AWS Cost Simulator

for Triton Server ML Architecture

Estimate hosting costs for NVIDIA Triton Inference Server workloads on GPU instances

Instance Type for ML/Triton Server

4 vCPU | 16 GB RAM | 1 NVIDIA T4 GPU (16GB) | 16 Gbps Network

Instance Count

1 5 10
5

Storage (GB)

50GB 500GB 1TB
100 GB

Data Transfer (GB/mo)

10GB 50GB 1TB
100 GB

User Count

1 10 25
7

Purchase Option

On-Demand 1-Yr RI 3-Yr RI
On-Demand

Cost Breakdown

Fetching AWS pricing data...
Using estimated pricing data
EC2 GPU Instances:
$383.95/month
Load Balancer:
$16.43/month
EBS Storage (GP3):
$8.00/month
Data Transfer:
$9.00/month
User Licenses (7 users):
$209.93/month
NVIDIA GPU Software:
$0.00/month

Total Estimated Cost:

$627.34/month

Cost Projection (12 months)

About NVIDIA Triton Inference Server

NVIDIA Triton Inference Server simplifies deployment of AI models at scale in production. This calculator estimates AWS resources needed for hosting Triton Server with various GPU options.