Dedicated IP vs Shared IP for Cold Email: When It Pays Off
Dedicated IP vs shared IP for cold email: the volume and consistency threshold that decides it, what each one costs and the signs you are ready to switch.
ON THIS PAGE · 7 SECTIONS
THE SHORT VERSION
- A dedicated IP only helps once you send enough volume, consistently, to keep it warm by yourself.
- SendGrid's own documentation puts the floor at 50,000 messages a month and recommends two dedicated IPs once you pass 250,000.
- Most cold email never gets near that. It runs through Google Workspace or Microsoft 365 inboxes, where you do not pick the IP at all.
- Below the threshold, a shared pool with a strong history beats a dedicated IP nobody has warmed yet.
- The real question is not who owns the IP. It is whether your sending pattern can keep one warm every day without a gap.
A dedicated IP pays off for cold email once you send a steady 50,000 or more messages a month through a platform built for IP level sending, not a Google Workspace or Microsoft 365 inbox. Below that volume, a shared pool with an established reputation delivers better than a dedicated IP a small team cannot keep warm on its own. In cold email, the IP is rarely the real bottleneck. Domain reputation, inbox count and send pattern usually matter more.
What a dedicated IP actually changes
A dedicated IP means your sending reputation lives on one address that nobody else touches. Every bounce, every spam complaint, every spike in volume is yours alone, for better or worse. A shared IP pools your reputation with every other sender on that same address inside an email service provider's infrastructure. Their good behavior buffers your mistakes and your mistakes can drag their average down. The tradeoff runs both ways, which is exactly why a well chosen shared pool often beats a brand new dedicated IP with no history behind it.
That choice only exists inside bulk sending platforms like SendGrid, Mailgun, Postmark or Amazon SES, where IP allocation is a setting you control. It does not exist inside Google Workspace or Microsoft 365. Google's own documentation caps a Workspace account at 2,000 messages a day per user and 1,500 for mail merge sends. Those messages route through Google's own IP infrastructure that no customer can dedicate or swap (Google Workspace Help). Most cold email, sent from a pool of warmed mailboxes rather than a bulk API, runs on exactly this kind of infrastructure. The same is true of Microsoft 365, where Microsoft manages the sending IPs behind the scenes and a customer account has no setting that touches them. The dedicated IP question only becomes real once sending moves off individual mailboxes and onto an ESP.
When is a dedicated IP worth it for cold email?
A dedicated IP is worth it when three things are all true at once: you send through an ESP that offers IP allocation, your monthly volume clears a real threshold and that volume arrives on a predictable schedule rather than in bursts. Twilio's SendGrid, one of the largest dedicated IP providers, documents 50,000 messages a month as the point where a dedicated IP starts to make sense and recommends splitting into at least two dedicated IPs once a sender passes 250,000 messages a month, separating transactional traffic from marketing or outbound sends (Twilio SendGrid). Amazon SES draws a similar line in its own guidance, pointing senders toward its managed warmup option below that volume and a self managed dedicated IP above it.
Below that line, a dedicated IP works against you. A fresh IP has no sending history, so every mailbox provider treats its early traffic with suspicion until weeks of consistent volume prove it is not a spam source. A team sending 10,000 emails a month on a dedicated IP spreads that volume so thin that the IP never builds enough signal to earn trust. Any gap in sending resets the clock. A cold email program with normal holiday or pipeline lulls hits that gap often. The same 10,000 emails on a shared pool inherit the pool's existing reputation from day one, with no warmup period to manage at all.
Source: Twilio SendGrid, Google Workspace Help and Amazon SES pricing.
Why cold email rarely touches this decision
Cold email runs on a different shape of infrastructure than a newsletter or a product email. You are not pushing one burst of volume through a bulk API. You are running a pool of mailboxes, each sending 15 to 30 messages a day, spread across several domains, warmed up over weeks. We walked through the inbox and domain math in how many inboxes and domains you need and the full setup in outbound infrastructure architecture. Neither depends on who owns the IP, because the mailboxes sit on Google or Microsoft's shared infrastructure by default.
THE USUAL WAY
- A vendor pitches a dedicated IP as a deliverability upgrade.
- The team buys one without checking monthly volume first.
- Sending is uneven while the campaign gets finished, so the IP goes cold for days at a time.
- The IP never builds a reputation and starts getting throttled.
THE SYSTEM WAY
- Match infrastructure type to the channel: mailbox pools for outbound, ESP IPs for bulk send.
- Check the monthly number against the 50,000 and 250,000 thresholds before buying anything.
- Confirm the schedule can hold steady for weeks, not just the launch week.
- Keep the dedicated IP question off the table until mailbox sending genuinely cannot scale further.
FROM THE YARD
We have seen a team buy a dedicated IP block because a provider pitched it as the fix for a flat reply rate, then leave it mostly idle while the actual problem, a stale list and a tired offer, sat untouched. The IP warmed for a week. Then sending paused for the holidays and it went cold again. The deliverability problem was never the IP. It was 40 days of silence on infrastructure that needs daily feeding to mean anything.
The decision rule
Run this before you sign up for anything. If you are below 50,000 messages a month, stay on a shared pool or on warmed Workspace and 365 mailboxes. If you are above it but your volume swings by more than 30% week to week, fix the schedule first, because an uneven dedicated IP performs worse than a steady shared one. Only move once both the volume and the consistency hold for at least a month.
| Factor | Shared IP | Dedicated IP |
|---|---|---|
| Minimum volume to justify it | None, works from day one | 50,000+ messages a month |
| Who controls reputation | Shared with other senders on the pool | Only you |
| Warmup needed | Already warm from pool history | 4 to 6 weeks, manual or automated |
| Cost | Included in most ESP plans | From $24.95 a month per IP on AWS SES |
| Risk if sending goes quiet | Low, the pool keeps it warm | High, a cold IP loses trust fast |
| Typical cold email fit | Mailbox pools on Workspace or 365 | Bulk lifecycle or transactional email at scale |
What changes between a shared and a dedicated IP and which one fits a typical cold email setup.
How to move to a dedicated IP without burning it
- Confirm the volume is real and steady. Pull three months of send history. One good month does not count, the pattern has to hold.
- Keep a shared IP as overflow during warmup. SendGrid's automated warmup requires at least one other warm IP to absorb the traffic your new IP cannot yet handle, so do not retire the shared option on day one.
- Follow a warmup schedule, do not skip it. A 4 to 6 week ramp, starting at a few hundred messages a day, is standard across SendGrid and Amazon SES. Cutting it short is the single fastest way to burn a new IP.
- Watch complaint and bounce rate daily, not weekly. Keep complaints near 0.1% and bounces at or below 1% through the whole ramp. A spike in week two means slow the ramp, not push through it.
- Never let the new IP go quiet. A gap of even a few days during warmup resets the reputation clock. If you cannot guarantee daily sending for the next two months, the IP is not ready yet.
A dedicated IP rewards consistency and punishes everything else. It does not fix a weak list or a flat offer.
What to monitor once you have made the switch
A dedicated IP does not remove the need to watch deliverability. It raises the stakes, because there is no pool left to absorb a bad week. Run the same weekly checks we laid out in deliverability monitoring: spam rate, bounce rate, inbox placement and blocklist status, now tracked per IP instead of per domain. If you are weighing whether this belongs on your roadmap at all, that is exactly the kind of call our system is built to make with you before you spend on infrastructure you do not need yet.
Questions about dedicated IP vs shared IP for cold email
Does a dedicated IP improve cold email deliverability?
Only above a real volume threshold, generally 50,000 or more messages a month sent consistently. Below that, a shared pool with an established history outperforms a cold, unproven dedicated IP.
What volume do I need before a dedicated IP makes sense?
SendGrid documents 50,000 messages a month as the point where a dedicated IP starts to help and recommends two dedicated IPs once you pass 250,000 a month. Most cold email teams never reach the first number, because a mailbox pool caps out well before that on purpose to protect domain reputation.
Can I get a dedicated IP with Google Workspace or Microsoft 365?
No. Both run mail through their own shared infrastructure and neither offers customers a dedicated sending IP. Dedicated IPs only exist as a setting inside bulk ESPs like SendGrid, Mailgun or Amazon SES.
What happens if a dedicated IP goes quiet for a few days?
It loses the reputation it built. Mailbox providers read a sudden stop as a sign the sender has gone dormant or been compromised, so a resumed send often gets treated like a brand new, unproven IP again.
Is a dedicated IP worth it for a single outbound campaign?
Almost never. A single campaign rarely hits the volume to warm an IP properly and the IP sits cold between campaigns. A warmed mailbox pool handles single campaigns far better, since it never needs a dedicated ramp in the first place.
Sources: Twilio SendGrid, dedicated IP addresses, Google Workspace Help, Gmail sending limits, Amazon SES pricing.