Savings Plans

What are Savings Plans?

Savings Plans are a flexible pricing model offered by AWS that provides a discount on cloud usage in exchange for a commitment to a consistent amount of spend over a one- or three-year term. The discount applies automatically to eligible usage, without needing to specify the exact instance types or regions in advance.

The trade-off is straightforward: commit to spending a fixed amount per hour, and AWS reduces the rate on qualifying compute usage.

How Savings Plans work

When you purchase a Savings Plan, you commit to a specific hourly spend, for example $10 per hour, for either a one- or three-year term. AWS applies the discounted rate to your usage up to that commitment level. Any usage beyond the commitment is charged at standard on-demand rates.

The commitment is to a dollar amount, not to specific resources. This is what makes Savings Plans more flexible than Reserved Instances, which require upfront decisions about instance type, size, and region.

Types of Savings Plans

AWS offers three types of Savings Plans:

  • Compute Savings Plans cover EC2, Lambda, and Fargate across any instance family, size, region, or OS. This is the most flexible type, though the discount rate is slightly lower than EC2 Instance Savings Plans.
  • EC2 Instance Savings Plans apply to EC2 only and are locked to a specific instance family and region. This is less flexible, but offers a deeper discount in return.
  • SageMaker AI Savings Plans are specific to SageMaker workloads and follow the same commitment model for ML infrastructure.
  • Database Savings Plans cover managed database services, such as Aurora, RDS, DynamoDB, ElastiCache, DocumentDB, and others, across any engine, instance family, size, or region.

Savings Plans in action: A real-world example

A company runs a mix of EC2 instances and Lambda functions across multiple AWS regions. Its average on-demand spend sits at around $15 per hour.

After reviewing 30 days of usage data, the FinOps team identifies a consistent baseline of $10 per hour that it is confident committing to. The company purchases a one-year Compute Savings Plan at that level.

On-demandWith Savings Plan
Hourly commitmentNone$10/hr
Discount on baseline usageNoneUp to 66%
Estimated monthly saving$0~$2,900

The discounted rate applies automatically across EC2 and Lambda without any further configuration.

Savings Plans vs. Reserved Instances

Both Savings Plans and Reserved Instances offer discounts in exchange for a usage commitment, but they work differently:

Savings PlansReserved Instances
Commitment typeHourly spendSpecific instance type and region
FlexibilityHighLower
Discount depthUp to 66%Up to 72%
Best forVariable or evolving workloadsStable, predictable workloads

Reserved Instances offer slightly deeper discounts but require more upfront decisions about exactly what you will be running. Savings Plans are generally the better starting point for most organizations, particularly those still evolving their infrastructure.

What to get right before committing

Using a Savings Plan is a financial commitment. Here are a few things worth doing before purchasing:

  • Analyze at least 30 days of usage data to understand your consistent spend baseline.
  • Commit only to the amount you are confident you will use. Unused commitment is not refunded.
  • Right-size workloads before committing—locking in Savings Plans on oversized instances means paying a committed rate on waste.
  • Consider the term length carefully. A three-year term offers a deeper discount but less flexibility if your infrastructure changes significantly.

How Savings Plans work across multiple AWS accounts

Savings Plans apply at the account level by default. In organizations running multiple AWS accounts under AWS Organizations, Savings Plans can be shared across the organization, which increases utilization and reduces the risk of unused commitment sitting in a single account.

Savings Plans vs. Azure and GCP equivalents

Savings Plans are an AWS-specific offering. Azure and GCP offer equivalent commitment-based discounts under different names:

  • Azure uses Azure Reservations for virtual machines and other services.
  • GCP uses Committed Use Discounts for Compute Engine workloads.

The mechanics of these discounts differ across providers, but the principle is the same: commit to consistent usage in exchange for a lower rate.