Postmark's KumoMTA Migration: Faster Queue Times and What They Mean for Your Stack

A password reset that takes eight seconds to arrive feels broken even when the email lands. Users hit “resend” twice, support tickets pile up, and the team that picked your transactional email provider hears about it. So when an email infrastructure company quietly rebuilds the engine that actually sends your messages, the right question is not “did the brand change” but “did the seconds change.”

On March 31, 2026, Postmark announced it had finished migrating every sending stream from PowerMTA to KumoMTA, an open-source mail transfer agent written in Rust. The published numbers show queue times dropping at all four major mailbox providers, with the largest absolute gains at Microsoft and Apple. If you send transactional email through Postmark — or you’re evaluating Postmark against SendGrid, Mailgun, Amazon SES, or Resend — this changes a few of the trade-offs.

This article walks through what KumoMTA is, what Postmark actually measured, what it means for senders, and where Postmark still falls short.

What Changed and Why It Matters

Postmark replaced PowerMTA, the proprietary MTA it had used for years, with KumoMTA, a free, Apache-2-licensed MTA built by veterans of the email infrastructure world. Every email Postmark sends now passes through the new engine. There was no migration step for customers — no API endpoint change, no SDK update, no DNS edit. The handoff happened on Postmark’s side.

The motivation, according to Postmark’s own engineering writeup, was traffic control. When Gmail or Microsoft tightens throttling on a particular IP, the sending platform needs to back off, reroute, and reshape traffic in seconds — not hours. PowerMTA’s configuration model made that slow. KumoMTA’s Lua-based config gives the deliverability team granular control over per-provider sending behavior, and its real-time event stream feeds straight into Postmark’s dashboards.

The secondary motivation was scaling cost. PowerMTA ran on aging hardware that was awkward to expand; KumoMTA runs on cloud infrastructure with auto-scaling.

The Numbers Postmark Published

Postmark’s blog post includes a chart of average queue time — the gap between the API accepting a message and the SMTP handoff to the recipient’s mail server. These are the figures from their internal testing:

• Gmail: ~1.2 seconds with KumoMTA, down from ~2 seconds
• Yahoo: ~3.2 seconds, down from ~4.6 seconds
• Microsoft: ~2.8 seconds, down from ~4.8 seconds
• Apple: ~6.3 seconds, down from ~8 seconds

Queue time is not the same as inbox time — DNS lookups, recipient server queues, and rendering delays all happen after handoff. But queue time is the part the sender controls, and a two-second reduction on Microsoft is meaningful when your user is staring at a login screen waiting for a magic link.

Bounce rates, according to Postmark, held steady or improved slightly across all four providers. The fact that bounce rates did not spike during the migration is arguably more important than the queue-time wins; it tells you the new system maintained the 16-year sender reputation Postmark has built.

What This Means If You Already Use Postmark

The practical change for existing Postmark customers is short:

• Your API keys, sending domains, and DKIM records are untouched
• Webhook payloads and event types are the same
• Pricing is unchanged
• Transactional emails arrive faster, especially to Microsoft and Apple addresses
• During provider throttling events, Postmark can now adjust sending in real time

If you’ve ever had to file a support ticket because a chunk of your sends got deferred when Gmail or Outlook tightened the screws, this is the part of the migration that matters most. The new traffic-shaping controls don’t eliminate provider-side blocks, but they shorten how long Postmark needs to react.

Pricing (verified May 2026):

• Developer (Free): 100 emails/month, permanent — useful for staging and side projects
• Basic: $15/month for 10,000 emails; $1.80 per 1,000 additional
• Pro: $16.50/month for 10,000 emails; $1.30 per 1,000 additional
• Platform: $18/month for 10,000 emails; $1.20 per 1,000 additional

The Basic-to-Pro jump is only $1.50/month and meaningfully reduces overage cost — most teams sending over 12,000/month should be on Pro. Annual billing is not offered, and unused emails do not roll over. Check the live numbers on Postmark’s pricing page before budgeting.

Where Postmark still falls short: The product is opinionated about what it does and does not allow. Promotional broadcasts and cold outreach are not permitted on Postmark — they belong on the Bulk API stream, and even there, the policy enforcement is strict. Marketers used to running newsletters through Mailchimp or ActiveCampaign cannot drop those campaigns onto Postmark. There is no built-in subscriber management, no automation builder, no landing pages. At very high volumes (millions of emails/month) Postmark’s overage rates make it more expensive than SendGrid Pro or Amazon SES. And the dashboard, while clean, has not seen a major UX refresh in a few years.

What This Means If You’re Evaluating Transactional Providers

The KumoMTA migration sharpens Postmark’s position as the deliverability-first option, but it doesn’t reset every comparison. Here’s how the trade-offs look in mid-2026.

Postmark vs SendGrid

SendGrid (now part of twilio.com after the February 2026 brand consolidation) remains the most widely integrated transactional API in the world. Its strengths are scale, ecosystem breadth, and the bundled Marketing Campaigns module for teams that want one platform for transactional and marketing email.

SendGrid pricing starts at $19.95/month for the Essentials plan covering up to 100,000 emails. The next tier, Pro, jumps to $89.95/month — there is no middle plan for teams sending 150,000–500,000 emails/month, so the per-email cost gets awkward in that range.

Where SendGrid loses to the new Postmark: SendGrid does not publish queue-time numbers and does not separate transactional from marketing traffic on shared IPs the way Postmark does. For pure transactional sends — password resets, receipts, two-factor codes — Postmark’s per-stream IP pools and stricter content policy give it consistently better inbox placement on shared infrastructure. The KumoMTA migration adds a queue-time advantage on top.

Where SendGrid still wins: If you need transactional and marketing email under one API, SendGrid does it. Postmark does not. SendGrid’s ecosystem also has more pre-built integrations and partner tooling, which matters for teams plugging email into a larger stack.

Postmark vs Mailgun

Mailgun (now under Sinch’s umbrella) shares the developer-API positioning with Postmark but takes a different stance: it allows promotional sending, it offers inbound email parsing as a first-class feature, and its pricing structure includes plans designed for higher-volume bulk.

Mailgun’s Basic plan starts at $15/month for 10,000 emails, with Foundation at $35/month for 50,000 and Scale at $90/month for 100,000. At higher volumes Mailgun’s overage rates ($1.10–$1.80 per 1,000 depending on tier) can beat Postmark’s.

Where Mailgun loses: Mailgun’s deliverability on shared IPs has historically been solid but not at Postmark’s level for pure transactional. Its shared pools mix transactional and bulk traffic, which is precisely the contamination Postmark’s policy is designed to prevent. Mailgun also does not publish per-provider queue times.

Where Mailgun wins: Inbound email parsing — the ability for your app to receive incoming email at an address like support@yourapp.com and process it via webhook — is built in. Postmark also supports inbound, but Mailgun’s routing engine is more flexible if your product needs to act on email replies.

Postmark vs Amazon SES

Amazon SES is the cheapest option at high volume and the most painful to operate. At $0.10 per 1,000 emails (charged per recipient), 1 million emails costs $100. The same volume on Postmark would run roughly $1,300/month at Pro overage rates ($1.30 per 1,000 above the included 10,000). The economics are not close above ~200,000 emails/month.

Where SES loses: SES is raw infrastructure. There is no template editor, no subscriber management, no out-of-the-box bounce dashboard. Before you can send to external addresses, you must exit the SES sandbox by submitting a support request and explaining your use case — approval can take 24–48 hours. Deliverability configuration (DKIM, DMARC, suppression lists, IP warmup) is entirely your problem. Meaningful support starts at $29/month on the AWS Developer plan.

Where SES wins: At enough scale, the operational cost of running SES yourself is less than the per-email premium you’d pay Postmark, SendGrid, or Mailgun. Teams sending 5+ million emails/month with a dedicated email-ops engineer have a real reason to look at SES.

How to Choose

The KumoMTA migration doesn’t rearrange the top of the transactional shortlist, but it does sharpen the case for Postmark in two specific scenarios.

Pick Postmark if transactional deliverability is the single thing you’re optimizing for, you send password resets or magic links where every second on Microsoft or Apple matters, and you’re fine with not running marketing campaigns on the same platform. The KumoMTA upgrade just made the strongest argument for Postmark stronger.

Pick SendGrid if you need transactional and marketing email under one API, you’re already in the Twilio ecosystem for SMS or voice, or you need a partner platform with many existing integrations.

Pick Mailgun if inbound email parsing is part of your product, your team is happy with developer tooling that’s slightly rougher than SendGrid’s, or your sending volume sits in a tier where Mailgun’s overage rates beat the alternatives.

Pick Amazon SES if you’re sending well over a million emails per month, you have engineering capacity to manage suppression lists and deliverability yourself, and you’re already invested in AWS.

For a wider scan of options, see the best transactional email services ranking. For a side-by-side on the two most-asked-about options, the Postmark review and SendGrid review both go deeper than this article does.

Open Source Under the Hood: Why It Matters Beyond Postmark

There’s a wider trend worth noting. PowerMTA is proprietary. KumoMTA is open source, Apache-2 licensed, with code on GitHub and a contributor community that includes some of the largest senders in the world. Postmark is not the only player betting on it — KumoMTA already powers infrastructure at multiple high-volume senders, and the project is positioned as a credible PowerMTA replacement industry-wide.

For senders, this matters in two ways. First, the talent pool for engineers who understand the MTA layer just got larger; anyone can read the code, run it locally, and debug behavior. Second, providers betting on KumoMTA are less locked into a single vendor’s roadmap, which historically has been one of the slower parts of email infrastructure to evolve.

Whether that translates to faster feature shipping at Postmark specifically is something to watch over the next 12 months. The Bulk API that Postmark launched alongside the migration — a way to send promotional and broadcast email through Postmark’s stricter content policy — is the first sign that the new foundation is being used for new product surface, not just performance.

The Bottom Line

The Postmark KumoMTA migration is not a marketing rebrand or a pricing change. It’s an under-the-hood swap that produced measurable queue-time improvements at every major mailbox provider, with the largest absolute gains where it arguably matters most: Microsoft (down 2 seconds) and Apple (down 1.7 seconds). Bounce rates held. Nothing on the customer side broke.

For Postmark users, the right action is no action: your emails are just faster now. For teams evaluating transactional providers, the migration tightens Postmark’s deliverability lead for pure transactional sends, while leaving its weaknesses — no marketing email, no automation, no high-volume cost story — exactly where they were.

If transactional inbox placement is the metric your product depends on, Postmark just got harder to beat. See the full Postmark pricing breakdown for cost at your specific volume.

Sources

1. Postmark — KumoMTA Migration Engineering Writeup — accessed 2026-07-23
2. KumoMTA — Official Project Site — accessed 2026-07-23
3. Postmark — Pricing — accessed 2026-07-23