Compute Comparison
vs
All comparisons →

Volta

NVIDIA V100 32GB

wins

1

The V100 was NVIDIA's flagship AI GPU before the A100.

Legacy ML trainingBudget HPCMature framework support

Ampere

NVIDIA A100 80GB

wins

4

The A100 remains one of the most widely available data center GPUs. Its broad provider support and mature software ecosystem make it a reliable choice for ML training and inference.

ML trainingInference at scaleHPC

Performance comparison

Visual bars — winner highlighted

NVIDIA V100 32GB

Metric

NVIDIA A100 80GB

FP16 TFLOPS

125 T
312 T

Bandwidth

900 GB/s
2.0k GB/s

VRAM

322 GB
802 GB

TDP (lower=better)

300 W
400 W

TFLOPS/watt

0.42
0.78

GB/s per watt

3.00
5.10

Full specification table

SpecNVIDIA V100 32GBNVIDIA A100 80GB
ArchitectureVoltaAmpere
VRAM32 GB HBM280 GB HBM2e
VRAM typeHBM2HBM2e
Memory bandwidth900 GB/s2,039 GB/s
FP16 TFLOPS125 TFLOPS312 TFLOPS
TDP300 W400 W
TFLOPS/watt0.42 T/W0.78 T/W
GB/s per watt3.00 GB/s/W5.10 GB/s/W
NVLinkYesYes
Release year20182020

Live cloud pricing

On-demand $/hr across providers — updated in real time

Loading prices…

Power efficiency analysis

TFLOPS/watt and GB/s/watt — critical for data center TCO

NVIDIA V100 32GB

FP16 TFLOPS125 TFLOPS
TDP300 W
TFLOPS/watt0.417 T/W
GB/s per watt3.00 GB/s/W

NVIDIA A100 80GB

FP16 TFLOPS312 TFLOPS
TDP400 W
TFLOPS/watt0.780 T/W
GB/s per watt5.10 GB/s/W

TFLOPS/watt measures compute efficiency — how much AI throughput you get per watt of power consumed. For data centers with PUE of 1.2–1.5, a 10% improvement in TFLOPS/watt translates directly to lower electricity costs and cooling requirements. GB/s/watt measures memory bandwidth efficiency, which is the binding constraint for memory-bound LLM inference workloads.

When to choose each GPU

Choose NVIDIA V100 32GB for:

  • Legacy ML training
  • Budget HPC
  • Mature framework support

Choose NVIDIA A100 80GB for:

  • ML training
  • Inference at scale
  • HPC