The week from November 9 to 16 became one of the most tense for America's digital infrastructure in the past year. Over seven days, the Downdetector platform recorded nearly 6 million outage reports in the US — the highest weekly figure in 12 months. At the same time, users temporarily lost access to bank accounts, passenger planes nationwide were grounded, and email and cloud services experienced disruptions. At first glance, the picture resembles a mass cyberattack scenario; however, as experts emphasize, this is currently a coincidence of circumstances rather than a coordinated cyber apocalypse.
What went down: banks, aviation, and cloud services
The list of incidents for the week looks impressive. On Monday, November 16, some US customers lost access to JPMorgan Chase banking services: the company attributed the issue to a software bug that delayed transactions involving other banks. The US Federal Aviation Administration (FAA) suspended departures nationwide for several hours — according to the agency, a surge of network traffic destabilized legacy router equipment. Meanwhile, Microsoft spent over a day resolving a large-scale outage in its systems, explaining it as a bug that arose several months ago, compounded by traffic growth and reduced server capacity.
AI as a trigger: the race against 'bugageddon'
According to US national security agencies and Wall Street specialists, the unifying factor behind all these disruptions was a connection to artificial intelligence. Large organizations have entered a total race to avoid the so-called 'bugageddon' — a scenario where hackers use AI to find and exploit vulnerabilities on the internet faster than they can be fixed. Mass attempts to quickly patch errors discovered by AI systems became the root cause of the instability. Homeland Security Secretary Markwayne Mullin suggested that AI systems capable of writing and modifying code may have contributed to the outages, although no direct evidence for this version has been found.
Statistics vs. panic: are these anomalies?
Looking at a broader sample, it is too early to panic. According to Downdetector, in the 30 days leading up to November 16, there were only three major outages in the US compared to the usual five, meaning no statistical anomalies are observed on a monthly horizon. The record weekly surge in problem reports has thus turned out to be a local peak rather than a systemic collapse. In most cases, as the companies themselves note, the outages are explained by routine software updates that turned minor technical glitches into large-scale issues.
What this means for users and business
Experts warn that AI has significantly accelerated the process of writing and improving code, creating new bottlenecks during security and reliability checks. Accelerated development leads to an increase in downtime, glitches, and unexpected failures — precisely the effect that manifested last week. For businesses, the conclusion is clear: the race to patch vulnerabilities discovered by AI at high speed requires not only rapid patches, but also enhanced testing so that the software update process itself does not become a source of new incidents.