A network engineer wearing a headset watches a wall of status monitors beside rows of servers in a data center at night

Opinion

When the Biggest Systems Fail, We Should Ask Harder Questions

Microsoft. The FAA. The largest bank in the country. Apple. All inside one week. I believe the explanations so far. I also know a first explanation is rarely the last one.

Published: November 18, 2026 · Last updated: November 18, 2026

Last Friday afternoon, every departing flight in the United States was held on the ground. It was only the second time that has happened since September 11, 2001. One day earlier, Outlook and Teams had gone dark for much of the working world. Three days later, the largest bank in America could not move money at its normal speed.

I have spent more than 45 years in technology. I have watched systems fail in ways that surprised the people who built them, and a few of those people were me. So I want to say something carefully, because this is the kind of week when careless talk travels fast.

I do not have evidence that any of these failures was an attack. Neither does anyone posting about it on social media. What I have is a lifetime of experience with how outages get explained, and that experience tells me the first statement and the final report are often two different documents. This is an opinion piece. Local Business VoIP is the small-business site of Carolina Digital Phone, Inc., and what follows is my own view.

The short version

Between November 10 and November 16, 2026, Apple, Microsoft, the FAA, JPMorgan Chase, United Airlines, and an internal NSA network all had disruptions. Officials and the companies say these were separate technical failures and not cyberattacks, and nothing public contradicts that. My argument is narrower than the rumors. "Not a cyberattack" on day one should open the conversation, and the public should expect full root-cause reports before it closes. Meanwhile, every business should plan as if the systems it depends on will fail again.

Ground stop
An FAA order holding aircraft at their departure airports.
Root-cause analysis
The engineering investigation, usually written up weeks later, that establishes exactly why a system failed.
Geo-redundant
Running the same service in data centers far enough apart that one event cannot take them all down.
BGP
Border Gateway Protocol, the routing system networks use to find another path when one carrier fails.

What failed, and what were we told?

Six organizations had trouble in seven days, and each one gave a different technical reason. Here is the week as it was reported.

One week of outages, November 10 to 16, 2026
Date What failed Explanation given
Tue, Nov 10 Apple iCloud mail, calendar, contacts, photos, and sign-in, for about two hours Logged and marked resolved on Apple's System Status page
Thu, Nov 12 Microsoft 365, including Outlook and Teams, for hours A problem from an earlier network capacity upgrade, according to Microsoft. BleepingComputer covered it as it happened.
Fri, Nov 13 FAA nationwide ground stop, about four and a half hours Telecommunications instability. The FAA posted its own questions and answers.
Mon, Nov 16 JPMorgan Chase transactions slowed, with delays reaching other banks A software bug in an on-premise hardware system, the Associated Press reported
Mon, Nov 16 United Airlines system outage delaying departures A technology issue. United said it saw no sign of a cybersecurity cause.
Same week An internal National Security Agency network Reported as a disruption under investigation

The FAA event deserves a second look, because it was the most serious. The Christian Science Monitor reported the order went out at 3:54 in the afternoon Eastern time. Transportation Secretary Sean Duffy later said aging switches and routers had been overloaded. FedScoop noted that it came in the middle of a massive overhaul of the FAA's technology.

On Tuesday, Secretary of State Marco Rubio addressed all of it at once. Speaking to reporters in the Philippines, he said the government has "zero evidence" that the incidents are related to one another or to hostile actors, as Yahoo News reported. He added that the Microsoft failure had nothing to do with the FAA's, and the FAA's had nothing to do with the banks'.

Why would anyone question a clear statement like that?

Because "zero evidence" describes what investigators have found so far, and several of these investigations are not finished. That is not a criticism of the Secretary. It is how incident response works.

Look at how the FAA story moved in four days. On Friday evening, the Homeland Security Secretary said there was no indication of a threat to the country, while declining to rule out any cause, a cyberattack included. Within days, the explanation had become overloaded equipment. Since then, SecurityWeek reported that the FBI was looking at a software update to the private network that links air traffic control, with the NSA, the Defense Department, and CISA also involved. Every one of those statements may turn out to be true. They are also three different levels of detail in four days, and the report is not written yet.

Microsoft's story moved too. A commentary in Fortune pointed out that engineers first found no recent change that matched the failure, then rolled one back, and only later named the capacity upgrade as the cause. I have been in that seat. That sequence is what honest troubleshooting looks like at three in the morning. It is also why I do not treat a statement issued during the outage as the last word on it.

I am not a conspiracy theorist. I am a practitioner. I have designed, run, and supported communications systems for decades, and I built redundancy into them for one reason: experienced engineers know things fail. Quick public reassurance has a real purpose. It keeps people from pulling cash out of banks and canceling trips they do not need to cancel. But calming people down and explaining what happened are two separate jobs, and the second one takes longer.

Is it reasonable to think this many failures in one week is a coincidence?

Yes, it is reasonable, and I think it is still the most likely answer. I say that as someone who finds the timing striking.

Most big outages really are ordinary. The last nationwide ground stop, in January 2023, came from a damaged database file and people not following procedure, according to the FAA's own findings. In July 2024, one faulty software update from a security vendor took down millions of Windows computers and grounded airlines around the world. Nobody attacked anyone. Big systems are complicated, they are changed constantly, and the same season of upgrades and year-end changes hits everyone at once.

There is another effect as well. Once the first outage makes the news, every later one gets noticed and counted. An Apple hiccup of two hours would not have made a headline in a quiet week.

So why not leave it there? Because history gives us the other example too. When the Stuxnet worm damaged Iran's uranium centrifuges around 2010, the control screens kept showing normal readings. The operators believed for months that they had bad equipment. That attack was designed to look like ordinary failure, and it worked. And in 2024, CISA warned that state-sponsored hackers had kept quiet access to American critical infrastructure networks for years, positioned to disrupt them later if they chose to.

I hold both of those thoughts at once. Most failures are accidents. Some attacks are built to look like accidents. The way you tell them apart is a finished investigation, and we do not have those yet.

What has changed about the threat?

The speed has changed. On November 16, in the middle of all this, Google's Threat Intelligence Group published a report under the headline "The era of autonomous attacks has arrived." It describes attackers using AI agents that watch for software updates, work out which flaw was just fixed, and build an attack for it with no person at the keyboard. BleepingComputer's summary says Google expects sophisticated, fully autonomous attacks to emerge within about six months.

I have seen nothing from Google connecting that report to last week's outages, and I am not connecting them either. My point is about the environment. Attacks at machine speed used to be a conference topic. Now they are a line in a quarterly report from one of the largest security teams in the world.

The world outside the data center is tense as well. CNN reported on Monday that new U.S. intelligence indicates Iran is developing plans to target senior administration officials in Washington with a dirty bomb or a chemical weapon. Officials are still debating whether Iran can or would attempt it, and the White House has said there is no sign of an imminent attack. China and Russia have deep cyber and AI talent of their own.

Let me be exact about what I am saying. No public evidence ties any of last week's outages to Iran, China, Russia, or anyone else. I mention the tension because it explains why so many Americans felt uneasy this week, and why "trust us" is a harder sell than it was ten years ago. When several critical systems stumble in quick succession during a period like this, people who run networks for a living want to see the homework.

What are the harder questions?

They are the same five questions I would ask my own engineers after an outage of ours. None of them accuses anyone of anything.

  1. Will there be a public root-cause report, and when? A statement is not a report. Good organizations publish one, with dates.
  2. Who checked the answer besides the organization that failed? An outside review is worth more than a press release.
  3. What changed right before it broke? An upgrade, a new vendor, a configuration push. There is almost always something.
  4. What would the evidence look like if this had been deliberate, and did anyone look for it? "We found no evidence" means more when you know what was searched.
  5. What is being done so the same failure cannot happen twice? If the answer is vague, the lesson has not been learned.

As citizens and as customers, we are entitled to ask those questions of an airline regulator, a software company, or a bank. Asking them is not panic. Taking a first statement at face value and never following up is the thing I would call careless.

What should a business owner do about it?

Plan as though the services you rely on will fail again, because they will. Americans now depend on a handful of platforms for work, travel, banking, and daily life. An outage of a few hours is a real cost. An outage of a day can be a real crisis for a small company.

The answer is not panic. It is redundancy, monitoring, and the discipline to take a failure apart afterward so the next one is smaller. For most businesses that comes down to a short list:

What we do at Carolina Digital Phone

We assume things will fail, because we have seen what happens when they do. Our platform runs in three geo-redundant data centers, in downtown Greensboro, Research Triangle Park, and Dallas. We use multiple upstream network paths with full BGP routing, so traffic can find another way around a failed carrier. We encrypt data, and we keep active cybersecurity threat monitoring in place.

We also have bad days. When we do, we post them on our public status page while they are happening. I would be a hypocrite to ask Microsoft and the FAA for transparency and not offer it myself.

Watch: We made a short video on how we build for weeks like this one. Secure Cloud Phone System with Geo-Redundant Data Centers | Local Support Since 2000 (YouTube).

Local accountability matters here. When something goes wrong, you can wait in a distant ticket queue, or you can call people in North Carolina who run the system and will tell you what they know. You can read more about why organizations choose Carolina Digital Phone, our history, and what we mean by a trusted communications partner. Our guides to geo-redundant VoIP service and telecom resilience go deeper on the engineering.

Where does that leave us?

It leaves us without the full story, and honest people should be able to say so. The explanations we have been given may be complete. I hope they are. What we do know is that our dependence on these systems is very high, the tools available to people who wish us harm are getting better, and being prepared is the only response that works whether last week was bad luck or something worse.

Outages will happen. Attacks will be attempted. The organizations that treat reliability and security as daily work, and that can explain what they are doing in plain language, will be the ones still operating when others are not.

A note on what this is. This is opinion. Every factual statement above about the outages comes from the linked reporting as of November 18, 2026, and the investigations are continuing. If the final reports confirm ordinary causes for all of them, I will be glad to say so here.

Frequently asked questions

Were the November 2026 outages at Microsoft, the FAA, and the banks a cyberattack?

According to federal officials and every company involved, no. Each organization gave a separate technical cause, such as a network upgrade, an overloaded router, or a software bug, and the Secretary of State said there is zero evidence the incidents are linked. Some of the investigations, including the FAA's, were still open when that statement was made.

What caused the FAA nationwide ground stop on November 13, 2026?

The FAA said its telecommunications system became unstable and it halted departures as a precaution. The Transportation Secretary pointed to aging switches and routers that became overloaded. The FBI and other agencies joined the review, and officials have said they found no sign of a hack.

What caused the Microsoft 365 outage on November 12, 2026?

Microsoft said the outage came from a problem introduced during an earlier network capacity upgrade that surfaced after a later network change. The company said it was not related to cyber activity.

What caused the JPMorgan Chase transaction delays?

JPMorgan Chase said a software bug in an on-premise hardware system slowed some transactions on November 16, 2026, and that it was not a cyber issue. Other banks saw delays only on transactions that involved Chase.

Why question the official explanation if there is no evidence of an attack?

Because a first statement is made with the information available that day, and root-cause work takes weeks. Asking for the full report, asking who verified it, and asking what changed just before the failure is ordinary engineering practice. It is not an accusation.

Can a cyberattack be made to look like an ordinary equipment failure?

Yes. The Stuxnet attack on Iran's uranium centrifuges, discovered in 2010, damaged equipment while reporting normal readings to the operators, who spent months believing they had a hardware problem. That history is why experienced operators keep an open mind until an investigation is finished.

How should a business prepare for major technology outages?

Assume the services you depend on will fail at some point. Keep your phone system separate from your email platform, make sure calls can reach cell phones if an office goes dark, keep a second way to reach your bank and your customers, and test the plan before you need it.

How does Carolina Digital Phone protect its phone service from outages?

Carolina Digital Phone runs its platform in three geo-redundant data centers, uses multiple upstream network paths with full BGP routing, encrypts data, and monitors for cybersecurity threats. It also posts incidents and maintenance on a public status page.

Want to compare notes on resilience?

If you want to talk through how your own communications are protected, or just compare notes on what resilience looks like in practice, we are here. Contact the experienced team at Carolina Digital Phone for solutions to your phone and messaging problems.

To talk to someone right away, call (336) 544-4000, or click below for a quick call back from a member of our team.

Prefer another route? Request a callback from a local onboarding engineer at HostedBusinessTelephone.com, or compare providers at LocalVoIPProvider.com.

Nicky Smith is the founder of Carolina Digital Phone, a North Carolina provider of cloud phone systems, business messaging, and AI reception services since 2000.