Spain’s Blackout: What Running a 78% Renewable Grid Actually Requires

Spain now has the most instructive operating experience in the world of running a very high renewable share grid, because it is the only large European system to have failed while doing so. The investigation produced a more precise answer than the initial public debate did, and the commercial consequences — new obligations on generators, new revenue for network owners, higher system costs — are already visible in company accounts. This analysis is part of the Spain Company Stories hub.

What was the system doing that day?

Operating in a configuration that is normal in Spain and unusual almost everywhere else. Shortly after midday, renewables accounted for around 78% of generation with solar alone near 60%, while conventional synchronous plant made up roughly 15% of the mix.

Spain reaches these shares routinely on sunny spring days when demand is moderate, temperatures are mild and solar output is high. That is a genuine achievement of two decades of deployment and it places the system in operating territory that few grids have experienced.

The vulnerability is that a system dominated by inverter-based generation behaves differently from one built around large rotating machines. Voltage support, reactive power provision, fault response and stabilising behaviour all depend on how inverters are configured and instructed, not on physics alone.

What did the investigation actually find?

A cascade rather than a single cause. 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, and each disconnection worsened conditions for the remaining units.

The panel identified insufficient reactive power reserves, renewable generation operating in fixed power factor mode rather than dynamically supporting voltage, divergent voltage regulation practices between actors, and market-driven schedule changes producing fast power ramps that stressed the system.

Critically, it tested and rejected the intuitive explanation. 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 faster than additional rotating mass could compensate.

What did it cost?

Billions across the peninsula in halted industrial production, spoiled stock, disrupted transport and telecommunications and lost commercial activity, plus fatalities recorded in circumstances connected to the outage.

The continuing cost is more instructive. After the event the Spanish system operator ran the grid with substantially more conventional generation online for voltage support and stability, which is expensive. Iberdrola booked around €177m of costs tied to reinforced Spanish system operation, and non-recurring ancillary service costs contributed to a 10% fall in its power division EBITDA.

Repsol’s Industrial division was the only part of that group not to improve in the first half of 2025, reflecting the disproportionate cost of an unplanned shutdown at a refinery or chemical complex. Continuous-process industry pays far more for a blackout than its share of consumption implies.

What changes for generators?

New obligations that reduce energy revenue slightly and improve system stability materially. Renewable plants will increasingly be required to provide dynamic voltage support, ride through disturbances, and operate in modes that sacrifice a small amount of output for grid service.

Compliance enforcement is the other consequence. The panel delivered a pointed message to national regulators about enforcement of grid code requirements, noting protection settings that had diverged from what the rules required — which implies audits, retrofits and penalties rather than merely new rules.

For developers this is a real cost. For owners of storage and synchronous condensers it is a revenue opportunity, and for network operators it is an investment case: reactive power compensation, wider-area monitoring and stability equipment are all regulated assets earning a return.

Why did it become a political argument so quickly?

Because Spain’s energy policy is contested and the event was immediately available as evidence for every existing position. Opponents of renewable expansion presented it as proof of instability; supporters presented it as a grid management failure; nuclear advocates linked it to the planned phase-out of Spanish reactors.

The Spanish government and grid operator published their reports in June 2025, attributing the event to overvoltage, insufficient voltage control capacity, a programming flaw and plants disconnecting inappropriately to protect their own installations. The ENTSO-E panel took considerably longer, publishing a factual report in October 2025 and the final report in March 2026.

The delay drew criticism and produced the more useful analysis. The additional time allowed the simulation work that disproved the inertia hypothesis, and policy built on the first available explanation would have mandated expensive synchronous capacity without addressing the actual failure mechanism.

What should other countries take from this?

That grid codes, protection philosophy and market design were built for a system that no longer exists, and that the transition has outpaced the rulebook. The panel noted that root causes closely resembled a 2024 incident in southeast Europe, indicating a structural gap rather than a Spanish failure.

The sequencing lesson is the important one. Renewable capacity has been added faster than the control and coordination capability needed to operate it, because capacity is subsidised, visible and politically rewarding while voltage control is none of those things.

The interconnection lesson is equally clear. The Iberian peninsula is effectively an electrical island attached to Europe through limited capacity toward France, so a severe disturbance has nowhere to dissipate. Increasing that interconnection is a resilience measure, not merely a market efficiency one.

What is the role of the system operator?

To balance the system in real time and to restore it when it fails, and Spain’s operator did the second part effectively. Restoration across the peninsula took most of a day, with substations progressively re-energised in a controlled sequence.

Black start — restoring a grid with no external power available — is one of the most demanding operations in electricity systems, requiring specific generating units capable of starting without supply and a carefully sequenced rebuild to avoid a second collapse.

The operational lesson is that black start capability must be maintained, tested and paid for even though it produces no value in normal conditions. As thermal plants close, the number of units able to perform this function falls, and it becomes a service that has to be procured deliberately.

How does Spain compare with other high-renewable systems?

It operates at shares that few large systems reach and with less interconnection than most. Denmark and Ireland run very high renewable percentages but are physically connected to far larger neighbouring systems that absorb disturbances.

Spain’s peninsula position means it must maintain its own stability with limited external support. That makes it a genuine test case: the technical challenges other systems will face at high penetration arrive in Spain first, and its operational experience is correspondingly valuable.

The favourable interpretation is that Spain is running the experiment that Europe needs someone to run. The unfavourable one is that it is doing so without the interconnection that would make failure recoverable, which is the argument that the April 2025 event settled in practical terms.

What does this mean for industrial location decisions?

That grid node strength has become a first-order site selection criterion rather than a technical footnote. Companies considering energy-intensive facilities in Iberia now examine local network capacity, reactive power support and restoration priority alongside land cost and labour availability.

For data centres in particular, where continuity requirements are extreme and load is very large, this changes the calculation. A location with cheap renewable power and a weak grid connection may be worse than one with slightly more expensive power and a strong node.

The broader consequence is that transmission capacity has become the binding constraint on industrial development in Spain, which is the same conclusion reached from the opposite direction in the analysis of electricity market economics.

Frequently Asked Questions