Dedicated IP vs shared IP: choosing your sending infrastructure

A dedicated IP is not automatically better. Below a volume floor it performs worse than a good shared pool, because reputation is built from consistent sending and a quiet IP looks like an unknown one. The real question is not dedicated versus shared — it is who else is on the pool.

Updated July 31, 202613 min read
  • The volume floor, explained
  • What "dedicated" means at different providers
  • Pool composition as the hidden variable
  • When isolation is required regardless of volume

Key takeaways

  • A dedicated IP needs consistent, near-daily volume. Roughly 100,000 messages a month is the practical floor; around 300,000 is where it reliably outperforms a good shared pool.
  • Below that floor, a quality shared pool usually delivers better, because you inherit established reputation instead of building one from nothing.
  • The decisive variable in a shared pool is not its size — it is who else is allowed on it.
  • "Dedicated IP" ranges from an address label on shared infrastructure to genuinely separate mail servers. Ask which one you are buying.
  • Some requirements — customer IP allowlisting, isolation guarantees in a security review — need dedicated infrastructure at any volume.

The dedicated-versus-shared question is usually framed as a question about control, which makes dedicated sound obviously preferable. It is not. A dedicated IP hands you sole responsibility for a reputation you then have to build from zero and sustain forever, and if your volume cannot sustain it, you have traded a pool with an established history for an address that looks perpetually unfamiliar to receivers.

The genuinely useful framing is different. Shared infrastructure means your reputation is partly a function of other people’s behaviour. Dedicated infrastructure means it is entirely a function of yours. Which is better depends on your volume, your consistency, and — the part most comparisons omit — on who those other people actually are.

What each one actually is

On a shared pool, your messages leave from IP addresses used by many senders at once. The reputation attached to those addresses is the aggregate of everyone’s sending. You inherit it on your first message, which is why a new account on an established pool can send successfully on day one without any warmup.

On a dedicated IP, your messages leave from an address nobody else uses. The reputation attached to it is built entirely from your own sending — which also means it starts at nothing, and receivers treat an unknown address conservatively until it has a history.

Shared poolDedicated IP
Reputation sourceEveryone on the poolYou alone
Time to trustedImmediate — you inherit it4–8 weeks of warmup
Volume floorNone~100k/month, sending near-daily
Effect of your own mistakeDiluted across the poolConcentrated entirely on you
Effect of someone else’s mistakeYou absorb part of itNone — there is no one else
Can you name your IPs to a customerNo — the range is shared and may changeYes
Goes cold if you pauseNoYes — dormancy erodes reputation within weeks
Operational burdenNoneWarmup, monitoring, consistency management

The volume floor and why it exists

Reputation is inferred from a pattern of behaviour over time. An IP that sends 50,000 messages one Monday and nothing until the following month has not given receivers a pattern — it has given them two anomalies. Sparse, bursty sending from a dedicated address reads as suspicious in a way the same volume on a busy shared pool never would.

That is why the guidance is expressed as a floor rather than a preference. Industry consensus puts the practical minimum around 100,000 messages a month with consistent near-daily sending, and around 300,000 a month for reputation that maintains itself comfortably. Below that, most senders get better results from a well-run shared pool.

Monthly volumeRecommendationReasoning
Under 100kShared poolNot enough consistent signal to build or hold a dedicated reputation
100k–300kDepends on consistencyViable if you send most days; erratic patterns still favour shared
300k–1MDedicated usually winsEnough steady volume to build and sustain reputation
Over 1MMultiple dedicated IPs, split by streamSeparate transactional from marketing at the IP layer as well as the domain layer

Transactional senders have a structural advantage here. Transactional volume tracks product usage, so it is naturally smooth and naturally daily — exactly the profile a dedicated IP wants. A marketing-heavy sender at the same monthly total has a much harder time keeping an IP warm, because their volume arrives in campaign-shaped bursts.

The variable nobody puts on the comparison page

Every shared-versus-dedicated comparison treats "shared pool" as a single, uniform thing. It is not. A pool carrying transactional mail from vetted, paying customers behaves nothing like a pool carrying free-tier signups, marketing campaigns, and whatever a self-serve account decided to send at 2am.

This is the question that actually determines whether shared infrastructure is safe for you, and it is almost never answered on a pricing page. Ask it directly.

  • Do free-tier accounts send from the same IPs as paid production traffic? An account that costs nothing to create and nothing to abandon has no incentive to protect a reputation you depend on.
  • Is marketing or bulk campaign traffic permitted on the pool? Campaign mail generates complaint volumes transactional mail never does, and complaints are the most heavily weighted negative signal there is.
  • Is cold outbound permitted anywhere on the infrastructure? Cold sending is the single fastest way to accumulate spam-trap hits and blocklistings.
  • How quickly is an abusive tenant detected and isolated? Specific answers — thresholds, automated cut-offs, notification policy — are a good sign. Vague ones are not.
  • Are transactional and marketing traffic separated at the IP layer? If the provider runs both, the pools should be distinct.

This is the constraint SuperSend TX is built on. The production pool carries transactional mail from paying customers only — no free tier on production infrastructure, no marketing campaigns, no cold outbound. It is a deliberate limit on who we will sell to, and it is the mechanism that produces pool quality rather than a claim about it.

What "dedicated" means at different providers

The word is sold at very different depths, and the differences are material. Three levels are common.

LevelWhat you getWhat you do not get
Dedicated IP add-onOne IP address assigned to your account, on the provider’s existing infrastructureServer isolation, control over warmup, usually any help with the ramp
Dedicated IP poolSeveral addresses reserved for you, often with stream separationServer-level isolation; you still share the underlying platform
Dedicated infrastructureMail servers provisioned for you, with IPs used only by your trafficNothing, if the provider also operates and warms it

The gap between the first and third rows is where most disappointment lives. A dedicated IP add-on that arrives cold, with no warmup plan and no operational support, is not a deliverability upgrade — it is a project you have just been handed, and it will perform worse than the shared pool you left for the first month or two.

The questions that separate them: are the IPs used only by my traffic? Are they named in the dashboard so I can give them to a customer? Who runs the warmup ramp? What happens if the reputation degrades — do I get told, and does someone help?

When dedicated is required regardless of volume

The volume analysis assumes the decision is about deliverability. Sometimes it is not, and then the floor stops being the deciding factor.

  • Customer IP allowlisting. The moment you sell to a company with a security review, someone will ask which IP addresses your mail arrives from so their gateway can allowlist them. "A shared range that may change" is not an answer that passes.
  • Reputation isolation as a contractual matter. Some buyers require that their mail not share infrastructure with unrelated senders, and want it in writing.
  • Regulated environments. Healthcare, finance, and government procurement frequently require documented isolation of the sending path.
  • Critical mail with a low tolerance for correlated failure. If a delivery incident caused by an unrelated tenant would be a serious business event, paying to remove that failure mode is rational even below the volume floor.
  • Very high volume with distinct streams. Above roughly a million messages a month, splitting transactional and marketing onto separate dedicated IPs is standard practice.

Making the decision

  1. 1

    Establish your real monthly volume and its shape

    Not the projection — the last three months. Then look at how many days a week you actually send meaningful volume.

  2. 2

    If under 100k and evenly spread, stay shared — but audit the pool

    Ask who else sends from it, whether free accounts share it, and whether marketing or cold traffic is permitted. That audit matters more than the shared/dedicated choice itself.

  3. 3

    If over 300k with near-daily sending, plan the move

    You have the volume to build and hold reputation. Budget four to eight weeks for the ramp and read the warmup guide before you start.

  4. 4

    If a customer or auditor needs nameable IPs, move regardless

    The requirement is not about deliverability and the volume floor does not apply to it. What matters is that the IPs are genuinely yours and someone competent is operating them.

  5. 5

    Separate the streams before you separate the IPs

    Sending transactional and marketing from distinct subdomains is cheaper, faster, and delivers more benefit than a dedicated IP. Do it first. Here is how.

  6. 6

    Never move both streams at once

    Migrate transactional first, stabilise, then marketing. If something degrades you want to know which change caused it.

How SuperSend TX handles both

Because we own and operate our sending infrastructure rather than reselling someone else’s, the shared and dedicated options are both things we control end to end.

The Pool carries production transactional mail from paying customers only. There is no free tier on it, marketing campaigns are not permitted, and cold outbound is not permitted anywhere on the infrastructure — our cold-email product runs on entirely separate systems and never touches this network. That constraint is the product.

Dedicated provisions mail servers and IP addresses for your workspace alone. We build them, run the full warmup ramp so you never receive a cold IP, monitor reputation continuously, and operate the stack. Your IPs are listed in the dashboard so you can hand them to a customer’s security team. Because the API contract is identical across Sandbox, Pool, and Dedicated, moving between tiers changes your bill and nothing in your application.

PoolDedicated
Who else sends from itPaying transactional customers onlyNobody
Marketing or cold trafficNot permittedNot applicable
WarmupAlready warm — inherit itWe run the full ramp for you
IPs nameable for allowlistsNoYes, listed in the dashboard
Operations and monitoringOursOurs
API contractPOST /emailsPOST /emails — unchanged

Dedicated servers and IPs, built and warmed by us

Not an IP label on shared infrastructure. Your own mail servers and addresses, provisioned, warmed through the full ramp, monitored, and operated — with the same API you already use.

Related reading

FAQ

Frequently asked questions

Do I need a dedicated IP for email?

Probably not below about 100,000 messages a month. A dedicated IP builds reputation only from your own consistent sending, so at low or bursty volume it looks perpetually unfamiliar to receivers and performs worse than a good shared pool that already carries established reputation. Volume, consistency, and isolation requirements decide it — not a preference for control.

How much volume do you need for a dedicated IP?

Roughly 100,000 messages a month is the practical floor, and around 300,000 is where a dedicated IP reliably outperforms a quality shared pool. Consistency matters as much as the total: you want meaningful volume on at least three or four days a week. A sender doing 200,000 in two batches will struggle where one doing 120,000 spread evenly will not.

Is a shared IP bad for deliverability?

It depends entirely on who else is on it. A pool carrying vetted, paying, transactional-only senders can outperform a poorly managed dedicated IP because you inherit established reputation with no warmup risk. A pool carrying free-tier accounts, marketing campaigns, and cold outbound is a different proposition. Ask the provider who your neighbours are.

What happens to a dedicated IP if I stop sending?

Reputation fades. An IP dormant for around 30 days is treated much like a new one, and resuming at previous volume will trigger throttling. Dedicated IPs need sustained, regular sending to stay warm, which is the main reason low-volume or seasonal senders are better off on a shared pool.

What is the difference between a dedicated IP and dedicated infrastructure?

A dedicated IP is usually a single address assigned to your account on the provider’s existing shared platform — you often warm it yourself. Dedicated infrastructure means mail servers provisioned for your traffic alone, with IPs used only by you. The practical questions are whether the IPs are yours exclusively, whether they are named in the dashboard, and who runs the warmup.

Can I get dedicated IPs below the usual volume threshold?

Yes, when the requirement is isolation rather than deliverability — customer allowlisting, contractual isolation, or a regulated environment. The caveat is that an unmanaged dedicated IP below the volume floor will underperform. What makes it workable is a managed tier where the provider warms the IPs, paces them for your actual volume, and monitors reputation continuously.

Critical email deserves infrastructure you can name

Start on Sandbox in minutes. Move to a transactional-only Pool for production, or to Dedicated servers and IPs that we build, warm, and manage for you.