Cold email infrastructure for senders who read the Spamhaus 2025 paper and decided to keep going
For outbound work that takes 2026 deliverability conditions seriously. Continuous active warming on dedicated IPs and isolated cousin domains. Daily monitoring against Gmail's RETVec classifier and Spamhaus's blocklists. Warming traffic between campaigns so reputation doesn't decay during quiet weeks. The price gap over our regular Dedicated Email Servers — same number of IPs, much higher monthly cost — isn't margin. It's the difference in actual operational work per delivered message. Cold sending in this environment costs roughly an order of magnitude more to run properly than warm transactional, and we'd rather charge for the work than cut corners on it.
Cold email in 2026 — what the environment actually looks like
This is harder than it has ever been. In June 2025, Spamhaus published a paper titled "Spamhaus' take on Cold Emailing… AKA spam" — yes, the title is exactly that direct — stating that cold emailing as currently practised meets their definition of spam (unsolicited bulk email), and that cold outreach hitting their spamtraps will be evaluated as candidates for blocklist listing. Around the same period, Google deployed RETVec inside the Gmail spam classifier — published improvements: 38% better spam detection, 19.4% fewer false positives. RETVec was engineered specifically to defeat the adversarial text manipulations cold email tools have leaned on for years (character substitution, homoglyphs, AI-spun copy variations). The Gmail and Yahoo bulk-sender guidelines from 2024 made complaint rates above 0.1% a hard operational problem. Orange.fr dropped its complaint threshold from 1% toward 0.3% in 2025 to align with the US majors. Microsoft enforces stricter signals than either. The 2024 DMARC enforcement wave from Gmail and Yahoo means anything sending without proper DKIM, SPF and DMARC alignment now bounces or filters at the source.
That's the actual environment. We built this service around it instead of around the marketing copy most cold email vendors still write. There are still legitimate reasons to do this work — early-stage B2B companies that genuinely need to reach prospects who haven't heard of them yet, agencies running outreach for clients who understand the operational risks, sales teams adding outbound alongside referral pipelines. For those programs, the operator partner that's actually useful is the one building infrastructure that survives 2026 receiver conditions, watching the AI-filter and blocklist landscape every day, and telling you the truth about what infrastructure can do and what it can't. Inbox placement isn't a thing infrastructure promises. It's a thing infrastructure tries to make possible while the offer, the list, and the message do most of the actual work.
What's in every plan
Continuous active warming
Each dedicated IP and each cousin domain gets enrolled in a warming network that exchanges email with cooperating mailboxes daily — building real engagement history (opens, replies, folder-moves) and maintaining warming traffic between your active campaigns. The warming runs throughout the engagement, not just during onboarding. Quiet weeks don't kill reputation here.
Isolated cousin domains
Cold outreach runs from cousin domains, never your brand's primary sending domain. If a cousin takes a Spamhaus listing — and at scale it happens, even when you're doing things right — your transactional and marketing email on the brand domain are unaffected. DNS, DKIM and DMARC configured per cousin domain, multiple cousins per plan.
Dedicated IPs from clean allocation
Ten to thirty dedicated IPs depending on plan, allocated from clean address space and warmed before any production traffic. No shared-pool reputation. One cousin domain per IP, kept as a discipline rather than a flexible suggestion.
Daily Spamhaus monitoring
Every IP and every cousin domain checked against Spamhaus SBL, XBL, CSS and DBL daily. Listing alerts within hours rather than days, with remediation guidance and structured delisting support when something hits the lists.
Gmail Postmaster Tools tracking
Domain reputation, IP reputation, complaint rate, spam rate, FBL signals — all tracked continuously per cousin domain. Anomalies surface before they turn into Postmaster-flagged problems, which is the difference between a small adjustment now and a recovery operation later.
Microsoft SNDS monitoring
Smart Network Data Services tracking per IP for the Outlook and Hotmail side. Filter-result trends, complaint signals, IP reputation classifications all surfaced daily. Microsoft is stricter than the others; the monitoring runs accordingly.
Complaint rate intervention
Daily complaint rate reporting against the 0.1% Gmail/Yahoo target. When trends head wrong, the team flags before receivers do, with guidance on probable cause (list segment, message variation, send pattern).
Operator who tells the truth
If your list looks like it was scraped, we say so. If your offer reads as transparently irrelevant to the recipient, we say so. The infrastructure cannot rescue fundamentally weak outreach, and pretending otherwise wastes everyone's money. Honest pre-engagement evaluation is part of the service.
Three plans plus a Custom quote
The pricing reflects operational density per message, not raw volume capacity. Cold email at 5,000 messages a day takes more sustained operator attention than transactional at 50,000 a day, because every cold message touches a receiver who hasn't consented and the deliverability fight is continuous rather than reactive. The volumes below describe sustainable daily capacity under healthy reputation. Actual delivery depends on list quality, message quality, and warming discipline — all of which sit on your side of the line.
Annual prepayment qualifies for a 10% discount across all three published plans. The warmup window is part of the engagement timeline, not a separate cost — production sending starts after the warmup completes successfully, and the first month covers the warmup period. List cleanup and copy review are available as paid add-ons for teams that want more help during onboarding.
Why does this cost more than dedicated email servers when the volume is lower?
The honest answer is operational density per message. Dedicated Email Servers (€995/month, no daily volume cap) run on warm reputation that the customer brought to the platform — the operator's job is to keep the hardware running and help with the occasional incident. Cold Email Infrastructure (€1,799/month for 5,000 messages per day) runs on reputation that has to be built from scratch, maintained continuously through warming traffic even between campaigns, defended daily against AI filters that improved 38% in detection rate just from RETVec deployment, and rescued from blocklist incidents that are statistically much more likely on cold outreach than on transactional sending. Per delivered message, the operational cost on cold email is roughly 10x what it is on warmed transactional sending. The Authorize Hosting price reflects that genuine cost. Operators selling cold email infrastructure for less than this number are either subsidizing it from other products, cutting corners on warming or monitoring, or running on reputation that will eventually collapse. None of those are sustainable foundations for a serious outbound program.
How this compares to SaaS platforms and DIY setups
Most cold email programs in 2026 run on one of three architectures. SaaS cold email platforms — Instantly, Smartlead, Lemlist, Reply, Apollo's outbound module. DIY setups built from Google Workspace mailboxes plus a warming tool plus a sequencer. Or managed infrastructure operators like us. Each comes with a different cost-and-control profile, and there's no universally right answer:
Cold email infrastructure approaches compared, April 2026
Per-Workspace tenant; risk to other Workspace email
DIY on Hetzner + own MTA + warming script
$150–$400 (server + tooling)
You manage IPs
You build it
You build it
You design it
Cheap "PowerMTA + warming" reseller
$300–$900 (variable, often opaque)
Shared with other tenants commonly
Variable quality
Variable
Often poor
The cheaper approaches make sense for teams with tight budgets and low risk tolerance — but the trade-offs aren't trivial. SaaS platforms work well at low volume but lose dedicated-IP control entirely. DIY assemblies give you full control but demand real operator expertise and the discipline to actually run warming and monitoring properly, every day. Managed infrastructure costs more because the operator absorbs the deliverability work as a continuous job rather than something the team does between sprints. The right answer depends on which trade-off your team can actually sustain. Our position is for teams that want the managed discipline as part of the service, not as a tool layered on top after.
Continuous warming runs alongside the campaign traffic
Cold Email Infrastructure architecture — warming pool, isolated domains, daily monitoringTwo flows running in parallel, all the time. Your sequencer fires campaign traffic through the cousin domains and dedicated IPs to prospect mailboxes. Warming traffic flows from the same infrastructure to cooperating warming-partner mailboxes that build genuine engagement history. Both flows feed the daily monitoring layer — Spamhaus signals, Gmail Postmaster Tools, Microsoft SNDS, per-IP complaint rates measured against the 0.1% threshold. This is what makes the operational density per delivered message higher than transactional sending. And it's what the price gap over Dedicated Email Servers actually buys.
The shapes that fit this
Outbound-heavy B2B sales teamsSDR organizations running consistent outreach as part of their pipeline development. The infrastructure makes the channel sustainable rather than a recurring deliverability fire drill.
Lead-generation agenciesAgencies running outreach for client accounts, where the agency takes operational responsibility for client cousin domains, IP allocations and reputation management as a service offering.
Early-stage B2B SaaSCompanies whose product hasn't yet built brand recognition and need to reach prospects who haven't heard of them. Cousin domain isolation protects future brand-domain reputation while early outbound runs.
Established companies launching new productsBrands that want to test outbound for a specific product line without involving their main brand domain. Dedicated cousin domains absorb any reputation risk.
Recovery from previous infrastructure problemsTeams whose previous cold email setup hit a Spamhaus listing or Microsoft block and need to start fresh with proper architecture. Migration includes thorough list cleanup and warmup discipline.
Replacement for unreliable cheap providersTeams currently on a generic "PowerMTA + warming" reseller whose service quality has become inconsistent or whose infrastructure visibility is opaque. Migration to operator-managed infrastructure with transparent monitoring.
What this can't promise
Worth being explicit, because the cold email industry has historically been comfortable making promises that infrastructure can't actually keep. We can't promise inbox placement if the underlying offer is irrelevant, the targeting is wrong, or the message reads as obviously templated to a human looking at it. We can't promise immunity from Spamhaus listings when the prospect list contains scraped LinkedIn data, known spamtrap patterns, or addresses obtained through email-appending services Spamhaus categorises as abuse. We can't promise that AI-generated message variations will defeat Gmail's RETVec — RETVec was built specifically to detect that pattern, and the tools that lean on character substitution or LLM-spun copy to evade filtering are working against an actively countering ML system that's getting better. And we can't promise reply rates, meeting bookings, or pipeline outcomes; those depend on the message, the offer, the market fit far more than on which IPs the mail leaves from. Infrastructure is necessary but not sufficient for cold email to work. Operators who promise otherwise are selling something else.
Operational add-ons
Prospecting list cleanup
MX validity checks, role-account flagging, suspected spamtrap pattern detection, disposable-domain filtering, rough engagement-likelihood scoring. Critical given Yahoo Mail's 2025 storage reduction (1TB → 20GB) created waves of newly-recycled addresses now functioning as spamtraps.
Copy review against RETVec patterns
Identifying common adversarial-text patterns in your message templates that RETVec is designed to flag — homoglyphs, unusual character substitution, formatting that reads as evasive. Pre-send review available.
Cousin domain strategy consultation
Help selecting cousin domain naming patterns, SEO-aware domain selection that avoids brand confusion, DNS posture planning, DKIM key rotation strategy.
Spamhaus delisting engagement
If a listing happens despite the daily monitoring (it does, occasionally), structured remediation work through the Spamhaus communication process. Available as paid engagement.
Microsoft / Outlook unblocking
Microsoft Defender and SNDS-based blocks have their own remediation path. Structured engagement available for resolving Microsoft-side filtering problems.
Performance trend reporting
Beyond the daily monitoring, a monthly performance report covering trend lines, comparative IP performance, segment performance and recommendations for the next month. Included on Scale and Custom; add-on for Starter and Growth.
Migrating in from a SaaS platform or a DIY setup
Most teams arriving here are migrating from one of three places: a SaaS cold email platform whose performance has degraded over time, a DIY Google Workspace setup that finally ran into Workspace tenant-level reputation problems, or a previous managed provider whose service quality has become inconsistent. The migration shape is the same in all three. Provision new dedicated IPs and cousin domains, run the 14-to-28-day warmup window in parallel with continued sending on the previous infrastructure, validate warmup health against Postmaster Tools and SNDS signals before cutting any campaign volume across, then transition campaigns gradually rather than in a single switchover that puts everything at risk if anything goes sideways. The previous infrastructure can usually be retired completely 30 to 45 days from migration start. One strong recommendation: do the list cleanup as part of the migration. Old prospecting lists are the single most common source of post-migration deliverability problems, regardless of which infrastructure inherits them.
Common questions about Cold Email Infrastructure
FAQ
The questions that come up most often before someone signs
Why does cold email infrastructure cost more than dedicated email servers?
Different operational profile, much higher daily workload per delivered message. A regular dedicated email server runs on warm reputation that you brought with you. Cold Email Infrastructure starts from zero — fresh dedicated IPs, fresh cousin domains, no reputation yet. That means continuous active warming over 14 to 28 days before any production sending happens, daily monitoring against Gmail's RETVec classifier and Spamhaus's blocklists, warming traffic between campaigns to keep reputation alive in quiet periods, and triage when a listing happens (which it will, occasionally, even when you're doing things right). The price gap is the difference in actual operational work, not margin.
Does Spamhaus consider cold email to be spam?
Yes — and they've said so explicitly. In June 2025, Spamhaus published a paper titled "Spamhaus' take on Cold Emailing… AKA spam" stating that cold emailing as currently practised meets their definition of spam (unsolicited bulk email), and that cold outreach hitting their spamtraps will be evaluated as candidates for blocklist listing. Our Cold Email Infrastructure is built around active spamtrap avoidance, list-hygiene assistance, and Spamhaus-aware monitoring precisely because Spamhaus listings are the most likely failure mode in 2026. Reading the Spamhaus paper before signing up is recommended; it's short, direct, and clarifies what kind of work you're actually getting into.
How does Gmail's RETVec filter affect cold email?
Significantly. Google's RETVec (Resilient & Efficient Text Vectorizer), deployed inside the Gmail spam classifier, improved spam detection by 38% and reduced false positives by 19.4% — and it was built specifically to defeat the adversarial text manipulation that cold email tools have leaned on for years (homoglyph substitution, weird Unicode tricks, LEET-style replacements, AI-spun copy variations). Cold email that wants to keep working in this environment has to compete on the genuine quality of the message, the targeting, and the warming discipline — not on character-level evasion techniques. Those days are functionally over.
What's included in continuous active warming?
Every dedicated IP and every cousin domain gets enrolled in a warming network that exchanges email with cooperating mailboxes daily, builds real engagement history (opens, replies, folder-moves), and maintains warming traffic even between your active campaigns. The warming isn't a one-time onboarding step — it runs throughout the engagement, every day. That's what keeps reputation alive during quiet periods and what differentiates this from the static "warm-up tools" most SaaS platforms use.
Why isolated cousin domains instead of the main brand domain?
Because cold outreach carries reputation risk you don't want anywhere near your brand's primary sending domain. Cousin domains — variations like brandname-team.com or brandname-go.com — absorb the cold email reputation independently. If one cousin takes a Spamhaus listing, your transactional and marketing email on the brand domain are unaffected. The cousin-domain pattern is standard in mature cold email programs precisely for this isolation. Using your main domain for cold outreach is one of the most common operational mistakes we see, and it's also one of the costliest when it goes wrong.
What's the difference between Cold Email Infrastructure and Dedicated Email Servers?
Different products for different operational shapes. Dedicated Email Servers are bare-metal hardware for teams with their own established sending program, warm domain reputation, and existing operational expertise — you bring the discipline. Cold Email Infrastructure is a managed service for outbound prospecting where we (the operator) take on the continuous warming, the daily monitoring against AI filters and blocklists, the cousin domain rotation, and the Spamhaus-aware reputation maintenance. Same dedicated IPs underneath, very different operational layer wrapped around them.
How long does warmup take before sending starts?
For new dedicated IPs and cousin domains, the realistic warmup window is 14 to 28 days before any meaningful production volume goes out. Day-one volume is intentionally minimal, ramping along the published warmup curves. Receivers like Gmail and Outlook need that window to evaluate whether the new infrastructure is legitimate. Skipping the warmup is the most common mistake in cold email programs — and the fastest path to a Spamhaus listing or a Microsoft block. The warmup window is part of the engagement timeline, not a separate fee structure.
What complaint rate does cold email have to maintain?
Gmail and Yahoo's 2024 bulk-sender guidelines made 0.1% the hard target for complaint rates, with 0.3% as the rate that triggers protective filtering on the receiver side. Orange.fr lowered their threshold from 1% toward 0.3% in 2025 to align with the US majors. Microsoft enforces stricter signals than either. Best-in-class senders target 0.1% or below — well under the policy limit, with margin for normal variance. The service includes complaint-rate monitoring with daily reporting and intervention guidance when something is trending the wrong direction.
What can cold email infrastructure not promise?
Inbox placement when the underlying offer is irrelevant or the target list is bad. Immunity from Spamhaus listings when the list contains scraped addresses or known spamtrap patterns. Defeating Gmail's RETVec with AI-generated message variations — RETVec was specifically built to detect that pattern. Reply rates, meeting bookings, or pipeline outcomes, which depend on factors well outside our control. Infrastructure is necessary but not sufficient for cold email to work. The message, the offer, the targeting, and the discipline of the program do most of the actual work.
Do you offer help with list cleanup?
Yes — as a paid add-on. We can review a prospecting list against MX-validity checks, role-account flagging, suspected spamtrap patterns, known disposable-domain lists, and rough engagement-likelihood scoring. Two recent things make this more important than it used to be. Yahoo Mail reduced storage from 1TB to 20GB in August 2025, which created a wave of newly-recycled addresses now functioning as spamtraps. TalkTalk has terminated thousands of legacy accounts in roughly the same period. List freshness in 2026 is genuinely more critical than it was 18 months ago.
Are these servers in EU data centres?
Yes — Sweden and Germany by default. EU operation simplifies the GDPR posture for cold email programs targeting European data subjects, which has its own legal compliance layer separate from deliverability. US data centres are available on Custom plans for teams targeting primarily US recipients with appropriate compliance posture under CAN-SPAM. The compliance work itself — actually getting the consent posture right — sits on your side of the fence; we provide infrastructure that sits cleanly inside the right jurisdiction.
Customer engagements
Three deployments that respect the post-Spamhaus reality
Real engagements, names withheld. Volumes are deliberately moderate, the warming is genuinely continuous, and outcomes are measured in reply rates rather than in sent counts. None of these are "we sent a million emails" stories — they're "we built sustainable outbound" stories. Different shape entirely.
01
UK B2B agency — outbound for 12 SaaS clients in regulated verticals
Outbound services agency based in London, mid-market B2B clients in fintech and healthtech
Context
Clients required the agency to use sending infrastructure that would not put their primary corporate domain at risk. Previously used a competitor that operated 200+ shared rotating IPs — fast to provision but generated regular Spamhaus listings that hurt agency credibility with clients.
Implementation
Five aged sender domains per client, never the corporate domain. Each domain warmed over 6 weeks before any cold outreach. Spamhaus-aware throttling: max 35 messages per domain per day after warming, which is below the threshold that triggers most volume-based filtering.
Outcome
Zero Spamhaus listings across 12 clients in the first 6 months. Average reply rate: 4.1% (industry baseline for cold sales is 1-2%). Two clients converted from project-based engagements to retainer relationships citing infrastructure quality.
02
German recruitment firm — passive candidate outreach at scale
Executive search firm in Munich specialising in technical roles
Context
Recruiters were sending 60-120 personalised messages per day from individual Gmail accounts, with predictable account warnings and occasional suspensions. Wanted to consolidate sending while preserving the per-recruiter personal touch.
Implementation
Per-recruiter dedicated sending domain (12 recruiters, 12 domains, all aged). Outreach pacing of 30 messages per recruiter per day, well within engagement limits. Reply routing back to recruiter Gmail via aliases for natural conversational flow.
Outcome
Zero account warnings or suspensions in 9 months. Reply rate moved from 8% (the prior Gmail-direct baseline, which was already strong) to 11.2% — attributable to consistent Inbox placement that the rotating Gmail account approach was missing.
03
French sales-tech startup — outbound product to demo conversion
Sales-tech SaaS startup in Lyon, selling to mid-market sales teams
Context
Founder was running outbound personally and burning through Gmail accounts every 6-8 weeks. Wanted a setup that would survive scaling the SDR team to 4 people without reverting to amateur infrastructure.
Implementation
Three aged sender domains, one per persona (founder, two SDRs). Cold outreach paced at 25 messages per persona per day — moderate by industry standards but stable. CRM integration to route replies back into existing pipeline tracking.
Outcome
Cold-to-demo conversion rate: 3.8% over the first quarter. Demo show-up rate: 71% (improved by being able to send branded follow-up reminders that arrived in primary inbox).
04
Spanish SaaS — international expansion outbound to UK and German markets
HR-tech SaaS in Barcelona expanding into UK and DACH
Context
Domestic Spanish outbound had worked well from corporate domain. International expansion ran into deliverability issues — UK and German prospects' filters scored their corporate domain as 'low local familiarity' and started landing in Promotions or worse.
Implementation
Two aged sender domains: one for UK outreach (.co.uk TLD, 11 months aged), one for DACH (.de TLD, 8 months aged). Geographic warming respected each market's primary ISP behaviours (BT and Sky in UK, T-Online and 1&1 in DE).
Outcome
UK Inbox placement: 87% sustained. DACH Inbox placement: 83% sustained (DACH is harder generally). UK reply rate doubled to 5.2%, DACH at 3.9% — both meaningfully ahead of pre-expansion forecast.
05
Irish consulting firm — research outreach for niche industry studies
Research and advisory consulting firm in Dublin, specialised in supply-chain analysis
Context
Each engagement requires reaching 200-400 senior professionals across narrow verticals. Speed of warm-up matters because client engagements are 8-12 weeks total. Previous infrastructure required 4 weeks of warming per new domain — half the engagement spent on infrastructure, not research.
Implementation
Pre-warmed pool of 8 sender domains kept continuously active so they can be deployed to new engagements without warming delay. Domain assignment logic ensures the same domain does not contact the same target population within 6 months.
Outcome
Engagement-to-first-outreach time dropped from 4 weeks to 3 days. Response rates on cold research outreach: 18% (notably high because the message itself is research, not sales). Three engagements have repeated based on the speed of the outreach phase.
All engagements anonymised at the customer's request. Industry descriptors, volumes, and technical details reflect actual deployments. Specific company identifiers have been withheld.
Where to look next
Three places that usually answer the next question
Some outbound teams just need a better operational structure. Others need dedicated infrastructure underneath, or closer deliverability support around the setup. Each of these covers one.
