The 2025 Iberian Blackout: What the Final Report Actually Found

The Iberian blackout is the most consequential electricity event in Europe this decade, and almost everything said about it in the first week was wrong. It was immediately framed as proof that renewables destabilise grids. The eighteen-month investigation that followed produced a more precise and more useful answer, one that changes how system operators, regulators and industrial energy buyers should think about a grid dominated by inverter-based generation. This analysis is part of the Portugal Company Stories hub.

What actually happened on 28 April 2025?

Shortly after midday the Iberian system was running with an unusually high share of inverter-based generation: renewables accounted for around 78% of electricity generation, with solar alone near 60%, while conventional synchronous plant made up roughly 15% of the mix. That configuration is not unusual in Iberia on a sunny spring day.

Voltage began behaving abnormally, oscillations developed, and a sequence of generation losses occurred in southwestern Spain. Protective systems disconnected plants to shield equipment from overvoltage, each disconnection worsening conditions for the remaining units. Spain lost roughly 15 GW of generation — about 60% of its supply — within seconds.

At 12:33 CEST the system lost synchronism and separated from the continental European grid. Portugal, which imports and exports across the Spanish interconnection and shares the same synchronous area, went down with it. Recovery took most of a day; REN restored 85 of Portugal’s 89 substations by late the same evening.

What did the final investigation conclude?

ENTSO-E published the expert panel’s final report on 20 March 2026, following a factual report in October 2025. The panel, comprising 49 members, concluded that the blackout resulted from many interacting factors: oscillations, gaps in voltage and reactive power control, differences in voltage regulation practices between actors, rapid output reductions and generator disconnections in Spain, and uneven stabilisation capabilities across the system.

The mechanism was overvoltage-induced cascading disconnection — a failure mode the existing defence architecture was not designed to interrupt. Critically, the panel’s simulations showed that even with higher system inertia, loss of synchronism would not have been avoided, because the cascade of generator trips reduced synchronising torque too fast for additional rotating mass to help.

That finding matters because it contradicts the most common public explanation. The intuitive story — not enough spinning machines, therefore not enough inertia, therefore collapse — was tested and rejected. The failure was in voltage and reactive power management, not in frequency stability.

Why does voltage control matter more than most people realise?

Because electricity systems require both active power, which does the useful work, and reactive power, which maintains voltage within operating limits. Reactive power cannot be transported far, so it must be produced locally by generators, capacitor banks, synchronous condensers or inverters configured to supply it.

The panel identified that renewable generation was operating in fixed power factor mode under grid code implementation, meaning inverters were not dynamically supporting voltage, and that reactive power reserves were insufficient with critical assets switched manually. Market design contributed too: schedule changes drove fast power ramps that stressed the system.

In other words, the equipment was largely capable of helping and was not configured or instructed to do so. That is a rules and coordination failure rather than a technology failure, which is why the remedies are largely regulatory.

What did the blackout cost, and who bore it?

The economic loss ran into billions across both countries: halted industrial production, spoiled refrigerated stock, disrupted transport and telecommunications, and cancelled commercial activity for the better part of a working day. Eight deaths were recorded across Spain and Portugal in circumstances connected to the outage, with more than twenty-five non-fatal injuries.

Losses fell largely on businesses and households rather than on the electricity sector, because standard supply contracts do not compensate for outages caused by system-wide events. This asymmetry is a structural feature of electricity markets: the party that suffers the loss is rarely the party that could have prevented it.

For Portuguese businesses specifically, the event exposed how completely the country’s system is integrated with Spain’s. Portugal did nothing wrong and lost power anyway. That dependence is the price of the interconnection that otherwise lowers Portuguese electricity costs.

What changes as a result?

The panel’s recommendations target strengthened operational practices, improved monitoring of system behaviour, closer coordination and data exchange among power system actors, and adaptation of regulatory frameworks to the evolving nature of the power system. It also delivered a pointed message to national regulators about enforcement of grid code requirements, noting protection settings that had diverged from what the rules required.

Practically, this means investment. Synchronous condensers, grid-forming inverters, dynamic reactive power compensation, wider-area monitoring and stricter compliance testing all cost money and all sit largely on the regulated network side — which is why the event strengthens the investment case for transmission operators such as REN.

It also means new obligations for generators. Renewable plants will increasingly be required to provide voltage support, ride through disturbances and operate in modes that slightly reduce energy revenue in exchange for system stability. That is a real cost to developers, including EDP Renováveis.

Does this argue against a renewables-heavy grid?

No, and the report is unusually direct about it. The ENTSO-E board chair stated that the problem is voltage control regardless of the type of generation. A system with 78% renewable share is operable; it requires different control architecture from a system built around large synchronous machines.

The genuine finding is that Europe’s grid codes, protection philosophy and market design were built for a system that no longer exists, and the transition outpaced the rulebook. Growing shares of inverter-based generation, declining synchronous capacity, deeper market integration and rising electrification place the system under conditions its defence architecture was not designed for.

The policy implication is sequencing. Renewable capacity has been added faster than the control and coordination capability needed to operate it, because capacity is subsidised and visible while voltage control is neither. Correcting that imbalance is the substantive work of the next decade.

What should companies actually do about grid risk?

Treat power continuity as an operational risk with a probability, not as a given. That means knowing your site’s exposure, sizing backup generation or storage against the loads that genuinely cannot stop, and testing restart procedures — the Iberian event revealed how many facilities had backup systems that had never been exercised under a real ten-hour outage.

It also means reading interruption clauses in supply contracts before an event rather than after. Compensation for system-wide outages is generally unavailable, so the mitigation has to be physical and on-site.

For companies making location decisions in Iberia — particularly data centres, which are driving the demand growth described in the Portugal hub — grid node strength and local reactive support have moved from a technical footnote to a first-order site selection criterion.

Why was Portugal dragged down by a Spanish event?

Because the Iberian Peninsula operates as a single synchronous system with limited connection to the rest of continental Europe. Portugal and Spain share a common electricity market, MIBEL, and are physically interconnected at multiple points, while the peninsula as a whole connects to France through interconnection capacity that has long been considered inadequate.

That weak French link is the structural vulnerability. In a well-connected system, a large generation loss is absorbed by neighbouring areas contributing frequency and voltage support. The Iberian peninsula is effectively an electrical island attached to Europe by a narrow bridge, so a severe disturbance has nowhere to dissipate.

The policy conclusion is that increasing France–Spain interconnection capacity is not merely a market efficiency measure that lowers prices; it is a resilience measure. Both arguments have been made in Brussels for two decades, and the blackout gave the second one considerably more weight.

How did the market and regulatory response unfold?

In stages, and with visible tension between speed and rigour. The Spanish government and grid operator published their reports in June 2025, with the ecological transition minister attributing the event to overvoltage, insufficient voltage control capacity, a programming flaw, and plants disconnecting inappropriately to protect their own installations.

The ENTSO-E expert panel took considerably longer, publishing a factual report in October 2025 and the final report with recommendations in March 2026. The delay drew criticism, but the additional time produced the simulation work that disproved the inertia hypothesis — a finding that would have been impossible to reach in the weeks after the event.

The sequencing lesson is one that applies to any major operational failure: the first official explanation is almost always incomplete, and building policy on it risks solving the wrong problem. In this case, an early response focused on mandating synchronous generation would have added cost without addressing the actual failure mechanism.

Frequently Asked Questions