Nvidia at the Center of the Generative AI Ecosystem—For Now

Abstract

This column explores several questions behind Nvidia’s extraordinary history and position in the center of the Generative AI ecosystem. Founded in 1993, Nvidia developed graphical processing units (GPUs) for gaming on PCs and other devices. The architecture of these devices from 2006 -- many simple compute cores running at very high speeds in parallel -- was very useful for graphics and, it turned out, very useful to power AI/ML software and neural networks in particular. Nvidia also developed proprietary software from 2006 to support these applications, and later optimized that software for Generative AI Large Language Models and inference engines.