Nvidia designed the GPU architecture behind the CUDA parallel-computing platform.
CUDA is a parallel-computing platform and programming model developed by Nvidia. It allows software developers to use compatible Nvidia graphics-processing units for general-purpose calculations rather than limiting them to displaying graphics.
Nvidia introduced CUDA in 2006. The technology became important in scientific computing, machine learning, and artificial intelligence because GPUs can perform many similar mathematical operations simultaneously. CUDA tools include libraries, compilers, and programming interfaces that help developers use Nvidia hardware.
Nvidia’s market capitalization expanded dramatically as demand grew for data-center accelerators used to train and run AI models. CUDA is not a separate chip company and is not the same as a graphics card model; it is Nvidia’s software and hardware computing ecosystem.