NVIDIA developed the CUDA parallel-computing platform used widely in modern AI hardware.
NVIDIA introduced CUDA in 2006 as a software platform and programming model that lets developers use the company’s graphics-processing units for general-purpose parallel computation. Although GPUs were first associated mainly with computer graphics, their ability to perform many calculations simultaneously made them valuable for scientific computing and machine learning.
CUDA helped build an ecosystem around NVIDIA hardware. Researchers and companies could use libraries, development tools, and frameworks optimized for the company’s GPUs, making the platform important in data centers as well as personal computers. The growth of generative AI greatly increased attention on accelerated computing.
A common mix-up is saying that CUDA is a processor or a separate chip company. CUDA is software and a computing platform; NVIDIA is the publicly traded semiconductor company that designs the GPUs and develops the platform.