIDIA Launches Cosmos 3 Edge for On-Device Physical AI

1 reports · First detected 2026-07-21 · Last active 2026-07-21

World models allow machines to interpret their surroundings, predict how scenes may evolve and select actions, making them a key building block for robotics, autonomous vehicles and smart infrastructure. NVIDIA’s Cosmos 3 platform spans text, images, video, ambient sound and action generation. Cosmos 3 Edge targets local deployment, where reducing reliance on cloud or data-center processing can improve latency, privacy and operational resilience for physical AI systems.

NVIDIA released Cosmos 3 Edge on July 20, 2026, as a 4-billion-parameter open model with a 2-billion-parameter reasoning module that can run independently. On Jetson Thor, it supports a 15-Hz real-time control loop, generating 32 actions per inference at 640-by-360 resolution. The model is optimized for Jetson, RTX PRO, GeForce RTX and DGX systems. Its weights are available on Hugging Face, while NVIDIA has published inference and post-training frameworks and recipes on GitHub.

All Coverage

1 original reports
MARKTECHPOST.COM 2026-07-21
NVIDIA Releases Cosmos 3 Edge: A 4B-Parameter Open World Model That Reasons and Generates Robot Actions On-Device

The Backstory

The history behind this event
NVIDIA Launches Cosmos 3 and Nemotron 3 Ultra to Advance Open-Source Physical AI Ecosystem2026-06-05 · 2 reports · similarity 0.81

Physical AI systems must enable robots, autonomous vehicles and visual agents to understand their surroundings and take action, but development has long been constrained by limited real-world training data and fragmented simulation tools. NVIDIA is making the weights, data and training recipes for its world models and agentic reasoning models openly available, seeking to lower development barriers and bring GPUs, DGX Cloud and software tools into a shared ecosystem.

NVIDIA unveiled Cosmos 3 at GTC Taipei on May 31, 2026, with native integration of text, images, video, ambient sound and actions. It also formed the Cosmos Coalition with companies including Runway. On June 4, NVIDIA released Nemotron 3 Ultra, which has 550 billion total parameters and 55 billion active parameters. The model delivers up to five times the inference throughput and can reduce agentic-task costs by as much as 30%.

Mark Radar｜MARK RADAR
Past Issues・by MarkReadFintech
All times are in Taipei time (GMT+8)