Case study: enterprise
40 AWS accounts. One renewal calendar nobody owned, and $250k a year nobody was watching for.
This is a real AWS Organizations footprint with 40 member accounts under one payer. Each account connects through its own read-only CloudFormation stack, so all 40 report into one view. We priced each gap at the rates these accounts pay for the exact classes below, in us-east-1, mid-2026. The customer asked us to anonymize the details. See the method notes at the bottom.
Finding 01: the renewal that slipped at scale
One rollout. One term. Forty accounts lapsed the same week.
Two years ago, a platform rollout gave every one of the 40 member accounts its own db.r6g.4xlarge Multi-AZ cluster. The team reserved all of them on the same one-year term, at the same time. That was a good, deliberate choice. A year later, every one of those terms ended in the same week. No single account owner had the full calendar. The central platform team assumed each account owner tracked their own renewal. A routine cross-account cost review caught the lapse after 14 days.
The lapse cost $48.00 an hour and $1,152 a day. Scale costs $5,988 a year. So one day of this lapse cost about a fifth of a year of Scale. Sixteen thousand dollars in two weeks is small against $3M of annual reservable compute, and that is how it hid. No single account's bill moved enough to trigger an alarm. Only a view across all 40 accounts made the pattern visible. Reserver's expiry radar flags every renewal, org-wide, before the term ends, on every plan. Scale connects all 40 accounts to one coverage map. You buy each renewal in the AWS console. Renewal autopilot is on the roadmap and is not built yet.
Finding 02: coverage drift across accounts
Every account looks fine alone. Together, coverage drifts down.
No single team here is careless. Each account owner reserves what they can see. But 40 decentralized teams have 40 review cadences and 40 ideas of "good enough." Nobody holds the org-wide number. So reservable coverage across the whole organization sits at 71%. A single coordinated view would reach a higher number.
Each account owner would say their own coverage is fine. None of them can see that the org as a whole leaves over a fifth of its reservable footprint on-demand. Each account connects through its own read-only CloudFormation stack. Reserver reads all 40 together and reports one coverage number. The person who must close the gap can then see it.
Finding 03: reservations stranded across accounts
The workload moved accounts. The reservations didn't.
A re-org consolidated a product line's compute into a shared account. The m6g.xlarge instances moved cleanly. The 45 Reserved Instances bought against them, in the account they left, did not move. By default, a Regional RI's discount follows matching usage to any account in the organization. A zonal RI never shares. Neither does an RI in an account where sharing is turned off. These 45 did not follow the workload.
Left uncaught, the same 45 reservations waste $36k/yr. The destination account also pays the full on-demand rate for capacity that a reservation already covers, two accounts away. The Scale plan is built for this. One scan covers every member account. It matches reservations to instances wherever either one lives, so a re-org cannot strand a purchase again.
The bill
$250,062 found across an organization that nobody was watching whole.
No one made a mistake they could own up to alone. A rollout renewed together. 40 reasonable local decisions did not add up to one good global decision. A re-org moved compute faster than anyone moved the paperwork. The common thread is scale. Each of these is invisible inside a single account and obvious in an org-wide view. The Scale plan gives you that view.
Method notes, for the skeptical reader (we hope that's you)
- This is a real AWS Organizations footprint across 40 member accounts. At the customer's request, we withhold the organization and any identifying details. The structure, reservation counts, and every dollar figure are its own.
- We priced every dollar figure at the accounts' own AWS On-Demand and 1-year Reserved Instance (No Upfront) rates, US East (N. Virginia), mid-2026, for the exact classes named: db.r6g.4xlarge (RDS MySQL, Multi-AZ) and m6g.xlarge (EC2 Linux).
- The on-demand-minus-RI premiums use the No Upfront payment option. It has the smallest discount of the three RI payment options, so every premium in this study is a low estimate. We use the same 33% discount ratio throughout, including for the org-wide coverage-drift estimate.
- We measured the 14-day renewal gap, the 5-month stranding window, and the 71%/93% coverage figures from the org's own scan and cost history.
- This study covers AWS only. The product does not cover Savings Plans or other cloud providers in v1. This study covers Reserved Instances across AWS Organizations member accounts.
- This page assumes no right-sizing, Spot, or workload changes. It covers purchasing only, on the fleet as described.
Stop overpaying for a steady fleet.
Reserver is in early access. Join the list, and we will onboard you when we have capacity. Your first recommendation lands within minutes of connecting.