ITIT Solution JournalEnterprise IT, reported straight

NVIDIA Expands Cosmos Platform With Physical AI Training and Simulation Tools

New blueprints, model variants, and faster inference target enterprise physical AI development workflows.

IT Solution Journal News Desk

Editorial illustration: A glowing golden mechanical robotic arm and hand floats above an aged blueprint scroll laid on a wooden surface, with two translucent ghost versions of the arm descending between the solid model and the flat technical schematic below, rendered in warm amber light against a dark blue background.

AI-generated editorial illustration.

NVIDIA announced expanded capabilities for its Cosmos platform at GTC Berlin, adding training blueprints, simulation checkpoints, and updated NIM microservices for physical AI development. Cosmos Data Factory now offers five cloud-native blueprints covering autonomous vehicle data, synthetic defect generation, and video augmentation, available on Azure, Nebius, and CoreWeave. Updated Cosmos 3 NIM microservices deliver up to 2x faster inference, per NVIDIA claims. Early adopters include Porsche, which NVIDIA says expects up to 90% fewer physical test drives, and ANYbotics, which generated over 1,500 synthetic training clips. The Metropolis VSS 3.3 blueprint reportedly enables monitoring of 40% more video streams per GPU.

Source: NVIDIA News. Original announcement: October 22, 2026.

Keep up with IT Solution Journal

Practical guidance and new coverage. You can withdraw your permission at any time.

Read our privacy and data-use policy.