NVIDIA · Kepler
NVIDIA K80 cloud pricing
Compare every cloud instance with NVIDIA K80 GPUs across AWS, Azure: 7 instance types with 1 to 16 K80s. Cheapest on demand: $0.90 per K80-hour on AWS p2.xlarge in US East (N. Virginia).
- on-demand / K80-hour
- $0.90
- spot / K80-hour
- $0.1404
- memory per device
- 12 GB
- instance types
- 7
AWS p2.xlarge
AWS p2.8xlarge
2 clouds
Every NVIDIA K80 instance
cheapest region per instance · Linux · on-demand per GPU-hour = hourly price ÷ GPU count · government regions excluded
| Provider | Instance | GPUs | GPU memory | vCPU | RAM | On-demand /hr | Per GPU-hr | Spot per GPU-hr | Cheapest region |
|---|---|---|---|---|---|---|---|---|---|
| AWS | p2.xlarge | 1 | 12 GiB | 4 | 61 GiB | $0.90 | $0.90 | $0.1578 | US East (N. Virginia) us-east-1 |
| AWS | p2.8xlarge | 8 | 96 GiB | 32 | 488 GiB | $7.20 | $0.90 | $0.1404 | US East (N. Virginia) us-east-1 |
| AWS | p2.16xlarge | 16 | 192 GiB | 64 | 732 GiB | $14.40 | $0.90 | $0.153 | US East (N. Virginia) us-east-1 |
| Azure | NC6prev | 1 | 12 GiB | 6 | 56 GiB | $0.90 | $0.90 | $0.1663 | East US eastus |
| Azure | NC12prev | 2 | 24 GiB | 12 | 112 GiB | $1.80 | $0.90 | $0.1663 | East US eastus |
| Azure | NC24prev | 4 | 48 GiB | 24 | 224 GiB | $3.60 | $0.90 | $0.1663 | East US eastus |
| Azure | NC24rprev | 4 | 48 GiB | 24 | 224 GiB | $3.96 | $0.99 | $0.183 | East US eastus |
Other GPUs
B300 $17.80B200 $14.24RTX PRO 6000 Blackwell $0.5596H200 $7.912Trainium2 $6.447Radeon Pro V710 $1.994H100 NVL $6.98L40S $1.861Inferentia2 $0.7582L4 $0.7045RTX 4000 Ada $0.52Qualcomm Cloud AI 100 $1.115H100 $6.88Trainium $1.344Habana Gaudi $1.639Radeon Pro V620 $2.54A10G $1.006A10 $2.20T4G $0.42A16 $0.471A100 80GB $2.397A40 $1.712A100 40GB $2.745Radeon Pro V520 $0.3785Inferentia $0.228Quadro RTX 6000 $1.50T4 $0.526V100 $3.06P40 $2.07P100 $2.07M60 $0.75GRID K520 $0.65
NVIDIA K80 FAQ
How much does an K80 cost per hour in the cloud?
On demand, from $0.90 per K80-hour (AWS p2.xlarge, $0.90 per hour for 1 K80 in US East (N. Virginia)). Spot capacity starts at $0.1404 per K80-hour on AWS p2.8xlarge.
Which clouds offer K80 instances?
AWS (from $0.90 per GPU-hour, e.g. p2.xlarge); Azure (from $0.90 per GPU-hour, e.g. NC6).
How much memory does the K80 have?
12 GB per GPU (Kepler architecture). Cloud instances combine 1 to 16 of them.