What letting a reservation lapse actually means

On AWS a Standard Reserved Instance and a Savings Plan both expire at the end of their one or three year term and simply stop billing. Letting one lapse means you choose not to repurchase coverage for that workload, so the capacity reverts to on demand or moves to a new commitment shaped around current demand.

This matters because commitments bill whether or not you use the capacity. The discount, roughly 20 to 72 percent against on demand depending on instrument and term, is the reward for carrying utilization risk. When the risk no longer pays, the right move is to stop carrying it.

The four signals that say do not renew

First, the workload is shrinking. If the service under the commitment is being consolidated or sunset, a renewal pays for instances that will go dark.

Second, the workload is migrating. A move to Graviton, to a managed service, or to spot for fault tolerant batch changes the instance shape, so renewing the old Reserved Instance strands the rate on hardware you are leaving.

Third, demand is volatile and falling. If the trough of the last several months sits below your current coverage, you are already paying for idle commitment, and renewing extends that waste.

Fourth, flexibility has value you are not capturing. A narrow Standard Reserved Instance that no longer matches your fleet is often better replaced by a flexible Compute Savings Plan sized to the real floor.

The effective savings rate test

Effective savings rate is the discount you actually realised, total on demand equivalent minus what you paid, divided by the on demand equivalent. A commitment only earns its keep while realised utilization stays high enough to beat on demand. Below that breakeven, lapsing and paying on demand is cheaper.

Indicative renewal decision by signal. Figures verified against client billing data and anonymized.
Situation at expiryForecast utilizationDecision
Steady base workload, no migration plannedAbove 90 percentRenew, ideally with a flexible Compute Savings Plan
Service consolidating onto fewer instances60 to 80 percentLet it lapse, recommit to the smaller floor
Fleet moving to Graviton this quarterBelow 60 percent on old shapeLet it lapse, commit on the new shape after migration
Seasonal demand now past peakBelow 50 percentLet it lapse, ride on demand until the floor is clear

A worked example

Indicative figures, verified against the client's billing data, anonymized. A scaling fintech held three year Standard Reserved Instances on a service it was moving to Graviton. Forecast utilization on the old shape for the next term was 55 percent. Renewing would have cost about 240,000 USD a year at an effective savings rate of roughly 12 percent, well below the 30 percent the same spend earned before the migration. Letting the reservations lapse and recommitting on Graviton backed instances after the move restored the rate and removed the stranded coverage.

Your next step

Before any renewal, pull utilization and effective savings rate for the covered workload and forecast the next term against it. For the full method read the AWS cost optimization guide, and for neighbouring detail see the AWS commitment renewal checklist and renewing AWS commitments from strength. To put it into practice, our AWS cost optimization service turns the forecast into verified savings, and you can request a free trial two ways.

Independent · buyer-side

Put a defensible number on your cloud spend.

No provider in the room, no published price list. Tell us your footprint and we will scope the savings against your billing data — we reduce your cloud spend or we reimburse our service fee.

Buyer-side intelligence, monthly.

The Cloud Spend Navigator: what changed in cloud pricing, commitments, and FinOps — no vendor spin.