AWS First Steps
LanguageEnglish
Cost Strategy

Cost Reduction Strategy - Comparing and Choosing Between Savings Plans and Reserved Instances

This article compares and explains the mechanisms / discount rates / flexibility of AWS Savings Plans and Reserved Instances (RI). It introduces how to use the four types, Compute / Database / EC2 Instance / Amazon SageMaker AI, for different needs, along with purchasing strategy.

About 11 min read

Basic Concepts of Commitment Discounts

In AWS pricing models, On-Demand pricing is the basis, but by committing to usage for a certain period, you can receive significant discounts. There are two mechanisms for commitment discounts: Savings Plans and Reserved Instances (RI). RI is a traditional discount offered since 2009, and you make a reservation by specifying a particular instance type / region / tenancy. Savings Plans is a new discount model introduced in 2019, in which you commit to a usage amount per hour (USD/hour). For example, when you declare "I commit to $10 of compute usage per hour," the discount is automatically applied to usage up to that amount. Savings Plans is more flexible than RI, and AWS recommends Savings Plans for new commitment discounts. However, this does not mean RI has been discontinued, and in certain use cases RI may also be able to provide a higher discount rate.

The Four Types of Savings Plans

Savings Plans offers four types (as of August 2026 / as stated on the official Savings Plans types page). Compute Savings Plans is the most flexible type and applies across EC2 instances, Fargate tasks, and Lambda functions. The discount continues even if you freely change the instance family, size, region, OS, and tenancy. The discount rate is up to 66% compared with On-Demand. EC2 Instance Savings Plans provides a high discount rate of up to 72% in exchange for committing to a specific region and instance family (for example, the m5 family in ap-northeast-1). Size changes within the same family (m5.large → m5.xlarge) and OS changes are unrestricted. Database Savings Plans is a type that applies to usage of database-family services and provides discounts of up to 35% (details in the next section). Amazon SageMaker AI Savings Plans (formerly SageMaker Savings Plans) is a type specialized for SageMaker AI instance usage. Discounts of up to 64% are applied across instance family / size / region / component (notebooks, training, inference, and so on), and it is used for cost optimization of ML workloads. The commitment term is 1 year or 3 years, with three payment options: All Upfront (maximum discount), Partial Upfront (intermediate), and No Upfront (minimum discount) (as an exception, Database Savings Plans is 1 year / No Upfront only). The 3-year / All Upfront option lets you obtain the highest discount rate, but it is a trade-off with flexibility.

Database Savings Plans and How to Choose for Database-Family Services

Database Savings Plans is a type that provides discounts of up to 35% on usage of database-family services (as of August 2026). The covered services are Aurora, RDS, DynamoDB, ElastiCache, DocumentDB, Timestream, Neptune, Keyspaces, DMS, and OpenSearch Service. For latest-generation provisioned instances, it is applied automatically regardless of engine, instance family, size, Availability Zone, or region, and it also applies to serverless usage. Therefore, the discount is maintained even across configuration changes such as migrating Aurora from db.r7g to db.r8g, migrating from RDS for Oracle to Aurora PostgreSQL-Compatible Edition, and moving workloads from RDS to DynamoDB. Commitment discounts for database-family services also include reserved instances (reserved instances for RDS / ElastiCache / OpenSearch Service / Redshift (called reserved nodes in ElastiCache and Redshift), and reserved capacity for DynamoDB). In exchange for fixing the target, these can in some cases provide a high discount rate for each service. One way to divide the roles is to use Database Savings Plans if you anticipate migrating engines or services or switching to serverless, and the reserved instances of the relevant service if you operate over the long term with the configuration fixed. Considering the options with this division makes the choice easier.

Comparison with RI and Choosing Between Them

The main difference between Savings Plans and RI is the balance between flexibility and discount rate. Because RI fixes the instance type (Standard RI) or the instance family (Convertible RI), it has difficulty adapting to workload changes, but on the other hand, Standard RI provides a discount rate of up to 72%. Convertible RI can be exchanged, but its discount rate drops to a maximum of 66%. The Compute type of Savings Plans has a discount rate equivalent to Convertible RI, but it excels in that it can be applied across regions and services (EC2/Fargate/Lambda). As a practical way of using them, an effective strategy is to aim for the maximum discount with EC2 Instance Savings Plans (or Standard RI) for instances you will definitely keep using, such as production DB servers running 24/7. Under the same strategy, for workloads whose configuration may change, such as development environments and batch processing, you secure flexibility with Compute Savings Plans. Savings Plans and RI can be used together, and discounts are applied in the order RI → Savings Plans.

Purchasing Strategy and Making Use of Cost Explorer

The most important thing when purchasing Savings Plans is determining an appropriate commitment amount. The Savings Plans recommendation feature in Cost Explorer analyzes usage patterns over the past 7 days / 30 days / 60 days and automatically calculates the optimal commitment amount and the estimated savings. Purchasing according to the recommendation is the safest, but there are several points to note. First, because it assumes that past usage will continue in the future, if a major workload change is planned, you need to adjust the recommended value. Next, the point is not to fix the purchase amount based on your own intuition. The recommendation screen shows side by side the monthly On-Demand spend based on usage over the selected period, the estimated monthly spend if you purchase the recommended amount, and the estimated monthly savings. This lets you understand before purchasing how an excess or shortfall in commitment appears in monetary terms. The same recommended values can also be obtained with GetSavingsPlansPurchaseRecommendation in the Cost Explorer API, and you can compare multiple options with different analysis periods and payment options. The unused portion of the commitment is not used for discounts and is wasted. Therefore, for workloads with deep troughs in usage, it is safer not to purchase the full recommended amount at once. The safer path is to commit to the base portion that will definitely remain and then make additional purchases after looking at actual results. You continuously monitor Savings Plans utilization and coverage with Cost Explorer and Budgets. If utilization is below 100%, it indicates over-purchasing, and if coverage is low, you need to consider additional purchases.

Pricing Comparison of Commitment Discounts

(The following discount rates are official values as of August 2026. Check the latest values on the Savings Plans types page of the Savings Plans User Guide and on the Savings Plans pricing page.) Compute Savings Plans applies across EC2 / Fargate / Lambda, with discounts of up to 66%. EC2 Instance Savings Plans is limited to a specific instance family and region, but its discounts are up to 72%. Database Savings Plans applies to database-family services such as Aurora / RDS / DynamoDB / ElastiCache, with discounts of up to 35%. Amazon SageMaker AI Savings Plans applies to SageMaker AI usage, with discounts of up to 64%. RI (Reserved Instances) is offered for EC2 and also as reserved instances for RDS / ElastiCache / OpenSearch Service / Redshift (called reserved nodes in ElastiCache and Redshift) and as reserved capacity for DynamoDB. For EC2, Standard RI offers up to 72% and Convertible RI up to 66%, a level equivalent to Savings Plans, but because the target is fixed, its flexibility is low. A combination in which EC2 Instance plans cover the stable baseline and Compute plans cover the fluctuating portion is effective.

Summary - Guidelines for Choosing Commitment Discounts

AWS Savings Plans, as a highly flexible commitment discount model, is the primary means of cost reduction in place of RI. Compute Savings Plans can be applied across EC2 / Fargate / Lambda and has flexibility that holds up well against workload changes. EC2 Instance Savings Plans provides a high discount rate in exchange for being limited to a specific instance family. For database-family services there is Database Savings Plans, and discounts of up to 35% are applied, including serverless usage. The approach of calculating the optimal commitment amount with Cost Explorer's recommendation feature, committing from the usage that will definitely remain, and making additional purchases while watching actual results is safe. The criterion for deciding on adoption is not the size of spending, but whether that usage is expected to continue stably in the future. There is no practical lower limit on the minimum commitment amount, and you can purchase even from a very small commitment per hour, so even at a small scale, if you have usage that runs constantly, it is worth considering.

References (Official AWS Resources)

The primary sources for this page are the official AWS website and documentation. Check the official pages below for the latest specifications and pricing.

If this page and the official documentation disagree, treat the official documentation as authoritative.

Share

Related Services

AWS Savings Plans SpecializedA pricing model offering up to 72% discount on EC2, Lambda, and Fargate with a 1- or 3-year usage commitmentAWS Cost Explorer PopularA service that visualizes and analyzes your AWS costs and usage to identify optimization opportunitiesAmazon EC2 EssentialA compute service that provides virtual servers in the cloud

Related Articles

Cloud Cost Budget Management - Spending Control and Alerts with AWS BudgetsThis article explains budget settings / cost alerts / automated actions with AWS Budgets. It introduces how to choose between Budgets and Cost Explorer, how to select budget types, and organization-wide budget management through Organizations integration.Up to 72% Cost Reduction with AWS Savings Plans - Choosing Between Compute and EC2 Instance PlansChoosing between Compute plans and EC2 Instance plans achieves cost reductions of up to 72%. This article introduces Cost Explorer's recommendation feature and the monitoring of utilization / coverage.Demand-Driven Infrastructure with AWS Auto Scaling - Designing and Optimizing Scaling PoliciesLearn how to use target tracking, predictive, and scheduled scaling policies effectively, and optimize costs with mixed instances policies that leverage Spot Instances.

More on This Topic

Inside AWS Pricing - How Per-Second, Per-Millisecond, and Per-Request Billing Actually WorksDiscover how AWS's diverse billing granularities, including EC2's per-second billing, Lambda's per-millisecond billing, and S3's per-request billing, are measured, aggregated, and invoiced through the metering system.90% of AWS Is Born from Customer Feedback - Working Backwards and Customer-Driven InnovationWe explain how the Working Backwards process and the Customer Obsession principle shape AWS service development. Learn about the innovation mechanism that starts with PR/FAQ documents and the process for incorporating customer feedback.AWS's Culture of Price Reductions - The Flywheel and Economies of Scale Behind 100+ Price Cuts Since 2006This article analyzes AWS's culture of price reductions, which passes the benefits of scale on to customers, and the economies of scale, hardware generation transitions, and competitive strategy behind it. It also introduces major past price reductions and their impact.Why AWS Pricing Is So Complex - The Economics and Game Theory Behind Pay-As-You-GoWe break down the three-tier pricing structure of On-Demand, Reserved Instances, and Spot Instances through the lens of airline yield management and demand forecasting economics, explaining the rational reasons behind the complexity of AWS pricing.Cost Management with AWS Budgets - Budget Alerts and Automated ActionsSet alerts on both actual and forecasted values, and use Budget Actions to automatically apply IAM policies or stop instances. Integrate with Slack to raise cost awareness across the team.Cloud Cost Budget Management - Spending Control and Alerts with AWS BudgetsThis article explains budget settings / cost alerts / automated actions with AWS Budgets. It introduces how to choose between Budgets and Cost Explorer, how to select budget types, and organization-wide budget management through Organizations integration.Why CloudWatch Logs Costs More Than You Expect - The Cost Structure of Ingestion, Storage, and AnalysisThis article explains the pricing structure where CloudWatch Logs ingestion costs over 20 times more than storage, the cost explosions caused by VPC Flow Logs and Lambda logs, and optimization strategies using the Infrequent Access class and S3 export.Resource Optimization with AWS Compute Optimizer - Instance RightsizingAnalyzes CloudWatch metrics with ML and generates recommendations that optimize resource sizes for EC2 / Lambda / EBS / ECS Fargate. This article explains how to use Graviton migration recommendations and improve accuracy with enhanced metrics.

Similar Articles and Services

Up to 72% Cost Reduction with AWS Savings Plans - Choosing Between Compute and EC2 Instance PlansChoosing between Compute plans and EC2 Instance plans achieves cost reductions of up to 72%. This article introduces Cost Explorer's recommendation feature and the monitoring of utilization / coverage.AWS Reserved Capacity Strategy Flexibility - How Savings Plans Cross-Service Coverage Changes Cost OptimizationAWS Savings Plans deliver cross-service discounts spanning EC2, Fargate, and Lambda, revolutionizing reserved capacity strategy. We analyze how to choose between Compute, EC2, and SageMaker Savings Plans, and compare the flexibility gap with Azure RI.AWS Savings PlansA flexible cost reduction model that discounts EC2, Fargate, and Lambda pricing in exchange for a 1-year or 3-year usage commitmentBreak-Even Analysis of Savings Plans vs Reserved Instances - Finding the Optimal Commitment Strategy with Real Pricing DataCompare discount rates, flexibility, and break-even points between Savings Plans and Reserved Instances using actual pricing data, and learn how to choose the optimal commitment strategy based on your workload characteristics.