AWS EC2 GPU instance price methodology
Ranks a fixed AWS EC2 GPU instance cohort by public Linux on-demand price per physical GPU.
- Version
- 1-us-east-1-linux-gpu-v1-2026-07-27
- Published
- Jul 24, 2026
- Live boards
- 1
- Method ID
- aws-ec2-gpu-instance-prices
Where the facts come from
- Source
ec2-ondemand-without-sec-sel
- Provider
Amazon Web Services
How the order is produced
public instance USD/hour divided by documented physical GPU count; ascending
When the method reruns
- daily
Boards using this version
Only live, quality-gated boards appear. Each board opens on the exact published table and evidence path.
| Board | Primary measure | Rows | Confidence | Freshness | Board version | Cadence | Evidence |
|---|---|---|---|---|---|---|---|
| AWS EC2 GPU Instance Pricesaws-ec2-gpu-instance-prices | aws-ec2-price-usd-gpu-hour | 9 | 100% | current | 1-us-east-1-linux-gpu-v1-2026-07-27 | daily | Open board → |
Complete measurement ledger
These fields are stored with the methodology version. A changed contract must publish a new version before comparable movement can resume.
Region
us-east-1
Source
ec2-ondemand-without-sec-sel
Currency
USD
Location
US East (N. Virginia)
Provider
Amazon Web Services
Selection
Exactly one public Linux on-demand price-list record per fixed reviewed EC2 GPU instance type.
Source File
Missing Data
Any missing, duplicate, malformed, non-positive, or specification-mismatched instance rejects the full cohort.
Not Measured
- GPU utilization
- GPU hours consumed
- capacity availability
- Spot prices
- Capacity Block prices
- Reserved Instance or Savings Plan discounts
- negotiated prices
Rank Formula
public instance USD/hour divided by documented physical GPU count; ascending
Purchase Option
on-demand shared tenancy
Operating System
Linux
Universe Version
us-east-1-linux-gpu-v1-2026-07-27
Instance Universe
- Vcpus
4
- Gpu Count
1
- Gpu Model
NVIDIA L4
- Entity Slug
aws-ec2-g6-xlarge
- Display Name
AWS EC2 g6.xlarge
- Instance Type
g6.xlarge
- Memory Per Gpu Gb
24
- Instance Memory Gib
16
- Accelerator Specification Url
- https://www.nvidia.com/en-us/data-center/l4/ ↗
- Vcpus
48
- Gpu Count
4
- Gpu Model
NVIDIA L4
- Entity Slug
aws-ec2-g6-12xlarge
- Display Name
AWS EC2 g6.12xlarge
- Instance Type
g6.12xlarge
- Memory Per Gpu Gb
24
- Instance Memory Gib
192
- Accelerator Specification Url
- https://www.nvidia.com/en-us/data-center/l4/ ↗
- Vcpus
4
- Gpu Count
1
- Gpu Model
NVIDIA L40S
- Entity Slug
aws-ec2-g6e-xlarge
- Display Name
AWS EC2 g6e.xlarge
- Instance Type
g6e.xlarge
- Memory Per Gpu Gb
48
- Instance Memory Gib
32
- Accelerator Specification Url
- https://www.nvidia.com/en-us/data-center/l40s/ ↗
- Vcpus
48
- Gpu Count
4
- Gpu Model
NVIDIA L40S
- Entity Slug
aws-ec2-g6e-12xlarge
- Display Name
AWS EC2 g6e.12xlarge
- Instance Type
g6e.12xlarge
- Memory Per Gpu Gb
48
- Instance Memory Gib
384
- Accelerator Specification Url
- https://www.nvidia.com/en-us/data-center/l40s/ ↗
- Vcpus
96
- Gpu Count
8
- Gpu Model
NVIDIA A100 40GB
- Entity Slug
aws-ec2-p4d-24xlarge
- Display Name
AWS EC2 p4d.24xlarge
- Instance Type
p4d.24xlarge
- Memory Per Gpu Gb
40
- Instance Memory Gib
1152
- Accelerator Specification Url
- https://www.nvidia.com/en-us/data-center/a100/ ↗
- Vcpus
96
- Gpu Count
8
- Gpu Model
NVIDIA A100 80GB
- Entity Slug
aws-ec2-p4de-24xlarge
- Display Name
AWS EC2 p4de.24xlarge
- Instance Type
p4de.24xlarge
- Memory Per Gpu Gb
80
- Instance Memory Gib
1152
- Accelerator Specification Url
- https://www.nvidia.com/en-us/data-center/a100/ ↗
- Vcpus
16
- Gpu Count
1
- Gpu Model
NVIDIA H100
- Entity Slug
aws-ec2-p5-4xlarge
- Display Name
AWS EC2 p5.4xlarge
- Instance Type
p5.4xlarge
- Memory Per Gpu Gb
80
- Instance Memory Gib
256
- Accelerator Specification Url
- https://www.nvidia.com/en-us/data-center/h100/ ↗
- Vcpus
192
- Gpu Count
8
- Gpu Model
NVIDIA H100
- Entity Slug
aws-ec2-p5-48xlarge
- Display Name
AWS EC2 p5.48xlarge
- Instance Type
p5.48xlarge
- Memory Per Gpu Gb
80
- Instance Memory Gib
2048
- Accelerator Specification Url
- https://www.nvidia.com/en-us/data-center/h100/ ↗
- Vcpus
192
- Gpu Count
8
- Gpu Model
NVIDIA H200
- Entity Slug
aws-ec2-p5en-48xlarge
- Display Name
AWS EC2 p5en.48xlarge
- Instance Type
p5en.48xlarge
- Memory Per Gpu Gb
141
- Instance Memory Gib
2048
- Accelerator Specification Url
- https://www.nvidia.com/en-us/data-center/h200/ ↗