Technology

Calculate Your AWS EC2 Monthly Costs

Q: How does AWS actually charge for EC2 — by the hour or by the second?

AWS bills EC2 in **per-second increments** with a 60-second minimum for Linux instances. This means a job that runs for 90 seconds costs the same as 90 seconds, not a full hour. However, for monthly budget planning, multiplying the hourly rate by hours used is accurate within <1% for workloads running more than a few minutes per session.

Q: What is a Reserved Instance and how much does it really save?

A Reserved Instance (RI) is a 1- or 3-year billing commitment to a specific instance type and region in exchange for a discount on the hourly rate. For a 1-year Standard RI with no upfront payment on a Linux instance in us-east-1, discounts average **~36–40%** off on-demand. A 3-year all-upfront Standard RI can save up to **62%**. You can purchase RIs through the AWS Management Console or Marketplace.

Q: What's the difference between Reserved Instances and Savings Plans?

Savings Plans (launched 2019) offer similar or greater discounts (up to 66% for Compute Savings Plans, up to 72% for EC2 Instance Savings Plans) but apply more flexibly — a Compute Savings Plan covers any EC2 instance regardless of family, size, region, or OS. RIs are locked to a specific instance type and region. AWS recommends Savings Plans for most new workloads due to their flexibility.

Q: Why does the same instance type cost different amounts in different AWS regions?

AWS prices vary by region based on local electricity costs, data center operating expenses, real estate, and taxes. us-east-1 (N. Virginia) is consistently the cheapest region. eu-west-1 (Ireland) runs ~10–15% higher, ap-southeast-1 (Singapore) ~15–20% higher, and ap-east-1 (Hong Kong) can be 30%+ more expensive than us-east-1 for the same instance type.

Q: Are there additional costs beyond the EC2 hourly compute price?

Yes — the compute price is often just 50–70% of your total EC2-related bill. Common additional charges include: **EBS storage** ($0.08–$0.10/GB-month for gp3), **data transfer out** ($0.09/GB for first 10 TB/month), **Elastic IP addresses** ($0.005/hr when not attached to a running instance), **NAT Gateway** ($0.045/hr + $0.045/GB processed), and **load balancer** hours. Always model these alongside compute.

Q: How do Spot Instances compare in cost to on-demand pricing?

Spot Instances use spare AWS capacity and can be **60–90% cheaper** than on-demand prices, but AWS can reclaim them with a 2-minute warning. For example, a `c6i.xlarge` on-demand costs $0.17/hr ($122.40/month 24/7); a Spot price for the same instance in us-east-1 has historically averaged around $0.05–$0.07/hr (~$36–$50/month). Spot is ideal for fault-tolerant batch jobs, CI/CD pipelines, and stateless web tiers.

Q: How accurate is the 40% Reserved Instance savings figure used in this calculator?

The **40% figure** is a conservative approximation for a **1-year Standard RI, no-upfront, Linux, us-east-1**. Actual discounts range from ~20% (1-year Convertible, no upfront) to ~62% (3-year Standard, all upfront). The exact discount depends on instance family, payment option, term length, operating system, and region. Always verify current RI pricing directly in the AWS Pricing Calculator at calculator.aws before making a commitment.

Q: What instance type should I start with for a basic web application?

AWS and the broader cloud community recommend starting with the **t3.medium** (2 vCPU, 4 GB RAM, $0.0416/hr ≈ $30/month 24/7) for small web apps or APIs. It supports CPU bursting, so it handles traffic spikes gracefully at a low base cost. If your app consistently pegs CPU above 40%, graduate to an `m6i.large` (2 vCPU, 8 GB RAM, $0.096/hr ≈ $69/month) for a fixed performance baseline without burst credits.

Calculator Free · Private

Was this calculator helpful?

This calculator estimates your monthly and annual AWS EC2 instance costs based on the on-demand hourly price, daily usage hours, and active days per month. The core formula is: Monthly Cost = Price/hr × Hours/day × Days/month. For example, a t3.medium at $0.0416/hr running 24/7 costs roughly $30/month on-demand. Use this tool when budgeting cloud infrastructure, comparing instance families, or evaluating whether a 1-year or 3-year Reserved Instance (RI) commitment — which saves up to 40–60% — makes financial sense for your workload.

Last reviewed: April 17, 2026 Verified by Martín Rodríguez Source: AWS EC2 On-Demand Pricing — Official AWS Pricing Page 100% private

When to use this calculator

Calculating the monthly bill for a production web server running 24/7 on a t3.large ($0.0832/hr) before committing to a Reserved Instance
Estimating costs for a nightly batch-processing job that runs only 6 hours/day on a c6i.2xlarge ($0.34/hr) to justify spot instance usage
Comparing total annual spend across three instance types (t3, m6i, c6i) to right-size a staging environment and cut cloud spend by 30%+
Building a startup's 12-month infrastructure budget forecast by projecting EC2 costs across dev, staging, and production tiers before a Series A pitch

Calculation example

$0.10 × 24 × 30
$72

Result: $72

How it works

3 min read

How It's Calculated

The fundamental formula powering this calculator is:

Monthly Cost ($) = Hourly Price ($/hr) × Hours per Day (hr/day) × Days per Month (days/mo)
Annual Cost ($)  = Monthly Cost × 12
Reserved Saving  = Monthly Cost × 0.40   // 1-year Standard RI average discount
RI Monthly Cost  = Monthly Cost × 0.60

AWS bills EC2 in per-second increments (with a 60-second minimum), but for monthly budgeting purposes, the hourly × hours × days model is the industry-standard approximation and matches AWS Cost Explorer projections within <1% for always-on workloads.

---

Reference Table — Common EC2 On-Demand Prices (us-east-1, Linux, 2025)

Instance Type	vCPU	RAM (GB)	On-Demand ($/hr)	30-day 24/7 Cost	1-yr RI ($/hr)	Monthly RI Cost
t3.micro	2	1	$0.0104	$7.49	$0.0063	$4.54
t3.small	2	2	$0.0208	$14.98	$0.0126	$9.07
t3.medium	2	4	$0.0416	$29.95	$0.0252	$18.14
t3.large	2	8	$0.0832	$59.90	$0.0504	$36.29
t3.xlarge	4	16	$0.1664	$119.81	$0.1008	$72.58
m6i.large	2	8	$0.096	$69.12	$0.058	$41.76
m6i.xlarge	4	16	$0.192	$138.24	$0.116	$83.52
m6i.2xlarge	8	32	$0.384	$276.48	$0.232	$167.04
c6i.xlarge	4	8	$0.17	$122.40	$0.102	$73.44
c6i.2xlarge	8	16	$0.34	$244.80	$0.204	$146.88
r6i.large	2	16	$0.126	$90.72	$0.076	$54.72
r6i.xlarge	4	32	$0.252	$181.44	$0.151	$108.72

> Prices reflect us-east-1 (N. Virginia) Linux on-demand rates. Windows adds ~$0.012–$0.20/hr depending on instance size. Prices vary by region — eu-west-1 is typically 10–15% higher.

---

Typical Use Cases with Real Numbers

Example 1 — Always-On Web Server

A startup runs a t3.large as their main application server 24 hours/day for a full 30-day month:

$0.0832/hr × 24 hr/day × 30 days = $59.90/month  (on-demand)
$59.90 × 12 = $718.80/year  (on-demand)
1-yr Standard RI annual upfront: ~$435  (~39% savings)

Break-even vs. on-demand: month 1 (upfront RI pays for itself within the first year by definition).

Example 2 — Nightly Batch Job

A data team runs a c6i.2xlarge ($0.34/hr) for 6 hours/night, every day of the month:

$0.34/hr × 6 hr/day × 30 days = $61.20/month

Because the instance runs only 25% of the day, a Spot Instance (typically 60–90% cheaper than on-demand) would cost ~$6–$24/month — a much better fit than an RI commitment.

Example 3 — Right-Sizing Comparison

A team is running a m6i.2xlarge ($0.384/hr, 24/7) but CPU utilization averages 15%. Downsizing to m6i.large ($0.096/hr):

m6i.2xlarge: $0.384 × 24 × 30 = $276.48/month
m6i.large:   $0.096 × 24 × 30 = $69.12/month
Monthly savings: $207.36  |  Annual savings: $2,488.32

---

Common Mistakes

1. Forgetting data transfer costs: EC2 compute is just one line item. Outbound data transfer costs $0.09/GB (first 10 TB/month). A server pushing 5 TB/month adds $450 — often larger than the compute bill itself.

2. Using on-demand pricing for always-on workloads: Running an instance 24/7 for 12+ months and not purchasing Reserved Instances wastes 40–60% of compute budget. AWS's own Cost Optimization Hub flags this automatically.

3. Comparing across regions without adjusting price: The same m6i.large costs $0.096/hr in us-east-1 but $0.114/hr in ap-southeast-1 (~19% more). Multi-region budgets must use region-specific pricing from the AWS Pricing API.

4. Ignoring EBS storage costs: A 100 GB gp3 EBS volume adds $8/month on top of compute — easy to overlook when an instance has multiple attached volumes.

5. Treating the 40% RI discount as fixed: The actual discount ranges from ~20% (convertible, 1-year, no upfront) to ~62% (standard, 3-year, all upfront). This calculator uses 40% as the conservative 1-year Standard RI average for Linux in us-east-1.

6. Not accounting for Savings Plans: EC2 Instance Savings Plans offer similar discounts to RIs (up to 72% vs. on-demand) with more flexibility — they apply to any instance in a family regardless of size or OS.

Frequently asked questions

How does AWS actually charge for EC2 — by the hour or by the second?

AWS bills EC2 in per-second increments with a 60-second minimum for Linux instances. This means a job that runs for 90 seconds costs the same as 90 seconds, not a full hour. However, for monthly budget planning, multiplying the hourly rate by hours used is accurate within <1% for workloads running more than a few minutes per session.

What is a Reserved Instance and how much does it really save?

A Reserved Instance (RI) is a 1- or 3-year billing commitment to a specific instance type and region in exchange for a discount on the hourly rate. For a 1-year Standard RI with no upfront payment on a Linux instance in us-east-1, discounts average ~36–40% off on-demand. A 3-year all-upfront Standard RI can save up to 62%. You can purchase RIs through the AWS Management Console or Marketplace.

What's the difference between Reserved Instances and Savings Plans?

Savings Plans (launched 2019) offer similar or greater discounts (up to 66% for Compute Savings Plans, up to 72% for EC2 Instance Savings Plans) but apply more flexibly — a Compute Savings Plan covers any EC2 instance regardless of family, size, region, or OS. RIs are locked to a specific instance type and region. AWS recommends Savings Plans for most new workloads due to their flexibility.

Why does the same instance type cost different amounts in different AWS regions?

AWS prices vary by region based on local electricity costs, data center operating expenses, real estate, and taxes. us-east-1 (N. Virginia) is consistently the cheapest region. eu-west-1 (Ireland) runs ~10–15% higher, ap-southeast-1 (Singapore) ~15–20% higher, and ap-east-1 (Hong Kong) can be 30%+ more expensive than us-east-1 for the same instance type.

Are there additional costs beyond the EC2 hourly compute price?

Yes — the compute price is often just 50–70% of your total EC2-related bill. Common additional charges include: EBS storage ($0.08–$0.10/GB-month for gp3), data transfer out ($0.09/GB for first 10 TB/month), Elastic IP addresses ($0.005/hr when not attached to a running instance), NAT Gateway ($0.045/hr + $0.045/GB processed), and load balancer hours. Always model these alongside compute.

How do Spot Instances compare in cost to on-demand pricing?

Spot Instances use spare AWS capacity and can be 60–90% cheaper than on-demand prices, but AWS can reclaim them with a 2-minute warning. For example, a c6i.xlarge on-demand costs $0.17/hr ($122.40/month 24/7); a Spot price for the same instance in us-east-1 has historically averaged around $0.05–$0.07/hr (~$36–$50/month). Spot is ideal for fault-tolerant batch jobs, CI/CD pipelines, and stateless web tiers.

How accurate is the 40% Reserved Instance savings figure used in this calculator?

The 40% figure is a conservative approximation for a 1-year Standard RI, no-upfront, Linux, us-east-1. Actual discounts range from ~20% (1-year Convertible, no upfront) to ~62% (3-year Standard, all upfront). The exact discount depends on instance family, payment option, term length, operating system, and region. Always verify current RI pricing directly in the AWS Pricing Calculator at calculator.aws before making a commitment.

What instance type should I start with for a basic web application?

AWS and the broader cloud community recommend starting with the t3.medium (2 vCPU, 4 GB RAM, $0.0416/hr ≈ $30/month 24/7) for small web apps or APIs. It supports CPU bursting, so it handles traffic spikes gracefully at a low base cost. If your app consistently pegs CPU above 40%, graduate to an m6i.large (2 vCPU, 8 GB RAM, $0.096/hr ≈ $69/month) for a fixed performance baseline without burst credits.

Sources and references

AWS EC2 On-Demand Pricing — Official AWS Pricing Page