Nvidia developed the CUDA parallel-computing platform, first released in 2007.
CUDA is a parallel-computing platform and programming model that allows software to use Nvidia graphics-processing units for general-purpose computation. Nvidia publicly introduced CUDA in 2006, and the CUDA software development kit became available in 2007.
The platform helped expand GPUs beyond image rendering and video games. Researchers and engineers could use many GPU cores simultaneously for workloads such as scientific simulation, data analysis, and machine learning. CUDA is not a standalone graphics card or operating system; it is a software ecosystem connecting programs with Nvidia hardware. Its adoption became especially important as artificial-intelligence systems required large amounts of parallel numerical computation.