On Thursday, JPMorgan Chase disclosed that third-party service providers were hit in a cyberattack that may have exposed customer names, email addresses, and some phone numbers. The incident was unrelated to its earlier banking outage, however.

Chase was just one of the string of outages. Multiple events caused the wave of outages: simultaneous software updates, emergency patching, and aging infrastructure failures – not a coordinated cyberattack or rogue artificial intelligence (AI), what the federal government now refers to as super intelligence (SI).

U.S. national security officials and cybersecurity experts confirmed that while the timing made the disruptions appear connected, these were actually separate technical incidents driven by organizations rushing to modernize and secure brittle, legacy systems.

What began with Microsoft 365, Outlook, Teams, and Copilot was triggered by an issue during a “network capacity upgrade.” The nationwide aviation ground stop initiated by the Federal Aviation Administration (FAA) was caused by a failed router and routine telecommunications maintenance on aging infrastructure. The outage at JPMorgan Chase was traced to a software defect in its on-premises equipment.

Other outages, including Apple’s iCloud and Apple Pay and Marriott’s website and app sign-in, were also due to internal updates. Even the disruption to the National Security Agency’s NSA Net resulted from routine maintenance.

“I don’t see evidence these outages are connected, and I wouldn’t mistake a bad week for proof that infrastructure is becoming less reliable. Uptime Institute’s research actually shows outage frequency per site declining for five consecutive years. What concerns me is how much disruption a single failure can cause when critical systems depend on shared infrastructure,” explained Ryan McCurdy, vice president of secure database provider Liquibase.

McCurdy told ClearanceJobs that in Microsoft’s case, a software defect sat dormant for months, then caused a global outage when traffic shifted, and server capacity was already constrained.

“Workday experienced a different problem when services in its disaster-recovery environment unexpectedly connected to production databases, leaving data in an inconsistent state. Neither incident required a cyberattack, yet both disrupted critical operations,” said McCurdy.

Systems at a Breaking Point

The recent string of high-profile outages across tech, aviation, banking, and government networks drew intense scrutiny, even though they appear unrelated. Still, the disruptions have put a spotlight on how fragile today’s interconnected digital infrastructure is.

“We’ve come to rely on extraordinarily complex, interconnected systems, and many of these are built upon decades of accumulated technical debt,” warned Donald McFarlane, advisory board member at cybersecurity provider Xcape, Inc.

“Our collective aging infrastructure, shared dependencies, and the loss of institutional knowledge all contribute to operational fragility,” McFarlane told ClearanceJobs. “What worries me is that we may be seeing pressure because several factors are converging: accumulated technical debt, accelerating work to remediate vulnerabilities, increasing cybersecurity threats, all coupled with the reality that in many of our critical infrastructure environments, even relatively modest failures can cascade into disruptions affecting millions of people and essential services.”

Outages are actually common. This week, for example, involved many high-profile companies. Moreover, past outages were often behind the scenes, but today, when a server goes down, you don’t just lose access to a website. Instead, you can’t board a flight or access your bank account.

The problem is that companies are constantly racing to patch security flaws before hackers can exploit them. Deploying urgent, complex code updates under tight deadlines often introduces accidental bugs that crash systems.

“What we’re seeing now is really fragility under change. Each of these failures happened while someone was fixing or upgrading something,” explained Denis Calderone, principal/CTO at cybersecurity provider Suzu Labs.

“We are seeing an uptick in outages and will continue to see more,” added Beau Bullock, director of emerging threats and advanced testing at Black Hills Information Security.

Bullock warned that many services can stop working at once because they depend on the same provider for hosting their services and systems. He told ClearanceJobs that even one underlying failure at a major cloud provider can feel like half the internet is down.

“The more we rely on the same cloud providers, the larger the potential blast radius becomes,” Bullock continued. “You can use several different vendors and still have them depend on the same underlying infrastructure. That means a failure completely outside your organization can bring your business to a halt.”

AI/SI Will Complicate Matters

From the beginning of the outages, questions arose about what role AI/SI played, but the answer isn’t straightforward.

“The velocity of the fix has changed, driven by the speed of AI vulnerability discoveries. The care and trepidation that up until recently were central to the fix cycle have shifted,” Calderone told ClearanceJobs. “Fixes have become more urgent, so there is less waiting. We’ve all understood that legacy interconnection was a problem for the last two decades, and now, addressing legacy fixes has taken on a whole new priority. The fix wave is emerging as the new outage vector.”

AI/SI could play a role in stopping the outages by discovering vulnerabilities earlier.

“The reported causes point mainly to operational failures, but I think cybersecurity pressure is part of the bigger picture,” added Bullock. “AI is helping uncover vulnerabilities in older systems/software, and the urgency to fix those problems can create operational risk even when nobody has been hacked.”

At the same time, AI/SI could increase the frequency of outages.

“This becomes even more challenging as AI accelerates software development and deployment,” suggested McCurdy. “Organizations need to match that speed with stronger validation, enforceable policies, and reliable evidence of what changed. For production databases, those controls need to apply whether a change comes from a developer, an automated pipeline, or an AI agent.”

Related News

Peter Suciu is a freelance writer who covers business technology and cyber security. He currently lives in Michigan and can be reached at petersuciu@gmail.com. You can follow him on Twitter: @PeterSuciu.