Nvidia is now a heterogeneous AI infrastructure platform company, tech analyst Patrick Moorhead said after hearing announcements at Nvidia\'s annual GTC 2026 conference. Nvidia unveiled its new Vera Rubin AI infrastructure platform, including chips and rack systems designed for mass production. It also highlighted integration with Groq inference hardware and growing demand for its Vera CPU. Nvidia detailed software moves with Dynamo 1.0 and a security-enhanced NeMo stack, and cited adoption across cloud, enterprise and robotics.

Nvidia will unveil at GTC 2026 a Physical AI full-stack strategy spanning industrial automation, autonomous driving, edge computing and space. It introduced a Physical AI Data Factory Blueprint built on the Cosmos world model and Osmo orchestrator to link data generation, simulation, evaluation and deployment. Microsoft Azure and Nebius are first cloud adopters. Nvidia also announced robotics and autonomous driving software updates and a space-optimised Vera Rubin Space Module aimed at real-time orbital operations.