Cloud ProviderEst. 2006
AWS Logo

AWS GPU Pricing

The dominant force in global cloud computing with deep integration of GPUs into its ecosystem for machine learning and other services.

23 Instances2 Regions19 Available

Lowest Price

$0.53/hr

NVIDIA Tesla T4

Live Inventory

Real-time GPU availability on AWS

23 offers available
AWS
AWS
Virginia
NVIDIA Tesla T4
16GB VRAM
4 vCPU
16GB RAM
$0.53/GPU/hr
AWS
AWS
Virginia
NVIDIA Tesla T4
16GB VRAM
8 vCPU
32GB RAM
$0.75/GPU/hr
AWS
AWS
Virginia
NVIDIA Tesla T44x
16GB VRAM
48 vCPU
192GB RAM
$0.98/GPU/hr
$3.91/hr total (4×)
AWS
AWS
Virginia
NVIDIA RTX A6000
48GB VRAM
4 vCPU
16GB RAM
$1.01/GPU/hr
AWS
AWS
Virginia
NVIDIA Tesla T4
16GB VRAM
16 vCPU
64GB RAM
$1.20/GPU/hr

GPU Models on AWS

AWS offers 7 GPU models. Select one for detailed pricing.

About AWS

Founded 2006
Seattle, Washington

Best For

Large-scale enterprises requiring deep integration with other cloud servicesOrganizations needing globally redundant availability zones

Unique Features

  • Proprietary silicon like Trainium and Inferentia chips
  • Fully managed ML development environment with SageMaker

Limitations

  • High cost relative to specialized clouds
  • Complexity of pricing including egress fees

Platform Features

Access Methods
FeatureAvailable
SSH Access
Jupyter Notebooks
Web Terminal
API Access
Kubernetes
Containers
Billing Options
FeatureAvailable
Billing Incrementper-second
Spot Instances
Reserved Instances
Prepaid Credits
Compliance
CertificationCertified
SOC 2
HIPAA
GDPR
ISO 27001
Support
FeatureAvailable
SLA Guarantee
Enterprise Support
Discord Community

AWS Pricing Breakdown

AWS offers 7 GPU types with on-demand pricing ranging from $0.53/hr for the Tesla T4 to $12.29/hr for the H100 SXM5. 19 of 23 instances are currently in stock across 2 regions.

Billing on AWS is per per-second. Spot instances are available for workloads that can tolerate interruption, typically at 50-80% lower cost than on-demand rates. Reserved capacity is available for committed usage, offering lower rates in exchange for a fixed term. AWS also accepts prepaid credits.

Alternatives to AWS

GPUPerHour tracks over 25 cloud GPU providers. AWS starts at $0.53/hr for the Tesla T4, but pricing and availability vary across providers. The best alternative depends on your requirements: GPU model, region, compliance certifications, and billing flexibility all differ between providers.

For head-to-head analysis, see: AWS vs Cirrascale, AWS vs CoreWeave, AWS vs Crusoe. The comparison tool covers all provider matchups.

Frequently Asked Questions

How much does AWS charge for GPUs?

AWS GPU pricing starts at $0.53/hr for the Tesla T4 and goes up to $12.29/hr for the H100 SXM5. Billing is per per-second. Prices are updated every 60 seconds on GPUPerHour.

What GPUs are available on AWS?

AWS offers 7 GPU types with 23 total instances. The most common models are: RTX A6000 (8 offers), Tesla T4 (6 offers), Tesla V100 16GB (3 offers), A100 SXM4 40GB (2 offers), H100 SXM5 (2 offers).

Does AWS offer spot pricing?

Yes, AWS offers spot instances with discounted pricing. Spot instances can be reclaimed when demand is high, making them best suited for fault-tolerant workloads like batch training or data preprocessing.

What regions does AWS operate in?

AWS has GPU instances available across 2 regions. Region selection affects both latency and pricing. Use the filters on GPUPerHour to find instances in a specific location.

How does AWS compare to alternatives?

AWS prices start at $0.53/hr, with 7 GPU types across 2 regions. GPUPerHour tracks pricing from over 25 cloud GPU providers, making it possible to compare AWS against alternatives on price, GPU selection, and availability in real time.

Does AWS have compliance certifications?

AWS holds SOC 2, HIPAA, GDPR, ISO 27001 certifications. Compliance certifications are important for organizations handling sensitive data or operating in regulated industries.

AWS GPU Pricing: 7 GPUs from $0.53/hr (2026) | GPUPerHour