Skip to content

Cold email · Infrastructure

Cold email infrastructure for teams that want to send, not to configure MX records.

Pre-warmed Google Workspace or Microsoft 365 mailboxes, isolated IP pools, domain rotation, OAuth for Instantly, Smartlead, Lemlist and Apollo. Built by operators who actually run cold outreach themselves — not by a hosting company that read about it.

  • Pre-warmed inboxes
  • OAuth integration
  • Google & Microsoft
  • Domain rotation
  • Daily seed testing
  • From €79/mo

Why dedicated cold email infra

Cold email in 2026 is an infrastructure problem, not a copy problem.

Five years ago you could run a cold email campaign from a single Google Workspace mailbox and hit inboxes. Today, Google's sender requirements and Microsoft's algorithmic filtering mean that a single mailbox sending 200 cold emails a day is a reputation risk. The operators winning in 2026 run pools of 10 to 100 pre-warmed mailboxes across multiple domains, rotate sending across them, monitor placement daily, and treat infrastructure as seriously as they treat copy.

Blue Spirit's Cold Email Infrastructure product is that infrastructure, delivered as a service. You plug the mailbox credentials into Instantly, Smartlead, Lemlist or Apollo and start sending. We own the boring parts: provisioning, warm-up, domain configuration, DNS records, mailbox replacement, inbox placement monitoring.

Pre-warmed mailboxes

Every mailbox arrives with 2 to 4 weeks of engagement history and proper SPF, DKIM, DMARC and MTA-STS published. You start sending from a healthy baseline.

Domain and IP diversity

Mailboxes spread across multiple sending domains; sending IPs isolated from other clients. No reputation contagion.

OAuth-ready integration

Credentials delivered as OAuth tokens for Google Workspace and Microsoft 365. Drop straight into Instantly, Smartlead, Lemlist, Apollo, Reply.io or Woodpecker.

Daily inbox placement testing

Test send to a private seed list every day. Inbox placement per ISP tracked, alerts on regressions, mailbox rotation when thresholds are crossed.

Suppression list sync

Global suppression list synced across all mailboxes so a bounce in one does not become a complaint in another.

White-label for agencies

On Agency plans, we can deliver the infrastructure under your brand: custom sending domains, branded reporting, separate client workspaces.

2026 reality

Cold email in 2026: the rules changed three times in 18 months.

If your cold email playbook was written before February 2024, it is dangerously out of date. Three regulatory waves reshaped the channel between early 2024 and mid-2025, and the industry has been catching up ever since. Understanding which wave you are operating in matters more than picking between Instantly and Smartlead.

The first wave was Google and Yahoo's bulk sender requirements, effective February 1, 2024. Every sender pushing more than 5,000 emails per day to Gmail must publish a DMARC policy, pass SPF and DKIM alignment, implement RFC 8058 one-click unsubscribe, and maintain a complaint rate sustainably under 0.3% (the public spec) or 0.1% (the operational reality). Domains failing those checks get throttled or filtered to spam regardless of content. Cold email programmes operating at meaningful B2B volume crossed the 5,000/day threshold easily, which means these requirements applied to you whether you knew it or not.

The second wave was Microsoft's bulk sender enforcement in May 2025, with comparable rules and shorter notice. Outlook.com and Office 365 tightened complaint thresholds, increased DMARC scrutiny and rolled out more aggressive backoff signaling. The 4.7.500–4.7.799 error code family expanded to include richer reputation telemetry that operators can use to diagnose, but only if they are looking. Most cold email tools are not surfacing this telemetry to users.

The third wave is per-mailbox volume detection. Both Google and Microsoft now apply pattern detection at the individual mailbox level, not just the domain or IP level. A mailbox sustainably sending 50+ cold emails per day to recipients who do not engage gets reputation-derated independently of how clean your DNS is. This is why "buy 5 Google Workspace seats and rotate them" stopped working as a strategy in 2025 — the pool count needed to support a meaningful cold programme tripled.

The practical consequence: cold email is now an infrastructure-first discipline. The teams winning in 2026 spend more on mailboxes, domains and monitoring than they do on the sequencing tool itself. The math has flipped from where it was in 2022.

Decision framework

Should cold email even be your channel? Decision tree.

Most cold email infrastructure articles assume cold is your channel. We do not. The decision tree below is the same one we walk through on the discovery call to determine whether cold infrastructure is the right answer or whether you should be doing something else. The honest version, including the branches where the answer is "do not buy this".

Should I run cold email infrastructure? Are your prospects identifiable B2B contacts with professional role relevant to your offer? No Cold is not your channel. Try LinkedIn ads, content, paid Yes More than 5,000 cold emails / month? No Do it yourself. 5 GW seats + Smartlead warmup Yes In-house engineer to manage 30+ mailboxes & rotation? Yes Mailforge or Inframail. Flat-rate, but you run warmup & ops No Agency reselling to multiple clients? Yes Blue Spirit Agency. Per-client isolation, white-label, branded reporting No Blue Spirit Growth.

Two sub-branches the diagram does not capture. First: even when DIY is the technically right path (under 5,000/mo + in-house resource), most teams underestimate the ongoing maintenance load — Google security challenges, mailbox health monitoring, replacement when one burns. The DIY math looks great in month one and degrades quickly. We have lost prospective clients to the DIY branch; many of them are back six months later. Second: the agency branch versus the Growth branch is not always obvious — if you run a single brand with multiple verticals or product lines that need reputation isolation, the Agency tier may be right even though you are not strictly "reselling". We figure that out together on the call.

How our Cold Email Infrastructure is architected

Every deployment starts with domain planning. We help you pick 5 to 20 sending domains that are orthogonal to your main brand domain, set up proper DNS (SPF, DKIM with rotating selectors, DMARC with a reporting endpoint we monitor, MTA-STS), and register them in Google Postmaster Tools. Sending domains are disposable by design; your main brand domain never enters the rotation. This is critical and often skipped — if your sending domain is your transactional domain, one bad cold campaign tanks your password resets and order confirmations along with it.

On top of those domains we provision Google Workspace or Microsoft 365 mailboxes — typically 2 to 5 mailboxes per domain, distributed across first-name aliases that look natural (john@, sarah@, alex@ rather than info@, sales@, contact@). Each mailbox enters a 2 to 4 week warm-up period on our network before you touch it. By the time you plug it into Instantly or Smartlead, it has a reputation profile consistent with a real person who uses email normally — sends a few emails a day, opens replies, marks some messages as important.

Sending IPs are the next layer. On Starter we share an isolated pool among cold email clients — isolated meaning you never share a pool with non-cold-email marketing senders, but pool reputation is shared. On Growth and Agency you get dedicated sending IPs within the pool, meaning your reputation is yours alone. Our Cold Email IPs live in US and EU datacenters and are never recycled from other sending purposes — every IP starts cold and we warm it ourselves before it touches your sending.

Finally, monitoring runs continuously. A private seed list spread across Gmail, Outlook, Yahoo and iCloud receives a daily test email from every mailbox. Inbox placement, spam placement and missing-entirely rates are tracked per ISP per mailbox. If placement drops below 70% on Gmail or Outlook for any mailbox, the system pauses sending from that mailbox and triggers a recovery warm-up. If recovery fails after a week, the mailbox is replaced from the pre-warmed pool — you keep sending; the infrastructure rotates underneath. This is what "managed" actually means.

Mailbox rotation pool design — why one box is dangerous, fifty is engineering

The core operational pattern of modern cold email is the rotation pool: a set of mailboxes across multiple domains that share the daily sending load such that no individual mailbox ever crosses the per-mailbox detection threshold. The diagram below shows how a properly designed rotation pool routes a single campaign across the underlying infrastructure.

Your campaign 30,000 cold emails / month Instantly / Smartlead / Lemlist Rotation orchestrator Distributes across pool 25 emails/day/mailbox cap Daily seed-list monitor Per-mailbox placement check Gmail · Outlook · Yahoo · iCloud 6 sending domains × 5 mailboxes per domain = 30-mailbox pool domain-1.outreach john@ · sarah@ · alex@ ben@ · mia@ 5 mailboxes · 125 emails/day domain-2.outreach jamie@ · ryan@ · sam@ erin@ · lee@ 5 mailboxes · 125 emails/day domain-3.outreach kim@ · drew@ · max@ tess@ · noah@ 5 mailboxes · 125 emails/day domain-4.outreach finn@ · iris@ · oscar@ paige@ · ross@ 5 mailboxes · 125 emails/day domain-5.outreach tom@ · lily@ · mark@ jen@ · zoe@ 5 mailboxes · 125 emails/day domain-6.outreach (reserve) 5 pre-warmed mailboxes held in standby replacement on degradation 2 dedicated IPs rotated across all 6 domains for diversity IP A · 198.51.100.10 Domains 1, 3, 5 IP B · 198.51.100.11 Domains 2, 4, 6 Automatic mailbox replacement loop Mailbox placement < 70% (Gmail/Outlook) for 2 consecutive seed-list days → pause sending, run recovery Recovery fails after 7 days → swap from reserve pool campaign continues uninterrupted

Three architectural decisions in this pool design that we get asked about most often. First: why 25 emails/day/mailbox and not 50. The honest answer is that 50/day worked in 2023 and stopped working in 2025. Per-mailbox volume detection at Google and Microsoft means sustained 50/day from a non-established address gets reputation-derated within 2-3 weeks. 25/day is the conservative-but-real ceiling for cold to fresh recipients; existing reply chains can go higher.

Second: why 5 mailboxes per domain and not 10. Above 5-6 mailboxes on a single domain, Google starts treating the domain pattern as bulk-marketing-grade, which lowers the per-mailbox volume tolerance further. The math says: more mailboxes per domain = lower per-mailbox cap = no actual capacity gain. Five is the sweet spot in our 2026 data.

Third: why we keep a reserve domain that never sends until something breaks. The economic argument against this is "you're paying for capacity you do not use". The operational argument for it is that a mailbox replacement that takes 2 weeks is a campaign gap; a mailbox replacement that takes 2 hours is invisible. The reserve is what makes "managed" different from "configured once and abandoned".

How many mailboxes do you really need? — by monthly volume

The biggest mistake new cold email operators make is under-provisioning. They look at a sending tool, see "Instantly handles up to 1000 emails/day per mailbox" and conclude that 5 mailboxes can handle 100K emails per month. The math is right; the deliverability reality is wrong. Real-world cold sending limits per mailbox sit at 25-30 emails per day before reputation degrades — half the marketing-quoted theoretical maximum. The chart below shows realistic infrastructure scaling from 5K to 500K monthly cold sends.

Cold email infrastructure needed by monthly volume

Realistic mailbox/domain/IP requirements at sustained 25-30 emails/day per mailbox cap.

Mailboxes, domains and IPs needed by monthly cold email volume
Categoría Mailboxes neededSending domainsDedicated IPs
5K/mo 831
10K/mo 1552
25K/mo 30103
50K/mo 50175
100K/mo 90308
200K/mo 1605312
500K/mo 35011720

Numbers reflect operational ratios from cold email programmes Blue Spirit operates in 2024-2026, assuming sustainable per-mailbox sending of 25-30 cold emails/day, 3-5 mailboxes per sending domain, and IP-to-domain ratios calibrated to keep per-IP volume under thresholds that trigger Gmail and Microsoft pattern detection. At 500K/month you cross into territory where dedicated PowerMTA infrastructure becomes more economical than mailbox-pool-based cold email — talk to us about hybrid architecture. The numbers are not minimums; they are sustainable steady-state pool sizes.

Two things to note on this chart. First, mailbox count grows roughly linearly with volume, but domain count grows more slowly — the per-domain mailbox count rises modestly with operational maturity. Second, dedicated IP count grows much slower than mailboxes — at 100K/month you need 90 mailboxes but only 8 IPs. That ratio is why our Agency plan is structured the way it is: lots of mailboxes, few IPs, isolation by client workspace.

If your cold email volume is above 500K/month, the math starts favouring dedicated PowerMTA over mailbox pools. At that scale you probably want IP-based sending with proper warmup (see our PowerMTA hosting) rather than mailbox rotation. We can design the hybrid for you.

What happens when you skip the warmup — the 30-day decay

The strongest argument for managed cold email infrastructure is a chart that nobody shows you when you sign up for Instantly or Smartlead. The chart below shows what happens to a brand-new mailbox sending cold from day one — placement starts at 100% (because the platform itself is reputation-neutral toward you), drops sharply once Gmail's filter activates, and lands somewhere near zero by week 4. This is not theoretical; it is the result of running this exact experiment on test domains we deliberately burned to verify the curve.

What happens when you skip warmup — 30-day reputation decay

Inbox placement % from new mailbox to burned domain.

Inbox placement percentage at each stage of skipped-warmup mailbox lifecycle
Categoría Inbox placement %
New domain registered 100
DNS configured 100
Mailbox provisioned 100
No warmup → first cold send 80
Day 7 — Gmail filter activates 50
Day 14 — domain reputation Low 25
Day 21 — most mail in spam 10
Day 30 — domain effectively burned 3

Inbox placement modelled from observed Postmaster Tools data and seed-list testing on test domains burned for measurement. Decay rates vary by sending volume and recipient quality — high-quality lists with low complaint rates may slow the decay curve by 1-2 weeks; low-quality lists or aggressive sending speeds compress it. The endpoint at day 30 is roughly where most domains stabilise: 0-5% inbox placement on Gmail, similar on Outlook. Recovery from this state requires a full domain migration (replacement domains plus 4-6 weeks of warmup), at which point you have spent more than the cost of doing it correctly the first time.

The lesson is simple: you cannot skip warmup. The choices are pay for managed pre-warmup (us), pay for software-driven warmup that runs slowly while you wait (Lemwarm, Warmup Inbox, MailReach — €15-25/mailbox/month, 2-4 weeks before sending), or burn the domain and start over. There is no fourth option that lets you cold send from day one and have it work; the decay curve above is what you get.

Five mistakes we keep finding in DIY cold email setups

Roughly half of our new clients arrive after running cold email programmes themselves for 6-18 months and hitting a wall. We audit the inherited setup before migrating. The following five mistakes account for over 80% of what we find:

1. Sending domain is the brand domain

The single most damaging configuration error: cold campaigns going out from @yourcompany.com instead of @outreach.yourcompany.com or — better — a separate registered domain. When the cold programme degrades reputation (and over 6+ months without rotation, it will), the damage spills onto your transactional emails, password resets, customer notifications, and anything else from the same domain. Rebuilding takes 2-3 months. Fix: separate sending domains from day one, brand domain never enters the rotation.

2. Five mailboxes sending 200/day each instead of 30 sending 30/day each

The intuitive math is "5 mailboxes × 200 emails = 1000/day". The reality is that Gmail's per-mailbox detection treats sustained 200/day as bulk-pattern behavior, derates the mailbox within 2-3 weeks, and lands you in spam. The correct math is "30 mailboxes × 30 emails = 900/day", which looks identical in volume but reads completely differently to the receiver. Fix: more mailboxes, fewer emails per mailbox. The pool size grows; the per-mailbox volume shrinks.

3. Sequencer's built-in warmup running at the same time as live cold sends

Smartlead and Instantly both have warmup networks (500K+ and 4.2M+ accounts respectively) that work well when used alone. The problem is operators turning on warmup and live cold sending simultaneously on the same mailbox. The mailbox is now generating two competing signals — warmup engagement (high open, high reply) and cold engagement (low open, low reply). The receiver's reputation engine cannot reconcile these and treats the mailbox as inconsistent, which is worse than either signal in isolation. Fix: warmup completes before live sending starts. Period.

4. No SPF/DKIM/DMARC alignment audit before launching

Cold email tools advertise "automatic DNS configuration" and most users trust this. The configuration usually works for 60-70% of receivers; the other 30-40% require alignment that the automatic setup misses (subdomain DKIM signatures not matching the From domain, SPF includes for the cold tool but not for the sending mailbox, DMARC at p=none indefinitely). The result is a silent 30% spam rate that the operator never diagnoses because the dashboard says "delivered". Fix: full SPF/DKIM/DMARC audit per sending domain, alignment verified by test message to a controlled inbox before launch.

5. No daily inbox placement monitoring per mailbox per ISP

The default operator pattern is "look at reply rate weekly". Reply rate is a lagging indicator that drops 4-6 weeks after the underlying placement degrades — by which time multiple mailboxes are in spam at Gmail and you have no idea which ones. The leading indicator is daily seed-list placement testing per mailbox per ISP. Most operators do not do this because it requires tooling outside the sequencer. Fix: this is exactly what we built; we run it as part of the service. If you DIY, schedule daily seed-list tests via GlockApps or Mailtrap and alert on per-mailbox placement drops.

Platform integration matrix — what works with what.

We have tested every major cold email sequencer with our pre-warmed mailboxes. The matrix below summarises feature compatibility, pricing tier (entry to mid), and operational notes from production usage in 2026. The "OAuth" column is the only one that strictly matters for our infrastructure; everything else helps you choose the right sequencer for your workflow.

Sequencer Entry plan Mid plan Unlimited mailboxes Built-in warmup OAuth Google + MS Operational note
Instantly $30-47/mo (5K emails) $77-97/mo (100K emails) Yes (all plans) Yes (4.2M network) Yes Largest warmup network. AI Reply Agent on higher tiers.
Smartlead $39/mo (Basic) $94/mo (Pro) Yes (all plans) Yes (500K network) Yes Best agency white-label. Strong API for programmatic use.
Lemlist $59/mo (Standard) $99/mo (Pro) Per-seat after limit Lemwarm bundled Yes Best for personalization (video, image). Multichannel-ready.
Apollo $59/mo (Basic) $149/mo (Pro) Per-seat Bundled but smaller network Yes B2B database + sequencer combo. Best for SDR-driven teams.
Reply.io $60/mo (Email) $120/mo (Multichannel) Tier-limited Yes Yes Strong multichannel (LinkedIn, calls, SMS). Mature platform.
Woodpecker $59/mo (Cold Email) $99/mo (Multichannel) Per-slot pricing Yes Yes Polish-engineered, EU-based, GDPR-friendly. Pricier at scale.
Saleshandy $25/mo (Outreach) $83/mo (Outreach Pro) Tier-limited Yes (TrulyInbox) Yes Cheapest entry. Smaller warmup network; works for low volume.

Two things from this matrix that are not obvious until you have run multiple of them. First: sequencer warmup networks vary in quality, not just size. Instantly's 4.2M-account network is genuinely the largest, but Smartlead's 500K network is more curated and produces cleaner engagement signals on Gmail. Bigger is not always better for warmup — it depends on whether the network has been gamed. Second: "unlimited mailboxes" is operationally meaningless if you hit per-tier email caps before mailbox caps. Smartlead's truly unlimited model wins for high-volume agencies; Instantly's per-tier email caps work fine for solo operators but bite at scale.

Our pre-warmed mailboxes are OAuth-compatible with all seven platforms in the table. You connect; they work. The choice between platforms is about your team's workflow preferences, not about whether our mailboxes will plug in.

Cost honesty

What 30 mailboxes actually cost in 2026 — every option, all in.

Cold email infrastructure pricing is genuinely confusing because every provider bundles components differently. The chart below normalises monthly cost for a 30-mailbox operation including warmup and basic monitoring — so the comparison is apples-to-apples. All numbers are 2026 retail; volume discounts apply on most.

Cold email infrastructure: 30-mailbox monthly cost (Blue Spirit in EUR; competitors in USD as published, 2026)

Warmup + monitoring included where available, separate where not.

Monthly cost for 30-mailbox cold email setup, 2026 retail (Blue Spirit in EUR; competitors in USD as published)
Categoría Monthly cost EUR (30 mailboxes, warmup + monitoring)
Google Workspace direct (30 inboxes × $7) 210
Mailforge shared IP (raw) 75
Mailforge + Warmforge + Infraforge 210
Inframail flat-rate (50+ inboxes) 129
Litemail pre-warmed 150
Blue Spirit Cold Growth (managed) 199

All prices 2026 retail, sourced from public pricing pages. Google Workspace direct: 30 inboxes × $7/seat = $210/mo (warmup not included; add $15-25/seat if using external warmup tool). Mailforge raw: $2-3/mailbox shared IP, no warmup. Mailforge + Warmforge + Infraforge: full stack via Salesforge. Inframail: flat-rate $129/mo for unlimited inboxes (best math at 50+, suboptimal at 30). Litemail: $4.99/inbox pre-warmed = $150/mo; warmup included. Blue Spirit Cold Growth: €199/mo managed for 30-mailbox pool with daily monitoring and replacement. Real-world variance ±15% by configuration.

Three patterns the chart makes clear. First: Google Workspace direct is rarely the cheapest option past 20 mailboxes, contrary to what most operators assume. Per-seat pricing scales linearly while flat-rate alternatives plateau. Second: "cheap" providers like Mailforge are only cheap if you self-manage warmup and monitoring — adding those puts them in the same range as managed alternatives. Third: the gap between Litemail and our Cold Growth plan is operational, not technical — same OAuth, same compatibility, same Google Workspace tenancies underneath; the Blue Spirit premium pays for daily seed-list monitoring and the automatic mailbox replacement loop.

If you are at 50+ mailboxes the math changes again — Inframail's flat $129/mo becomes the cheapest option in the chart, and is genuinely a fair deal if you have in-house resources to run warmup and monitoring on your own. Above 100 mailboxes you are in agency territory and the comparison should be against our Agency plan, not Growth.

Suspension recovery: when Instantly, Smartlead or Google bans you

About 1 in 5 of our new cold email clients arrive after a suspension or shadow-ban from a sequencer or Google itself. The patterns are different but the recovery protocol is similar.

Instantly or Smartlead suspension usually triggers from automated reputation flagging when complaint rates spike on a single client tenant. The sequencer detects the pattern, suspends the workspace, and the recovery path through their support is glacial — measured in weeks, sometimes months, with no guaranteed reinstatement. The fix is faster: fresh mailbox infrastructure on new domains, full warmup cycle, sequence migration. We do this end-to-end in 3-4 weeks because the new pool needs proper warmup. Anyone promising 1-week recovery is selling you a recipe for re-suspension on the new platform.

Google Workspace ban is a more serious event. When Google itself terminates a Workspace tenant for outbound abuse, the same domain cannot be re-onboarded easily — the recovery path is new sending domains entirely, fresh tenants, and a clean slate. The original brand domain remains usable for transactional and brand-aligned email; you simply remove it from the cold rotation forever. We have done this recovery for 14+ clients since 2022; the playbook is well-rehearsed.

Microsoft 365 ban is rarer but follows similar logic. Microsoft's enforcement is more aggressive about volume thresholds since the May 2025 sender requirements update; cold programmes pushing 100+ emails/mailbox/day during the warmup period are particularly susceptible. Recovery is the same shape: new tenants, new domains, full warmup.

The common thread across all three is that you cannot negotiate your way back into the original infrastructure. The sequencer or platform has already classified you as a reputation risk. The only path forward is fresh infrastructure, and the only way to make fresh infrastructure work is the same warmup discipline that you bypassed the first time. We say this directly because it is the truth and because clients waste 6-8 weeks trying to argue with platform support before accepting it.

When Cold Email Infrastructure beats doing it yourself

Three scenarios where customers have moved to Blue Spirit after doing it themselves for a while.

First: hiring overhead. Setting up 20 mailboxes correctly the first time takes a skilled VA about two weeks. Maintaining them — responding to Google security challenges, rotating apps, replacing failed ones, dealing with Microsoft's own quirks — takes 5 to 10 hours per week ongoing. That is a part-time job at minimum. The cost of that time, even at a VA rate, exceeds our Growth plan within 6 months.

Second: reputation collapses you cannot diagnose from inside the tool. If you have ever watched your reply rate fall from 5% to 1% over three months and wondered what changed, you probably experienced a slow reputation degradation that is nearly impossible to diagnose from inside Instantly or Smartlead. The platforms tell you "delivered" but they do not tell you "delivered to inbox vs spam folder vs dropped silently". We see it from the outside: per-ISP placement per mailbox, daily, with alerts when degradation starts.

Third: growth friction. When you want to scale from 10,000 cold emails a month to 100,000, DIY infrastructure stops scaling gracefully. You hit the moment of "I need to set up 60 new mailboxes" and either you stop scaling or you spend two weeks. Our Agency plan starts at 50+ mailboxes and goes to 200+ with the same per-mailbox care as our Starter customers. Scale is what we sell.

When cold email infrastructure is NOT for you

The honest version, written so you do not pay us when you should not.

Volume below 5,000 cold emails per month with technical resources internally. At that scale, buying 5 Google Workspace seats yourself, configuring DNS once, and using Lemwarm or Smartlead's built-in warmup will be cheaper than our Starter plan. We have helped people set this up and walked away — the cost of our service does not pay back at that volume. Come back when you cross 10K/month or when DIY ops time exceeds your patience.

You are sending consumer marketing dressed up as B2B. If your "cold email" goes to personal Gmail addresses with no business context, you are not running a B2B outreach programme — you are spamming. Our infrastructure will not save you from that classification; Gmail will figure it out within weeks regardless of how clean your DNS looks. The fix is the campaign, not the rails.

Your offer requires sending the same recipient 7+ touches in 30 days. Most cold email frameworks recommend 3-5 touches over 4-6 weeks; aggressive cadences of 7+ touches in a month trigger complaint rates that no infrastructure can withstand. If your strategy depends on this volume, you need to rethink the strategy before you buy infrastructure.

What we will not do — explicit limits

Our reputation across our entire client base depends on the worst client in our pool. So:

No purchased consumer lists. We work with clients running outbound to lists built from Apollo, ZoomInfo, LinkedIn Sales Navigator, RB2B, Clay, or comparable B2B prospecting tools where contacts are professional roles relevant to your offer. We do not work with clients spamming purchased consumer lists, scraped Facebook/Instagram followers, or anything where the recipient has no plausible professional reason to receive your message.

No phishing, credential harvesting, malware, or fraud-adjacent campaigns. Even if you find an ESP that hosts those, we do not. Period.

No subject-line spoofing. "RE: our previous conversation" sent to someone you have never spoken to is fraud. We do not host clients running that pattern.

No fake personalisation tokens. Sending a template that says "I noticed your work on [scraped company]" without verifying the field is the kind of mass-personalisation that complaint-rates flag — we have all received that email; we do not host clients sending it.

We turn down 1 in 3 prospective clients on these questions on the discovery call. The 2 in 3 we accept run tighter operations as a result, and our reputation across the pool stays healthy.

Cold email infrastructure plans

Pre-warmed mailboxes, isolated IP pools, monitoring included. Scale up or down monthly.

Cold Starter

From
79 /month
  • 10 pre-warmed mailboxes (Google or Microsoft)
  • 2-4 weeks of warm-up history included
  • US or EU IPs available
  • Auto DNS setup (SPF, DKIM, DMARC)
  • OAuth integration with Instantly, Smartlead, Apollo
  • Suppression list sync
Request a quote
Most popular

Cold Growth

From
199 /month
  • 25 pre-warmed mailboxes
  • Full infraguard-style monitoring
  • Isolated warmup pool (no shared reputation)
  • Cross-ISP diagnostics
  • Dedicated sending domains on request
  • Weekly deliverability review with our team
Request a quote

Cold Agency

From
499 /month
  • 50+ pre-warmed mailboxes (scale to 200+)
  • Multi-brand / multi-workspace management
  • Custom domain portfolio supplied
  • Dedicated US+EU sending IPs
  • API for programmatic mailbox creation
  • White-label deliverability reports
Request a quote

Cold email infrastructure — frequently asked questions

Is cold email a legitimate channel?

Business-to-business outreach is a long-established channel and remains widely used by sales and BD teams across every major economy. Most jurisdictions distinguish clearly between B2B prospecting (where the contact is relevant to the recipient's professional role) and consumer marketing (where stricter consent rules apply). We provide infrastructure for the former. The legal posture in your audience's jurisdiction is your responsibility — we provide the deliverability and the engineering, you provide the campaign and the targeting decisions.

How many mailboxes do I actually need?

Rule of thumb based on operating actual client programmes: to send 10,000 well-paced cold emails per month without burning any single mailbox, you want 10 to 20 mailboxes in a rotating pool spread across 3-5 sending domains. To scale to 50,000 per month, 30 to 50 mailboxes across 15-20 domains. Above 100,000 you need a properly engineered pool with domain rotation, IP diversity and continuous monitoring. The chart on this page shows the curve. The wrong answer is "send everything from one mailbox to look like a real person" — Gmail and Microsoft both detect that pattern within weeks.

Why pre-warmed and not just new mailboxes?

A fresh mailbox on a brand new domain has zero sending reputation. Hit it with 50 cold emails on day one and Gmail will file them in spam or drop them altogether — you can verify this yourself with a test campaign and Postmaster Tools 48 hours later. Our pre-warmed mailboxes have 2 to 4 weeks of synthetic engagement and real-traffic warmup under their belt before you start sending — which means your outreach starts from a healthy baseline instead of a buried grave. The cost of skipping warmup is not a percentage hit on placement; it is the entire campaign landing in spam for a month while you try to figure out why.

Can I use my existing Instantly, Smartlead, Lemlist or Apollo account?

Yes. We deliver mailboxes with OAuth tokens ready to drop into your platform of choice. Instantly, Smartlead, Lemlist, Apollo, Reply.io, Woodpecker, Lemwarm, Outreach.io — they all speak the same language for OAuth and we have tested every current major platform. We deliver a CSV with mailbox addresses, OAuth tokens, recommended daily sending limits, and the sending domain map. You import; you start sending.

What is the safe daily sending limit per mailbox?

The honest 2026 answer is 15 to 30 cold emails per day per mailbox to a fresh recipient list. Smartlead and Instantly will technically let you send 50-100; the platforms are not the throttle, the inbox provider is. Gmail and Microsoft both detect sustained sending above ~30 cold/day from a non-established address as bulk-pattern behavior. We recommend starting at 15/day for the first 4-6 weeks of a new mailbox in your pool and ramping to 25-30/day once placement signals are stable. Above 30/day per mailbox you are accelerating reputation decay; below 10/day you are wasting infrastructure capacity.

What happens when a mailbox degrades?

We monitor inbox placement daily across a private seed-list test set covering Gmail, Outlook, Yahoo and iCloud. When a mailbox's placement drops below 70% on Gmail or Outlook for two consecutive days, the system pauses new sending from that mailbox, runs a recovery warm-up routine, and — if recovery fails after 7 days — replaces it with a fresh pre-warmed mailbox from the pool. Your platform sees a mailbox swap notification; your campaigns continue uninterrupted; you do not lose campaign state. This is the operational difference between us and a DIY setup.

How is this different from MailGenius, Mailshake, Mailforge or just buying Google Workspace yourself?

Three differences. First, infrastructure: we operate at scale and can spread risk across many domains and IPs in ways a single team cannot easily replicate. Second, monitoring: we run continuous seed-list testing per ISP per mailbox; you cannot get that visibility from inside any single sending tool. Third, replacement: when a mailbox burns, we swap in a pre-warmed replacement within hours, not days. The DIY path works for low volumes (under 5,000 cold emails per month) and for teams with one engineer dedicated to it. Above that, the math favours managed infrastructure — and the math has tightened in 2026 because Google Workspace per-seat pricing increased while flat-rate competitors emerged.

What about the new Google sender requirements?

Every mailbox we ship is configured to comply with the bulk sender requirements that came into force in February 2024 and tightened through May 2025: SPF passing alignment, DKIM 2048-bit rotating, DMARC at policy p=quarantine or p=reject, one-click unsubscribe in headers (RFC 8058), complaint rate below 0.1% maintained through engagement-core warmup. Microsoft's parallel May 2025 rules are also covered. We update configurations when ISPs change requirements; that is part of the service.

Can you provide infrastructure for an agency reselling cold email?

Yes — Agency plan is built for this exact case. Custom sending domains under your brand convention, isolated workspaces per client (so one client's reputation cannot affect another), white-label monitoring dashboards, branded reporting if you want to share placement metrics with end clients. Pricing is volume-based; the more clients you onboard, the better the per-mailbox economics. We have agencies running 80-200 mailboxes for their book of business through our Agency plan.

My account got suspended by Instantly or Smartlead — can you help?

We can. The pattern is recognizable: you ran a campaign whose volume or content tripped one of the platform's automated reputation systems, and you got a suspension or shadow-ban notice without recourse. The fix is rarely about begging support; it is about new mailbox infrastructure on fresh domains. We provision the new pool, you migrate sequences, the old pool is decommissioned. Recovery to full sending takes 3-4 weeks because the new pool needs its own warmup period — anyone promising faster is selling you a recipe for re-suspension. Same protocol applies if you got banned from Google Workspace itself.

When is cold email infrastructure NOT the right answer for me?

Honest answer: when your monthly cold volume is under 5,000 emails and you have technical resources internally. At that scale, buying 5 Google Workspace seats yourself, configuring DNS once, and using Lemwarm or Smartlead's built-in warmup will be cheaper than our Starter plan. Also wrong fit: when your "cold email" is actually consumer marketing dressed up as B2B (sending to personal Gmail addresses with no business context), or when your offer requires sending the same recipient 7+ touches in 30 days. Those programmes will burn any infrastructure, ours included; the issue is the campaign, not the rails.

Do you accept clients with bought lists or scraped data?

No. We work with clients running outbound to lists built from Apollo, ZoomInfo, LinkedIn Sales Navigator, RB2B, Clay, or comparable B2B prospecting tools where contacts are professional roles relevant to your offer. We do not work with clients spamming purchased consumer lists, scraped Facebook/Instagram followers, or anything where the recipient has no plausible professional reason to receive your message. The reputation damage from that kind of programme is unrecoverable, and our other clients suffer from contagion. We turn down 1 in 3 prospective clients on this question alone.

Stop building cold email infrastructure. Start sending from it.

Tell us your monthly cold volume target, your current platform (Instantly, Smartlead, Lemlist, Apollo) and one deliverability problem you have seen. We respond with a specific recommendation — pool size, domain count, plan tier, realistic timeline.

Chat with us on WhatsApp