The argument has merit, but the headline overstates it. The data behind it points to wider impact when something fails, not clearly to more frequent failures. Here's what the evidence supports and what Windows and Microsoft 365 admins should take from it.
The Microsoft 365 outage: three faults that lined up
Microsoft's preliminary post-incident report is dated November 15, 2026 and filed under incident ID EX1550543/MO1550821. It gives the start as Thursday, November 12, 2026, at 4:20 PM UTC and the end as Friday, November 13, 2026, at 4:10 PM UTC. That's just under 24 hours. Microsoft says the incident was caused by three main factors that together caused a global outage for the networking gateway that fronts many Microsoft 365 services.
Factor one: a dormant defect. According to the report, a latent defect caused servers to become overwhelmed and unresponsive under certain load conditions. Data Studios' summary of the report says the problem dates to a software change made in July 2026 to the network gateway servers that accept and route connections for Exchange Online. It describes the mechanism as lock contention in a Windows process on gateway servers. The change sat in production for more than three months without causing problems before a combination of extra traffic and reduced capacity exposed it.
That Windows detail matters to this audience, so here is Microsoft's exact scope: This defect was the result of a fix that was prepared specifically for Microsoft 365 networking gateway servers; neither the underlying defect nor the fix applies to broader Windows deployments. Your own Windows Server fleet and Windows 11 desktops did not receive this change. Nothing needs patching on your side.
Factor two: a failover that added load. Microsoft says it runs two independent and heterogenous network gateway systems designed for isolation and high availability. Then, approximately fifteen hours before the outage, Teams failed over from the alternative networking gateway to the 'primary' gateway that also supports Exchange Online.
Factor three: less capacity than usual. WindowsForum's earlier coverage of the report noted that nearly 14% of gateway capacity was unavailable because of a separate problem.
None of these would have taken the service down alone. Together they did. The defect then spread across Microsoft's gateway fleet in less than two hours and that, at the worst point, more than 80% of the servers were unresponsive.
When failover makes things worse
The detail Around Prague highlights is the most useful lesson here. Automated traffic management moved users off failing servers and onto healthy ones, as designed. That pushed the same workload onto fewer machines, and those machines hit the same bug. Imagine the fire evacuation route running straight through the kitchen that's on fire.
The damage wasn't uniform, which helps admins understand what their users saw. Exchange Online took the main hit: mailbox access and delayed mail delivery. Other services held up better because they could move off the faulty gateway. WindowsForum noted that Microsoft says Teams chats and meetings worked without issue for most of the outage because Microsoft 365 runs two independent gateway systems and many services can fail over between them. Around Prague also mentions that some Copilot features were affected. Copilot is on the report's list of impacted services.
Section summary:
- The window was November 12, 16:20 UTC, to November 13, 16:10 UTC. The incident IDs are EX1550543 and MO1550821.
- The cause was a July 2026 gateway defect, exposed by Teams failover traffic and missing capacity.
- Automated rerouting helped spread the failure to more than 80% of gateway servers.
- The defective fix was specific to Microsoft 365 gateway servers, not general Windows.
Read the preliminary report with care
This is still a preliminary account. Coverage said Microsoft says its final post-incident report is due on November 20, 2026. Admins should open the Microsoft 365 admin center, look up the incident IDs and compare the final report with the preliminary one. Microsoft's corporate vice president Charles Lamanna posted an abridged version and wrote: "We are deeply sorry for the impact to all our customers and recognize the impact that this has on their normal business & operations." An apology helps. A final list of fixes would help more.
JPMorgan Chase: you can be affected without being down
Around Prague's second example is a software error at JPMorgan Chase that delayed payments, including for customers of other banks. Those banks were working normally. The only requirement was that the other side of the transaction held a Chase account.
The point is narrow but important. Your organization can be hit through a counterparty's systems while your own are perfectly healthy. A green status page doesn't help if the business at the other end of the transaction is having a bad day.
The two incidents do not appear to be connected. Around Prague itself reported that US Secretary of State Marco Rubio said the recent outages at Microsoft, in US aviation and in banking were unrelated, each with its own cause and no signs of hacking. Microsoft's report likewise blames internal software and configuration, not attackers.
Do the numbers say outages are becoming more common?
This is where the headline and the data part ways. Around Prague cites Uptime Institute's May report, and that report cuts against the "more frequent" framing. In its Annual outage analysis 2026, published May 11, Uptime found that outage frequency per site fell for the fifth year running. The pace of improvement has slowed, though, and around one in ten respondents still said their most recent outage had serious or severe impacts.
Other findings in the same report back the broader concern:
- External infrastructure failures, including fiber and connectivity problems, are becoming more prominent in publicly reported outages. They are also more likely to cause extended disruption.
- Costs keep rising. 57% of respondents to Uptime's 2025 survey said their most recent major outage cost more than $100,000, and one in five reported an impactful outage costing over $1 million.
- Resiliency assessments still focus more on internal systems than on external and systemic risk.
A fair reading: individual data centers are getting more reliable, but when a shared platform fails, more organizations fail with it. Around Prague's illustration makes the point. Three online shops use three different AI tools from three different vendors, but if all three run on the same cloud service, one fault takes down all three shops. That's a hypothetical, not a documented case, but anyone who has mapped SaaS dependencies has seen something like it.
AI, automation and keeping humans able to step in
Around Prague also argues that AI can help detect faults and shorten outages. But an AI system may lose the data it needs mid-incident, and if it can change settings on its own, a bad decision can widen the failure. Microsoft's outage shows this without any AI involved: ordinary automated rerouting was enough to start the chain.
The article points to the NIST AI Risk Management Framework, which NIST describes as voluntary. Its core is organized into four functions: Govern, Map, Measure and Manage. The framework covers human intervention when an AI system can't detect or correct its own errors. It doesn't guarantee that a person can restore service during an outage. That depends on whether staff can still reach the underlying data.
Around Prague also reports that after the Microsoft outage, the Australian Computer Society renewed its call for a national review of which critical systems depend on a single supplier and where real alternatives exist.
What Windows and Microsoft 365 admins can do
These are practical suggestions based on the incidents, not requirements from Microsoft or NIST:
- Map your dependencies, including indirect ones. List which business processes depend on Exchange Online, Teams, SharePoint and Copilot, and which third-party SaaS tools run on the same hyperscaler.
- Check whether your failover is truly separate. Microsoft's own isolation between gateway partitions didn't stop the cascade. Ask whether your backup path shares hardware, identity, DNS or a gateway with the primary.
- Plan for email outages specifically. Exchange suffered most because it couldn't fail over at scale. Decide in advance how urgent messages will go out when Outlook is down, whether by Teams chat, SMS or phone trees, and test it.
- Keep humans able to reach the data. If an AI tool sorts tickets or complaints, make sure staff can open the original messages through a different route.
- Be careful with automated fixes during incidents. Self-healing scripts and auto-scaling are great until they pile load onto the last healthy servers.
- Track the final report. Compare the November 20 final report with the preliminary findings, and adjust your playbooks if the causes or fixes change.
Bottom line
Around Prague is right that dependence on shared infrastructure turns local faults into widespread outages. "More frequent" is the wrong word, though, and its own source, Uptime Institute, shows per-site outages still falling. The better word is "bigger." The Microsoft 365 outage showed a safety feature spreading a dormant bug across a global fleet. The Chase incident showed that a counterparty's failure can become yours. Neither is a reason to leave the cloud. Both are good reasons to find out where your organization has a single point of failure.
References
- IT outages: why major system failures will become more common - Around Prague Around Prague · 2026-11-21T03:46:15+00:00
- Microsoft 365 outage cause: July gateway server bug, 80% of servers down datastudios.org
- Preliminary Post Incident Report For Microsoft 365 Report Date: November 15, 365fanatic.com