NVIDIA Pascal

NVIDIA P100

Tesla P100 introduced Pascal Tensor performance and HBM2 memory, remaining popular for legacy ML training and HPC workloads.

Launch year
2016
Memory
16 GB HBM2
Memory bandwidth
720 GB/s
Peak FP16 / FP32
21 TFLOPS · 10.6 TFLOPS

Market snapshot

$0.17 /hr

Range $0.17 – $14.46

Catalog coverage

1392 live offerings

Across 2 providers · 52 regions

  • HBM2 memory
  • NVLink connectivity
  • Mixed-precision acceleration
NVIDIA P100 render

Last refreshed Oct 20, 2025, 1:53 AM

Performance snapshot

Normalized versus NVIDIA A100 (=1.0). Values use public reference benchmarks for training and inference workloads.

  • AI Training×0.25
  • AI Inference×0.20
  • Memory Bandwidth×0.30

Provider availability

Price bands per provider pulled from the live catalog.

  • Google Cloud logoGoogle Cloud1272 offers
    $0.17 – $14.46 /hr
  • Oracle Cloud logoOracle Cloud120 offers
    $0.64 – $2.55 /hr

Popular regions

  • asia-east1-a106 offers
  • asia-east1-c106 offers
  • australia-southeast1-b106 offers
  • europe-west4-a106 offers
  • europe-west1-d106 offers
  • europe-west1-b106 offers
  • us-east1-b106 offers
  • us-west1-b106 offers

Weighted average price $3.14 /hr · median $2.36