Ask five deliverability engineers "how many emails can I send from one IP per day?" and you'll get five ranges, all defensible, none of them a real ceiling. That's not evasiveness — it's an honest answer to a question that's shaped wrong. Mailbox providers don't publish a number. There is no register anywhere that reads "2,000,000" and stops you at 2,000,001.
What actually happens is closer to a dynamical system than a spec sheet. Volume, complaints, authentication, engagement, and provider throttling all feed back into each other, and the system's behavior over time is sensitive to its starting conditions. Two senders with identical hardware and identical daily volume can have completely different outcomes, because their histories — their reputations — diverged weeks earlier.
So instead of a ceiling, think in bands. Here's the planning range that reputable senders actually use.
Capacity is earned, not configured
An IP's safe volume grows with its track record, not with a setting. New IPs get almost no benefit of the doubt; established ones with clean signals get a wide berth. The bands below are the ranges senders plan around at each stage — read them as trust bands, not thresholds a switch flips at.
Planning ranges by IP maturity
View as table
| Stage | Daily volume | Approach |
|---|---|---|
| New / cold IP | 100 – 5,000 | Start conservative, build gradually |
| Warming IP | 5,000 – 250,000 | Raise only while metrics stay healthy |
| Established IP | 250,000 – 2,000,000 | Common range for reputable senders |
| 2M+ per day | 2,000,000+ | Time to add another IP, not push harder |
Notice the axis is logarithmic — the jump from "warming" to "established" is a 100x swing in volume, not a modest bump. That's the shape of trust-based systems generally: progress is slow to start and compounds once the signals hold.
A daily total hides the number that actually throttles you
"1.2 million a day" sounds like one number. It's actually two, and the second one is the one mailbox providers watch. A daily total spread across 24 hours behaves nothing like the same total crammed into a six-hour send window — the second scenario multiplies your per-second rate by four, and per-second rate is exactly what triggers provider throttling.
1.2M emails/day, sent inside a 6-hour window
112 sends a second is a materially different conversation with a mailbox provider than 56. This is why two senders who both claim "1.2 million a day" can have completely different deliverability outcomes: one paced it evenly, the other detonated it in a morning send window. Daily volume and throughput are different capacity questions — plan both.
A new IP is a butterfly-effect problem
This is where chaos theory earns its place in the name. A new IP has no history for a mailbox provider to trust, so it starts small — recent, engaged recipients only — and the ramp speed is set entirely by how the provider responds. Push too fast and a small early signal (a bounce spike, a complaint tick) compounds into a throttled or blocklisted IP weeks later. The graph below is illustrative, not a fixed schedule: the real pacing is "as fast as acceptance allows," which is different for every sender.
A graduated warm-up curve
View as table
| Week | Daily volume |
|---|---|
| 1 | 250 |
| 2 | 750 |
| 3 | 2,500 |
| 4 | 8,000 |
| 5 | 25,000 |
| 6 | 70,000 |
| 7 | 150,000 |
| 8 | ≈280,000 |
What actually throttles you (it isn't the network)
Nothing about your server hardware caps how many emails you can send. The ceiling is a trust ceiling, enforced by mailbox providers reading signals in real time. In practice, inbox risk shows up long before any technical limit does.
Throttling responses
4xx replies from the receiving server are a live rate limiter — the provider slowing you down mid-send.
Reputation signals
Trust and placement move continuously based on how recipients treat your mail, not on a static score.
One provider dominating
If a single mailbox provider is most of your list, their tolerance becomes your effective ceiling.
Authentication gaps
DMARC, SPF, or DKIM failures undercut trust before volume is even a factor.
Slow queue drain
A queue that can't finish inside its delivery window forces uncomfortable choices under pressure.
Complaint rate
The single fastest signal to move against you — see the thresholds below.
Inbox risk usually appears before the network limit, through throttling, deferrals, spam placement, complaints, or a queue that won't drain.
Where the guardrails sit
Complaints below 0.1% is where most providers keep extending trust. Cross 0.3% and expect active suppression — throttling, filtering to spam, or a stalled ramp — regardless of how clean your authentication is otherwise.
The checklist that comes before the number
None of the bands above are safe to lean on if the fundamentals underneath them are shaky. Work through this before raising volume on any IP, new or established.
| Check | What "healthy" looks like |
|---|---|
| Authentication | DMARC, SPF, and DKIM all passing at a high, stable rate |
| Complaints | Below 0.1%; treat 0.3% as a hard stop, not a target |
| Unsubscribe | Visible link, one-click functional, honored immediately |
| Blocklist status | IP and domain both monitored continuously |
| DNS configuration | Forward and reverse DNS match and resolve cleanly |
| SMTP security | Encrypted transport, bounces handled and fed back into the list |
Who's actually controlling the chaos
The IP itself is just the address the reputation attaches to. Whether you or your provider controls the variables feeding that reputation is a separate decision.
Provider manages the pool
- Best fit for inconsistent or lower volume that can't sustain a dedicated IP's reputation alone
- Provider routes volume and manages overall pool health
- Your reputation is partly influenced by other senders in the pool
You control the pattern
- Sender controls volume and sending patterns directly
- Requires consistent volume — an idle dedicated IP loses trust over time
- Better isolation between mail streams, but a higher monitoring burden
When to add a second IP
A simple rule of thumb: if one IP carries more than a million emails a day and the business needs predictable delivery windows, start planning the next IP before it's needed — not after the queue is already backing up.
- Queue delays — campaigns aren't finishing inside their scheduled window
- Provider deferrals — recurring pushback during peak sending periods
- Reputation isolation — transactional and marketing mail need to stop sharing risk
- Growth forecasts — projected volume will exceed the current IP's comfort band
- Provider-specific bottlenecks — one mailbox provider needs its own throttling plan
An IP limit isn't a mailbox limit
Everything above describes IP-level capacity — what a mailbox provider will tolerate from a sending address.
That's a different question from what a hosted mailbox account allows. Most email service providers cap sending per-user, per-minute, or per-tenant, and those account-level caps often trigger long before the IP's technical ceiling is anywhere in sight. Check both, separately, for every platform you send through.
Benchmarks in this piece are planning ranges compiled from general deliverability practice, not guarantees from any mailbox provider. Actual capacity depends on your sending reputation, authentication health, and recipient engagement — measure your own signals rather than sending to a number.