Nvidia developed the CUDA platform, helping make it a leading artificial-intelligence chipmaker.
CUDA is Nvidia's parallel-computing platform and programming model. Nvidia introduced CUDA in 2006, allowing software developers to use compatible graphics-processing units for general-purpose calculations rather than only for graphics.
That capability became especially valuable for machine learning, where large numbers of calculations can be performed in parallel. Nvidia's GPUs, software libraries, and developer tools became foundational to many training and inference systems used in modern artificial intelligence.
CUDA is not a chip and is not an operating system. It is a software platform tied closely to Nvidia's GPU ecosystem. Competitors such as AMD and Intel produce their own accelerator hardware and software tools, but CUDA remains one of Nvidia's most important competitive advantages and a major reason investors have assigned the company an exceptionally high market capitalization.