On April 28, 2025, the power grid serving continental Spain and Portugal went down, causing gridlock in cities, cutting communications networks, and stranding people on trains, in airports, and in elevators all across the Iberian peninsula and briefly in a small area in southwest France close to the Spanish border. The unprecedented, massive blackout lasted as long as 12 hours in some areas, including in the capital city, Madrid. Not surprisingly, placing blame for the outage was rapid. Quick reactions pointed to cyberattack, sabotage, and natural phenomena such as solar flares. But such theories were quickly laid to rest, and a panel of experts was formed to determine exactly what caused the blackout. On the one-year anniversary of the outage—and after much analysis by many experts—there isn’t a simple answer: In short, no one technology was to blame. While solar and wind generation was high, experts agree that the renewables weren’t at fault. For an explanation, we turned to Pablo Duenas-Martinez, a research scientist at the MIT Energy Initiative and an assistant professor at Universidad Pontificia Comillas in Madrid, Spain.
Q. First, how does a proper, well-functioning power grid behave and what does the system operator do to help?
A. There are two components to the flows on a power grid. One is “active power”—the part that lights up our light bulbs and runs our engines. With active power, the demand on the grid must always equal supply. The other component is “reactive power,” the part we can’t see but controls the voltage at which the power is delivered so it suits our devices. If voltage is too low, lights will flicker. If voltage is too high, devices may not only fail to work but be damaged beyond repair.
The operator of the transmission system—the TSO—must control both components, and that can be tricky. Active power supply and demand are largely coordinated through markets. But controlling reactive power is harder. The main way the TSO can control it is to call on operators of conventional power generators, so generators burning natural gas or coal or nuclear plants. Those systems can be adjusted to either absorb or inject reactive power as needed to control voltage on the power grid—indeed, they are typically required by law to provide “reactive power control.”
In contrast, solar and wind generators always absorb reactive power. The large solar and wind sources can provide reactive power control when it’s needed, but doing so is costly for them—and in Spain, unlike in most countries, it’s not mandated by law, so they typically don’t do it. Meanwhile, there are many small solar systems—imagine lots of rooftop solar installations and small solar farms. Those small systems are directly connected to the distribution system. As a result, they’re not controlled by the TSO; the TSO may not even know whether they’ve shut down or are still running and absorbing reactive power.
Sometimes, fluctuations in voltage called “oscillations” can happen on a power grid, for example, when a transmission line or a generator is connected or disconnected. Oscillations can increase and decrease the voltage rapidly, and if voltage gets too high, generators and user devices can start “tripping,” that is, automatically disconnecting to prevent being damaged. Operators have standard protocols to follow to bring oscillations under control.
Q. So what happened on April 28?
A. First, I’ll describe the setup. The Spanish grid is loosely connected to the French grid and in practice is merged with the grid serving Portugal. Within Spain, we have many large solar and wind farms and lots of small installations of solar systems, many located in the southwestern area of the country. On April 28—as on most spring days, when demand is low—about two-thirds of the power on the grid came from renewable sources. The rest came from a mix of nuclear and natural gas plants.
The day before the blackout, the TSO confirmed that there were no conventional generators scheduled to run. So, to ensure safe operation the next day, the TSO took steps that included dispatching 12 conventional generators, 10 of them to provide reactive power control. One of the units in the south called him back and said, “I won’t be available. I cannot switch on tomorrow.” The TSO thought he had things under control and continued operations with only nine units available to provide reactive power control.
At 12:19 p.m., a major oscillation was detected on the grid, again coming from Europe. In response, the TSO—again following standard protocol—reduced exports to Portugal, switched the flows to France from alternating current to direct current, and connected five more transmission lines within Spain. While those steps stabilized the voltage, the TSO recognized that there was now limited capacity on the system to control voltage. So, he called on a different conventional generator to begin running. But that unit couldn’t be available for an hour.
Suddenly, as a consequence of the previous actions, the voltage increased dramatically, and generating units began to trip. Within half a second, many of the small solar generators—especially prone to damage from high voltages—automatically shut down. Twenty milliseconds later, a big solar plant in southwestern Spain tripped. Because the solar plants were no longer absorbing reactive power, voltage on the system went up even more, and more systems shut down. The grid went into what some have called a death spiral, resulting in a total blackout across the Iberian peninsula and some areas of southern France.
Q. What have we learned from the Iberian blackout, and have changes been implemented to ensure that the same won’t happen again—or happen elsewhere?
A. A resilient power system must prevent, mitigate, respond, and recover. In this case, the first three components clearly failed. Preventive mechanisms were insufficient; they initially mitigated the oscillatory events but left the system in a weakened state, and the response triggered the death spiral that led to the final blackout.
The good news is that the recovery was quick. The northern and southern sections of the peninsula had power back within a few hours. I live in the suburbs of Madrid, and I had power back just six hours later. My parents live downtown, so that was far more challenging—a big city with a large, complex load. Even so, they had power back in 12 hours—and 12 hours is quick for such a major, widespread blackout.
In the end, experts and analysts have agreed that the blackout was caused by a series of events that were all happening in the same place, at the same time. And the experience did provide a number of valuable learnings.
Learning #1:
The experience clearly demonstrated the importance of having a sufficient number of conventional power plants prepared to provide reactive power control or to turn on right away when called on. There’s a recommendation calling for a set ratio between conventional generators and renewables on a power grid. Conventional facilities such as nuclear, hydroelectric, and fossil fuel plants rely on heavy metal wheels to generate electricity. Those massive rotating wheels have high inertia, so they’ll keep running and can help stabilize frequency and voltage even when solar and wind plants shut down. Before the blackout, Spain had a sufficient number of “rotating units” to meet the recommended ratio. However, in southern Spain, there was just one such unit—well below the recommended number, given the huge number of small solar units plus several large solar units in the ar
The message here is that you can’t just look at the country as a whole. You have to look at regions. Voltage is a local problem that can propagate at the system level. Before the blackout, southern Spain typically had at most three conventional power plants. Now the region usually has six or seven at the ready to help with reactive power control.
Learning #2:
The rules or protocols for controlling reactive power and dealing with oscillations were not well designed. By law, rotating generators must automatically—and without being paid— do a defined amount of reactive power control. But making the needed operational change costs money, and a plant can do less than the required amount and not incur any kind of penalty. However, the TSO doesn’t know in advance how much reactive power control a given plant will actually do. Now that loophole in the law has been reviewed by the regulator.
The main rules have been updated and now also require large solar and wind power plants—those above 5 megawatts—to provide reactive power control. More importantly, voltage control will be auctioned and remunerated, incentivizing rotating conventional generators and bringing in a new money stream for solar and wind power plants. Those power plants that do not upgrade their installation for voltage control might be disconnected by the TSO if local voltage issues arise.
Learning #3:
Another learning concerns the many small solar power generators and the protections that cause them to trip. The TSO doesn’t know in advance when this may happen because the small solar sources are directly connected to the distribution system and therefore are under the umbrella of the distribution system operator. So, the learning here is that there should be more communication and coordination between the operator of the transmission system—the TSO—and the operator of the distribution system.
Learning #4:
In most countries, laws dictate a range of voltage that is approved. In Spain, the upper limit is high—in fact, it’s very near a voltage at which equipment may be damaged. And the Spanish grid tends to hover close to that upper limit, even during normal operation, and that can be a big problem: If there are strong oscillations—as there were leading up to the blackout—voltage can reach that upper limit, and protections on devices will automatically trip. The panel of experts has strongly recommended to lower this upper limit in Spain and align it with the rules in neighboring countries, including Portugal and France. The TSO is still studying the recommended change.
Learning #5:
During normal operation, the TSO controls voltage by activating rotating generators that can provide reactive power control. But as we saw in conditions leading up to the blackout, the TSO doesn’t always have rotating generators available.
Theoretically, TSOs have two more ways to control voltage. They can connect a device called a shunt reactor, which absorbs reactive power—a means of dealing with voltage rise. And they can regulate voltage directly using a “STATCOM,” a special device that provides rapid, dynamic voltage control.
However, neither the shunt reactors nor the STATCOM could help prevent the blackout. The shunt reactors available at that time were operated manually, and collapse of the grid happened so quickly that the TSO didn’t have time to connect them. And at that time, there was a single STATCOM device on the Spanish system. Planning was under way to install three more devices—and that installation is being rapidly completed.
From newspaper articles and off-the-record conversations, I’ve learned that the system has—due to similar external circumstances—been close to blackout again during the past year. But in part due to the learnings and to changes that have been implemented as a result, it didn’t happen again.