NVIDIA (NVDA. O) CEO Jen-Hsun Huang announced the launch of a complete line of models based on Llama, using the NVIDIA Llama Nemo Tron language base model called Nemotron. In addition, NVIDIA launched NVIDIA Cosmos, a base model designed to understand the real world, using the NVIDIA Llama Nemo Tron language base model. Cosmos models can take cues from text, images, or video to generate virtual world states as video outputs for the unique needs of autonomous driving and robotics applications. Developers can use Cosmos to generate AI feedback for reinforcement learning to improve policy models and test performance in different scenarios. Nvidia is working with a global developer ecosystem to drive the next wave of advances in physical AI. Cosmos is the world's first foundational model, based on 20 million hours of video training, with a focus on physical dynamics and human interaction.