NVIDIA is Monday opens a bunch of models, new libraries and other infrastructure for the robotic, the most famous language of the vision “reason” reason “models and robots.
Also join a group that exists Model World Cosmos Whether Transfer Cosmos-2, which can speed up the generation of synthetic data from the 3d simulation scene or spatial control input to move more optimized for speed.
During the announcement at the conference Sigrg in Monday, Nvidia noted that the model should be used to create text data, image and synthetic pictures and agent AI.
The reason for the cosmos, per nvidia, allows the robot agent and ai “reason” thanks to memory and physical understanding, which allows “a planning model for the reason you can do.” The company can be used for lower data, robot plans, and analytics video.
This company also faced the nerve reconstruction of the nerve, which includes one for rendering techniques that allow the simulate developers provide the real world in 3D using Sensor data. The rendering capability is also integrated into open the Carla simulator, a popular development platform. In fact there are updates to the Omniverse Software Authenticate.
There is a new server for the robotic workflow as well. The Blackwell Nvidia Server RTX Pro offers a single architecture for workloads with robotic, while NVIDIA DGX Cloud is a cloud-based management platform.
The announcement is a semiconductor giant undergo robotics because it appears in the case of a large usage for GPUs exceeded AI AI AI
TechCrunch Events
San Francisco
|
October 27, 2025