AWS EC2 Monthly Cost Calculator
The AWS EC2 Monthly Cost Calculator estimates your on-demand monthly bill for any EC2 instance type by multiplying the published on-demand hourly rate by 730 hours (the AWS standard for one month). For example, a t3.medium in US East (N. Virginia) runs at $0.0416/hr × 730 = ~$30.37/month and delivers 2 vCPUs + 4 GB RAM. Use this tool when budgeting cloud workloads, comparing instance families (T3 vs. M6i vs. C6g), or validating AWS Cost Explorer estimates before committing to a Reserved Instance or Savings Plan.
When to use this calculator
- Estimating the monthly bill for a t3.medium web server running 24/7 in us-east-1 before launching a production environment.
- Comparing the cost of a compute-optimized c6i.xlarge ($0.204/hr → ~$148.92/mo) vs. a general-purpose m6i.xlarge ($0.192/hr → ~$140.16/mo) for a CPU-intensive batch job.
- Calculating region price differences — the same m5.large costs $0.096/hr in us-east-1 but $0.124/hr in ap-southeast-2 (Sydney), a 29% premium.
- Validating whether purchasing a 1-year Standard Reserved Instance (up to 40% discount) is worth it based on the on-demand baseline monthly cost.
- Projecting infrastructure spend for a startup's Series A pitch deck using realistic AWS line-item estimates per instance.
- Determining the break-even point between on-demand and Spot pricing (typically 70–90% cheaper) for fault-tolerant workloads like data pipelines.
Calculation example
- t3 Medium, US East
- USD ~30
How it works
3 min readHow It's Calculated
AWS publishes on-demand pricing per instance type per region. The monthly estimate uses AWS's own 730-hour month convention:
Monthly Cost (USD) = Hourly On-Demand Rate × 730 hours
Example — t3.medium, us-east-1, Linux:
$0.0416/hr × 730 = $30.37/month
Example — m6i.2xlarge, us-west-2, Linux:
$0.384/hr × 730 = $280.32/month> Note: 730 = 365 days × 24 hours ÷ 12 months. AWS Cost Explorer and billing statements also use this convention.
---
Reference Table — Common Instance Types (Linux, On-Demand, us-east-1, 2025)
| Instance | vCPU | RAM (GB) | $/hr | $/month (~730 hr) | Family Use Case |
|---|---|---|---|---|---|
| t3.nano | 2 | 0.5 | $0.0052 | $3.80 | Dev/test, micro apps |
| t3.micro | 2 | 1 | $0.0104 | $7.59 | Low-traffic sites |
| t3.small | 2 | 2 | $0.0208 | $15.18 | Small web servers |
| t3.medium | 2 | 4 | $0.0416 | $30.37 | General-purpose apps |
| t3.large | 2 | 8 | $0.0832 | $60.74 | Mid-traffic workloads |
| t3.xlarge | 4 | 16 | $0.1664 | $121.47 | Business apps |
| t3.2xlarge | 8 | 32 | $0.3328 | $242.94 | Large web backends |
| m6i.large | 2 | 8 | $0.096 | $70.08 | General-purpose |
| m6i.xlarge | 4 | 16 | $0.192 | $140.16 | App servers |
| m6i.2xlarge | 8 | 32 | $0.384 | $280.32 | Databases, caches |
| m6i.4xlarge | 16 | 64 | $0.768 | $560.64 | In-memory workloads |
| c6i.large | 2 | 4 | $0.085 | $62.05 | Compute-intensive |
| c6i.xlarge | 4 | 8 | $0.17 | $124.10 | Batch, HPC |
| c6i.2xlarge | 8 | 16 | $0.34 | $248.20 | Video encoding, ML |
| r6i.large | 2 | 16 | $0.126 | $91.98 | Memory-intensive DBs |
| r6i.xlarge | 4 | 32 | $0.252 | $183.96 | SAP, Redis, Memcached |
| r6i.2xlarge | 8 | 64 | $0.504 | $367.92 | Large in-memory stores |
| g4dn.xlarge | 4 | 16 | $0.526 | $383.98 | ML inference, GPU |
| p3.2xlarge | 8 | 61 | $3.06 | $2,233.80 | Deep learning training |
Prices as of mid-2025. Always verify at aws.amazon.com/ec2/pricing.
---
Typical Use Cases (with real numbers)
Case 1 — Small SaaS Web App
A startup runs one t3.medium (app server) + one t3.small (background worker) in us-east-1:
Case 2 — Data Processing Pipeline
A data engineering team uses a c6i.2xlarge for nightly ETL jobs that run ~200 hours/month (not 24/7):
Case 3 — Memory-Optimized Database
A PostgreSQL replica on r6i.xlarge running 24/7 in eu-west-1 ($0.2772/hr):
---
Common Mistakes
1. Using 720 instead of 730 hours — AWS bills on 730 hours/month (~30.42 days average). Using 720 underestimates by ~1.4%, which compounds across many instances.
2. Ignoring region multipliers — Prices vary significantly: us-east-1 is the cheapest AWS region. The same t3.medium costs $0.0464/hr in eu-west-1 (London) vs. $0.0416 in us-east-1 — an 11.5% difference.
3. Forgetting that EC2 pricing covers the instance only — EBS storage, Elastic IP addresses, NAT Gateway, data transfer, and load balancers are billed separately. A "t3.medium at $30/mo" server realistically costs $50–80/month fully loaded.
4. Confusing T3 burstable credits with constant compute — T3 instances earn CPU credits over time and burst above baseline. Under sustained 100% CPU load, a t3.medium is throttled to a 20% CPU baseline unless "unlimited" mode is enabled (which adds cost).
5. Comparing on-demand to Reserved without accounting for upfront payment time value — A 1-year All Upfront RI requires paying ~$263 today for a t3.medium vs. $30.37/mo × 12 = $364.44 on-demand. The savings are real (~28%), but capital is locked for 12 months.
6. Assuming Windows costs the same as Linux — Windows Server licensing is baked into the EC2 price. A t3.medium Windows instance costs $0.0544/hr in us-east-1 vs. $0.0416 for Linux — a 31% premium.
---
Related Calculators
Explore more tools on Hacé Cuentas to complete your cloud budget analysis. (Related calculators coming soon in the cloud cost category.)
Frequently asked questions
How many hours per month does AWS use to calculate monthly EC2 costs?
AWS uses 730 hours as the standard monthly equivalent (365 days × 24 hours ÷ 12 months = 730.42, rounded to 730). This is the same figure used in AWS Cost Explorer, the Pricing Calculator, and your monthly bill. Using 720 hours will slightly underestimate your cost.
What is the monthly cost of a t3.medium in us-east-1?
A t3.medium (2 vCPU, 4 GB RAM) in us-east-1 on Linux on-demand pricing costs $0.0416/hr × 730 = $30.37/month. For Windows, the rate is $0.0544/hr → $39.71/month. These figures cover the instance compute only — EBS, network, and other services are billed separately.
How much cheaper are Reserved Instances compared to on-demand?
AWS Reserved Instances (RIs) offer up to 72% savings vs. on-demand for 3-year All Upfront commitments, and about 40% savings for 1-year All Upfront. A t3.medium on-demand costs ~$365/year; a 1-year Standard RI (All Upfront) in us-east-1 costs approximately $263 upfront — saving ~$102/year. Convertible RIs offer slightly less savings (~31–54%) but allow instance family changes.
How do Spot Instance prices compare to on-demand for EC2?
Spot Instances typically cost 60–90% less than on-demand pricing, but can be interrupted by AWS with a 2-minute warning when capacity is needed. For a t3.medium in us-east-1, Spot prices have historically ranged from $0.0060–$0.0125/hr (vs. $0.0416 on-demand). They are ideal for batch jobs, CI/CD pipelines, and stateless, fault-tolerant workloads.
Does the region affect EC2 pricing significantly?
Yes — region is one of the biggest pricing variables. us-east-1 (N. Virginia) is consistently the cheapest AWS region for EC2. The same m5.large costs $0.096/hr in us-east-1, $0.111/hr in us-west-2 (Oregon, +16%), and $0.124/hr in ap-southeast-2 (Sydney, +29%). Running workloads in Asia-Pacific or South America regions can add 20–40% to your EC2 bill compared to US East.
Are T3 instances always the cheapest option for low-traffic workloads?
T3 instances are generally the cheapest for variable, low-to-moderate CPU workloads due to their burstable credit model. However, if your workload sustains high CPU usage (>20–40% baseline), T3 instances in 'unlimited' mode can become more expensive than a fixed-performance M6i or C6i instance. For steady-state workloads above the CPU baseline, M6i or C6g (Graviton) instances often offer better price-performance.
What additional costs should I budget beyond the EC2 instance hourly rate?
The EC2 hourly rate covers compute only. A realistic total cost typically adds: EBS storage (gp3 = $0.08/GB-month, so 50 GB = $4/mo), data transfer out (first 100 GB/mo free, then $0.09/GB), Elastic IP ($0.005/hr if not attached to a running instance), and NAT Gateway ($0.045/hr + $0.045/GB). A 't3.medium at $30/mo' can realistically reach $55–80/month fully loaded.
What is the difference between t3 and t3a instance types?
t3a instances use AMD EPYC processors instead of Intel Xeon and are priced approximately 10% lower than equivalent t3 instances. For example, a t3a.medium costs $0.0376/hr vs. $0.0416/hr for t3.medium in us-east-1 — saving about $2.92/month per instance. Performance is comparable for most general-purpose workloads, making t3a a straightforward cost optimization when processor vendor is not a constraint.
How does AWS Graviton (C6g, M6g, R6g) compare in cost to Intel-based instances?
AWS Graviton3-based instances (e.g., c6g, m6g, r6g) are priced ~10–20% cheaper than equivalent Intel-based instances and often deliver 20–40% better price-performance for supported workloads. An m6g.large costs $0.077/hr vs. $0.096/hr for m6i.large in us-east-1 — a 20% cost reduction. They require ARM-compatible software (Linux works well; most open-source stacks support ARM64 natively).