Updated Jul 2, 2026
TL;DR: For most cold email, you don't choose your IP at all: Workspace and Microsoft 365 mailboxes send over the provider's managed shared pools. A dedicated IP only pays off when you run your own SMTP at high, steady volume, above roughly 100,000 emails a year. Below that, shared wins.
Search "dedicated IP vs shared IP for cold email" and you'll find a hundred articles arguing both sides as if the choice is yours to make. Here's the part most of them skip. If you run cold outreach the standard way, through Google Workspace or Microsoft 365 mailboxes, you don't pick your IP. Those mailboxes send over the provider's own managed IP pools, which are shared by design and not something you can swap for a dedicated address.
The dedicated-versus-shared decision only becomes real when you step off provider mailboxes and onto your own SMTP server, or onto an email service provider that sells dedicated IPs as a paid add-on. And even then, for the volumes most cold senders run, a dedicated IP is usually the wrong call. It sounds like control. In practice it's a reputation you have to feed, and cold email rarely feeds it enough.
This guide draws the line clearly: who actually has the choice, what each option does to deliverability, and the volume threshold where a dedicated IP finally earns its keep.
Key Takeaways
- Most cold senders never make this choice. Workspace and Microsoft 365 mailboxes send over the provider's managed shared IPs, and you can't replace them with a dedicated one.
- A dedicated IP only helps at high, consistent volume. SendGrid's own threshold is roughly 100,000 emails a year before a dedicated IP is worth it, and cold email's low-per-inbox model rarely hits that on a single IP.
- A starved dedicated IP is a liability, not an asset. Providers want steady sending patterns, so an IP that trickles out 30 emails a day builds a thin, fragile reputation.
- For Gmail, your domain reputation is the primary signal anyway. That's the thing to protect, with authentication, warmup, and clean lists, not the IP.
- Shared pools carry one real risk (noisy neighbors), but the giant provider-managed pools behind Workspace and Microsoft 365 are a different animal from a sketchy bulk-ESP pool.
The short answer: who actually gets to choose
Before debating which is better, figure out which scenario you're in. There are three, and only two of them involve a real decision.
Scenario 1: Workspace or Microsoft 365 mailboxes (the typical cold email setup). You provision mailboxes on real Google or Microsoft tenants, connect them to your sequencer, and send. Your mail leaves over Google's or Microsoft's IP pools. Those are shared across millions of senders, and you have zero say over which IP your message uses. There is no dedicated-IP option here, full stop. If this is you, the rest of the dedicated-versus-shared argument is academic. Skip to the section on protecting domain reputation, because that's the lever you actually control.
Scenario 2: An ESP or relay that offers both. Services like SendGrid, Mailgun, or Amazon SES let you send through a managed shared pool or pay for a dedicated IP. Here the choice is real, and the volume math below decides it.
Scenario 3: Your own mail server (self-hosted MTA). You run Postfix or a similar stack on a server with its own static IP. That IP is dedicated by definition, and every bit of its reputation is yours to build and yours to ruin.
Most cold operators live in Scenario 1 and think they're in a position to choose. They aren't. That single distinction settles the question for the majority of readers. The genuine decision belongs to a smaller group running their own infrastructure or an ESP relay, and that's who the rest of this is written for. For the broader question of which provider to send from in the first place, our email provider comparison breaks down Workspace, Microsoft 365, and the alternatives.
Dedicated IP vs shared IP, defined
A shared IP is one address (or a pool of them) used by many senders at once. The pool operator manages it, keeps it warm, and spreads reputation across everyone on it. A dedicated IP is a static address used by exactly one sender. Its reputation reflects only your sending, nothing else.
That sounds like a clean win for "dedicated," and that's the trap. Sole ownership of a reputation cuts both ways. Nobody else can drag it down, but nobody else holds it up either. On a dedicated IP, the entire weight of building and maintaining a trusted reputation falls on your volume and your hygiene alone.
Here's the honest tradeoff side by side:
Factor | Shared IP (managed pool) | Dedicated IP |
|---|---|---|
Who controls reputation | Pool operator plus every other sender | You alone |
Warmup burden | Low: the pool is already warm | High: weeks of ramp on a cold IP |
Volume needed to stay healthy | None in particular | Steady, high, and ongoing |
Cost | Usually included | Paid add-on, billed per IP |
Best for | Most cold senders, low or variable volume | High, consistent volume on your own MTA |
Main risk | Noisy neighbors | Under-sending your own IP |
Google is blunt about the shared-pool dynamic in its sender guidelines: "The activity of any senders using a shared IP address affects the reputation of all senders for that shared IP address." That line is usually quoted as the case for a dedicated IP. Read it again with cold email volumes in mind, and it's also the case against one, because it means a dedicated IP only protects you if you can carry the reputation by yourself.
IP reputation for cold email needs volume you probably don't have
This is the core of the whole decision, so let's be precise about it. A dedicated IP earns trust the same way a domain does: through consistent, predictable sending that mailbox providers can learn to recognize. Starve it, and the reputation never sets.
SendGrid puts a number on the floor. Its guidance is that a dedicated IP starts to make sense once you're sending over 100,000 emails a year, and that below that, "it might not be worth moving to a dedicated IP." The reason is reputation mechanics, not billing. As SendGrid explains, "ISPs like to see consistent sending patterns, so if you don't have high enough activity levels for a dedicated IP, this could impact your reputation and result in low email delivery rates."
Now hold that against how cold email actually works. Safe per-mailbox volume is small on purpose. A warmed cold-email inbox sends on the order of dozens of messages a day, not thousands, and you scale by spreading sends across many mailboxes rather than blasting from one. That distributed model is the right way to protect deliverability, and we cover the mechanics of it in our guide to inbox rotation. But it's the exact opposite of what a dedicated IP wants. A dedicated IP wants concentrated, steady, high volume. Cold email gives it thin, fragmented volume scattered across inboxes.
Run the math. Even 100,000 emails a year is roughly 274 a day, every single day, with no gaps. Plenty of cold campaigns send in bursts: heavy on launch weeks, quiet between. A dedicated IP punishes that rhythm. Google's own advice is to "start with a low sending volume to engaged users, and slowly increase the volume over time" and to avoid sending in bursts. A managed shared pool absorbs your uneven cadence inside its much larger steady flow. A dedicated IP doesn't get that cushion. Every quiet stretch reads as a reputation that went cold.
There's a second reason this matters less than the IP-obsessed crowd claims. For Gmail, the primary reputation signal is your domain, not your IP. That's why authentication and domain warmup do more for cold-email placement than any IP decision. Set up SPF, DKIM, and DMARC correctly, warm the domain properly, and keep the list clean, and you've moved the levers that actually decide whether Gmail trusts you.
Shared IP pool deliverability risks (and which ones are real)
The standard argument for a dedicated IP is the noisy-neighbor problem: if you share an IP with spammers, their behavior can tank a reputation you didn't earn. That risk is real, but it's not uniform across every shared pool, and conflating them is where a lot of bad advice comes from.
Two very different things both get called "shared IP":
- Provider mailbox pools (Google, Microsoft). Enormous, heavily managed, and trusted by default across receiving servers. Google and Microsoft have every incentive to police abuse on their own ranges, and they do. This is the pool your Workspace and Microsoft 365 cold-email mailboxes use, and it's about as safe as shared gets.
- Bulk-ESP shared pools. Smaller, more variable, and only as clean as the operator's vetting. A reputable ESP curates these tightly and quarantines bad senders fast. A cheap, lax one is where the horror stories come from.
So "shared IP pool deliverability risks" is a fair concern, but the fix usually isn't a dedicated IP. It's choosing a well-managed pool. If you're on provider mailboxes, you're already on the best-managed shared infrastructure available. If you're on an ESP, the quality of the operator's abuse controls matters far more than whether your specific IP is shared or dedicated.
When does the noisy-neighbor risk justify isolation? When you can't trust the pool and you have the volume to stand alone. M3AAWG frames the isolation case well: "When an entity's mail or list quality is unknown or poorer than average, the sender or ESP may wish to isolate the mail from other entities to protect them." Note the framing. Isolation protects other senders from a risky stream as much as it protects the risky stream from them. That's a deliberate infrastructure decision, not a default.
When to use a dedicated sending IP
So when does a dedicated IP actually win? The honest list is short, and every item assumes you've left provider mailboxes behind and run your own MTA or an ESP relay.
- High, consistent volume. You're sending well past the 100,000-a-year floor on a single stream, with steady daily cadence, not lumpy bursts. SendGrid recommends scaling further as you grow: "When your volume reaches 250,000 messages each month, allocate a minimum of two dedicated IP addresses." If you're nowhere near that on one IP, this case doesn't apply to you.
- You want full control and accountability. A dedicated IP's reputation is purely a function of your own sending. For senders who instrument everything and want a clean, single-cause signal to monitor, that isolation is worth the warmup cost.
- You're isolating a known-risky or unusual stream. Per the M3AAWG logic above, if one segment of your sending is experimental or higher-risk, putting it on its own IP keeps a problem contained instead of letting it bleed into everything else.
- Compliance or contractual requirements. Some enterprise and regulated senders are required to send from a known, fixed IP for allowlisting or audit reasons. That's a hard requirement, not a deliverability optimization.
Notice what's not on the list: "I want better cold-email open rates." A dedicated IP, run below its volume floor, will more often hurt placement than help it. The control you gain is also the responsibility you inherit, and for most cold programs that's a bad trade.
The warmup tax nobody mentions
Switch to a dedicated IP and you don't inherit a clean slate, you inherit a blank one. A brand-new IP has no reputation at all, and mailbox providers treat unknown IPs with suspicion until they've watched them behave. So before that IP carries real campaigns, you have to warm it: ramp volume gradually so providers see steady, predictable growth instead of a sudden spike.
In practice that's commonly a two-to-six-week ramp, often starting around 50 to 100 messages a day and climbing slowly as the numbers hold. The principle is the same one Google states for any sender: start low, increase over time, and don't send in bursts. The methodology is identical to warming a domain, which our complete guide to email warmup walks through in full.
Here's the kicker. A managed shared pool is already warm. Move to it and you borrow an established reputation from day one. Move to a dedicated IP and you pay the warmup tax up front, then keep paying it forever in the form of the steady volume the IP needs to stay warm. For a high-volume sender, that's a one-time cost against a long-term gain. For a cold program sending in waves, it's a recurring penalty with no payoff.
One more wrinkle specific to provider mailboxes: tenants on official Microsoft or Google IP ranges warm up faster than a fresh self-hosted IP, because the ranges already carry trust. The IP isn't the thing you warm in that setup. Your domain and mailboxes are. That's another reason the provider-mailbox path sidesteps the whole dedicated-IP burden.
The decision framework
Strip away the noise and the call comes down to your infrastructure and your volume. Match your setup to the row:
Your setup | Recommendation |
|---|---|
Workspace or Microsoft 365 mailboxes (typical cold email) | Provider shared pool. No dedicated IP is available. Put your effort into domain reputation. |
ESP relay, under ~100,000 emails/year | Managed shared pool. You lack the steady volume to keep a dedicated IP healthy. |
ESP relay or own MTA, 100,000+/year and consistent | A dedicated IP starts to make sense. Warm it properly first. |
250,000+ messages/month | Multiple dedicated IPs, split by stream or message type. |
A risky, experimental, or compliance-bound stream | Dedicated IP for isolation, regardless of volume. |
If you read that table and landed on the first or second row, you have your answer, and it's shared. The dedicated-IP question is a high-volume, own-infrastructure question wearing a cold-email costume.
How the IP decision fits the rest of your stack
The IP is one layer in a sending system, and it's rarely the layer that decides your fate. Reputation isolation, for instance, is usually better handled at the domain level than the IP level. Sending cold outreach from a separate subdomain shields your primary brand from any damage, and that structural choice does more for most senders than a dedicated IP would. We cover the tradeoffs in subdomain vs root domain for cold email.
Whatever you land on, you can't manage what you don't measure. A dedicated IP especially demands monitoring, because you're the only one watching it. Track reputation and complaint signals on a cadence, which our guide to domain health monitoring lays out. And if you're assembling the whole sending operation from scratch, the cold email infrastructure guide puts the IP question in the context of domains, mailboxes, authentication, and routing, where it belongs.
The throughline: for cold email, deliverability is won at the domain and content level far more often than at the IP level. Pick the path that lets you spend your attention there.
Common questions
Does a dedicated IP improve cold email deliverability?
Usually not, for the typical cold sender. A dedicated IP only helps when you can feed it high, steady volume, and most cold programs can't, because safe per-mailbox volume is deliberately low and sending is often bursty. Below SendGrid's rough 100,000-emails-a-year floor, an under-used dedicated IP tends to build a weaker, more fragile reputation than a warm managed pool would give you. For Gmail in particular, domain reputation is the primary signal, so that's where the deliverability gains actually live.
Can I get a dedicated IP with Google Workspace or Microsoft 365?
No. When you send cold email through Workspace or Microsoft 365 mailboxes, your mail leaves over the provider's own shared IP pools, and you can't request or assign a dedicated IP for those mailboxes. Dedicated IPs are only an option if you send through an ESP that sells them as an add-on, or if you run your own mail server with a static IP. For the standard cold-email setup, the IP is the provider's call, not yours.
How much volume do I need to justify a dedicated IP?
As a floor, plan for steady volume above roughly 100,000 emails a year on a single stream before a dedicated IP is worth it, and aim higher than that for comfort. The "steady" part matters as much as the total: an IP that sends 274 emails most days holds a reputation far better than one that sends 2,000 on Monday and nothing for a week. At 250,000 messages a month, SendGrid recommends moving to at least two dedicated IPs.
Are shared IPs safe for cold email?
For provider mailbox pools, yes. The shared ranges behind Google Workspace and Microsoft 365 are large, tightly managed, and trusted by default, which makes them the safest shared infrastructure you can be on. Bulk-ESP shared pools vary more, so the operator's abuse controls matter. The real risk in any shared pool is a noisy neighbor whose bad sending bleeds onto everyone, since Google notes that "the activity of any senders using a shared IP address affects the reputation of all senders" on it. With a well-run pool, that risk is small and well worth the warm, ready-made reputation you get in return.
Is IP reputation or domain reputation more important for cold email?
For Gmail, domain reputation is the primary driver, and it follows your domain across whatever IP you happen to send from. That's why a dedicated IP isn't the deliverability shortcut it's sold as. The work that moves the needle is domain-level: authentication, a properly warmed domain, clean lists, and low complaint rates. The IP is a supporting signal, not the headline.
The bottom line
Dedicated versus shared is a real decision for a narrow group: high-volume senders running their own MTA or an ESP relay, with the steady cadence to keep a dedicated IP warm. For everyone else, and that's most cold senders, the question is settled before it's asked. Workspace and Microsoft 365 mailboxes put you on managed shared pools you can't change, and those pools are the right place to be.
If you do run your own infrastructure, use the volume floor as your gate: roughly 100,000 steady emails a year on a single stream before a dedicated IP pays off, and a second IP past 250,000 a month. Under that, a dedicated IP is a reputation you can't afford to feed, and a warm shared pool will outperform it.
Either way, point your effort where deliverability is actually won: a clean domain, real authentication, disciplined warmup, and a list that doesn't generate complaints. MailBeast handles the warmup, rotation, and monitoring side of that so the IP question stops being something you lose sleep over. Get the domain layer right, and the IP mostly takes care of itself.
Sign up for free
