Nvidia launches generative physical AI platform to transform factory automation at CES

Nvidia introduced a new generative physical AI platform at CES in Las Vegas, focusing on automation in factories and warehouses. This technology aims to enhance robots' ability to understand and interact with three-dimensional environments, moving beyond traditional language and image models. The company unveiled its "three computer solution," which includes Nvidia DGX for AI training, Nvidia Omniverse for simulation, and the new Nvidia Cosmos for generating synthetic data. Cosmos will help developers create realistic training scenarios for robots and autonomous vehicles more efficiently. Nvidia also launched the Nemotron model family for agentic AI, designed to optimize performance and reduce costs. New AI blueprints were introduced, including tools for converting documents into podcasts and analyzing video data for better search and reporting capabilities.