IP warmup in 2026: what works, what is outdated, and the real schedule

Most IP warmup guides you will find online are copy-pastes of advice that was good in 2018. The email landscape changed enough during 2024-2025 that the old templates now under-perform, sometimes dangerously: Gmail and Yahoo enforced bulk-sender authentication in February 2024, Microsoft followed on May 5, 2025 with hard 550 5.7.515 rejection for non-compliant 5,000+/day senders, PCI DSS v4.0 made DMARC effectively mandatory for cardholder-data entities as of March 31, 2025, and DORA put email authentication into ICT third-party risk for every EU financial entity from January 17, 2025. A warmup schedule that was state-of-the-art in 2018 will produce a stuck IP at “average sender” reputation in 2026, where ML-driven engagement weighting at the major receivers rewards consistency and engagement quality over volume. Here is what we do in 2026 and why.

Start from proper authentication, not volume

The single biggest change since the old warmup guides: authentication is no longer optional, and p=none minimum DMARC is mandatory at all three major US receivers from email number one on a new IP. The Microsoft May 2025 enforcement closed the last major receiver gap; the next industry shift will be receivers raising minimum policy from p=none to p=quarantine for bulk senders, telegraphed by Gmail postmaster communications as “late 2026 or 2027” without firm dates.

The first thing you do on a fresh IP is not “send 50 emails”. It is the authentication baseline below — and we deploy it 48 hours before the first warmup message so receivers see stable DNS records during the entire reputation-accumulation arc:

• Publish SPF with an include: of your sending host and ~all initially (soft fail), promote to -all (hard fail) after warmup completes and aggregate DMARC reports confirm zero unauthorised legitimate sources.
• Generate a fresh 2048-bit DKIM key pair (1024-bit deprecated since RFC 8301 in 2018, and Microsoft has flagged 1024-bit signers with reduced trust scores since Q3 2025). Publish at selector._domainkey.domain.tld, sign every outgoing message.
• Publish DMARC starting at p=none; rua=mailto:[email protected]; fo=1; adkim=r; aspf=r so you actually receive the aggregate XML reports; monitor for at least 14 days before moving toward p=quarantine.
• Set proper rDNS on the IP matching your sending domain. Forward-confirmed reverse DNS (FCrDNS) is required by Gmail and increasingly by Microsoft; without matching rDNS, you start at lower trust regardless of warmup discipline.
• Enrol the domain in Google Postmaster Tools and Microsoft SNDS on day -2 (two days before first send) so historical data is available to read against.
• Publish MTA-STS in testing mode (promote to enforce after week 2) and TLS-RPT, especially if you expect Gmail-heavy traffic — these signals contribute to faster reputation accumulation in 2026.
• For European-heavy lists, verify your relay supports ARC sealing on outbound forwarded mail. Orange.fr has required ARC sealing on forwarded messages since September 2024; T-Online.de, GMX and Web.de increasingly enforce on aggregated chains.

Without those pieces in place, your warmup will under-perform no matter how disciplined your volume curve is. Validity’s January 2025 industry analysis found that 84% of domains used in From addresses do not have a published DMARC record at all, and 7.64% of those that do have invalid records — which means roughly 6 in 7 senders we encounter at audit need authentication remediation before any warmup work begins.

Pre-warmup checklist — the 14 items operators consistently miss

Before the first warmup message goes out, the checklist below is what we walk through at Blue Spirit during managed warmup onboarding. Most operators starting DIY warmup miss at least 4-6 of these, and the ones they miss are usually the ones that determine whether the warmup arc lands at “high reputation” or stalls at “average sender”. Work through them sequentially before day 1.

1. rDNS / FCrDNS verified. dig -x your.ip.address returns a hostname that matches your sending domain (e.g., mail1.yourdomain.com), AND that hostname’s A record resolves back to the same IP. Without forward-confirmed reverse DNS, Gmail and Microsoft start you at lower trust regardless of authentication completeness. Most managed PowerMTA hosts set this automatically; generic VPS providers often do not.
2. MXToolbox blacklist check clean. Run the IP through MXToolbox blacklists lookup AND the Spamhaus + SURBL + Barracuda + Talos individual lookups. Any listing on a major DNSBL means you cannot start warmup until delisting is complete (typically 48-72 hours for most lists).
3. SPF record published with single TXT. dig +short txt yourdomain.com returns exactly one SPF record (v=spf1 ...). Two or more v=spf1 records is a configuration error that most receivers treat as permerror.
4. SPF lookup count under 10. Use the SPF flattening check from our 10-lookup limit guide. If at 8+ lookups, flatten before starting; the limit will hit the moment you onboard one more sending service mid-warmup.
5. DKIM 2048-bit key generated and published. Verify with dig +short txt selector._domainkey.yourdomain.com returns a record starting with v=DKIM1; k=rsa; p= followed by a long base64 key (1024-bit keys are short enough to fit in the visible terminal output; 2048-bit keys span multiple TXT chunks).
6. DMARC at p=none with rua to a parsed endpoint. _dmarc.yourdomain.com TXT record exists, p=none, rua=mailto: points to a real mailbox you have configured to parse XML reports. Postmark DMARC Digest is free for 1 domain and adequate for warmup.
7. MTA-STS in testing mode published. Three artefacts: TXT record at _mta-sts.yourdomain.com, TXT at _smtp._tls.yourdomain.com, and policy file served at https://mta-sts.yourdomain.com/.well-known/mta-sts.txt in mode: testing (will promote to enforce post-warmup).
8. Gmail Postmaster Tools verified domain. Domain ownership confirmed via DNS TXT record at least 48h before day 1 so historical data has time to populate.
9. Microsoft SNDS access requested. Even though SNDS access takes 5-10 business days for new senders, request access on day -10 so the signal source is available by mid-warmup if not from day 1.
10. List validated within 30 days. Run the warmup audience through NeverBounce, ZeroBounce or BriteVerify; suppress addresses with confidence below 90%. Hard bounce rate at day 1 should be under 0.5%, not 2-3%.
11. Engagement segments computed and confirmed. The Tier 1 / Tier 2 / Tier 3 / Tier 4 cohorts are computed on the actual list, sized appropriately (Tier 1 should be 8-15% of healthy list — if it’s only 2-3%, your list is degraded and needs validation before warmup).
12. Content templates frozen for the warmup window. The two or three templates you will use during warmup are finalised; no template swaps mid-warmup. Template changes after day 7 trigger reputation re-evaluation that costs 3-5 days of progress.
13. Send-time pattern documented. The 2-3 specific hours daily during which warmup sends will go out, in the timezone of your audience majority. Send-time consistency is one of the engagement signals receivers track during reputation accumulation.
14. Recovery plan documented before sending. Specifically: who pauses the warmup if Gmail Postmaster drops, what the criteria are, who decides on IP swap. Operations during a stall need a pre-agreed plan because real-time decision-making while reputation is degrading produces worse outcomes than pre-planned response.

The checklist looks long because it is — but each item costs 5-30 minutes to complete and absorbing all 14 prevents 80% of the warmup failures we see at audit. The compressed version, for senders working alone: items 1-9 are non-negotiable (cannot start without them); items 10-14 are recoverable mid-warmup but expensive to retrofit.

The actual 2026 volume curve

The old “day 1: 50; day 2: 100; day 3: 200” schedule assumes you are ramping to a few thousand per day. For anyone sending at real volume, the curve is steeper and the focus is different. Here is what we use as a starting template for marketing sends on Gmail-heavy lists, pushing toward a target of 100,000 emails per day:

Day	Target sends	Gmail cap	Microsoft cap	Yahoo cap	Apple iCloud cap	Notes
1	100	50	25	25	25	Highest-engagement segment only (top 5%)
2-3	250	125	75	50	50	Same segment, plus known opens last 14 days
4-7	1,000	500	250	200	200	Expand to last-30-day actives
8-14	5,000	2,500	1,500	1,000	1,000	Expand to 90-day actives
15-21	15,000	8,000	4,000	3,000	3,000	Start mixing in normal campaigns
22-28	40,000	20,000	10,000	8,000	8,000	Full campaign mix at reduced send
29-35	75,000	40,000	20,000	15,000	15,000	Approaching full volume
36+	100,000	55,000	30,000	20,000	20,000	Target reached; steady state

Three things matter more than the exact numbers. First, segment quality: only send to engaged recipients in the first two weeks. Open rates in week one should be above 25% or you are warming incorrectly. Second, content stability: do not warm a new IP with one content template and then swap it for a different template after day 30. Third, time-of-day consistency: send at the same handful of hours daily during the warmup window. Change that only after the IP is established.

The chart below visualises the exponential ramp shape and the per-provider divergence — Gmail tolerates the steepest ramp because its ML reputation system rewards consistency-with-engagement over absolute volume; Microsoft requires the most conservative pace because its filter is the most reactive to volume jumps without engagement signal; Apple iCloud sits between them with conservative ramping but Microsoft-comparable steady-state caps.

Per-provider daily volume cap during 35-day warmup to 100K/day target

Categoría	Gmail cap	Microsoft cap	Yahoo cap	Apple iCloud cap
1	50	25	25	25
3	125	75	50	50
7	500	250	200	200
14	2500	1500	1000	1000
21	8000	4000	3000	3000
28	20000	10000	8000	8000
35	40000	20000	15000	15000
36+	55000	30000	20000	20000

Volume caps reflect Blue Spirit operational guidance derived from 100+ managed warmup engagements through 2025-2026. The Gmail-Microsoft divergence is structural, not coincidence: Gmail's ML reputation system (rolled out incrementally 2022-2024) rewards engagement consistency relatively independently of absolute volume, while Microsoft's filter (updated significantly post-May-2025 enforcement) reacts faster to sudden volume increases without proportional engagement signal. Apple iCloud's caps are derived from Apple Mail Privacy Protection era observations where open-rate metrics are unreliable — caps trail Microsoft because Apple's ML signal is harder to read in real-time. The log scale shows the exponential nature of the ramp — linear schedules produce stalled IPs at receivers using ML reputation models.

Monitoring, not just sending

Warming is not a send-and-pray operation. You need real-time feedback on how the warmup is going, per-ISP. The three signals we watch:

Gmail Postmaster Tools “IP reputation” and “domain reputation”. These are blunt (High/Medium/Low/Bad) but they tell you when Gmail has decided it does not like you. Our rule: if IP reputation drops to Medium during the warmup, pause all non-engagement traffic for 72 hours. If it drops to Low, pause everything and investigate. Postmaster Tools v2 (released October 2024) added more granular daily metrics including engagement scoring, complaint rate trends and authentication compliance per-day; we use the v2 daily metrics for early-warning signals before the rolled-up reputation indicator catches up.

Microsoft SNDS “Filter result”. Microsoft is less generous with detail than Gmail, but the green/yellow/red colour code on SNDS is a reliable early indicator. A yellow day during warmup means slow the ramp by 50% for three days. JMRP (Junk Mail Reporting Program) complaints are visible per-IP within 24-48 hours and should be the secondary monitoring source — JMRP rate above 0.1% during warmup days 1-7 is a warning, above 0.2% sustained is stop-and-investigate.

Your own seed list. A privately-maintained seed list across Gmail, Outlook.com, Yahoo, iCloud, Fastmail, ProtonMail, and a representative sample of European regional providers (Web.de, GMX, T-Online.de, Orange.fr, Free.fr, Libero.it) receiving one of your real emails on each campaign. We check placement daily. Below 80% inbox in week two is a warning; below 70% is a stop-and-investigate threshold. The European seed addresses matter because Postmaster Tools and SNDS only cover Gmail and Microsoft — a warmup that looks healthy at Gmail/Microsoft can be quietly junked at the European regional providers and you only see it in seed list data.

What to do when the warmup stalls

Warmups do not fail catastrophically, they stall. Common symptoms and fixes:

Gmail placement drops at day 10. Usually means you expanded your segment too fast. Go back to the previous segment for three days, then expand more gradually. The Gmail ML reputation model evaluates “audience expansion velocity” as a signal — adding 10× more recipients overnight even with engaged subscribers triggers conservative filtering until consistency stabilises.

Microsoft starts junking you around day 14. Microsoft is more sensitive to volume jumps than Gmail. Cap the Microsoft percentage of your send at 20% of your total until day 21, then scale up. The Microsoft reactivity changed materially after May 5, 2025 enforcement — non-compliant senders are now hard-rejected (550 5.7.515 Access denied) but compliant senders still face stricter ML evaluation than pre-May 2025.

Hard bounce rate climbs to 3%+ in the first week. This is a list-hygiene problem, not an IP problem. Stop warming, validate the list (NeverBounce, ZeroBounce, BriteVerify), and restart the warmup from day 1 on the cleaned list. Continuing to warm with bounce >3% damages reputation faster than the warmup curve builds it.

Everything looks fine but reply rates are anaemic. Possibly not an IP problem at all — check your content, your from name, and whether you are passing DMARC alignment consistently (not just DKIM or SPF individually, but alignment per the DMARC RFC rules). Read the <auth_results> section of your DMARC rua aggregate XML reports to verify the authenticating domain matches your visible From: header.

Orange.fr / T-Online.de start junking around day 21 even though Gmail/Microsoft look healthy. European-specific issue: ARC sealing requirement at Orange.fr (Sept 2024) and stricter throttle behaviour at T-Online.de mean that a forwarder in your sending chain may be breaking authentication for European recipients only. Check that your outbound relay ARC-seals; if you operate forwarders, confirm they ARC-seal correctly per our OpenARC + Postfix guide.

When to abandon a warmup

Sometimes the right call is to stop. If after two weeks the Gmail reputation is Low and nothing you try brings it back, the IP has probably acquired a bad history you cannot see and you should swap it for a clean one. We include IP rotation at no charge on our managed warmup products for exactly this reason; if you are warming an IP you bought from a generic VPS provider and it is stuck, contact the provider about a swap or cut your losses. The cost of continuing to warm a stuck IP — calendar time, list damage from sustained junk-folder placement, opportunity cost vs starting over — exceeds the cost of swap-and-restart 95% of the time.

Cost of warmup — the operator economics in 2026

Warmup is the most under-budgeted line item we see during deliverability audits. Operators consistently underestimate the actual fully-loaded cost — calendar time, engineering attention, list-segmentation work, monitoring tool subscriptions, opportunity cost of holding back campaigns during the 35-day arc. The honest economics, by approach:

Approach	Out-of-pocket cost	Calendar time	Engineering hours	Risk profile	Best fit
DIY on owned IP, no tools	€0 (existing infrastructure)	35-45 days	60-90 hours total (segmentation, monitoring, recovery)	High — no early warning system, late stall detection	Senders with operational deliverability expertise in-house
DIY + free monitoring (Postmaster Tools, SNDS, free DMARC reporting)	€0-50 (seed list service)	35 days standard	40-60 hours	Medium-high — better early warning but still manual	Senders with one engineer dedicated for the duration
DIY + paid monitoring (PowerDMARC, EasyDMARC, GlockApps, Litmus)	€200-800/month during warmup	35 days standard	25-40 hours	Medium — automated alerting, faster stall detection	Senders building in-house deliverability function long-term
Managed warmup service (Blue Spirit, Mailgun Pro, SendGrid)	€1,500-5,000 one-time or included with hosting	35 days standard	5-10 hours (monitoring oversight only)	Low — managed by team that does this weekly	Senders without dedicated deliverability engineering capacity
Skip warmup, use established ESP shared pool (Mailchimp, Klaviyo, Brevo, Sendinblue)	€0.0008-0.002 per message ongoing	0 days (instant)	0 hours	Low risk, but no own-IP control + per-message premium permanently	Senders below 100K/month where shared pool economics work

The hidden cost most operators miss: the opportunity cost of campaigns held back during warmup. A B2C operator with a 500K list normally sending 4 campaigns/month at 8% conversion to €40 average order value is generating ~€640K/month from email. Holding back those campaigns for 35 days during warmup is not literally €640K of lost revenue (some campaigns can run on the existing IP infrastructure during warmup, others can be deferred), but the realistic foregone revenue during a warmup window is typically 15-25% of normal monthly email revenue. For a sender at €640K/month email revenue, that is €100K-160K of opportunity cost — a number that puts the €1,500-5,000 managed warmup fee into perspective.

The honest operator framework: if your monthly email revenue is below €20K, DIY warmup makes sense because the opportunity cost is small enough to absorb. Above €100K/month, managed warmup is almost always the right ROI even at premium pricing — the calendar acceleration and risk reduction pay back many times over. Between those bounds, the decision depends on whether you have dedicated engineering capacity for the 35-day window. We see the worst outcomes from senders in the €50K-100K/month band who try DIY warmup with insufficient engineering time, stall around day 15, and lose 3-4 weeks of revenue while debugging.

The closest thing to free warmup that actually works in 2026 is starting on an established ESP shared pool (Mailchimp Standard, Klaviyo, Sendinblue) and migrating to a dedicated IP only after volume justifies the per-message premium. The shared pool gives you 0-day warmup at the cost of permanent per-message pricing; the migration to dedicated IP eventually requires its own warmup but you arrive at it with better list hygiene, established subscriber engagement patterns, and an authentication baseline already operational.

Maintenance after warmup

A warmed IP is not a permanent achievement. The reputation decays if you stop sending (noticeably after 30 days of inactivity, severely after 90) or if you change sending patterns drastically. Treat the post-warmup period as a low-risk operating zone, not a finish line — continue monitoring, avoid sudden volume changes, and if you need to pause sending for more than a couple of weeks, plan a small re-warm when you come back.

The maintenance cadence we recommend post-warmup: weekly Gmail Postmaster check, weekly Microsoft SNDS check, monthly seed list verification across all major providers, quarterly DMARC report review for unauthorised-source detection. DKIM key rotation annually minimum (per our rotation guide) with two-selector overlap window. SPF record review whenever you onboard a new sending integration.

Recovery from a stalled or failed warmup — the 4-step protocol

When a warmup stalls badly enough that pushing through is not an option, the structured recovery protocol below works for roughly 70% of cases we see. The remaining 30% require IP swap (covered in the abandonment section above) — but before that, work through the protocol:

Step 1 — Diagnose what stalled, by reading the data not by guessing. Pull Gmail Postmaster Tools for the warmup window and identify the day reputation degraded. Pull Microsoft SNDS Filter Result history. Pull seed list placement data. Cross-reference with what you sent that day — volume jump, segment expansion, content change, time-of-day shift, or pattern that did not exist on prior days. Stalls always have a proximate cause; “warmup just stalled randomly” is the diagnosis of an operator who has not read the data.

Step 2 — Identify whether the stall is reputation-driven or list-driven. Reputation-driven stalls (Gmail drops to Medium, Microsoft junks unexpectedly, Apple shows decreased engagement) are recoverable by pausing volume and re-establishing engagement quality. List-driven stalls (hard bounce >3%, complaint rate >0.3%, unsubscribe rate >0.5%) are NOT recoverable by warmup tactics — they require list validation and segmentation work, then a fresh restart on the cleaned list. Continuing to warm with a list-driven stall damages reputation faster than the curve repairs it; the 70% recovery rate above includes only reputation-driven stalls correctly diagnosed.

Step 3 — Execute the recovery cycle. For reputation-driven stalls: pause for 72 hours (volume zero), then resume at 50% of the volume on the day before the stall, holding constant for 5 days, then attempt to resume the curve at the stalled day’s level. For list-driven stalls: full pause, validate list with NeverBounce or ZeroBounce, suppress everything below confidence 90%, restart the warmup from day 1 on the cleaned subset. Document the stall and recovery in your runbook; recurring stall patterns indicate structural issues with your sending operation that no warmup tactic will fix.

Step 4 — Decide whether to continue or swap IPs. After the recovery cycle, evaluate honestly: is reputation actually rebuilding (Gmail Postmaster trending toward High, SNDS green, seed list inbox >85%) or stuck? If still stuck after the recovery cycle, swap the IP. The cost of running 2-3 recovery cycles on a stuck IP is calendar time you cannot get back; the cost of swap-and-restart is 5-7 days of lost volume but a fresh reputation slate. We see operators run 4-5 recovery cycles in a row hoping the IP will “come around” — it rarely does, and they end up where they would have been after a swap, just six weeks later.

The one operational rule that captures the recovery protocol: the second consecutive recovery cycle without measurable improvement is the signal to swap, not the third or fourth. Operators who absorb this rule lose 7-10 days during stalls; operators who don’t absorb it lose 4-6 weeks waiting for a recovery that won’t come.

Engagement segmentation — the technique that makes warmup actually work

The single technique that separates successful warmups from failed warmups is engagement segmentation. The volume curve in the table above is the easy part; getting the segmentation right is what determines whether the curve survives contact with reality. Here is how we do it in practice across paying client warmups.

For a list of 500K subscribers we are warming a fresh IP onto, the segmentation we run looks like this:

• Tier 1 (week 1 audience) — subscribers who opened or clicked in the last 14 days. Typically this is 8-15% of a healthy list, so 40K-75K subscribers.
• Tier 2 (week 2 audience) — subscribers who engaged in the last 30 days. Typically 25-35% of healthy list.
• Tier 3 (week 3 audience) — last 60 days engagement.
• Tier 4 (week 4+) — last 90 days.
• Subscribers who have not engaged in 90+ days do not get included in the warmup at all; they get sunset-policy-treated separately after the IP is established.

The math works because Gmail and Yahoo both weight engagement signals heavily during the first 30-90 days of an IP’s reputation accumulation. If your first 1,000 messages on a new IP have 40% open rates and 5-8% click rates, the IP’s reputation arc starts strongly. If those same 1,000 messages were sent to random subscribers with 12% open rates, the IP starts at “average sender” rather than “engaged sender”, and recovery to high reputation takes months instead of weeks.

The mistake operators consistently make: treating engagement segmentation as “nice to have” and warming on full lists because “the volume math works out”. We see it about every other warmup that comes to us already broken. The volume math may work; the reputation math does not. The IP gets to volume but never gets to high reputation, and inbox placement plateaus around 75-82% rather than the 92-96% you can hit with proper segmentation.

The diagram below maps the segmentation tier flow that makes the warmup curve actually work — what subset of your list enters the warmup at each week, what gets excluded entirely, and how the IP’s reputation arc maps onto the audience expansion. The shape of this is what receivers see, and it is what determines whether the IP lands at “engaged sender” or at “average sender” steady state.

Choose your warmup strategy — the decision tool

The standard 35-day warmup is the safe path, but operational reality often imposes calendar pressure or list-quality constraints that change the optimal approach. Use the tool below to get a calibrated strategy recommendation; the math behind it reflects 100+ managed warmup engagements through 2025-2026 and will reject combinations that we know operationally fail.

The tool’s logic, in summary:

• Standard 35-day — fits most senders with comfortable calendar and reasonable list health. Highest steady-state reputation outcome.
• Compressed 21-day — accelerated 1.5× pace for excellent or good lists with moderate calendar pressure. Tighter monitoring required.
• Aggressive 14-day — only for excellent lists with known IP source and tight calendar pressure. Daily checkpoints non-negotiable; abort-and-restart plan must exist.
• Multi-IP cohort (4-5 months) — required for volume targets above 5M/day where pool warmup with sequential cohorts is the only reputation-safe path.
• Reject — combinations of degraded list + recycled IP + urgent calendar that will produce stuck reputation rather than working IP. Tool will recommend restructuring the project rather than force-warming.

ESP-specific quirks that matter in 2026

Each major ESP has handling quirks that affect warmup decisions. The Gmail/Yahoo/Microsoft framing in the volume table covers the basics; the operational nuances below are what experienced operators adjust for, organised by provider.

Provider	Behaviour during warmup	Operator adjustment
Gmail / Workspace	ML-driven engagement weighting since 2022-2024. Rewards consistency over volume; ratio of engagement to total send is the primary signal.	Do not exceed segmentation capacity even if volume curve says you can. Gmail watches the ratio, not the absolute numbers. Postmaster Tools v2 daily metrics surface stalls 24-48h faster than rolled-up reputation.
Microsoft (Outlook.com / Hotmail / Live)	Most reactive to volume jumps; junk-folder placement does not flag itself in any visible way (only via SNDS or seed lists). May 5 2025 enforcement makes non-compliant senders hard-rejected (550 5.7.515).	Cap Microsoft delivery at 15-20% of daily volume during first 21 days regardless of subscriber distribution. Hold back gradual flow-through.
Yahoo Mail (incl. AOL)	Got more aggressive on complaint rate since the Feb 2024 bulk sender requirements. 0.3% threshold enforced more aggressively at Yahoo than Gmail.	Monitor Yahoo complaint rate hourly during warmup (FBL signals within 30-60min). If sustained climb above 0.2% for 2 days, pause Yahoo for 48h and investigate.
Apple Mail / iCloud	Mail Privacy Protection (rolled out 2021-2022) makes traditional open-rate metrics unreliable for Apple users — pre-fetching inflates apparent open rate.	Use click rate as leading indicator for Apple-heavy lists. Supplement with reply tracking when campaign type allows.
Web.de / GMX (United Internet)	Reputation-led acceptance with stable behaviour. FBL available for established senders.	Builds trust steadily over 21-28 days; standard 35-day curve fits well. Verify FBL signup post-warmup.
Orange.fr / Wanadoo	ARC sealing required on forwarded mail since September 2024. Conservative throttling, stricter rate limits than US providers.	Verify outbound relay ARC-sealing before warmup day 1. Cap Orange.fr at 12% of daily volume during weeks 1-2.
T-Online.de / Magenta (Deutsche Telekom)	Region-specific filtering; reputation builds slowly. DMARC enforcement and RDNS critical.	Standard schedule works but may need 25% extra time for Deutsche Telekom-heavy German B2C lists.
Free.fr	Applies stricter throttles to non-ARC-sealed traffic; less predictable than Orange.fr.	ARC sealing mandatory in practice. Monitor seed list at @free.fr addresses daily.
Libero.it / Virgilio.it	Italian regional providers; ML signal is harder to read in real-time. Less aggressive than the major providers but reputation builds slower.	Add 25% buffer in schedule for Italy-heavy lists. Standard schedule typically lands around day 42-45 for Italian audiences vs day 35 for US-heavy.
ProtonMail	Privacy-focused, full RFC 8617 ARC verification, conservative on signing requirements.	DKIM 2048-bit and aligned DMARC mandatory; 1024-bit signers see materially reduced acceptance.
Fastmail	Reference for small EU receivers; ARC since 2021. Stable behaviour, predictable acceptance after baseline reputation.	No special adjustment needed; if Fastmail looks healthy, similar small EU receivers usually do too.
Yandex / Mail.ru	Russian receiver with geographic filtering; conservative on non-Russian senders.	Treat as bonus channel; non-Russian senders typically see 6-12 week trust accumulation independent of main warmup.
Regional US ISPs (Comcast, Charter, AT&T, Verizon)	Less ML-driven, more rule-based. Respond predictably to RBL listings and DNS configuration changes.	Pre-warmup MXToolbox / Spamhaus / SURBL / Barracuda blacklist check. Once clear, behaviour is consistent throughout warmup.

Pool warmup — when you are warming multiple IPs simultaneously

Single-IP warmup is the simple case. The harder operational scenario is warming an IP pool of 8, 16, or 32 IPs simultaneously for high-volume operators. The math changes in two important ways.

First, per-IP volume needs to be lower for the same total volume. If you are warming a 16-IP pool toward 1.6M emails/day, each IP individually warms at 100K/day equivalent — not 1.6M divided by 16 with each IP at full pool volume. The receivers see each IP independently, so each IP’s reputation arc must work independently. This means a 16-IP pool warmup is not 16× faster than single-IP warmup; it is roughly the same calendar duration with parallel reputation building.

Second, rotation discipline during warmup matters more than during steady-state operations. During warmup, sequential IPs in the pool should warm with overlapping segments rather than handing off cleanly between IPs. The reason: if IP-1 warms on the engaged segment and then hands off entirely to IP-2 in week 2, IP-2 sees the engaged segment as “first impression” and IP-1 sees a sudden volume drop that registers as suspicious. Better pattern: weighted rotation where IPs are introduced gradually (IP-1 at 100% week 1, then IP-1 at 75% + IP-2 at 25% week 2, IP-1 at 50% + IP-2 at 50% week 3, etc.).

For very large pools (32+ IPs), use the staggered cohort approach: warm 4 IPs to completion over 35 days, then add 4 more IPs in the next cohort. This trades total calendar time (a 32-IP pool takes 4-5 months) for reduced operational risk during any given week. Acceptable when you have time; not acceptable when business needs the full pool capacity quickly.

For senders combining transactional and marketing on separate IP segments — which is the production-grade architecture for any sender above 500K/day — see our transactional vs marketing IP separation guide for the segmentation pattern that survives reputation contagion between traffic types.

Five myths about IP warmup that still circulate in 2026

Because warmup advice on the open internet is mostly stale, the same myths keep recurring. Here are the five we see most often, and why they are wrong.

Myth 1 — “Send to your most engaged list and you can skip warmup”. No. The most engaged list helps reputation accumulate faster but does not eliminate the need for the volume ramp. Gmail’s reputation system specifically penalises IPs that go from zero traffic to high volume within a 24-72 hour window regardless of engagement quality. The volume curve is the discipline; the engagement segmentation is the multiplier on that discipline.

Myth 2 — “Warmup is only needed for cold outreach, not for opt-in marketing”. Wrong. The IP reputation system does not distinguish between cold outreach and opt-in marketing — it sees engagement patterns. A cold IP sending opt-in marketing to engaged subscribers needs warmup just as much as a cold IP sending cold outreach. The difference is that opt-in lists give you better engagement during warmup, so the curve completes faster, but the curve still has to happen.

Myth 3 — “If I use a ‘pre-warmed’ IP from my hosting provider, I can skip warmup”. Mostly wrong. A “pre-warmed” IP from a hosting provider has reputation history with their previous customer’s content and audience, not yours. The IP starts your traffic with some accumulated trust, which helps, but Gmail will still notice the change in sending patterns and adjust reputation accordingly. Treat pre-warmed IPs as “warmup-shortened” not “warmup-eliminated” — typically you can compress a 35-day warmup to 14-21 days, but you still need the ramp.

Myth 4 — “BIMI and MTA-STS speed up warmup”. Untrue but understandable confusion. BIMI and MTA-STS improve inbox placement and brand visibility for senders with established reputation, but they do not accelerate reputation accumulation on a new IP. They are post-warmup optimisations, not warmup accelerators. Operators who publish BIMI on day 1 of warmup will not see faster placement improvement than operators who publish BIMI on day 60. Both will see equivalent placement bumps once their reputation reaches “established” status. (See our BIMI guide for the actual cost-benefit math on VMC and CMC.)

Myth 5 — “Once warmed, the IP is permanent — I can change anything”. False. A warmed IP is contextually warmed: warmed for the content type, sending pattern, audience profile, and authentication configuration that produced the warmup. Drastic changes to any of these — switching from marketing to cold outreach, swapping content templates entirely, abruptly changing audience composition — can degrade reputation rapidly. Treat the post-warmup state as “warmed for this specific operating context” and re-warm gradually if you are making structural changes.

What the next 12 months will likely require

Three trends I see emerging in deliverability work that will probably affect warmup advice during 2026-2027.

First, DMARC enforcement at p=reject is moving from optional to expected for serious senders. Microsoft already weights DMARC enforcement higher than Gmail does; Google has been signalling tightening enforcement requirements through Postmaster Tools communications in late 2025. PCI DSS v4.0 made DMARC effectively mandatory for cardholder-data entities from March 31, 2025, and DORA put email authentication into ICT third-party risk for EU financial entities from January 17, 2025. For 2026 warmups, I am increasingly recommending publishing DMARC at p=quarantine within 30 days of warmup completion (rather than the 90-day-comfort zone we used to suggest), and moving to p=reject within 90 days. The reputation bonus for being on enforcement-grade DMARC is real and growing.

Second, TLS-RPT and MTA-STS are becoming entry-level requirements rather than advanced configurations. Operators warming new IPs in 2026 should publish both from day 1 of warmup, even before sending the first message. The signals these provide to ISPs about sender sophistication contribute to faster reputation accumulation, particularly with Gmail. Five years ago, advanced operators published these; now everyone serious should. (See our MTA-STS implementation guide for the deployment specifics.)

Third, AI-assisted content quality scoring is increasingly weighted in ISP filtering decisions. Gmail and Microsoft are both reportedly using ML models that evaluate message body quality alongside reputation and engagement signals. The implication for warmup: content quality matters more than it did even two years ago, and operators warming IPs with AI-generated low-effort content will see slower reputation accumulation than operators warming with human-quality content. This is observation rather than published policy from any ESP, but the pattern is consistent across multiple warmups we have run in 2025-2026.

The fundamentals of IP warmup — gradual volume ramp, engagement-first segmentation, per-ISP monitoring, recovery discipline when stalls happen — have not changed and will not change. What changes is the surrounding requirements: authentication standards get stricter, ML-driven filtering gets more sensitive to engagement signals, and the cost of skipping warmup discipline keeps going up. Operators who warm IPs properly in 2026 will find the practice rewards them with better placement than the same operators got in 2022; operators who try to shortcut warmup in 2026 will find the failures more catastrophic than they were in 2022. The proper authentication baseline laid out in our email authentication 2026 guide is the foundation that warmup discipline builds on top of — neither alone is sufficient.

---

Warmup done-for-you is included with our dedicated IP and email warmup plans. We do this every week for paying clients.