GET and PUT requests can outgrow storage for small-object workloads.
For millions of small objects, S3 request fees rival storage — and writes cost 12× reads. Front hot objects with a CDN. Model the storage side on the object storage cost calculator and cache savings on the CDN cost calculator.
The S3 Request Cost Calculator estimates what you pay for S3 request operations — the per-request charges that sit alongside storage and data-transfer fees and are easy to overlook when budgeting. You enter your monthly GET requests (in millions) and your monthly PUT/POST requests (in millions), and the tool multiplies each volume by its per-operation rate to show the monthly cost. The main driver is request count, not object size: because PUT and POST typically cost several times more per thousand requests than GET, a write-heavy workload can run up a larger bill than a read-heavy one at the same volume. Splitting the two inputs makes it clear which operation type is actually driving your spend.
The practical thing to watch is that request costs scale with how your application accesses data, independent of how much data you store. An app that writes many small objects, or reads the same object millions of times without caching, can pay more in request fees than in storage. Batching writes, adding a cache or CDN in front of frequently read objects, and consolidating tiny files all cut request counts and therefore cost. Run the numbers before you assume storage is your biggest line item — for high-traffic or write-intensive workloads, the per-operation charges are often the part of the bill that grows fastest.
Yes. On top of per-GB storage and egress, S3 bills per request: roughly $0.0004 per 1,000 GET requests and $0.005 per 1,000 PUT/POST/LIST requests. For apps handling millions of small objects, request fees can exceed the storage cost.
Put a CDN in front to absorb repeat GETs, batch small objects into larger files where possible, avoid unnecessary LIST operations, and cache. Since PUT costs ~12× a GET, reducing write frequency (buffer then flush) helps write-heavy workloads most.
Cost at different volumes:
| Volume | Monthly | Yearly |
|---|