The Fossil Fuel Side of the Blackout

There is no darkness deeper than that which refuses to change.

The blackout as a starting point: a grid that is no longer the same

On April 28, 2025, at 12:33 p.m. CEST, the power systems of peninsular Spain and Portugal suffered a total blackout. It was the most serious incident recorded in the European power system in over two decades and, given the sequence of events, an unprecedented episode in the synchronous area of continental Europe. The final report by the ENTSO-E expert panel, published in March 2026, ruled out a single-cause explanation and described a combination of interacting factors: oscillations, deficiencies in voltage and reactive power control, differences in regulation practices, rapid reductions in generation, generation outages, and uneven stabilization capabilities. The combination of these phenomena led to rapid voltage increases and a cascade of outages that ultimately caused the collapse of the Spanish and Portuguese systems [1]. 

That assessment is also important because of what it omits. The blackout does not allow us to conclude that a particular generation technology is, in and of itself, responsible for the collapse, nor that the solution simply lies in reverting to the electricity mix of previous decades. What it does show is that stability depends on a network of services, controls, and dynamic responses that must remain available even as the way we generate electricity changes profoundly.

For decades, much of these services were built almost automatically into the system’s architecture. Large thermal, nuclear, and hydroelectric power plants used synchronous machines directly coupled to the grid. While generating power, these machines simultaneously provided rotational inertia, voltage and reactive power control, high current during certain short-circuit events, and a robust electromagnetic reference for the rest of the system. In many thermal power plants, the primary energy came from coal, fuel oil, or gas, so fossil fuel generation and certain stability services have historically been linked. 

Hence the point of talking about the “fossil fuel side of the blackout.” This is not about attributing blame to fossil fuels or automatically proposing solutions involving them. It is about understanding which part of the grid’s technical backup has historically been linked to fossil-fuel facilities, which part actually come from the synchronous machine, and what technological alternatives exist to maintain or improve those capabilities in a progressively decarbonizing system. 

The blackout on April 28 thus serves as a starting point for a much broader question: Are we transforming only the energy sources, or are we also transforming, at the same pace, the mechanisms that physically sustain the grid? Replacing a conventional power plant with renewable generation replaces its MWh, but not necessarily all the functions the facility provided to the system while it was connected. 

“The ‘fossil’ aspect of the blackout isn’t about singling out a particular technology, but about understanding which stability services we must preserve as the way electricity is generated changes.”

From a predominantly synchronous system to an increasingly renewable mix

The scale of this transformation is evident in the evolution of Spain’s energy mix. In 2010, renewable energy accounted for 33.2% of national electricity generation [2]. Ten years later, in 2020, its share had already reached 44% [3]. In 2024, a new all-time high was set: 56.8% of all electricity generated in Spain came from renewable sources [4]. 

The trend continued in 2025. According to Red Eléctrica, renewables produced 55.5% of Spain’s electricity, and when self-consumption estimates were factored in, that share rose to 56.6%. Taking these self-consumption installations into account, nearly 69% of installed generation capacity at the end of 2025 was renewable. Photovoltaics ranked as the technology with the highest installed capacity, and wind power led generation excluding self-consumption. Combined-cycle power plants, though accounting for a smaller share than in previous periods, contributed 16.8% of annual generation, illustrating the coexistence of a growing renewable energy fleet and manageable thermal generation [5]. 

Europe is following the same path. In 2014, renewables accounted for 28.6% of the European Union’s gross electricity consumption; by 2024, that figure had reached 47.5% [6]. Looking specifically at European electricity generation, renewables were already the leading source in 2024, accounting for 47.3%, compared to 29.2% from fossil fuels and 23.4% from nuclear power [7]. This is, therefore, a structural transformation and not a phenomenon limited to a specific country. 

But energy percentages and dynamic characteristics are not equivalent concepts. A grid with 55% renewable generation is not necessarily less stable than one with a lower percentage; what changes is the nature of much of the equipment connected to it. Much of solar photovoltaic power, modern wind power, and energy storage is integrated through power electronics converters. Consequently, an growing portion of the grid’s response is no longer determined by rotating masses directly coupled to the frequency but instead depends on control algorithms, electrical measurements, current limits, and energy availability on the DC side. 

This change requires us to separate two discussions that have long been intertwined. The first is about energy: where electricity comes from and what emissions are generated in its production. The second is systemic: which resources provide frequency, voltage, reserve capacity, grid strength, damping, and the ability to respond to disturbances. The energy transition can only be considered complete when progress is made on both fronts simultaneously. 

“The transition is not just about replacing fuels. It also requires replacing, redesigning, or eliminating the services that conventional machines implicitly provided.”

Fossil fuel support: energy, power, and stability are not the same thing

The word “backup” is often used as if it described a single need, but from the perspective of the electric power system, it is important to distinguish at least three dimensions. The first is energy backup: having sufficient energy available during periods when variable renewable generation is low. The second dimension is power backup or adequacy: having available capacity to respond to demand peaks or unexpected losses in generation. The third is stability backup: having resources capable of maintaining frequency and voltage, providing reactive power, contributing to grid strength, and responding appropriately during faults and disturbances. 

Natural gas, particularly through combined-cycle power plants, continues to play a significant role in the first two dimensions and also provides certain electricity services through its synchronous generators. The IEA has noted that, in systems with a growing share of wind and solar generation, gas-fired power plants can play a particularly important backup role during periods of low renewable output. In the first half of 2025, for example, lower wind and hydroelectric generation in Europe led to increased gas use in the electricity sector, underscoring that thermal flexibility remains valuable when weather conditions are unfavorable [8].

This role, however, establishes a link between electricity security and fuel availability. When a system relies on gas-fired generation to respond during critical moments, the security of the electricity supply becomes tied to the capacity for transporting, storing, and delivering the gas itself. The IEA has been highlighting this interaction between gas and electricity security for years, precisely because greater penetration of variable renewables can reduce the number of operating hours for thermal power plants while, at the same time, maintaining their value during certain events [9]. 

The economic paradox is clear. A power plant that operates fewer hours may be less significant in terms of energy production, yet it remains valuable as available capacity. As its capacity factor decreases, it becomes more difficult to recoup fixed costs solely through energy sales. This shifts the debate toward mechanisms that properly value capacity, reserves, flexibility, and ancillary services, rather than compensating solely for the MWh generated. 

It is also important to distinguish between the fuel and the machine. A gas-fired power plant does not provide inertia because it burns gas; it provides inertia because it uses a synchronous machine with a rotating mass that is electrically connected to the grid. Similarly, a synchronous condenser can provide inertia, reactive power, short-circuit current, and grid strength without producing electricity or consuming fuel to generate power. The renewed interest in this equipment in systems with a high proportion of converter-based resources shows that part of the backup historically attributed to conventional generation can be decoupled from fossil fuel production [10]. 

The relevant question is therefore no longer “How much fossil fuel-based generation do we need?” but is instead phrased more precisely as: “What services does the system need, for how long, in what location, and through which technology is it most efficient to provide them?” This reframing opens the door to batteries, synchronous condensers, interconnections, demand flexibility, managed renewable generation, and advanced power electronics. 

From Physical Stability to Designed Stability

Wind and photovoltaic resources, as well as BESS systems, are typically connected to the grid via converters. This interface largely decouples the behavior of the primary source from what the system perceives electrically. The consequence is profound: many features that were inherent in a synchronous machine can be defined via software in a converter. 

For years, renewable energy integration has relied primarily on grid-following converters (GFLs). Simply put, these devices measure the existing grid voltage, synchronize with it—typically using a PLL or other synchronization mechanisms—and control the active and reactive power they feed into the grid. It is a mature architecture that is extraordinarily effective when the grid provides a sufficiently robust reference. 

A conceptual problem arises when the proportion of units that need to “follow” a reference increases, while the number of machines that historically set that reference decreases. In weak grids—those with low short-circuit power or strong interactions among multiple controls—the dynamics of the converters can play an increasingly significant role in overall stability. Recent literature specifically incorporates categories such as converter-driven stability and resonances to describe phenomena that do not fully fit within the classical synchronous machine paradigm [11]. 

Hence the growing interest in grid-forming converters (GFCs). Rather than simply tracking an external voltage, a GFC locally controls a voltage and frequency reference and can behave, within its physical limits, as a voltage source. Recent reviews highlight their potential to provide frequency and voltage support, operate in weak grids, and participate in isolated operation or restoration processes [11][12]. 

One of the most intuitive examples is what is known as synthetic or virtual inertia. A converter does not contain a large mechanical mass from which to extract kinetic energy, but it can rapidly modify active power according to a defined control law. To do so, it naturally requires an energy source: a battery, a renewable resource operating below its maximum power, the DC link, or another system capable of temporarily supplying that energy. The response is no longer dictated by mass but becomes, at least partially, a design variable. 

This opens up opportunities beyond simply “imitating” a synchronous machine. Electronic control makes it possible to adjust damping, frequency response, reactive power, and behavior in response to grid changes in ways that are not available with the same degree of flexibility in an electromechanical machine. Stability can gradually become a programmable feature. 

But programmable does not mean unlimited. Semiconductors can withstand much lower overcurrents and for a shorter duration than a conventional synchronous machine. When a GFM reaches its current limit during a fault, its ability to maintain the expected behavior as a voltage source may degrade, and the current-limiting strategy becomes part of the transient stability itself. A 2024 IEEE review specifically identifies overcurrent limitation as a critical area for the large-scale deployment of GFMs [13]. 

Power electronics, therefore, do not eliminate physical constraints: they shift and transform them. They replace rotating mass with semiconductors, stored energy, and algorithms; they replace mechanical constants with control parameters; and they require stability, protection, and coordination to be evaluated from a different perspective. 

“Power electronics don’t have to exactly replicate the power grid we’ve inherited: they allow us to design responses that were previously dictated by the physics of the machine.”

A hybrid system rather than a system of extremes

It would be simplistic to imagine the future as a binary choice between synchronous power plants and converters. For years—probably decades—synchronous generators, combined-cycle plants, hydropower, nuclear power (where it remains operational), wind and solar farms, energy storage, synchronous condensers, HVDC, flexible demand, and distributed resources will coexist. The challenge will be to coordinate this diversity and use each technology for what it can contribute most efficiently. 

Spain already provides a snapshot of this coexistence. In 2025, more than half of the electricity came from renewable sources, but combined-cycle power plants continued to account for nearly one-sixth of annual generation [5]. At the same time, storage systems—pumped-storage and batteries—helped integrate 9,213 GWh, 6.2% more than in 2024. This figure is significant because it shows that the evolution of the energy mix is not solely about installing renewable generation; it requires expanding resources capable of shifting energy and providing flexibility [5]. 

In the short term, thermal backup will continue to be useful during certain periods of low renewable production or high demand. But as storage, interconnections, active demand management, and flexible renewable generation continue to develop, the need to use fuels to fulfill functions that can be provided in other ways will decrease. In its 2026 analysis, the IEA notes that the simultaneous growth of solar power, wind power, battery storage, electric vehicles, heat pumps, and large loads is substantially increasing the system’s flexibility requirements [14]. 

Demand can also cease to be a passive element. Digitalization, aggregators, energy management systems, and market signals make it possible to shift certain consumption patterns or adjust them based on the state of the grid. The IEA estimates that demand flexibility can reduce costs, facilitate the integration of renewables, and alleviate grid constraints, transforming consumption that traditionally only required energy into resources capable of contributing to grid balance [15]. 

Renewable generation itself can play a more active role. In 2025, ENTSO-E advocated for the need to improve observability, controllability, and incentives so that renewable resources can provide flexibility and system services, rather than treating them solely as sources that generate power whenever resources are available [16]. This represents a cultural and technical shift: renewables are no longer elements for which the grid must “make room,” but are gradually becoming active participants in its operation. 

The same shift is taking place in technical codes. In 2024, ENTSO-E published an initial framework of comprehensive requirements for grid-forming capabilities in wind farm modules, and in November 2025, a consolidated second phase intended to serve as the basis for future implementation guidelines related to the evolution of the Network Code Requirements for Generators. The stated objective is to ensure the availability of capabilities that enable the stable and safe operation of a European grid with a growing share of non-synchronous generation and storage [17][18]. 

This regulatory process is particularly significant: it confirms that GFM is evolving from a purely research-based concept into a capability that operators, manufacturers, and regulators will need to specify, verify, and coordinate. This does not mean that a universal solution already exists, but it does mean that the European system is beginning to formally define the performance it expects from the new generation of converters. 

The real transition: replacing functions, not just fuels

The blackout on April 28 served as a stark reminder that a power grid does not function simply because it has sufficient installed capacity. It requires dynamic behavior, coordination, regulation capabilities, monitoring, and adequate physical margins. The ENTSO-E final report specifically emphasizes strengthening operational practices, improving monitoring of system behavior, refining coordination and data exchange, and adapting regulatory frameworks to a changing grid [1]. 

This learning process is directly linked to the transformation of the Spanish and European energy mix. In Spain, we have gone from a 33.2% share of renewable energy in 2010 to more than half of current generation [2][5]. Europe is also approaching a point where one out of every two units of electricity consumed comes from renewable sources [6]. The grid that must support this production can no longer rely exclusively on the mechanisms that worked when most of the generation was tied to large synchronous machines. 

Fossil fuel generation will continue to play a backup role during the transition wherever it provides capacity, manageable energy, or services that cannot yet be efficiently met by other resources. But its future value will become increasingly less tied to the volume of electricity it produces. Part of what it used to provide can be maintained through synchronous condensers; part can be shifted to storage; part can be absorbed by demand; and another part will depend on advanced converters and regulatory frameworks that allow for their effective use. 

This even changes the way we compare technologies. The cost per MWh will remain important, but it is not enough to represent the systemic value of a facility. It will be necessary to evaluate available capacity, storage duration, location, response time, voltage support, frequency contribution, behavior during faults, black-start capability, robustness against weak grids, and other services that until now may have been implicit in the operation of a conventional power plant. 

The opportunity lies in the fact that power electronics allow these functions to be separated. A single battery can provide energy during some hours, deliver rapid frequency response during others, and—with proper control—contribute to grid formation. A renewable energy fleet can operate at maximum output for much of the time and reserve capacity for when the system requires an active response. An HVDC interconnection can transport energy between regions and simultaneously participate in control. A microgrid can decouple from the main grid and maintain critical loads. The new architecture does not have to replicate the components of the previous system on a one-to-one basis. 

That is why “There is no darkness deeper than that which refuses to change” serves here as more than just a slogan. The lesson of the blackout is not to return to a past system or to rush a transition while ignoring its physical needs. It is to accept that the electric grid is taking on new dynamics and that its stability tools must evolve along with them. 

The question for the future is no longer simply what percentage of generation will come from renewables or how much fossil-fuel backup will remain online. The relevant question is what the grid needs to remain stable at every time scale and what combination of resources can provide those capabilities in the safest, most efficient, and most sustainable way. 

Replacing fuels is a fundamental part of decarbonization. Replacing, redesigning, and improving the services that for decades remained hidden behind synchronous machines is the true engineering challenge. And it is precisely there—in the combination of infrastructure, storage, flexibility, and power electronics—that the stability and resilience of the grid we are building will be determined. 

References

1. ENTSO-E (2026). Final Report on the Grid Incident in Spain and Portugal on 28 April 2025. Expert Panel Final Report, 20 March 2026.
2. Red Eléctrica (2020). Renewables Put a Green Stamp on the Decade. Growth Data 2010–2019.
3. Red Eléctrica (2021). 2020: The “Greenest” Year for Energy Thanks to Record Wind and Solar Photovoltaic Generation.
4. Red Eléctrica (2025). Renewable Energy Production in Spain Grows by 10.3% in 2024, Reaching Record Levels.
5. Red Eléctrica (2026). The Spanish electricity system in 2025: electricity demand, generation, and installed capacity all increase.
6. Eurostat (2026). 2024: nearly 50% of EU electricity came from renewables. Dataset nrg_ind_ured.
7. Eurostat (2025). Renewable energy supply grew by 3.4% in 2024. Preliminary annual energy supply statistics.
8. International Energy Agency (2025). Gas Market Report, Q3-2025. IEA, Paris.
9. International Energy Agency. Power Systems in Transition: Electricity Security Matters More Than Ever. IEA, Paris.
10. Soleimani, H.; Habibi, D.; Ghahramani, M.; Aziz, A. (2024). Strengthening Power Systems for Net Zero: A Review of the Role of Synchronous Condensers and Emerging Challenges. Energies, 17(13), 3291. DOI: 10.3390/en17133291.
11. Mirmohammad, M.; Azad, S. P. (2024). Control and Stability of Grid-Forming Inverters: A Comprehensive Review. Energies, 17(13), 3186. DOI: 10.3390/en17133186.
12. Khan, M.; Wu, W.; Li, L. (2024). Grid-forming control for inverter-based resources in power systems: A review on its operation, system stability, and prospective. IET Renewable Power Generation, 18(6), 887-907. DOI: 10.1049/rpg2.12991.
13. Baeckeland, N.; Chatterjee, D.; Lu, M.; Johnson, B.; Seo, G.-S. (2024). Overcurrent Limiting in Grid-Forming Inverters: A Comprehensive Review and Discussion. IEEE Transactions on Power Electronics, 39(11), 14493-14517. DOI: 10.1109/TPEL.2024.3430316.
14. International Energy Agency (2026). Electricity 2026 – Flexibility. IEA, Paris.
15. International Energy Agency (2026). Scaling Up Demand Flexibility: From Peak Management to Efficient System Operation. IEA, Paris.
16. ENTSO-E (2025). Flexibility from Renewable Energy Sources (RES): Report and Position Paper.
17. ENTSO-E (2024). Grid Forming Capability of Power Park Modules – First Interim Report in Technical Requirements.
18. ENTSO-E (2025). Grid Forming Capability of Power Park Modules – Phase II Technical Report.

Photo of the Zwinger Palace | Dresden, Germany