CUDA (Compute Unified Device Architecture) is a closed-source low-level API that interfaces software with NVIDIA GPUs.
CUDA is a major moat for NVIDIA. It’s part of why NVIDIA GPUs command such a premium over other hardware (and are perpetually in short supply).
A few reasons why the monopoly exists:
* Hardware/software synergy. NVIDIA has consistently shipped the fastest hardware) and software. It’s been difficult for other companies to build this flywheel (software companies don’t have the
It sucks. Rocm is the best alternative and theoretically can run cuda tasks but it’s got a huge uphill battle in terms of documentation and industry adoption. It’s gotta be better or significantly more cost effective to get people to switch, and right now its not cost effective enough to have people switch en masse. So the documentation and community will stay small.