NVIDIA · Ampere Architecture
Rent NVIDIA A100 SXM (80GB) in the Cloud
Previous-generation workhorse for AI training and inference. Still widely available at lower cost than H100.
VRAM
80 GB HBM2e
Bandwidth
2,039 GB/s
FP16
312.0 TFLOPS
FP32
19.5 TFLOPS
TDP
400W
Architecture
Ampere
Cheapest On-Demand
$1.10/hr
Average On-Demand
$1.67/hr
Providers
6
Compare NVIDIA A100 SXM (80GB) Cloud Pricing — 6 Providers
On-Demand
| Provider | Price / GPU / hr | Availability | Notes | Action |
|---|---|---|---|---|
Vast.ai
United States
|
$1.10/hr CHEAPEST | Available | Marketplace avg | Visit Provider |
Massed Compute
United States
|
$1.23/hr | Available | SXM4 | Visit Provider |
RunPod
United States
|
$1.49/hr | Available | Secure Cloud, SXM | Visit Provider |
Novita AI
United States
|
$1.60/hr | Available | SXM | Visit Provider |
Cherry Servers
Lithuania
|
$2.18/hr | Available | Bare metal | Visit Provider |
Vultr
United States
|
$2.40/hr | Available | PCIe | Visit Provider |
Spot / Preemptible
| Provider | Price / GPU / hr | Availability | Notes | Action |
|---|---|---|---|---|
|
Novita AI
United States
|
$0.79/hr CHEAPEST | Available | Spot, 1hr guaranteed | Visit Provider |
|
RunPod
United States
|
$1.39/hr | Available | Community Cloud | Visit Provider |
Prices last verified: April 13, 2026
NVIDIA A100 SXM (80GB) Technical Specifications
Manufacturer
NVIDIA
Architecture
Ampere
VRAM
80 GB HBM2e
Memory Bandwidth
2,039 GB/s
FP16 (Tensor)
312.0 TFLOPS
FP32
19.5 TFLOPS
TDP
400W
Release Year
2020
Segment
Data center
Memory Type
HBM2e
Best For
Compare with Other GPUs
See how NVIDIA A100 SXM (80GB) stacks up against other popular cloud GPUs in specs, pricing, and availability.