A100
A100 is a NVIDIA GPU based on the Ampere GA100 architecture, first released in 2020.
Built from BenchGecko data as of April 14, 2026 · updates when the data changes
A100 is a NVIDIA GPU based on the Ampere GA100 architecture, first released in 2020.
Basic
The NVIDIA A100 is an AI GPU made by NVIDIA, codenamed Ampere. Peak throughput is 312 TFLOPS at FP16/BF16. It carries 80 GB of HBM2e with 2.04 TB/s of memory bandwidth. Rated power (TDP) is 400 W. It is built on N7 at TSMC.
Deep
In the BenchGecko dataset it sits in the legacy tier: end of life or being phased out. It is followed by the H100. Disclosed buyers include Meta. Specs come from manufacturer datasheets; the spec card on this page shows the dataset values and their as-of date.
Expert
That is 0.78 FP16 TFLOPS per watt. Memory bandwidth per unit of compute is 6.54 GB/s per FP16 TFLOP: a higher ratio helps memory-bound inference, while peak TFLOPS matter more for compute-bound training. Fabrication: Taiwan. Real throughput depends on the model, precision, batch size and software stack, so compare tokens per second per dollar for your own workload.
Depending on why you're here
- ·A100: 312 TFLOPS FP16/BF16
- ·Process: N7 at TSMC
- ·Tier: legacy
- ·80 GB of HBM2e per chip
- ·Cloud availability in the dataset: AWS, Lambda and RunPod
- ·Compare it with other chips on /hardware/a100
- ·Disclosed buyers: Meta
- ·NVIDIA · fabbed at TSMC
- ·Supply tightness is tracked on /hardware
- ·A100 is a chip built to run AI, made by NVIDIA
- ·It has 80 GB of fast memory so large models fit
- ·Companies buy or rent these to train and serve AI models
Frequently Asked Questions
Read the primary sources
- A100 datasheetwww.nvidia.com