What is Amazon SES?

The Short Answer

Email infrastructure, not an email product

Amazon SES, launched in 2011, is the same idea as the rest of AWS applied to email: instead of running your own mail servers — or paying an email platform a per-contact subscription — you call an API (or connect over SMTP) and AWS's sending infrastructure delivers the message. It handles the hard parts of getting mail accepted at scale: connection management with mailbox providers, authentication, feedback loops, and IP reputation on its shared sending pools.

What it doesn't do is anything a marketer would recognize as an email tool. There's no campaign editor, no audience list, no meaningful analytics screen — just a console for configuration and an endpoint for sending. That split is the whole story of SES: world-class pipes, bring your own everything else. It's why publicly documented users include Netflix, Duolingo, and Reddit, and also why plenty of teams bounce off it in the first week.

Less well known: SES also receives email. You can point a domain's MX records at SES and have inbound mail delivered to S3 or trigger Lambda functions — useful for support inboxes, reply processing, and email-driven workflows.

What It's Used For

The three jobs SES does

Almost every SES workload falls into one of three buckets:

The pattern across all three: SES supplies the delivery mechanics, and your application supplies the intent — what to send, to whom, and when.

How It Works

The lifecycle: verify, send, listen

1. Verify a domain. Before SES sends anything as you, you prove you own the domain by publishing DNS records. Verification sets up DKIM — a cryptographic signature added to every message so receiving servers can confirm it wasn't forged. Optionally, you configure a custom MAIL FROM domain so SPF alignment points at your domain rather than AWS's. The full walkthrough is in our Amazon SES setup guide.

2. Send via API or SMTP. Two doors into the same pipe. The SESv2 API is what application code and AWS SDKs use — a SendEmail call with your content and recipients. The SMTP interface exists for everything that already speaks SMTP: frameworks, CRMs, legacy systems, that one cron job from 2019. By default your mail leaves on shared IP addresses whose reputation AWS manages across many senders; high-volume senders can lease dedicated IPs instead.

3. Listen to events. After a send, SES emits events — deliveries, bounces, complaints, and (if you enable tracking) opens and clicks. You route these through a configuration set to SNS, EventBridge, or Kinesis Firehose, and your systems act on them: suppress the address that hard-bounced, log the complaint, update the dashboard. This step is not optional in practice — SES holds you accountable for bounce and complaint rates (roughly 5% and 0.1% respectively as of July 2026, covered in our SES deliverability guide), and ignoring the event stream is how accounts end up under review.

Pricing

Pricing in one sentence

As of July 2026, outbound email costs $0.10 per 1,000 messages, plus $0.12 per GB of attachment data — no monthly fee, no per-contact charge, no minimum. A startup sending 100,000 transactional emails a month pays about $10; the same volume on a typical email platform runs into the hundreds. Extras like dedicated IPs and inbound receiving carry their own line items.

The honest caveat: the sticker price excludes the tooling you'll build or buy around SES — and for marketing use cases, that's where the real cost lives. The full breakdown, including the extras and the hidden costs, is in our Amazon SES pricing guide.

Sandbox & Limits

The sandbox, quotas, and reputation

Every new account starts in the sandbox: 200 messages per 24 hours, 1 message per second, and — the real constraint — you can only send to addresses you've verified. It exists so spammers can't spin up throwaway AWS accounts. Getting out requires a short production-access request where you explain what you send and how recipients opted in.

In production, two limits apply: a sending quota (messages per rolling 24 hours) and a sending rate (messages per second). Both start modest and grow as you demonstrate consistent, clean sending — SES raises them automatically for accounts with healthy bounce and complaint rates, or you can request more. And those reputation metrics are enforced: cross the thresholds and SES can pause your sending regardless of your quota. The details — defaults, throttling errors, and how to raise limits — are in our SES sending limits guide.

What's Not Included

SES is not an ESP — here's the gap

An email service provider (ESP) like Mailchimp or Brevo bundles sending with everything around it: audience management, a campaign editor, automations, unsubscribe pages, analytics. SES sells only the sending. That's not a flaw — it's the product decision that makes $0.10 per 1,000 possible — but it means every SES adopter has to answer the same question: who's building the rest?

Teams close this gap one of three ways: build the missing layer in-house (common for pure transactional sending), pay for a full ESP and forgo SES pricing, or run a tool on top of their own SES account. That third path is what SendOps is: broadcasts, audiences, templates, unsubscribe handling, and analytics that connect to your SES account — so you keep SES's pricing and infrastructure while getting the ESP layer SES never shipped. There's a free plan, and here's how it works with SES.

Who Should Use SES

A good fit — and a bad one

The middle ground — a team that wants SES economics and marketing-tool ergonomics — is exactly the gap the previous section described, and it's increasingly served by tooling that runs on top of your own SES account rather than replacing it.

FAQ

Frequently asked questions