
Jensen Huang's keynote at GTC revolves around NVIDIA's advancements in AI infrastructure, highlighting the shift towards AI factories and the importance of token generation. He emphasizes the necessity of vertically integrated, horizontally open computing platforms, showcasing NVIDIA's CUDA and CUDAx libraries. Huang details the evolution of NVIDIA's hardware, from GeForce to the new Vera Rubin and Feynman architectures, designed for agentic AI. He introduces OpenCLAW, an open-source framework for building AI agents, and Nemo Claw, NVIDIA's enterprise-ready version. Huang also discusses NVIDIA's partnerships and initiatives in autonomous vehicles and robotics, including collaborations with Uber and Disney. The keynote stresses the increasing demand for computing power driven by AI inference and the need for efficient, high-throughput token factories.
Sign in to continue reading, translating and more.
Continue