Suped

Exchange Online starts Direct Send blocking rollout for new tenants

News
Published 16 Nov 2026
Updated 16 Nov 2026
7 min read
Summarize with
Exchange Online Direct Send blocking rollout for new tenants.
17 November 2026 (Australia/Sydney).
Microsoft is rolling out mandatory Reject Direct Send for new Exchange Online tenants, enabled by default without an option to disable it. The rollout is underway, rather than complete across all new tenants. Existing tenants remain opt-in, with the setting off by default.
The change blocks anonymous inbound messages using a tenant accepted domain in the envelope sender when no mail-flow connector is attributed. I treat this as a mail acceptance policy change: legitimate devices and applications need their delivery routes checked, even when their email authentication is correct.

What changes for new and existing tenants

Microsoft published its original announcement on 16 November at 16:14 UTC and revised it at 18:59 UTC. Both times fall on 17 November in Sydney. The announcement says the mandatory default for new tenants is being deployed at publication; it gives no completion date.
The distinction matters for administrators managing several organizations. An existing tenant does not become subject to mandatory blocking simply because the new-tenant rollout has started. The policy is available in the worldwide environment, but is unavailable in GCC High and DoD. There is no announced universal migration deadline.

Tenant group

Default or availability

Control

New worldwide tenants
Blocking rolling out
Mandatory; no disable option
Existing worldwide tenants
Off by default
Opt-in
GCC High
Unavailable
Not applicable
DoD
Unavailable
Not applicable
Tenant treatment in the 16 November announcement.

The envelope sender determines the scope

Reject Direct Send evaluates P1 MAIL FROM, the SMTP envelope sender, against the tenant accepted domains. It rejects matching inbound traffic when the setting is active and Exchange Online cannot attribute the message to a mail-flow connector. In this context, anonymous means unattributed to a connector.
The visible P2 From header is not evaluated by this setting. Subdomains fall within an accepted domain's coverage when MatchSubDomains is true. The check works regardless of where the public MX record points, so routing ordinary inbound mail through an intermediary does not remove this condition. Microsoft sets out these boundaries in its policy explanation.
An external sender whose envelope domain is not accepted by the tenant is outside this particular condition. Reject Direct Send is therefore not a blanket block on external mail, nor a complete defense against visible From impersonation.
Own-domain envelope sender
Envelope sender: scanner@example.com. Visible From: scanner@example.com. If example.com is accepted and there is no connector attribution, this message is in scope for blocking.
Own-domain visible From only
Envelope sender: bounce@outside.example. Visible From: scanner@example.com. The visible From alone does not trigger Reject Direct Send. Other email security checks still apply.

Recognize the specific rejection

An affected sender receives SMTP status 550 5.7.68 with the TenantInboundAttribution diagnostic. This is a permanent rejection for the attempted delivery, rather than a request to retry later. Repeated attempts over the same unattributed route do not correct the cause.
Capture the complete response in the sending application or SMTP logs, together with the timestamp, recipient and envelope sender. A visible From address or a general delivery failure screenshot is insufficient to establish that Reject Direct Send caused the failure. Investigate the actual envelope and connection used. If the response has a different status or diagnostic, investigate that failure separately rather than assuming it belongs to this rollout.
Exact SMTP rejection
550 5.7.68 TenantInboundAttribution; Direct Send not allowed for this organization from unauthorized sources

Inventory dependencies before enabling an existing tenant

I start with a read-only inspection in an existing Exchange Online PowerShell session. The following command displays the organization setting; it does not change mail acceptance or configure a connector.
Then establish which systems send using your accepted domains into tenant mailboxes. Include multifunction printers, network appliance alerts, scheduled application jobs and third-party business notifications. Ask the owners for real message samples instead of relying only on their configured visible sender addresses.
Inspect the existing tenant settingpowershell
Get-OrganizationConfig | Format-List RejectDirectSend
A False result means the existing tenant has not enabled this setting. A True result means it is configured on. Neither result proves that a particular application has a working connector: inspect the route and test that application separately.
Use sending logs and message trace to build the inventory, with SPF records as another starting point for identifying infrastructure. Cover scheduled and low-volume jobs across a normal business cycle. An empty sample over a quiet afternoon is weak evidence that no dependencies exist. For infrequent jobs, arrange a test run with the owner rather than assuming that no recent messages means the system is unused.
  1. Ownership. Record the application or device owner and its business purpose.
  2. Envelope. Capture the actual MAIL FROM domain and affected tenant recipients.
  3. Connection. Record the sending IP or TLS certificate and intended inbound connector.
  4. Forwarding. Identify messages sent outside the tenant and forwarded back into it.

Verify the approved connector route

For a legitimate source, Microsoft recommends a partner inbound connector that matches the sender's TLS certificate, with sending IP addresses as another option. Scope the connector to the actual approved source. A broad shared IP range or certificate is a poor substitute for understanding who can use that infrastructure.
For each sender, I want a fresh test delivered to the affected tenant and evidence that Exchange Online attributes it to the intended connector. Creating a connector object alone is insufficient. Use the test message ID in message trace, allowing time for trace data to appear.
Inspect attribution for a test messagepowershell
Get-MessageTraceV2 -MessageId "<message ID>" | Get-MessageTraceDetailV2 | Format-List
Review the trace details for the intended connector and retain the result with the sender inventory. Repeat the test for distinct sending routes, including failover infrastructure. A successful test on one route does not establish attribution for another. Record configuration alongside observed traffic: a device's documented IP is insufficient if its actual SMTP connection uses a different outbound gateway.
If a sender cannot provide a suitable certificate or stable sending IPs, resolve its delivery route with the owner before enabling blocking on an existing tenant. An SPF include or a familiar sender address does not supply the missing connector identity. Connector attribution also does not promise inbox placement. Investigate other filtering or delivery errors when a traced message passes this check but still does not arrive as expected.

Test forwarding out of the tenant and back

Microsoft identifies a specific forwarding problem: a tenant user sends outside the organization, then the recipient's provider forwards the message back without Sender Rewriting Scheme (SRS). The returning message retains the original own-domain envelope sender. With Reject Direct Send on and no partner connector, that return is rejected.
Test this round trip explicitly, including external support workflows and automated forwarding. These messages already fail SPF in the scenario Microsoft describes, but the new setting adds outright rejection. SRS rewrites the envelope sender; its presence does not establish connector attribution or guarantee acceptance under other policies. The forwarding warning makes this a separate acceptance test.

Authentication checks have a separate job

This is not a new DMARC, SPF or DKIM standard. Passing SPF, signing with DKIM or publishing a DMARC policy does not establish Exchange Online connector attribution. Fixing authentication remains useful, but it does not clear this rejection when the message still meets the blocking condition.
I keep recipient acceptance tests separate from authentication tests. Send to the affected tenant to check the connector route. Use Suped's email tester on a separately sent sample to inspect authentication; that result does not prove the tenant will accept the message.
Our DMARC monitoring in Suped helps investigate reported sending sources and authentication failures. Aggregate reports are not a complete inventory of internal Direct Send dependencies. Match that evidence with sender logs and Exchange Online trace details before deciding a route is ready.

Make connector evidence the readiness test

For new tenants, check critical device and application deliveries during onboarding because the mandatory default is rolling out now. For existing tenants, finish the sender inventory and verify approved routes before choosing to enable the setting.
Keep one acceptance record per route: the test message ID, attributed connector and delivery result. Include an external forwarding round trip. Resolve unexplained failures with the sending owner before treating the application as ready.

Frequency asked questions

No. Mandatory blocking for new tenants is rolling out as of the 16 November announcement. Existing worldwide tenants remain opt-in and off by default. Microsoft has not announced a universal completion date.

No. SPF authorization does not attribute a message to an inbound connector. An own-domain envelope sender without connector attribution still meets the blocking condition when Reject Direct Send is on.

No. The setting checks P1 MAIL FROM, not the visible P2 From header. A visible own-domain sender alone is insufficient to trigger this policy.

No. Reject Direct Send works regardless of the MX target. Verify connector attribution for the actual inbound route, including forwarded messages that return with an own-domain envelope sender.
DMARC monitoring

Start monitoring your DMARC reports today

Suped DMARC platform dashboard
What you'll get with Suped
Real-time DMARC report monitoring and analysis
Automated alerts for authentication failures
Clear recommendations to improve email deliverability
Protection against phishing and domain spoofing