AWS
Microsoft Azure
Google Cloud
What are AWS Savings Plans?
AWS Savings Plans are a commitment-based pricing model. You commit to a specific dollar amount of eligible usage per hour, and AWS applies lower Savings Plans rates to qualifying usage.Usage beyond the commitment is billed at the applicable On-Demand rate unless another discount applies.
The hourly structure matters. A commitment that seems reasonable against average monthly spend can still be oversized during low-usage hours.
Size a Savings Plan based on hourly eligible usage and its durable floor, not simply your average monthly AWS bill.
For the underlying mechanics, see the AWS Savings Plans user guide.
What are the four types of AWS Savings Plans?
| Savings Plan | Maximum advertised savings | Coverage | Typical fit |
|---|---|---|---|
| Compute Savings Plans | Up to 66% | EC2, Fargate, Lambda | Changing or distributed compute |
| EC2 Instance Savings Plans | Up to 72% | One EC2 family in one Region | Stable EC2 usage |
| SageMaker AI Savings Plans | Up to 64% | Eligible SageMaker AI instance usage | Predictable ML workloads |
| Database Savings Plans | Up to 35% | Eligible supported database usage | Stable eligible database workloads |
Choose this plan when
Choose Compute Savings Plans when eligible compute may move between instance families, Regions, Fargate, or Lambda.
Choose EC2 Instance Savings Plans when you have durable usage in one EC2 family and Region and can trade flexibility for potentially deeper rates.
Choose SageMaker AI Savings Plans when you have stable, eligible SageMaker AI usage.
Choose Database Savings Plans when eligible database usage meets AWS's supported service and generation requirements.
Database Savings Plans need extra eligibility checks
Database Savings Plans cover eligible usage across supported AWS database services.AWS allows commitments to follow eligible serverless usage and the latest provisioned instance generations across supported engines, instance families, sizes, deployment options, and Regions.
Eligibility matters. Do not assume every older provisioned database instance qualifies.
AWS also documents service-specific interactions with existing discounts. Before adding a Database Savings Plan, validate whether RDS Reserved Instances or DynamoDB reserved-capacity discounts already cover the same workload.
See AWS Database Savings Plans pricing for current eligibility and rates.
Database Savings Plans advertise savings of up to 35% for eligible serverless database usage and up to 20% for eligible provisioned database instances, compared with On-Demand pricing.
Not every database configuration qualifies. Check the supported services, instance generations, and eligible usage on the official AWS Database Savings Plans pricing page before purchasing.
AWS Savings Plans terms and payment options
Compute, EC2 Instance, and SageMaker AI Savings Plans are available in 1-year or 3-year terms.AWS supports three payment options:
| Payment option | How it works | Best consideration |
|---|---|---|
| No Upfront | No initial payment; recurring commitment charges | Preserves cash |
| Partial Upfront | Pay part initially and the remainder over the term | Balances cash flow and rates |
| All Upfront | Pay the commitment cost at purchase | Often offers the lowest effective rate |
A 3-year commitment may look attractive on rate alone, but the extra discount is less valuable if the workload changes significantly during the term.
Database Savings Plans differ: they currently use a 1-year term with No Upfront payment.
The right choice depends on both the AWS rate and your confidence that the underlying usage will remain.
How AWS Savings Plan pricing works
Savings Plan commitments are measured at Savings Plans rates, not your current On-Demand spend.Simplified pricing example
Assume an eligible workload costs $10 per hour On-Demand and the applicable Savings Plans rate is 40% lower.$10 × (1 - 0.40) = $6 per hour
What determines AWS Savings Plan pricing?
AWS Savings Plan rates depend on the plan type, eligible service, instance configuration, Region, commitment term, and payment option.For example, Compute Savings Plans can cover eligible EC2, AWS Fargate, and AWS Lambda usage, while EC2 Instance Savings Plans apply to a selected instance family in a specific Region.
Compare current rates using the official AWS Savings Plans pricing page before purchasing. Advertised maximum discounts do not represent the savings every workload will receive.
Real AWS environments contain multiple usage types with different Savings Plans rates, so do not apply one headline discount percentage to your entire AWS bill.
Compute Savings Plans vs EC2 Instance Savings Plans
| Factor | Compute Savings Plans | EC2 Instance Savings Plans |
|---|---|---|
| Maximum advertised savings | Up to 66% | Up to 72% |
| EC2 family flexibility | Yes | No |
| Region flexibility | Yes | No |
| Fargate | Yes | No |
| Lambda | Yes | No |
| Typical fit | Changing infrastructure | Stable family and Region |
An EC2 Instance Savings Plan can offer a better rate, but only if sufficient usage continues to match the selected family and Region.
Compute Savings Plans exchange some maximum discount potential for broader flexibility.
If you are also evaluating RIs, see our AWS Savings Plans vs Reserved Instances guide.
Both plans use a fixed hourly spending commitment, but their flexibility differs. Compute Savings Plans can apply across eligible EC2 instance families and Regions, as well as Fargate and Lambda. EC2 Instance Savings Plans are limited to one instance family in one Region but can offer larger discounts for consistently matching usage.
How to size an AWS Savings Plan
1. Isolate eligible hourly usage
Do not size a commitment based on total AWS spend or monthly averages.Identify the usage a proposed Savings Plan can actually cover, then inspect its behavior hour by hour.
2. Account for existing commitments
Review existing Reserved Instances, Savings Plans, expiration dates, sharing settings, and queued purchases.AWS applies applicable EC2 Reserved Instances before Savings Plans. EC2 Instance Savings Plans then apply before Compute Savings Plans.
See the AWS Savings Plans application rules for the full order.
3. Check organization-level sharing
In an AWS Organization or consolidated billing family, Savings Plan benefits can apply across eligible linked-account usage when discount sharing is enabled.Your commitment baseline may therefore need to be evaluated at the management-account level rather than for one account in isolation.
Before sizing a new purchase, confirm which accounts participate in Savings Plans discount sharing and where existing commitment benefits are being consumed.
4. Identify your durable baseline
Separate usage likely to remain from usage that may disappear.Review:
planned migrations
rightsizing programs
application retirements
temporary projects
seasonal demand
instance family changes
regional moves
Example: Choosing an hourly commitment
Suppose your eligible AWS compute usage costs $10/hour at Savings Plans rates during quiet hours and $18/hour during busy hours.Assume existing commitments consistently cover $4/hour of the same eligible usage, after AWS applies the relevant discount rules.
| Metric | Quiet hours | Busy hours |
|---|---|---|
| Eligible usage at Savings Plans rates | $10/hour | $18/hour |
| Existing commitment coverage | $4/hour | $4/hour |
| Remaining eligible usage | $6/hour | $14/hour |
This is an illustrative example, not an AWS pricing quote or a final purchase recommendation. Actual sizing requires hourly billing data, the applicable benefit order, existing commitments, and expected workload changes.
5. Review AWS recommendations
AWS Savings Plans recommendations use historical usage and support 7-, 30-, or 60-day lookback periods.AWS states that these recommendations do not forecast future usage.
A recommendation cannot account for plans to shut down an application next month or move a workload next quarter.
Use a recommendation as an input, not an automatic purchase instruction.
See the AWS Savings Plans recommendation methodology.
If usage is already declining, read our guide on buying more AWS Savings Plans when usage falls.
6. Model the purchase with Purchase Analyzer
AWS Purchase Analyzer lets you model a proposed commitment against historical usage.You can evaluate variables including plan type, term, payment option, lookback period, commitment amount, and target coverage.
Use it to understand how a proposed purchase would have performed against historical usage, not as a prediction of future demand.
See the AWS Purchase Analyzer documentation.
7. Stress-test the downside
Do not model only the expected case.Test what happens if eligible hourly usage falls by 10%, 20%, or 30%.
A commitment that looks attractive today can become less valuable after a migration, rightsizing project, or application shutdown.
The goal is not maximum coverage. It is the highest coverage you can support without unacceptable underutilization risk.
If forecast confidence is limited, our staggered vs one large AWS Savings Plan comparison explains the tradeoff between capturing savings sooner and preserving future purchasing decisions.
When should you avoid buying more Savings Plans?
Buying more commitment is not always the right move.Consider waiting when:
a major migration or shutdown is approaching
a significant rightsizing program is still underway
existing commitments are already underutilized
uncovered usage is mostly seasonal or temporary
architecture changes could move workloads outside a narrower plan
you cannot confidently explain the stable hourly usage floor
Coverage vs utilization
Coverage measures how much eligible usage receives Savings Plans pricing.Utilization measures how much of your purchased commitment is actually consumed.
High utilization with low coverage can mean an existing commitment is working well while additional eligible usage remains uncovered.
Low Savings Plan utilization indicates unused commitment, while high coverage shows that much of the eligible usage receives discounted pricing. Review both metrics over the same reporting period and scope before deciding whether a commitment is oversized.
No universal target works for every AWS environment.
Review both metrics alongside workload volatility, expected changes, expirations, actual rates, and unused commitment cost.
Our AWS Savings Plans coverage vs utilization guide covers this analysis in more detail.
What happens if AWS usage drops?
A Savings Plan commitment does not automatically shrink when infrastructure usage falls.If usage declines, part of the hourly commitment may remain unused while AWS continues billing according to the plan terms.
AWS provides a limited correction window for purchase errors. A Savings Plan with a commitment of $100 per hour or less may be returned within the previous seven days, in the same calendar month, subject to AWS requirements.
Treat this as an error-correction mechanism, not an ongoing exit strategy.
See the AWS Savings Plan return policy.
AWS Savings Plan purchase checklist
Before purchasing, confirm:eligible hourly usage is correctly isolated
monthly averages are not hiding low-usage hours
existing RIs and Savings Plans are included
organization-level discount sharing is understood
commitment expirations are accounted for
migrations and rightsizing plans are reflected
current AWS Savings Plans rates are being used
the term and payment option fit your workload confidence
coverage and utilization are reviewed together
a realistic downside scenario has been modeled
purchasing permissions and approval controls are clear
How Usage.ai Helps Optimize AWS Savings Plans
Usage.ai helps businesses identify savings opportunities and manage eligible AWS commitments based on actual usage and existing coverage.With Flex Insured Commitments, eligible teams can access commitment-based discounts while reducing financial exposure. If a covered commitment costs more than equivalent On-Demand usage, Cashback covers the qualifying difference, subject to program terms.
Teams can review recommendations through CoPilot or automate eligible purchases with Autopilot.
Conclusion
AWS Savings Plans can materially reduce eligible AWS costs, but the largest advertised discount is not automatically the best purchase.Start with stable hourly usage. Account for commitments you already own, check sharing across accounts, and remove workloads likely to change. Review AWS recommendations as historical evidence, model the proposed purchase, and stress-test the downside before committing.
The goal is sustainable realized savings, not maximum commitment.
Disclosure
We publish this guide and provide cloud commitment optimization services. Our Cashback protection, eligibility, fees, and program terms depend on the applicable customer agreement. AWS pricing, eligibility, discount rates, and Savings Plan rules can change, so verify current AWS documentation and your own AWS account before making a commitment purchase.
Frequently asked questions
What are the four AWS Savings Plan types?
AWS currently offers Compute, EC2 Instance, SageMaker AI, and Database Savings Plans. They differ in eligible services, flexibility, commitment structure, and available rates.
Which AWS Savings Plan offers the largest advertised discount?
EC2 Instance Savings Plans advertise savings of up to 72% compared with On-Demand rates. Compute Savings Plans advertise up to 66%, SageMaker AI Savings Plans up to 64%, and Database Savings Plans up to 35%. Actual savings depend on the usage covered and the applicable rate.
Do AWS Savings Plans cover RDS?
Compute and EC2 Instance Savings Plans do not cover RDS. Eligible RDS usage may qualify for Database Savings Plans, while RDS Reserved Instances remain another commitment option.
Can AWS Savings Plans be cancelled?
Generally, no. A Savings Plan is a term commitment. AWS provides a limited return exception for qualifying plans with commitments of $100 per hour or less that were purchased within the previous seven days and remain within the same calendar month, subject to AWS restrictions.
How should I choose an AWS Savings Plan commitment amount?
Start with stable eligible hourly usage, not your monthly average. Account for existing commitments, organization sharing, planned workload changes, and term and payment choices.
Then review historical AWS recommendations, model the proposed purchase with Purchase Analyzer, and test the economics under a realistic decline in usage before buying.
