Dedicated IP vs shared IP for cold email is a question most senders should refuse to answer, because it is being asked from the wrong decade. The choice comes from the world of email service providers, where transactional and marketing platforms sell IP addresses to customers pushing six figures of monthly volume. Cold email in 2026 does not run on that infrastructure. It runs from Google Workspace and Microsoft 365 inboxes, which send through enormous shared pools you cannot opt out of, and which are judged almost entirely on domain reputation. The verdict up front: if you send cold outreach from Google or Microsoft inboxes, the IP question is already answered for you, and answered correctly. And if anyone offers you a dedicated IP for a cold email program, the volume math below says it would hurt you, not help you.
That answer comes from sending, not from theory. Our parent agency, Referral Program Pros, has booked more than 7,000 meetings across 4,000+ outbound campaigns, and not one of those campaigns ran on a dedicated IP. GTM Bud is built on the same playbook. IP choice has never been one of them.
What is a dedicated IP, and what is a shared IP?
A dedicated IP is a sending IP address used by one sender alone, so every message that shapes the address’s reputation is yours. A shared IP is an address an email service provider pools across many customers, whose combined traffic builds one collective reputation. The decision rule follows directly from volume: an IP’s reputation is built by consistent sending history, so a dedicated IP only outperforms a shared pool when you send enough mail, steadily enough, to write that history yourself. Below that volume the shared pool wins, because the pool is already warm and already trusted at a scale no small sender reaches alone. This is why the choice is an ESP-era question. It assumes you are pushing bulk mail through SendGrid, Mailgun, Postmark, or Amazon SES, platforms where IP addresses are part of the product. Cold email sent from Google Workspace or Microsoft 365 inboxes never touches that decision at all.
Keep that definition pair in hand, because every vendor pitch for a dedicated IP leans on the word control. Control of an IP’s reputation is only valuable if you can feed the IP enough consistent volume to give it a reputation worth controlling. The next two sections put numbers on exactly where that line sits.
Does the dedicated IP question even apply to cold email senders?
For most cold email senders the dedicated IP vs shared IP decision does not exist, because most cold email is sent from Google Workspace and Microsoft 365 inboxes, and neither provider sells you a sending IP. Mail from a Workspace inbox leaves through Google’s own shared infrastructure, mail from a Microsoft 365 inbox leaves through Microsoft’s, and both providers pool enormous numbers of customers across those addresses. An IP shared by that many senders tells an inbox provider almost nothing about any one of them, so Google and Microsoft weigh the signal that actually distinguishes you: your sending domain. Google’s own limits close the loop. A Workspace account can send at most 2,000 messages in a rolling 24 hours, per Google’s published Gmail sending limits, a ceiling that keeps every individual sender far too small to register against a pool-sized address. The reputation you can build, and the reputation you can burn, belongs to your domain.
This is not a limitation to route around. The provider pools are the best-policed, most-trusted IP space in email, and sitting inside them is an asset a cold sender could never buy. Which provider’s pool suits your recipient list better is a separate, smaller question, and our comparison of Google Workspace vs Microsoft 365 for cold email covers it with verified pricing and the placement evidence. The scoring system those pools feed into, spam complaints, engagement, authentication, and the rest, is mapped in our cold email deliverability guide.
When does a dedicated IP actually make sense?
The ESPs publish their own thresholds, and they disagree with each other by a factor of six. SendGrid’s IP warm-up guide recommends a dedicated IP only for senders above 50,000 emails per month. Mailgun’s help center sets the line at a consistent 100,000 emails per month, and suggests roughly one dedicated IP per million monthly emails as volume grows beyond it. Postmark will not sell a dedicated IP at all below 300,000 emails per month, at 50 dollars per IP per month on its higher plans, because in its reasoning that is the volume required to establish an IP’s reputation with each receiving provider. The spread itself is the message: even the ESPs that profit from selling dedicated IPs place the entry point somewhere between 50,000 and 300,000 monthly emails. A cold email program measured in thousands of monthly sends is not in the conversation.
Note who these thresholds are written for. The senders clearing 100,000 emails a month on one platform are sending receipts, password resets, newsletters, and product announcements: mail that recipients signed up for, concentrated through a single pipe. That concentration is what makes a dedicated IP coherent. The sender has the volume to warm the address, the consistency to sustain it, and a complaint rate low enough to keep the reputation they build worth having. Cold email fails all three conditions by design, which is the subject of the next section.
Why a dedicated IP actively hurts a low-volume sender
A dedicated IP is not a better IP. It is an empty one. Whether that emptiness is an asset or a liability is decided entirely by how much mail you can pour into it, and how steadily.
A new dedicated IP arrives with zero sending history, so inbox providers treat its first mail with suspicion until a track record exists. Building that record is a formal process: ESP warm-up schedules ramp a fresh dedicated IP over weeks, which is why SendGrid publishes an entire guide on the procedure, and why our own data on how long it takes to warm up a new email domain shows vendor timelines spanning two weeks to three months for the matching domain-level process. A shared pool skips all of it. The pool is already warm the day you join, kept warm by everyone else’s traffic.
Worse, a low-volume sender can never finish the job. Mailgun’s guidance states that sending too few messages from a dedicated IP can be as damaging as sending too many, and that mailbox providers treat volume spikes from an address with thin history as a spam signal. Cold email is structurally low-volume and structurally spiky: a few thousand sends a month, paused for list rebuilds, redistributed across domains the moment anything degrades. On a dedicated IP, every one of those healthy cold email habits reads as erratic behavior from an unknown address. The sender who bought the IP for control ends up with an address whose reputation never consolidates, carrying mail that the warm shared pool would have delivered better on day one.
Dedicated IP vs shared IP for cold email at a glance
| Sender profile | Typical monthly volume | Right answer | Why |
|---|---|---|---|
| Cold email sender on Google Workspace or Microsoft 365 | 1,000 to 10,000 per domain | Provider shared pool (no choice exists) | Neither provider offers dedicated IPs; domain reputation carries the entire signal |
| Product sending transactional mail through an ESP | Under 50,000 | ESP shared pool | Below every published ESP threshold; the pool is already warm |
| Marketing or transactional sender scaling up through an ESP | 50,000 to 300,000 | Evaluate a dedicated IP | SendGrid’s guide starts the conversation at 50,000, Mailgun’s at a consistent 100,000 |
| High-volume platform or bulk sender | 300,000 and up | Dedicated IPs, possibly several | Postmark’s minimum to buy one at all; Mailgun suggests about one IP per million monthly emails |
Read the first row twice, because it is where nearly every cold email operation lives, including every campaign we run. The rows below it describe a different business: products and marketing programs pushing consented mail through sending APIs. If your cold outreach somehow routed through that stack, the volume alone would disqualify a dedicated IP, and the complaint profile of unsolicited mail would make one a liability even past the thresholds.
What actually moves deliverability at cold email volume
Run the math on a serious cold email program. Even 200 sends a day, which is more than the 100 to 150 per day a single warmed domain should carry per our sizing guide on how many domains and inboxes you need for cold email, totals about 6,000 emails a month: roughly an eighth of SendGrid’s lowest dedicated IP threshold and a fiftieth of Postmark’s. At that scale the levers that decide inbox placement are the ones scored against your domain. Warm-up discipline before full volume. Authentication that passes and aligns. Per-inbox daily caps in the 20 to 50 range. List quality tight enough to keep complaints near zero. Lookalike sending domains so your root domain never carries the risk. None of these involve an IP address you can choose.
This is exactly the layer GTM Bud operates for you. A cold email automation tool built on agency practice provisions the sending domains, manages warm-up, holds the volume caps, and rotates inboxes, so the infrastructure questions that actually matter are handled before your first campaign, and the one that does not matter never comes up.
Frequently asked questions about dedicated IPs and shared IPs for cold email
Can you get a dedicated IP with Google Workspace or Microsoft 365?
Not from the providers themselves. Both send all customer mail through their own shared infrastructure and offer no dedicated sending IP option. Workarounds exist that route Workspace or Microsoft 365 mail through an external SMTP relay selling dedicated IPs, but that trades the warm, trusted provider pool for an empty address you must warm yourself, which is a downgrade at cold email volume. Treat the IP layer as settled and put the effort into your domains.
Do you need to warm up a shared IP before sending cold email?
No. A shared pool is already warm, because the combined traffic of every sender on it maintains a continuous reputation history, a point Mailgun’s shared IP guidance makes explicitly. What still needs warming is everything that identifies you: the sending domain and the individual inboxes, which start from zero history of their own. Our guide to how long domain warm-up takes covers the timelines and the ready signals.
Does a shared IP expose you to other senders’ bad behavior?
Partially, and far less than the fear suggests. On an ESP pool one spammer can degrade the address for everyone, which is why reputable ESPs police their pools aggressively. On Google and Microsoft infrastructure the pools are so large, and so central to email itself, that providers score the sending domain instead. Cold campaigns live or die on their own domain reputation, list quality, and volume discipline, not on pool neighbors.
How much does a dedicated IP cost in 2026?
Amazon SES lists a standard dedicated IP at 24.95 dollars per month per IP on its published pricing page, with warm-up left to you, and Postmark charges 50 dollars per IP per month, sold only to senders at 300,000 monthly emails or more. The subscription is the small part. The real price is weeks of ramped warm-up and the permanent obligation to keep volume high and steady enough to sustain the address’s reputation.
Should an agency running cold email for many clients buy dedicated IPs?
No. Agency cold email scales horizontally across client domains and inboxes, never vertically through one high-volume pipe, so no single sending identity approaches the volumes where ESP thresholds begin. Each client gets separate lookalike domains and warmed inboxes with per-domain caps, which isolates every client’s reputation. That is the architecture behind our cold email for agencies setup, and a shared dedicated IP across clients would collapse it in one bad campaign.
Spend the worry on your domains, not your IPs
The dedicated IP vs shared IP question for cold email resolves in one move once you see whose question it is. It belongs to transactional and marketing senders concentrating six figures of monthly volume through an ESP, where the published entry points run from SendGrid’s 50,000 emails per month to Postmark’s 300,000, and where an empty IP can earn its keep. Cold email senders on Google Workspace and Microsoft 365 are never offered the choice, would be an order of magnitude short of the thresholds if they were, and are scored on domain reputation either way. The correct amount of time to spend on IP strategy is zero, reinvested into warm-up, domain architecture, and list quality.
If you would rather not spend the time on those either, GTM Bud provisions the sending domains, runs the warm-up, and executes campaigns on the playbook that booked 7,000+ meetings. Start with done-for-you outbound and let the infrastructure questions stay answered.