Electricity overcapacity in France: is the real issue production or demand management?

On April 1, 2025, the French electricity system experienced an episode that deserves better than a binary reading. As market prices became negative, some renewable energy producers reduced or interrupted their production for economic reasons. The rapid decline in supply has highlighted an obvious fact often lost in energy debates: an electricity network cannot be managed only with installed megawatts. It is managed with coherent market signals, means of flexibility, robust operating rules and demand capable of absorbing production at the right time.

This episode fueled a strong thesis: France would have entered a situation of electrical overcapacity. This reading is based on a real paradox. On the one hand, renewable capacities are progressing. On the other hand, consumption remains lower than predicted by several electrification trajectories. When wind, sun and nuclear power produce simultaneously in a context of low demand, prices can fall, incomes compress and the individual decisions of producers weaken the collective balance.

But the word overcapacity can become misleading if it describes a photograph as a lasting reality. The right question is not only: do we have too much capacity today? It is also: do we have the appropriate capacity, in the right place, at the right time, for the uses that will develop tomorrow?

Supply uncorrelated with demand

The first lesson concerns synchronization. Electricity is still difficult to store on a very large scale and at a completely neutral cost. The value of a megawatt hour therefore depends on its production time, its location and the capacity of the system to transport or move it over time.

Adding capacity that produces in the same windows can cause revenue cannibalization. This phenomenon does not mean that renewables are useless. It means that energy policy can no longer be a policy of volume only. Each new capacity must be thought of with the network, storage, erasure, interconnections, industrial uses and territorial strategy.

The second lesson concerns incentives. If a producer has an interest in suddenly disconnecting when prices become negative, the problem is not only their behavior. It also lies in the design of the market. An efficient architecture must align the optimization of each actor with the security of the system. This requires clear flexibility mechanisms, remuneration for services provided to the network and obligations compatible with real technical constraints.

Shift priority towards uses

France has a strategic advantage: largely carbon-free electricity production and an industrial base capable of designing, building and operating complex infrastructures. However, this advantage will only create value if the electricity finds productive uses.

The electrification of mobility, heat and certain industrial processes constitutes a first path. It can reduce imports of fossil fuels, support competitiveness and better distribute the fixed costs of the system. But it requires more than goals. It requires strengthened local networks, financing solutions, available equipment, skills and coordination between the State, communities, manufacturers and network managers.

Flexibility must become an infrastructure in its own right. Batteries, consumption management, electrolysis, thermal storage, intelligent vehicle charging and demand response contracts can convert temporarily excess production into deferred value. The subject is therefore not to oppose production and consumption. It is to build a coherent chain between capacity, network, flexibility and use.

The unknown of data centers and artificial intelligence

One variable can profoundly change the equation: the growth of data centers, supercomputing and artificial intelligence. There are numerous investment announcements, but their energy translation remains uncertain. Not all announced projects will be carried out, not all will connect at the same pace and their needs will strongly depend on the efficiency of equipment, cooling and the evolution of digital models.

It would therefore be unwise to consider each announcement as a firm request. It would be equally imprudent to build a long-term strategy while ignoring this new burden. Data centers are looking for stable power, fast connection, high availability and, increasingly, traceable, low-carbon electricity. France has the assets to welcome them, but this opportunity raises questions of location, water, network, sovereignty and sharing of value.

The right method is to integrate several demand scenarios, with decision milestones. Investments must be able to evolve as projects are confirmed. This scenario-based management avoids two symmetrical errors: overinvesting on the basis of hypothetical growth or underinvesting to the point of creating a bottleneck.

Moving from the capacity debate to portfolio governance

French energy policy would benefit from being managed like an infrastructure portfolio. Each decision should be evaluated according to five dimensions: contribution to safety, complete cost for the system, flexibility provided, industrial and territorial value, compatibility with demand trajectories.

This approach requires more integrated governance. Production, network, storage and industrial location decisions can no longer be made on separate schedules. They must share assumptions, maturity thresholds and regular reviews. When a demand assumption changes, the portfolio must be recalibrated rather than defended by inertia.

It is also necessary to distinguish the cost of an asset from the cost of the system. A technology may appear competitive in isolation while generating additional network or flexibility needs. Conversely, a more costly investment per unit can create significant value if it reduces a local constraint or secures a critical peak.

The truly strategic question

Are we overcapacity? At certain times and in certain areas, the answer may be yes. Are we permanently protected from a deficit? There is nothing to say with certainty. Industrial electrification, mobility, heating, hydrogen and digital can transform demand. The pace, location and simultaneity of these uses, however, remain difficult to predict.

The responsible response is therefore neither a general cessation of investments, nor a blind race for capacity. It lies in adaptive planning: producing when the system value is demonstrated, accelerating uses, investing in flexibility, strengthening networks and reviewing choices based on verifiable data.

The incident of April 1, 2025 is not definitive proof of a France that produces too much. It is the signal of a system that must better coordinate its decisions. Tomorrow's energy performance will not only be measured in installed capacities or tonnes of carbon avoided. It will be measured in our ability to transform each carbon-free kilowatt hour into security, competitiveness and useful value for society.