NVIDIA
NVIDIA A100
Ampere-generation data-center accelerator that became a major pre-Hopper standard for AI training and HPC, with Multi-Instance GPU support on supported configurations.
- Architecture
- Ampere
- Memory
- 80 GB HBM2e
- Memory bandwidth
- Over 2 TB/s
- Release
- Introduced 2020
Main use cases
AI training, inference, HPC and data analytics
Cloud pricing
Full tracker →No tracked rates for this accelerator yet.
Still widely available in cloud fleets; compare against newer H100/H200 pricing because older hardware is not always cheaper.
NVIDIA A100 in the news
GPU news →No coverage yet.