Field report SE-82 / L6 · sheet /operations
Operations & RestructuringFiled from the territory
Four US disruptions in six days expose software patching strain
Microsoft, an FAA grounding, bank-access failures and an NSA outage in six days exposed patch-management strain, while official findings rejected a systemic attack.
- Grid
- SE-82 / L6
- Surveyed
- November 19, 2026
- Reading
- 5 min read
- Correspondent
- Sophie Lindqvist

Field notes
- Four major U.S. disruptions occurred within six days.
- Microsoft 365's outage lasted nearly 24 hours, while the FAA ground stop exceeded four hours.
- More than 9,100 flights were canceled after the nationwide FAA grounding.
- Published CVEs passed 100,000 on Nov. 13, 2026.
- Median exploitation time fell from 120 days in 2025 to 80 in the first half of 2026.
By Sophie Lindqvist
Four major U.S. disruptions in six days point to operational failures, not a confirmed systemic cyberattack. The sequence exposed Microsoft 365, nationwide aviation, banking access and a classified government network.
Microsoft 365 suffered an outage lasting nearly 24 hours. The Federal Aviation Administration (FAA) grounded every flight nationwide; FlightAware data cited by Fox Weather showed more than 9,100 cancellations that Friday. Chase and Wells Fargo customers could not access their money, and the National Security Agency’s (NSA’s) classified internal network went dark.
On Tuesday, prominent cybersecurity specialist Nicole Perlroth said she was answering hundreds of texts and direct messages asking whether the world was ending. In a long post on X, she wrote: “We are entering a period of unprecedented technological instability.” She called this a “vulnpocalypse.”
Perlroth said the reports she heard did not contradict the official line that most breakage came from “screwy updates.” Her warning concerned fragility in routine software operations, not a proven coordinated offensive.
What broke, and what explanations followed?
Microsoft said a bug introduced in a July update to its networking gateway servers caused the nearly 24-hour Microsoft 365 outage. The FAA ground stop lasted more than four hours, the longest nationwide halt since the Sept. 11, 2001 shutdown.
- Transportation Secretary Sean Duffy blamed “routine maintenance” by a contractor replacing copper wiring with fiber.
- Chase cited a “technical issue.”
- Apple said a faulty internal update knocked out iCloud and other customer services on Tuesday.
- The NSA called its network failure a “temporary degradation” during routine maintenance.
No company or agency described any incident as an attack. Most disruptions occurred while someone was changing or fixing something.
Was the United States under systemic attack?
On Wednesday in Manila, Secretary of State Marco Rubio rejected that interpretation: “We are not under systemic attack,” and said there was “zero evidence whatsoever” that the disruptions were connected. He added: “Every indicator we have is that every single one of these events has an explanation or a cause.”
Rubio described the mechanism: when companies patch vulnerabilities, they can cause a “short-term disruption in the way your system works.” President Donald Trump offered a less technical assessment: “It’s like when you’re big and powerful, and you have all the money. Things like that happen. It’s fine.”
Does AI explain the vulnerability surge?
The count of published vulnerabilities, recorded as Common Vulnerabilities and Exposures (CVEs), passed 100,000 on Nov. 13, 2026, according to VulnCheck. That was more than double the full 2025 total.
A recount of CVE Project records found 89,849 disclosures through October, against 39,638 in the same months of 2025. October alone produced 17,049, nearly four times October 2025. IBM’s X-Force team counted a 578% jump in disclosures across Microsoft, Oracle, Chrome and Apache through September.
“There’s no denying the fact that vulnerability hunting with AI is driving a majority of what we’re seeing,” VulnCheck’s Josh Bressers told The Register. The data do not prove that claim.
CVE records do not identify whether AI found each flaw, so they cannot isolate AI’s contribution from the total. IBM said its data “does not establish a direct causal relationship.”
A recount found little growth in flaws designated for urgent patching, either because they appeared on the Cybersecurity and Infrastructure Security Agency’s exploited list or had high exploitation odds. Its author warned: “Twice as many CVEs does not mean twice the danger.”
Perlroth’s claim that high and critical vulnerabilities had tripled also remains unverified by those datasets. Those datasets do not separately report that category.
How short is the patching window?
Perlroth wrote that mean time to exploit, or the interval before attackers use a flaw, had fallen to “negative seven days.” That indicates attackers used some flaws before a fix existed. Public measurements are less dramatic.
VulnCheck found the median interval from CVE publication to exploitation dropped from 120 days in 2025 to 80 in the first half of 2026. CrowdStrike found 88% of exploitation using public proof-of-concept code (a demonstration of how to exploit a flaw) occurred within 48 hours of release.
More flaws require more patches. Faster exploitation leaves less time to test them, and customers can detect failures before vendors do. Rubio put the banking risk plainly: if users cannot enter an app, “they’re going to notice that very, very quickly.”
What has this pipeline cost before?
A faulty CrowdStrike update crashed more than 8.5 million Windows machines in July 2024, without hacker involvement. Insurer Parametrix estimated $5.4 billion in direct losses for Fortune 500 companies, with cyber insurance covering only 10% to 20%.
Yale sociologist Charles Perrow named this class of failure in 1984 as “normal accidents”: failures built into complex, tightly coupled systems that no one fully understands. Perlroth said “almost nobody understands end to end” the software systems companies now run.
At Fortune’s Innovation Forum in Detroit on Monday, Honeywell Technologies chief executive Vimal Kapur said 50 years of added software and infrastructure had made industrial systems too complex for humans to keep up with. “What happened in the last two weeks is an example of complexity,” he said.
What does this change for technology risk teams?
For technology buyers and risk teams, the question changes from whether an attack caused each outage to whether routine software change remains reliable across complex dependencies. More vulnerabilities and shorter testing windows increase exposure to vendor outages even when no attacker is present.
Perlroth called last week the “opening chapter”: “This, I’m sorry to say, is just the opening chapter.” The next test already sits on the calendar: Microsoft’s next Patch Tuesday security-update cycle on Dec. 8, 2026.
via fortune.com/feed/ (Source)
More from Sophie Lindqvist
Show full bio
Correspondent covering money, funding and market shifts around corporate strategy and chief-executive leadership at CEO Frontier.
469 articles
Adjacent sheets
- One Router, 18,000 Cancelled Flights: America's Aging Tech Fails
- Microsoft Restores Global Services After Multi-Hour Outage Hit 74 Countries
- U.S. Flight and Cloud Outages Demand Answers, Not Unsupported Blame
- DHS Chief Points to 'Super Intelligence' Behind Five Days of U.S. Outages
- Three U.S. Outages in Five Days Raise Unanswered DDoS Questions