Microsoft Is Blocking Cold Email Accounts in 2026: Limits, ML Detection, and How to Survive

Short answer: Microsoft went from tolerant to hostile toward cold email in 2026. Reports across the outbound industry describe millions of M365 accounts blocked in Q1 alone, sharply tightened sending limits from newly-created domains, and (since spring) machine-learning detection keyed on engagement patterns rather than just volume.

What triggers Microsoft's blocks now

• Volume from young domains/tenants — fresh domain + immediate outbound volume is the classic flag. Ramp over weeks.
• Engagement-pattern ML — low reply rates, high delete-without-read, and burst-pattern sending now score against you even under volume limits.
• Authentication gaps — missing or misaligned SPF/DKIM/DMARC compounds every other signal. Microsoft's filters treat unauthenticated bulk mail as presumptively abusive.
• Blacklist echoes — listings at Spamhaus/Barracuda feed corporate filter decisions across the Microsoft ecosystem.

How outbound teams survive it

1. Authenticate everything, verify continuously — a drifted DKIM key or broken SPF include is now an account-level risk, not just a spam-folder risk. Grade your domains.
2. Ramp like it's 2026, not 2021 — weeks-long warmup on new domains, conservative daily volume per mailbox, consistent (not bursty) sending windows.
3. Earn engagement — smaller, better-targeted lists beat volume; the ML rewards replies and punishes apathy.
4. Monitor the leading indicators — auth health and blacklist status degrade before blocks land. Teams that catch a Barracuda listing on Tuesday don't lose an M365 tenant on Friday.

Aurelius watches those leading indicators across all your sending domains — free for 2 domains.