Amazon SES sending limits come down to two numbers per account, per AWS Region: the sending quota (how many recipients you can mail in a rolling 24-hour window) and the maximum send rate (how many emails SES accepts per second). In the sandbox they are fixed at 200 per day and 1 per second; with production access they depend on your use case and grow over time. You can see both on the SES Account dashboard and raise them through the Service Quotas console, where AWS aims to answer within 24 hours.

This guide explains how each limit is counted, where to read them, how increases work, which errors your app will hit when it goes over, and how to plan a 100,000-recipient send with the arithmetic shown.

Key takeaways

  • Two limits: a daily quota measured over a rolling 24 hours, and a per-second send rate. Both count recipients, not messages.
  • Per Region: limits in us-east-1 and eu-west-1 are completely separate.
  • Increases: AWS raises limits automatically for good senders near their quota; otherwise request one in Service Quotas.
  • Over the limit, the message is dropped: SES returns a throttling error and does not retry for you, so your app must queue and resend.
  • Reputation can stop you sooner: a 5% bounce rate or 0.1% complaint rate puts the account under review; 10% or 0.5% can pause sending.

The two Amazon SES sending limits: daily quota vs send rate

Sending quota (the "daily" limit)

The sending quota is the maximum number of emails you can send in a 24-hour period, calculated on a rolling window. Every time you send, SES counts what you sent in the previous 24 hours. There is no midnight reset: if you sent 40,000 emails between 09:00 and 10:00 on Monday, that capacity comes back gradually between 09:00 and 10:00 on Tuesday.

Maximum send rate

The send rate is the maximum number of emails SES accepts from your account per second. AWS says you can exceed it for short bursts but not for sustained periods, and that the rate at which SES accepts your mail can be lower than your maximum. Treat the number as a ceiling, not a target.

How recipients are counted

Both limits count recipients. One email with 10 addresses across To, CC and BCC uses 10 units of quota. A single message can carry at most 50 recipients, and AWS recommends one SendEmail call per recipient anyway, because one bad address can get the whole multi-recipient call rejected.

The full set of limits that matter for bulk sending, from the SES service quotas page:

LimitValueCan you raise it?
Sending quota200 per 24 hours in the sandbox; varies by use case in productionYes
Sending rate1 per second in the sandbox; varies by use case in productionYes
Recipients per message50 (To + CC + BCC)No
Message size40 MB with the v2 API or SMTP; 10 MB with the v1 API (after base64 encoding)No
Other SES API calls1 request per second for actions other than the send operationsNo
Verified identities10,000 per RegionYes

Two smaller notes: emails to the SES mailbox simulator count against your send rate but not your daily quota, and messages SES rejects for containing a virus still count against the quota.

Where to see your current limits

In the console

Open the SES console in the right Region and choose Account dashboard. Your quotas are under Sending limits; emails sent, remaining sends and the percentage of quota used are under Daily email usage. The same page shows your Account status (Healthy, or under review/paused).

With the API or CLI

The SES v2 GetAccount operation returns a SendQuota object plus your enforcement status. Run it per Region:

aws sesv2 get-account --region us-east-1

{
  "ProductionAccessEnabled": true,
  "SendingEnabled": true,
  "EnforcementStatus": "HEALTHY",
  "SendQuota": {
    "Max24HourSend": 50000.0,
    "MaxSendRate": 14.0,
    "SentLast24Hours": 12873.0
  },
  ...
}

(The numbers are an illustration, not AWS defaults.) The older v1 call aws ses get-send-quota returns the same three fields: Max24HourSend, MaxSendRate and SentLast24Hours. A sending app should read these before each campaign and pace itself from MaxSendRate rather than a hard-coded number, because AWS can change your limits at any time.

Watch EnforcementStatus, not SendingEnabled: AWS's enforcement FAQ says an account-level pause shows up in EnforcementStatus (HEALTHY, PROBATION or SHUTDOWN). SendingEnabled only reflects whether you switched sending off, and SES emits no EventBridge event for an account-level pause, so poll it.

How AWS raises Amazon SES sending limits automatically

Once you are out of the sandbox, AWS may raise your quotas without a request, often before you need them. According to the SES quota increase documentation, all of these have to be true:

  1. You send content recipients want. AWS explicitly says to stop mailing people who do not open.
  2. You send real production mail. Tests to fake addresses or only to internal staff do not demonstrate anything.
  3. You send near your current quota. Your daily volume should regularly approach the maximum without exceeding it.
  4. Bounce and complaint rates are low. High rates can lower your quotas instead.

In practice this favors steady senders. A newsletter that mails 5,000 people once a week sits far below a large quota six days out of seven and gives AWS little reason to raise it; that is when a manual request makes sense.

How to request a sending limit increase in Service Quotas

  1. Open the Service Quotas console.
  2. Select the Region you send from in the top-right Region menu.
  3. Choose AWS services, then Amazon Simple Email Service (SES).
  4. Choose the quota (Sending quota or Sending rate) and request the new value.

AWS's stated turnaround for these requests is up to 24 hours, longer if they need more information, and they can decline a request that does not fit their policies. Support for these requests is in English only.

How to size the request

Ask for what your sending pattern needs plus headroom, and be ready to justify it:

Daily quota needed = (largest single campaign)
    + (transactional email per day)
    + 25% headroom

Example: 100,000 newsletter + 4,000 transactional = 104,000
104,000 x 1.25 = 130,000 per 24 hours

Send rate needed = recipients / seconds you are willing to spend
Example: 100,000 in 1 hour = 100,000 / 3,600 = 27.8 per second, ask for 30

If the increase is part of a new use case, say so and describe list source, consent and bounce handling the same way you would for getting out of the Amazon SES sandbox. Requests backed by a clean sending history and specific numbers are the easiest to approve.

Throttling errors and how your sending app should handle them

When you go over either limit, SES drops that message and does not attempt to redeliver it. It returns an error instead, and the wording depends on the interface (from the SES quota errors page and the SMTP response code list):

InterfaceError you will seeMeaningWhat to do
API / SDKThrottlingException: "Maximum sending rate exceeded"Too many per secondSlow down, back off, retry
API / SDKThrottlingException: "Daily message quota exceeded"Rolling 24-hour quota used upStop and reschedule; retrying in a loop will not help
SMTP454 Throttling failure: Maximum sending rate exceededToo many per secondRetry with increasing delays
SMTP454 Throttling failure: Daily message quota exceededQuota used upPause the campaign until capacity frees
SMTP421 Too many concurrent SMTP connectionsToo many open connectionsReduce worker count
API v2 / v1SendingPausedException / AccountSendingPausedSending disabled for the accountDo not retry; check EnforcementStatus and the support case

Exception class names differ between SDKs and API versions (the v2 API also defines TooManyRequestsException, HTTP 429, for request throttling), so match on the message text as well as the class name.

A safe sending loop

  1. Pace on the client side. Read MaxSendRate and send at about 80 to 90% of it with a token bucket shared by all workers. Ten workers each sending at "the rate" is ten times the rate.
  2. Check remaining quota before a campaign. Max24HourSend - SentLast24Hours must cover the whole send plus your transactional traffic. If not, split the campaign.
  3. Back off on rate errors. For the API, AWS says to wait for an interval of up to 10 minutes and retry. For SMTP 4xx errors, AWS suggests progressively longer waits (5 seconds, 10 seconds, 30 seconds, then 20 minutes before repeating). AWS SDKs already retry HTTPS calls with backoff.
  4. Do not retry quota errors immediately. Requeue the remaining recipients for when the rolling window frees capacity.
  5. Record every accepted message ID. AWS notes that a timed-out request may already have been accepted, so a blind retry can send a duplicate. Mark recipients as sent only after you get a message ID, and use message tags to match events.
rate = 0.85 * MaxSendRate    # e.g. 0.85 x 14 = 11.9/s
for recipient in queue:
    bucket.wait_for_token(rate)
    try: send(recipient); mark_sent(recipient, message_id)
    except "Maximum sending rate exceeded": sleep(backoff()); requeue(recipient)
    except "Daily message quota exceeded": pause_campaign_until(window_frees)
    except SendingPaused: stop_all_and_alert()

Worked example: sending to a 100,000-contact list

Suppose your Account dashboard shows a quota of 50,000 per 24 hours and a rate of 14 per second (illustrative numbers; yours will differ). You want to send one newsletter to 100,000 subscribers.

Step 1: does the quota cover it?

No. 100,000 recipients is twice the 50,000 quota, and that ignores the transactional mail (password resets, receipts) that shares the same quota. You have three options:

  • Split across two days: send 45,000 on day 1 (leaving 5,000 for transactional), the rest 24 hours later. Simple, but half your list gets the email a day late.
  • Request an increase first: using the formula above, ask for about 130,000 per day and 30 per second, a few days before the send.
  • Split across sending accounts or providers: for example, route part of the list through a second Region or a second provider. This also gives you failover; see multi-server email routing.

Step 2: how long will it take?

Duration is recipients divided by the rate you actually send at. Pacing at 85% of the maximum leaves room for transactional mail and avoids bursts:

Max send rateAt 100% (100,000 / rate)At 85% pacing
14 per second7,143 s = 1 h 59 min100,000 / 11.9 = 8,403 s = 2 h 20 min
30 per second3,333 s = 56 min100,000 / 25.5 = 3,922 s = 1 h 5 min
50 per second2,000 s = 33 min100,000 / 42.5 = 2,353 s = 39 min
100 per second1,000 s = 17 min100,000 / 85 = 1,176 s = 20 min

A two-hour send is usually fine for a newsletter. Speed matters more for time-boxed promotions; mailbox providers also notice sudden spikes from a young domain, so faster is not automatically better (our warm-up guide covers ramping up).

Step 3: what does it cost?

As of November 2026, on the Essentials plan new SES accounts start on, 100,000 emails cost 100,000 / 1,000 x $0.16 = $16. On à la carte pricing at $0.10 per 1,000, the same send is $10, plus attachment data if any. Splitting the send across days does not change the price. The plan differences are covered in Amazon SES pricing.

Bounce and complaint thresholds that can pause sending

Your quota is not the only thing that can stop a campaign. AWS reviews account reputation continuously, and these rates matter more than any quota. From the SES sending review FAQ:

MetricAWS recommends belowAccount placed under review atSending may be paused at
Bounce rate2%5% or greater10% or greater
Complaint rate0.1%0.1% or greater0.5% or greater
  • Not a fixed time window. AWS calculates both rates over a "representative volume" that differs by sender, so you cannot reproduce them from daily CloudWatch counts. Watch the Reputation.BounceRate and Reputation.ComplaintRate metrics instead and alarm well below the review levels.
  • What counts: only hard bounces to domains you have not verified count toward the bounce rate; complaints count only from mailbox providers that send feedback to SES.
  • Big sends amplify old data. Mailing 100,000 addresses that have not heard from you in a year is the fastest way to hit 5% bounces in one afternoon. Clean or re-permission old segments first.
  • Gmail is stricter on complaints. Google asks bulk senders to keep the spam rate in Postmaster Tools below 0.3%, and recommends below 0.1% (Google sender guidelines).

The automation that keeps you under these numbers (suppressing hard bounces and complainers the moment SES reports them) is covered step by step in Amazon SES bounce and complaint handling.

If you would rather not build pacing, retries and suppression yourself, eMailBase sends through your own SES account and lets you add several sending servers with weighted routing and failover, so a large list can be spread across accounts or providers. It is one option among several; the arithmetic above applies whatever tool you use. For the broader setup, see our guide to Amazon SES email marketing.

Frequently asked questions

Does the Amazon SES daily quota reset at midnight?

No. It is a rolling 24-hour window. Each send is checked against everything you sent in the previous 24 hours, so capacity returns exactly 24 hours after the original sends, not at a fixed time of day.

Are SES sending limits shared across AWS Regions?

No. Quotas, send rates, sandbox status and reputation are tracked separately per Region. Using a second Region gives you a second set of limits, but also a second account reputation to manage and a separate set of verified identities.

Do failed or rejected sends count against my quota?

A call that SES refuses with an error (for example a throttling error) does not send the email and is not billed. Messages SES accepts and then rejects, such as ones containing a virus, do count against your daily quota and are billed.

What is the highest sending limit Amazon SES allows?

AWS does not publish a maximum. Production limits "vary based on your specific use case", and AWS raises them for accounts with steady volume and low bounce and complaint rates. Large senders typically reach higher limits through a history of clean sending plus specific Service Quotas requests.

Why did my send rate drop even though my quota did not change?

AWS notes that SES may accept mail at less than your maximum rate. Check EnforcementStatus and your reputation metrics first; if the account is healthy, reduce parallel workers and keep pacing below the maximum.

Amazon SES sending limits are predictable once you treat them as two separate budgets: a rolling 24-hour recipient quota and a per-second rate, both per Region. Read them from GetAccount before every campaign, pace below the rate, requeue on throttling instead of losing mail, request increases a few days ahead of big sends, and keep bounces and complaints far below the 5% and 0.1% review levels that can stop you faster than any quota.