On this page
Microsoft has now explained the Microsoft 365 outage cause: a "latent defect" in an earlier fix for its networking gateway servers, which was set off when Teams traffic moved onto servers already running at reduced capacity. Microsoft executive Charles Lamanna set out the account on X, and outlets reported it from 16 November 2026. Separately, on 17 and 18 November, US officials and lawmakers began linking a run of outages, Microsoft's among them, to a rush to patch software flaws found by AI. Microsoft itself has not made that link.
This is a follow-up to our earlier report on the Microsoft 365 outage that hit Exchange Online on 12 and 13 November 2026. At that point Microsoft had said only that a network change and lost server capacity were involved, and that the underlying reason had not been disclosed.
What is new since 13 November?
- 16 November 2026: reports of Microsoft's explanation appear, drawn from a preliminary post-incident report and Lamanna's post on X. BleepingComputer says the preliminary report was shared in the Microsoft 365 Admin Center, with a final one expected later that week.
- 17 to 18 November 2026: administration officials tell Congress and the press that cyberattacks almost certainly did not cause the week's outages, and some point to patching as a factor, according to CyberScoop and AP, carried by News4JAX.
We could not open Microsoft's own post-incident report, which sits behind the admin centre. Everything below about Microsoft's explanation comes from outlets quoting it and from Lamanna's post, which we could not load directly either.
What caused the Microsoft 365 outage, according to Microsoft?
iTnews reports Lamanna's account as follows:
- A lock contention bug in Microsoft's internal networking gateway servers made them unresponsive under certain load conditions.
- The bug came from a fix for those servers that began rolling out in July and ran worldwide for more than three months.
- Testing missed it because labs used weaker hardware with fewer cores. The problem showed on newer machines with more cores.
SlashGear quotes Lamanna describing a "latent defect" and a "cascading failure".
How did the Teams traffic switch trigger it?
Two things coincided, per iTnews and BleepingComputer:
- About 15 hours before the outage, Teams failed over from a separate gateway system onto the primary gateway, which also serves Exchange Online. That added load and changed traffic patterns.
- Nearly 14 per cent of gateway capacity was already out of service because of an unrelated issue.
Servers began failing during peak US traffic, and traffic moved to the remaining machines, which then failed too. More than 80 per cent of the gateway's servers became unresponsive in under two hours. Isolation between capacity units and regions should have contained the damage but did not, because of a separate issue Lamanna did not describe, iTnews says.
Why was Exchange Online hit hardest?
According to iTnews, Teams, OneDrive, SharePoint and Office apps failed out of the faulty gateway and recovered earlier. Exchange Online took longer, and servers brought back were immediately swamped by queued traffic. BleepingComputer adds that Microsoft says it has increased capacity, improved monitoring and added safeguards against traffic redirection causing a repeat.
Why could Microsoft not simply roll back?
Lamanna said that because the main defect was a latent bug from several months earlier, normal rollback, flighting and feature flags could not quickly revert it, iTnews reports. Engineers instead built an update that returned the code to its previous state and applied it in large blocks of servers. Lamanna also said neither the defect nor the fix applies to broader Windows deployments.
Are AI-found flaws behind the wave of outages?
The wider claim is more contested. The AP reports a run of incidents: the Microsoft 365 disruption; a Federal Aviation Administration network malfunction on 13 November that caused the longest nationwide ground stop in 25 years; a United Airlines outage and a JPMorgan Chase software bug on 16 November; and Apple, Marriott and NSA disruptions on 17 November.
What officials said, per the sources we could open:
- AP reports Director of National Intelligence Jay Clayton told Fox News that patching after last month's PHP security alert had left some systems failing to restart as expected, and that this happened at the NSA on 17 November.
- SiliconANGLE reports Clayton said officials ruled out cyberattacks and attributed the outages to companies patching vulnerabilities on a much larger scale than before.
- CyberScoop reports Clayton, the Homeland Security Secretary and the FBI deputy director said there was no evidence of malicious attacks.
- Senator Mark Warner called the disruptions a "wake-up call", AP reports, saying AI is advancing faster than people can manage its consequences.
Reports from 19 November, including one carried by Business Standard that draws on the Financial Times, describe US officials worrying that the rush to patch AI-detected flaws is feeding the outages. We could not open that article, so we have not relied on it. The sources we did read do not name AWS in this run of incidents.
Analysis: two different stories, one uncomfortable theme
This is our analysis, not a finding from any source. Microsoft's own explanation points to a code defect, a capacity loss and a traffic shift. It does not mention AI-found vulnerabilities, and AP notes Microsoft said the outage did not stem from its AI or agent systems. Treating the Microsoft failure as proof of the patching theory would go beyond the evidence.
Still, the common thread is change. A fix shipped in July sat dormant until conditions changed in November. If faster patch cycles mean more changes reach production with less testing, latent faults of that kind could surface more often. Sauce Labs chief executive Prince Kohli made a similar point to SiliconANGLE about fixes deployed without full understanding of their effects. The CyberScoop report also notes the link to AI is largely lawmakers' and unnamed briefing participants' view, not a formal administration finding.
What to watch: Microsoft's final post-incident report, any specific named cause for the FAA stoppage, and whether officials publish evidence tying outages to AI-driven patching.
Sources
- SlashGear: Microsoft November 2026 outage cause explained
- BleepingComputer: Microsoft says months-old bug caused global Microsoft 365 outage
- iTnews: Microsoft says three-month-old bug behind global M365 outage
- AP via News4JAX: Tech outages are piling up, and AI is raising the stakes
- CyberScoop: Consensus mounts that tech outages were fueled by patching of AI-discovered vulnerabilities
- SiliconANGLE: Recent spate of outages attributed to technical issues amid accelerated AI bug discovery
FAQ
Frequently asked questions
Microsoft says a latent defect from a fix for its networking gateway servers, rolled out from July, made servers unresponsive under load. A Teams traffic failover and about 14 per cent of capacity already being offline triggered it.
Microsoft attributes the outage to a latent software defect, as described above. US officials quoted by CyberScoop and SiliconANGLE also say there was no evidence that attacks caused the week's outages.
No. Microsoft's explanation, as reported, concerns a code defect and capacity. The link between AI-detected flaws and outages comes from US officials and lawmakers talking about the wider run of incidents.
It began on Thursday 12 November 2026 and recovery continued through 13 November, according to BleepingComputer. See our [original report](/microsoft-365-outage-exchange-online-12-13-november-2026) for the timeline.
Spotted an error? Tell us. Read how we report and correct stories in our editorial policy.
Written by
Mainstream Tech
Newsroom
News and analysis on technology, AI, business, politics and international affairs, from our newsroom in London.
More from Mainstream




