.png)
Key takeaways
1
DynamoDB Reserved Capacity can reduce eligible provisioned throughput costs by up to 54% for a one-year term or up to 77% for a three-year term.
2
It applies only to provisioned capacity using the Standard table class, and the reservation is tied to the AWS Region where it is purchased.
3
Reservations are purchased separately in blocks of 100 WCUs or 100 RCUs. Unused committed capacity is still billed.
4
Since AWS cut DynamoDB on-demand throughput pricing by 50% in November 2024, compare capacity modes before committing.
5
Database Savings Plans provide up to 18% savings on DynamoDB on-demand throughput and up to 12% on provisioned throughput with a one-year term and no upfront payment.
DynamoDB Reserved Capacity, which AWS now calls DynamoDB reserved provisioned capacity, is a billing discount for stable provisioned throughput. You commit to a minimum level of read or write capacity for one year, or three years in selected Regions, and receive discounted pricing. The purchase does not change table performance, availability, or configuration.
Reserved Capacity can cover eligible provisioned RCUs and WCUs for tables and secondary indexes. It does not cover replicated WCUs used by Global Tables. AWS applies the discount automatically to eligible usage, including eligible linked-account usage under consolidated billing.
For the broader commitment model, see our AWS Savings Plans guide.
Short answer
DynamoDB Reserved Capacity is best suited to predictable, long-running provisioned workloads using the Standard table class that are unlikely to move Regions during the term. AWS publishes savings of up to 54% for one year and 77% for three years. The main risk is overcommitting because reservations cannot be cancelled or transferred, and unused capacity is still billed.What is DynamoDB Reserved Capacity?
AWS describes reserved provisioned capacity as a commitment to a minimum level of provisioned throughput for a defined term in exchange for lower pricing. You purchase read and write reservations separately in 100-unit blocks. See the AWS DynamoDB pricing page for the current rules and rates.Reserved Capacity can cover eligible provisioned RCUs and WCUs for tables and secondary indexes. It does not cover replicated WCUs used by Global Tables. AWS applies the discount automatically to eligible usage, including eligible linked-account usage under consolidated billing.
For the broader commitment model, see our AWS Savings Plans guide.
What does provisioned capacity cost?
For DynamoDB Standard in US East (N. Virginia), AWS currently uses $0.00065 per WCU-hour and $0.00013 per RCU-hour in its pricing examples.| Capacity | Standard hourly rate | Approx. monthly cost per 100 units* |
|---|---|---|
| 1 WCU | $0.00065/hour | $47.45 per 100 WCUs |
| 1 RCU | $0.00013/hour | $9.49 per 100 RCUs |
*Illustrative monthly values use 730 hours. Actual charges vary by Region and workload.
A WCU costs five times as much as an RCU at these reference rates. For a workload with equally stable read and write capacity, WCUs therefore create the larger dollar-saving opportunity.
A WCU costs five times as much as an RCU at these reference rates. For a workload with equally stable read and write capacity, WCUs therefore create the larger dollar-saving opportunity.
FinOps Tip:
Prioritize the capacity type driving the most spend, but size the reservation from a durable provisioned baseline supported by actual consumption and autoscaling behavior.
Prioritize the capacity type driving the most spend, but size the reservation from a durable provisioned baseline supported by actual consumption and autoscaling behavior.
What are the eligibility constraints?
Reserved Capacity has several hard rules:- Provisioned capacity mode only. On-demand tables are ineligible.
- DynamoDB Standard table class only. Standard-IA is excluded.
- Single-Region scope. The discount applies only in the purchase Region.
- No replicated WCU coverage for Global Tables.
- 100-unit purchase blocks, with RCUs and WCUs purchased separately.
- Fixed commitment. AWS does not allow cancellation, resale, Region transfer, or account transfer after purchase.
How did the 2024 price cut change the decision?
AWS reduced DynamoDB on-demand throughput prices by 50% effective November 1, 2024 and now recommends on-demand for most scenarios. Provisioned capacity can still fit steady, predictable throughput, but the first decision should be capacity mode, not reservation term. AWS’s November 2024 pricing announcement explains the change.AWS cost-optimization guidance also notes that on-demand can be cheaper when average provisioned-capacity utilization is below about 35%, with workload shape affecting the crossover point.
For a deeper comparison, see our DynamoDB on-demand vs. provisioned guide.
Reserved Capacity vs. Database Savings Plans
Database Savings Plans changed DynamoDB commitment planning in December 2025. They use a one-year, no-upfront hourly spend commitment and can apply flexibly across eligible database usage and Regions. AWS currently publishes up to 18% savings for DynamoDB on-demand throughput and up to 12% for provisioned throughput.| Factor | Reserved Capacity | Database Savings Plans |
|---|---|---|
| DynamoDB mode | Provisioned only | On-demand and provisioned |
| Maximum DynamoDB savings | Up to 54% for 1 year; 77% for 3 years | Up to 18% on-demand; 12% provisioned |
| Commitment unit | 100 RCUs or 100 WCUs | Consistent $/hour spend |
| Upfront payment | Partial upfront | No upfront |
| Scope | Specific Region | Flexible across eligible services and Regions |
| Term | 1 or 3 years | 1 year |
The discounts do not stack on the same usage. AWS applies Reserved Capacity first to eligible provisioned usage, after which remaining eligible provisioned capacity can receive Database Savings Plans coverage.
Read our Database Savings Plans guide for the broader eligibility and sizing model.
Read our Database Savings Plans guide for the broader eligibility and sizing model.
Should you buy Reserved Capacity?
Use this matrix as a screening framework, not as an AWS eligibility test.| Condition | Strong signal to consider it | Signal to avoid or delay |
|---|---|---|
| Capacity mode | Provisioned | On-demand |
| Table class | Standard | Standard-IA |
| Usage pattern | Stable, predictable throughput floor | New, seasonal, or highly variable workload |
| Workload horizon | Same Region for the commitment term | Migration, decommissioning, or Region change likely |
Apply this decision per workload and Region. A single account can contain tables that deserve different purchasing strategies.
At AWS’s reference rates:
These are ceiling illustrations, not purchase quotes. Use the actual upfront fee and discounted hourly rate shown for your Region to calculate total cost and break-even.
Worked example
Assume a Standard table in us-east-1 runs continuously at 1,000 provisioned WCUs and 500 provisioned RCUs, and exclude any applicable DynamoDB free-tier capacity from this illustration.At AWS’s reference rates:
- WCU cost: 1,000 × $0.00065 × 8,760 = $5,694/year
- RCU cost: 500 × $0.00013 × 8,760 = $569.40/year
- Total standard provisioned cost: $6,263.40/year
These are ceiling illustrations, not purchase quotes. Use the actual upfront fee and discounted hourly rate shown for your Region to calculate total cost and break-even.
Calculate your break-even
AWS recommends comparing standard provisioned pricing with the effective hourly rate of reserved provisioned capacity.
Effective reserved hourly cost
=
upfront fee
÷
term hours
+
reserved hourly charge
Compare that result with the standard provisioned cost for the same capacity. For autoscaled tables, reconcile provisioned RCUs and WCUs with consumed demand, autoscaling minimums and maximums, target utilization, and expected workload changes. Commit only when the provisioned baseline is durable enough for the reserved cost to remain lower.
How to size Reserved Capacity correctly
Start with both consumed and provisioned throughput.- Pull 30 to 90 days of ConsumedReadCapacityUnits, ConsumedWriteCapacityUnits, ProvisionedReadCapacityUnits, and ProvisionedWriteCapacityUnits from CloudWatch by table, GSI, and Region.
- Compare consumed demand with the configured provisioned floor. For autoscaled tables, review minimum and maximum capacity settings and target utilization so temporary scaling behavior is not mistaken for a long-term baseline.
- Identify a sustained low-traffic baseline that is supported by both demand and configured provisioned capacity. A lower-percentile consumption measure such as P10 can be an internal heuristic, but AWS does not prescribe P10.
- Round conservatively to usable 100-unit blocks rather than rounding up for headline coverage.
- Model RCUs and WCUs separately for each Region.
- Recheck planned migrations, table-class changes, decommissions, and major architecture work before selecting a term.
- Compare the final model with current on-demand pricing and Database Savings Plans.
Pre-purchase checklist
Confirm provisioned mode and Standard table class.
Confirm the Region and desired term availability.
Review consumed and provisioned RCU and WCU history.
Review autoscaling minimums, maximums, and target utilization if enabled.
Exclude replicated WCUs from Global Tables.
Model RCUs and WCUs separately in 100-unit blocks.
Compare Reserved Capacity with on-demand and Database Savings Plans.
Confirm the workload should remain in the Region for the full term.
Obtain finance approval for the upfront payment and commitment.
How to purchase Reserved Capacity
In the DynamoDB console, select the correct Region, open Reserved Capacity, choose Purchase Reserved Capacity, select the capacity type and term, enter the quantity, then review the upfront and hourly charges before confirming. AWS documents the workflow on its DynamoDB pricing page.AWS supports console purchases up to a combined 1,000,000 reserved provisioned capacity units. Larger purchases require a service-limit increase.
How Usage.ai fits
At Usage.ai, we focus on the commitment-management layer after a stable baseline is identified. For eligible recommendations, we analyze usage and present a savings recommendation for customer approval.After approval, we initiate the commitment purchase through the cloud provider API and manage it as a Flex Commitment. With Flex Insured Commitments, teams can get the up to 57% savings of a 3-year AWS commitment with none of the commitment.
We charge a percentage of realized savings. If an eligible Flex Commitment creates a loss because its cost exceeds equivalent On-Demand usage, we calculate that loss and provide cashback according to current program terms. Our pricing documentation and cashback documentation explain the current mechanics.
For teams comparing commitment horizons, our 1-year vs. 3-year AWS commitment guide explains the tradeoff between discount depth and lock-in risk.
Frequently asked questions
Can you cancel or transfer DynamoDB Reserved Capacity?
No. AWS states that DynamoDB Reserved Capacity cannot be sold, cancelled, transferred to another account, or moved to another Region. The upfront payment is non-refundable.
Does Reserved Capacity apply to secondary indexes?
Yes. Eligible provisioned read and write capacity includes tables and secondary indexes. Replicated WCUs used by Global Tables remain excluded.
What happens when Reserved Capacity expires?
The discount applies only for the purchased term. Before expiry, review current usage and pricing rather than automatically replacing the same quantity.
Can Reserved Capacity and Database Savings Plans work together?
Yes, but not as a double discount. Reserved Capacity applies first to eligible provisioned usage. Remaining eligible provisioned capacity can then receive Database Savings Plans coverage.
Is three-year Reserved Capacity available everywhere?
No. Three-year Reserved Capacity is limited to selected Regions. One-year offerings are available in all AWS Regions. Check the current DynamoDB pricing and offer details before purchasing.
Share
Facebook
X
LinkedIn
Reddit
