Three blackouts and a submerged infrastructure. The pattern is almost always the same: a local event affects an interconnected system, and what seemed redundant ceases to be so.

Major infrastructure crises do not necessarily begin with a large-scale incident. Sometimes it all starts with a voltage fluctuation, a fire at a substation, vegetation growing too close to a power line, or a damaged cable on the seabed. The real problem arises later, when that local incident interacts with a highly interconnected system and the consequences begin to spread.

April 28, 2025: Spain and Portugal in the dark

At 12:33 p.m. on April 28, 2025, the power grids of peninsular Spain and Portugal suffered a total collapse. A small part of southwestern France experienced brief outages, while the rest of the European grid remained largely stable. Restoration took approximately 12–16 hours. ENTSO-E reconstructed the incident with an international group of experts and, in its final report published on March 20, 2026, concluded that the blackout did not have a single cause but resulted from the interaction of multiple factors. Oscillations, problems with voltage and reactive power control, differing control strategies, rapid reductions in generation, and generator trips led to very rapid voltage increases and a cascade of outages. The lesson for those concerned with resilience is clear. In complex systems, it does not take a single component to “fail and shut everything down.” All it takes is a combination of conditions that, taken one by one, seem manageable, but together exceed the system’s ability to stabilize itself.

Heathrow: Three Power Sources, One Dependency

A few weeks earlier, on the evening of March 20, 2025, at 11:21 p.m., one of the three transformers at the North Hyde substation (275 kV, in west London) failed and caught fire. The other two went offline shortly afterward, and the substation lost all power. The fire was not inside the airport and did not affect any runways or terminals. Nevertheless, Heathrow suspended operations for most of March 21. This case is instructive for three reasons.

• The redundancy was only apparent. Heathrow considered itself protected: it drew power from three substations, each with multiple transformers and multiple connections. The Kelly Review, commissioned by the airport itself and published on May 28, 2025, found that the loss of an entire power supply was considered a low-probability event. According to the review, nearly all emergency power sources connected to the North Hyde power supply would have been exhausted before the grid was restored.
• The defect was known. The final report by NESO (National Energy System Operator), which the British government had commissioned to investigate, points to moisture entering a high-voltage bushing as the likely cause. The issue had been detected in 2018 but had not been resolved. The substation’s fire suppression system had been out of service since at least 2022.
• No one could see the addiction. According to NESO, neither the transmission system operator nor the distribution system operator was aware of the extent to which the loss of a power supply point could affect the airport’s critical systems.

It’s the difference between redundancy and independence, as applied to deferred maintenance. Three power supplies on paper; in reality, a weak point that no one had linked to the service it supported.

A multinational blackout caused by vegetation

On June 21, 2024, a major power grid incident struck southeastern Europe: blackouts in Albania, Bosnia and Herzegovina, and Montenegro, and a partial blackout in Croatia. The final ENTSO-E report, published in February 2025, identifies two short circuits caused by vegetation growing too close to the power lines as the initial causes. The root of the problem, the report states, is that the vegetation had not been pruned within operational limits: the risk of a short circuit had been underestimated. Subsequent outages involved several operators. When the voltage began to drop, the available measures were insufficient to prevent the collapse. The report recommends that operators review their pruning procedures and expand monitoring to include neighboring grids, in order to obtain a more complete picture in real time. It is difficult to find a more telling example of how a trivial cause can lead to a systemic incident. Not a sophisticated cyberattack, not an earthquake, not sabotage: an underestimated routine maintenance risk in an interconnected grid. In complex systems, the severity of the consequences is not proportional to the spectacular nature of the initial event.

Beneath the Baltic Sea: A Different Case

The Baltic Sea is not a case of a cascading failure, but rather of infrastructure that no one sees. According to figures reported by the AP in January 2025, at least eleven undersea cables in the region—including communications and power cables—had been damaged since October 2023. In January 2025, NATO launched the Baltic Sentry mission to monitor cables and pipelines. The causes are not always clear. In the waters of Northern Europe, damage typically stems from fishing gear and anchors, and in several cases, investigations have focused on vessels suspected of dragging their anchors; the suspicion of sabotage remains under debate. Not all of these incidents have resulted in serious disruptions, precisely because the networks have a certain degree of redundancy. However, a significant portion of Europe’s communications and energy connections passes through physical infrastructure on the seabed, which is difficult to monitor. Fiber-optic cables in the Baltic Sea have typically been repaired within weeks. For power cables and pipelines, it can take months.

The pattern is always the same

These cases are very different from one another. In the Iberian blackout, there was a complex interplay between operating conditions and grid dynamics. At Heathrow, a fire at an outdoor substation—caused by a known, unresolved defect—brought an international airport to a standstill. In the Balkans, two short circuits caused by vegetation contributed to a multinational blackout. In the Baltic region, the vulnerability of submerged infrastructure—on which communications and energy rely—became apparent. The underlying mechanism is similar: a local event affects an interconnected system, and the consequences ripple out. This is the principle of “cascading failure,” and it is one of the most difficult risks to manage, because it calls into question the logic of plans developed for isolated scenarios. Many organizations have separate plans for the loss of their building, the unavailability of IT systems, a power outage, or the loss of a supplier. These are useful scenarios, but a significant infrastructure crisis rarely stops there. A blackout can affect telecommunications, transportation, payments, and people’s mobility. A telecommunications problem can render the cloud unusable. A transportation crisis can compromise the fuel supply for generators. The risk is not only that a resource will become unavailable, but that the planned solution itself depends on another piece of infrastructure affected by the same event.

The True Story of Horror

Modern infrastructure is designed with redundancies, safeguards, and emergency procedures, which is one reason why events of this magnitude remain relatively rare. But it is precisely real-world incidents that reveal the limitations of an approach based solely on the probability of a single failure. The real horror story isn’t that something might go wrong—we know it will happen sooner or later. It’s discovering, in the midst of the incident, that systems considered independent were actually linked by the same invisible dependency. When that dependency breaks, we don’t just lose a piece of infrastructure. At the same time, we also lose some of the alternatives upon which we had built our resilience.

Questions for Resilience Managers

• Which of our “redundant” power supplies, lines, or sites share the same single point of failure, including the maintenance performed by their operators?
• Are we aware of the known, unresolved issues with the infrastructure we depend on, and do we know who is managing them?
• How long does our emergency battery life last compared to a realistic recharge time, in hours or days?

Sources

• NESO, North Hyde Review – Final Report
• Kelly Review
• ENTSO-E, Final Report on the Iberian Blackout
• ENTSO-E, Final Report on the Incident in Southeastern Europe
• AP/Defense News, Baltic cables