We cut a client's AWS bill by 41% without touching a single feature
Right-sizing, storage tiering and one very expensive NAT gateway.
- BrandKul Cloud
- 19 May 2026
- 5 min read

Cloud waste is rarely dramatic. It is a hundred small defaults nobody revisited after launch day.
Compute right-sizing gave us the first 18%. Instances provisioned for a launch-day traffic spike had been running at 9% utilisation for fourteen months.
Storage tiering gave us another 11%. Logs and media older than ninety days moved to infrequent-access classes with no change to the application.
The last 12% was a single NAT gateway routing internal traffic that should never have left the VPC. One route table change, thousands of dollars a year.
Building something where this matters? We scope in writing before any code.
Start a conversation
