AWS S3

AWS S3 — The API everything else copies, and by far the most expensive egress in this roster. S3 defines the interface, the semantics and the edge cases every other provider on this page implements against, and it remains the most complete: storage classes, Object Lock, replication, Access Points, event notifications, and IAM condition keys nobody else matches. It is also the reason this comparison exists — at USD 0.09 per gigabyte, egress on a read-heavy workload costs more than everything else combined. Built by Amazon Web Services. Founded 2006.

We track it in 1 comparison — S3-compatible storage — so every claim below is a cell in a table you can open and check rather than an impression. Across those rosters it sits against 11 other tools, and what follows is where it visibly separates from them.

Where it wins.

• S3 API coverage — 100 %, the best figure of the 12 here.
• TTFB from Sydney — 25 ms, 2nd of 10.
• IAM granularity — 5, tied for best across 2 of 12 tools.
• Official MCP server — Yes, which only 3 of 12 tools here manage.

Each of those is ranked against the whole roster on its category page, not against a hand-picked subset, so a first place here means first of everything we list.

Where it gives ground.

• Egress $/TB — $90 /TB, the weakest of the 11 here. Cloudflare R2 records $0 /TB.
• True cost /mo (USD) — $472.6 /mo, the weakest of the 9 here. Hetzner Object Storage records $12.47 /mo.
• Storage $/TB/mo — $23 /TB/mo, the weakest of the 11 here. Storj records $4 /TB/mo.
• True cost /mo (AUD) — A$718.35 /mo, the weakest of the 9 here. Hetzner Object Storage records A$18.95 /mo.

None of these disqualify it on their own. They are the fields to check against your own requirements before you commit, because they are the ones where a competitor genuinely does better.

Provenance. 27 of 28 tracked fields carry a value for AWS S3, and 16 of those cite a document you can open. Last verified 2026-07-23. Every figure keeps its own provenance — measured by us, claimed by the vendor, inferred, or community-reported — and we would rather print a dash than a guess.

Its nearest neighbour in our data is Backblaze B2. AWS S3 is ahead on Sydney region (Yes against No) and Official MCP server (Yes against No). Backblaze B2 takes True cost /mo (USD) ($26.95 /mo against $472.6 /mo) and Egress $/TB ($10 /TB against $90 /TB). That pattern repeats across the rest of the roster — see AWS S3 alternatives for the other rivals, each compared the same way.

Company

Amazon Web Services

Founded

2006

Fields we track

27 of 28

Last verified

2026-07-23

Still the most capable object store by a wide margin, and thirty-eight times the cost of Hetzner on this scenario almost entirely because of bandwidth. Stay on S3 if you genuinely use Object Lock compliance mode, Access Points, cross-region replication or IAM condition keys; otherwise you are paying $441 a month for an API you can get elsewhere.

Egress $/TB

$90 /TB11th of 11

Vendor-claimedverified 2026-07-23source

True cost /mo (USD)

$472.6 /mo9th of 9

Inferredverified 2026-07-23source

S3 API coverage

100 %1st of 12

Inferredverified 2026-07-23

Storage $/TB/mo

$23 /TB/mo11th of 11

Vendor-claimedverified 2026-07-23source

TTFB from Sydney

25 ms2nd of 10

Inferred

True cost /mo (AUD)

A$718.35 /mo9th of 9

Inferredverified 2026-07-23

IAM granularity

51st of 12

Vendor-claimedverified 2026-07-23source

Class A $/M ops

$5 /M ops11th of 12

Vendor-claimedverified 2026-07-23source

AWS S3 is the right call in these situations, each one drawn from a field we actually record:

• Workloads already inside AWS where traffic stays on the internal network.
• Regulatory requirements needing Object Lock compliance mode with a real SLA.
• Anything depending on IAM condition keys, Access Points or S3 Tables.
• S3 API coverage is your binding constraint. AWS S3 records 100 %, the best figure in the S3-compatible storage roster.
• TTFB from Sydney is your binding constraint. AWS S3 records 25 ms, 2nd of 10 in the S3-compatible storage roster. We define that field as typical time to first byte for an uncached GET of a small object from a client in Sydney, using the provider's nearest available location.
• IAM granularity is your binding constraint. AWS S3 records 5, the best figure in the S3-compatible storage roster.
• Official MCP server is your binding constraint. AWS S3 records Yes, which only 3 of the 12 tools in the S3-compatible storage roster do.

Reach for something else when any of the following is a requirement rather than a nice-to-have:

• Egress $/TB. AWS S3 records $90 /TB, the weakest of the 11 tools in the S3-compatible storage roster. Cloudflare R2 records $0 /TB on the same field. If that is a hard requirement rather than a preference, start elsewhere.
• True cost /mo (USD). AWS S3 records $472.6 /mo, the weakest of the 9 tools in the S3-compatible storage roster. Hetzner Object Storage records $12.47 /mo on the same field. If that is a hard requirement rather than a preference, start elsewhere.
• Storage $/TB/mo. AWS S3 records $23 /TB/mo, the weakest of the 11 tools in the S3-compatible storage roster. Storj records $4 /TB/mo on the same field. If that is a hard requirement rather than a preference, start elsewhere.
• True cost /mo (AUD). AWS S3 records A$718.35 /mo, the weakest of the 9 tools in the S3-compatible storage roster. Hetzner Object Storage records A$18.95 /mo on the same field. If that is a hard requirement rather than a preference, start elsewhere.
• Class A $/M ops. AWS S3 records $5 /M ops, 11th of 12 in the S3-compatible storage roster. Backblaze B2 records $0 /M ops on the same field. If that is a hard requirement rather than a preference, start elsewhere.

We publish this block because a comparison that only lists what a tool is good at is marketing. Every figure above sits on the same page as its source, and the field definitions are on the category tables if you want to check how we measured them.

What we don't know. 1 of the 28 fields we track for AWS S3 are still blank: Egress cap (x stored). Those render as dashes rather than as zeroes or assumptions, because an empty cell and a bad cell are not the same thing and only one of them is honest. If you know any of these figures and can point at a document, tell us.