Nvidia created the CUDA parallel-computing platform for its graphics processors.
Nvidia introduced CUDA in 2006, allowing developers to use compatible graphics-processing units (GPUs) for general-purpose computing rather than only rendering images. The platform supports programming languages and libraries that distribute suitable workloads across many GPU cores.
CUDA became especially important for scientific computing, machine learning, and artificial intelligence. Nvidia’s hardware and software ecosystem helped make GPUs central to training and running modern AI models, alongside their original role in computer graphics.
CUDA is sometimes confused with Nvidia’s GPU architecture names, such as Ampere or Hopper. CUDA is the software platform and programming model; those names describe hardware generations. Nvidia’s corporate name is commonly styled NVIDIA in its branding.