Zero Evidence, Maximum Exposure: Is America’s Week Of Outages A Concentration Risk Time Bomb?

Blog
Corporate Risk Leaders
19 Nov, 2026

Between November 12 and 17, the US suffered a run of technology failures that would test anyone’s faith in coincidence. Microsoft 365 was disrupted for nearly 24 hours, router problems at the Federal Aviation Administration grounded flights nationwide, and Workday – the HR and payroll platform used by about 11,500 organizations – saw a widespread outage. Within days, Bank of America, Chase and Wells Fargo customers had reported online banking problems, and even the National Security Agency (NSA) logged disruptions.

Senior US officials, including Secretary of State Marco Rubio, insist there is no evidence that the incidents are linked or the work of a malicious actor. Their explanations range from a software defect and faulty routers to a bank’s software release and routine maintenance.

The lesson for risk leaders worldwide is that a link does not need to be malicious to be material. When critical services share the same cloud platforms, software components and patch cycles, they can fail together whatever the trigger. Risk registers that log each vendor outage as an isolated incident will consistently underestimate this correlated exposure.

Shared dependencies are multiplying

Two trends are amplifying the problem. The first is AI-driven vulnerability discovery. In October 2026, two critical AI-discovered flaws in the widely used PHP programming language forced governments and firms to take systems offline while they patched – US intelligence officials have linked the NSA disruption to patching a PHP vulnerability. As AI tools uncover flaws in thousands of systems at once, organizations are pushed into simultaneous patching that strains systems never designed for that pace.

The second is infrastructure concentration. Amazon, Google and Microsoft together held 63% of the global cloud infrastructure market in the second quarter of 2026, and a handful of cloud software platforms such as Workday run core HR and payroll processes for thousands of organizations. A defect at one provider can cascade through its customers and their suppliers within hours.

The economy has the same single point of failure

Concentration extends well beyond IT. US GDP grew at an annualized 3.4% in the third quarter of 2026, with investment in computers and related equipment up 37.4% on the back of AI spending. Alphabet, Amazon, Meta and Microsoft planned capital expenditure of around $650 billion for 2026. As of early November 2026, the 10 largest US stocks accounted for a record 39% of S&P 500 market capitalization, 12 percentage points above the peak of the dot-com bubble. These estimates will change following NVIDIA’s latest quarterly revenues, which doubled compared with Q3 2025, beating market expectations.

The firms that run America’s digital infrastructure are the same ones underpinning its economic growth and equity returns. Verdantix research highlights how capital and dependency are clustering around a small number of chipmakers, hyperscalers and frontier model labs (see Verdantix Market Insight: AI Bubble Risk And Capital Cycles). An operational shock at one hyperscaler could quickly become a financial one.

Concentration risk needs to move up the board agenda

Whether or not the outages were linked, risk and third-party risk management (TPRM) leaders should:

  • Map dependencies beyond the first tier.
    TPRM programmes should identify which critical services rely on the same cloud provider or software component, including fourth and nth parties. The EU’s Digital Operational Resilience Act is already pushing buyers towards this multi-tier visibility (see Verdantix Strategic Focus: Innovations In Third-Party Risk Management).
  • Stress-test correlated failures.
    Scenario plans should assume several critical providers fail in the same week, as in the US in November 2026. Operational resilience software can help model these scenarios (see Verdantix Smart Innovators: Operational Resilience Software).
  • Treat patching as a resilience risk.
    AI-driven vulnerability discovery will compress patch windows, so firms need staged rollouts and manual workarounds for critical processes.
  • Quantify exposure to a tech correction.
    Boards should ask how a sharp fall in a handful of AI-linked stocks would hit their investment portfolios and pension schemes.

 

Marco Rubio may be right that nobody orchestrated the November outages. That conclusion should worry risk leaders. An economy that can stumble this badly without an adversary has a structural concentration problem, and no intelligence briefing will fix it. To read more on interdependencies, resilience and TPRM, check out Verdantix Insights. 

Discover more Corporate Risk Leaders content
See More