Jensen Huang

NVIDIA's pivot from gaming GPUs to general-purpose accelerated computing (GPGPU) began in the mid-2000s with CUDA. By 2012, when AlexNet won ImageNet running on two NVIDIA GPUs, Huang's decade-long bet on GPU computing for AI was vindicated. NVIDIA's GPU dominance in AI training is now near-total.

At CES 2023 and GTC keynotes, Huang framed generative AI as a platform shift as significant as the smartphone. His argument: every industry will be reinvented by AI, and accelerated computing is the foundational infrastructure. NVIDIA's market cap went from ~$300B to over $3T on the strength of this thesis.

Announced NVIDIA's major push into physical AI and robotics (Cosmos platform, Isaac Robotics). Argues that the next wave of AI is not language but embodied systems that understand and act in physical space — autonomous vehicles, industrial robots, humanoid robots. NVIDIA positions itself as infrastructure provider for this wave too.

Argues that AI sovereignty — nations having their own computing infrastructure, training data, and AI models — is becoming a national security imperative. NVIDIA has signed deals with governments worldwide on this basis. Positions AI infrastructure like highways or power grids: critical national assets.