S3 storage class calculator that includes the cost of the move

Every archive class advertises a low price per gigabyte. Three other charges decide whether you actually save, and they are easy to miss.

• •Reading data back costs money. Glacier Instant Retrieval charges $0.03 for every GB you read. If a class is cheap to store and dear to read, data that is read often can cost more there than on Standard.
• •Small objects are billed as bigger ones. Standard-IA and Glacier Instant Retrieval bill every object as at least 128 KB. A library of 43 KB thumbnails pays for almost three times the bytes it holds.
• •Moving the objects is a charge of its own. A lifecycle rule pays a request fee for every object it moves: $0.05 per 1,000 into Deep Archive. On tens of millions of small files that one-off fee can be larger than a year of saving.

Enter your numbers below. The three examples show a records archive, an image library served to a website, and a pile of small files.

It prices your data on S3 Standard and on each of the three classes people usually move to, using AWS's own list prices for US East (N. Virginia), read from the AWS price list on 1 October 2026.

• •Storage on the new class, after the minimum object size is applied. Deep Archive adds 8 KB (billed as Standard) and 32 KB (billed as Deep Archive) of metadata to every object instead.
• •Retrieval: the share of your data you read back each month, times the class's per-GB fee. Enter 100 if you read all of it once a month, and more if a website or CDN reads the same files repeatedly.
• •The one-off move fee, and how many months of saving it takes to repay it.
• •The minimum storage period. Delete or overwrite an object before 30, 90 or 180 days and the days that were left are still billed.

Standard is priced at the first-tier rate of $0.023 per GB-month. Estates over 50 TB pay slightly less, so the saving shown for them is a little high. GET requests, data transfer and Glacier Flexible Retrieval are not modelled.

The input that matters most is how much of your data is read, and a storage console does not show it. S3 records when an object was last changed, not when it was last read.

For an archive those are the same date. For images behind a website they have nothing to do with each other: the files are never modified and are read all day. That is why "move anything untouched for 90 days to a cheaper class" is good advice for some buckets and expensive advice for others.

If you do not know the access pattern, S3 Intelligent-Tiering moves objects for you based on how they are used and has no retrieval fee. It is often the safer choice before you have measured anything.

A calculator needs you to type the shape of your data. BucketPilot reads it.

• •The real size distribution of every bucket, so the 128 KB floor is applied to your actual objects and not to an average.
• •The Savings Finder names cold data, old versions and duplicate copies with the monthly saving on each, from the index of every object you connect.
• •Lifecycle simulation. Write the rule, press Simulate, and see which objects it would move or delete before it touches anything.

Connecting a bucket is read-only and changes nothing. See the cost optimization page for the full list.

1. 1.Use the calculator above with your own numbers, or one of the three examples.
2. 2.Connect a bucket under Credentials with a read-only role you can revoke.
3. 3.Press Index bucket, then open Analytics to see the Savings Finder with figures from your real objects.

Every account starts with a 14-day trial, and the trial shows the headline figure before you decide.