Parallel computing platform and programming model for GPU-accelerated applications across HPC, AI, and graphics workloads.
CUDA is NVIDIA's general-purpose parallel computing platform and programming model that lets developers use NVIDIA GPUs for compute-intensive workloads spanning high-performance computing, artificial intelligence, scientific simulation, and accelerated graphics.
The CUDA Toolkit bundles a C/C++ compiler, GPU-accelerated libraries, runtime, and Nsight debugging and profiling tools, and extends to higher-level surfaces such as CUDA Tile C++ and cuTile Python for tile-based programming across data center, workstation, edge, and embedded targets.
| Attribute | CUDA | Blackhole, Tenstorrent Galaxy, TT-QuietBox | Metal | ROCm | Snapdragon, Snapdragon Digital Chassis | |
|---|---|---|---|---|---|---|
| Provider | ||||||
| Founded | 1993 | 2016 | 1976 | 1969 | 1985 | |
| Sells To | Consumers, Enterprise | Enterprise | Enterprises, Individuals | Consumers, Enterprise | Consumers, Enterprise | |
| Pricing Model | Licensing, Product Sales, Subscription | Transactional | Services, Subscription | Software, Transactional | Licensing, Recurring, Transactional | |
| Ownership | Public | Venture-backed private company | Public | Public | Public |
Provider NVIDIAnvidia.com$24.1B raised · Public
Founded 1993
Sells To Consumers, Enterprise
Pricing Model Licensing, Product Sales, Subscription
Ownership Public
Provider 
Founded 2016
Sells To Enterprise
Pricing Model Transactional
Ownership Venture-backed private company
Provider 
Founded 1976
Sells To Enterprises, Individuals
Pricing Model Services, Subscription
Ownership Public
Provider 
Founded 1969
Sells To Consumers, Enterprise
Pricing Model Software, Transactional
Ownership Public
Provider Qualcommqualcomm.comPublic
Founded 1985
Sells To Consumers, Enterprise
Pricing Model Licensing, Recurring, Transactional
Ownership Public
arm.com

GeForce
nvidia.com
HGX
nvidia.com
DGX
nvidia.com