France has nearly 33 GW of industrial projects waiting to connect to the grid. Queuing rules constitute a brake, but the network’s physical capacities and reinforcements remain equally decisive. A project reserves its capacity upon acceptance of the grid connection offer, under conditions of maturity and financial commitment, and not solely upon filing a request, but insufficiently advanced projects could subsequently block capacity. RTE opened a consultation on January 27, 2026 and was to brief the CRE by October 1, 2026 at the latest on an initial reform stemming from this consultation, shifting toward a simple principle: only those who are truly ready advance.
The essentials
- France has accumulated nearly 33 GW of industrial projects awaiting connection to the high-voltage grid. The bottleneck has a significant institutional dimension, but it is not exclusively institutional.
- Unfinished files monopolized reserved capacities without ever being realized, paralyzing the queue for real projects.
- RTE announced a reform to be implemented by the end of 2026. Deliberation 2026-54 concerns RTE’s investment program for 2026.
- In Europe, according to Ember, nearly 700 GW of renewable projects were in queue across eight reporting countries: France is not an isolated case, but its reform can serve as a model.
- Once the queue is purged, real tension begins: data centers, heavy industry and renewables will compete for the same freed-up capacity.
The queue as an institutional trap
To connect a factory or solar farm to the French electrical grid, you must file a request with the relevant network operator: RTE for the transmission network, Enedis or a local distribution company for the distribution network. Acceptance of the grid connection offer, under certain conditions, allows entry into the queue and reservation of associated power, sometimes before the project is fully financed or technically defined. The result: a significant portion of reserved capacity was blocked on paper, occupying a place in the queue without ever materializing.
Analyses of grid connections in France have documented this phenomenon. Analysis of grid connections in France shows that saturation results simultaneously from allocation rules and physical capacity deficits, which vary by zone. A phantom project filed in 2021 can block a real industrial project filed in 2024.
This mechanism is not unique to France. Ember counts nearly 700 GW of renewable projects awaiting connection in eight reporting countries, representing comparable saturation in several European states. Spain, the United Kingdom, and Germany each accumulated queues of several hundred gigawatts before undertaking reforms. France arrives later in this cycle, but it has the advantage of observing its neighbors’ mistakes and successes.
Thirty-three gigawatts in numbers
Thirty-three gigawatts is roughly the power output of thirty large nuclear power plants. If put into service, these projects could electrify a significant share of French reindustrialization. But the unit of measurement conceals a diversity of files: electrolyzers to produce green hydrogen, electric arc furnaces to decarbonize steel, data centers, battery production sites, solar or wind farms.
This heterogeneity is the first challenge of the reform. All these projects do not deserve the same priority, and all do not have the same risk profile for the network. A data center consumes continuously, twenty-four hours a day. An industrial arc furnace can absorb massive spikes. A wind farm injects power intermittently.
Putting everyone in the same queue, even reformed, amounts to ignoring these structural differences.
That is precisely why the RTE reform is not limited to the principle of “first ready, first served.” It also introduces project maturity criteria: RTE was considering a maturity milestone such as the filing of a building permit or the purchase of equipment related to the project. The rules aim to prevent capacity from being captured by insufficiently advanced projects, according to the procedures proposed by RTE.
Deliberation 2026-54 approves RTE’s investment program for 2026, without providing the alleged tools for purging the queue.
New rules introduced by the reform
Before the reform, first come, first served applied, regardless of project solidity. An experienced developer could file a dozen speculative projects to reserve capacity, even if only one or two materialized. This practice, common in several European countries, is sometimes called “queue speculation” in sectoral analyses.
The reform introduces a milestone logic: the project sponsor must pass documented steps at specific dates, on pain of losing their place in the queue. The system favors projects demonstrating their administrative, land-related, and financial advancement, without systematically requiring all permits and full financing. For a battery factory whose owner has signed contracts and obtained permits, it is a significant improvement. For a promoter betting on multiple projects simultaneously, it is a new constraint.
The United Kingdom undertook a similar reform in 2023, under the name “Connections Reform.” Its electricity transmission network, National Grid, purged several hundred gigawatts from its queue in less than a year. The British experience shows that the initial purge can be politically brutal, as it involves eliminating project sponsors who had invested time and money to position themselves, but it then measurably accelerates actual connections.
Data centers against heavy industry: the arbitrage to come
Once the queue is purged, tension sets in. Data centers have benefited so far from accelerated treatment in several countries, under pressure from digital operators who argue the urgency of their deployments. Five major data center projects can benefit from an accelerated procedure on sites identified by the state.
A data center consumes between 50 and 200 MW continuously, mobilizes few direct jobs, and produces concentrated added value. A decarbonized foundry or chemical plant can absorb comparable power, employs hundreds of workers, and anchors an industrial value chain over several decades. Both need the network. The proposed reform introduces criteria adapted to demand profiles beyond equal priority among qualified candidates.
The reform also has political scope. American reindustrialization mobilizes billions in subsidies to attract factories but runs up against network and labor constraints that money alone does not resolve. France, by correcting its institutional bottleneck without massive subsidy, pursues a different approach: creating conditions for equitable access to public infrastructure. This formal equity remains to be tested against competitors relying on other levers.
Can heavy industry be electrified in time
The real test of the reform will lie between 2027 and 2032. RTE estimates that substantial realization of electrification projects by 2030 is necessary for a rapid decarbonization trajectory. An electric arc furnace does not connect quickly: connection delays remain significant, even with improved rules.
This temporal constraint raises a question that queue reform alone does not resolve. Purging phantom files frees reserved capacity, but does not build new lines. The historic industrial zones of northern and eastern France, heirs to coal economies and heavy industries such as steel, chemicals, and textiles, have a network whose topology was designed for uses different from massive heavy industry electrification. Shifting to heavy electrification sometimes requires revisiting the topology of transformer stations themselves.
RTE is investing in this reinforcement, as part of its ten-year network development plan. But network condition audits show accumulated maintenance backlog in certain zones, a legacy of years when network investment was squeezed. The cost of maintenance neglect is measured in additional delays and higher connection costs for project sponsors. Reformed access rules can clear the queue; they do not replace missing cable.
Two trajectories are emerging for the coming five years. In the first, queue reform is accompanied by acceleration of targeted network investments in zones with strong industrial potential, and the arbitrage between data centers and heavy industry is made explicit by public authorities rather than left to project maturity logic alone. In this trajectory, a significant portion of the thirty-three gigawatts actually connects by 2030, and reindustrialization finds its infrastructure. In the second, network investments remain insufficient, arbitrage remains implicit, and the best-financed projects—namely data centers and large digital platforms—capture most of the freed capacity. Heavy industry finds itself then in a shorter queue, certainly, but still blocked, for physical reasons this time.
The signal to watch is the pace of actual industrial project commissioning after the queue purge. If, a year after the fourth quarter of 2026, the number of actual industrial connections has not progressed measurably, the bottleneck has simply shifted, from institutional to physical.
Europe facing the same problem, with uneven reform progress
France invents nothing. It joins a movement running through Europe, at different speeds depending on the country. The United Kingdom reformed its queue in 2023. Spain put similar maturity criteria in place the same year. Germany is working on a reform of its distribution network, distinct from transmission, which also blocks hundreds of small-scale solar and wind projects.
The projects awaiting connection counted by Ember in several European countries represent systemic waste. Better allocation of existing capacity can accelerate connections, but does not eliminate capacity and reinforcement needs identified by Ember to reach 2030 climate targets. It is the order of magnitude that analyses from Zenon and Ember converge in suggesting, subject to these reforms being actually implemented and respected.
This European convergence also opens an avenue for cooperation. Network access rules remain national, and a cross-border project can find itself stuck on one side of a border even if capacity exists on the other. The European Commission has undertaken work on harmonizing connection rules, but the timeline remains uncertain. In the meantime, each country reforms on its own, with different criteria and timelines, which complicates cross-border industrial projects, notably in hydrogen and steel.
For France, reform of the queue is a necessary condition of its industrial strategy, not a guarantee. It removes a real, measurable obstacle that inherited rules had created. It does not lift questions of network investment, arbitrage between competing uses, and European coordination. These questions remain open, and their answer will determine whether the thirty-three gigawatts currently blocked ever power arc furnaces, electrolyzers, or server rooms.
Sources
- Zenon NGO, “Grid Connection: The Queue, True Bottleneck of the Transition”: https://www.zenon.ngo/insights/raccordement-au-reseau-la-file-attente-vrai-goulot-de-la-transition
- Ember, European analysis of connection queue backlogs (April 2026): Ember Climate (ember-climate.org), report unlinked to avoid any unverifiable URL
- Energy Regulation Commission, CRE Deliberation 2026-54: CRE (cre.fr), official deliberation
- RTE, official statements 2026 on queue reform: RTE France (rte-france.com)
- Knowledge of Energies, Enerzine, Selectra: French sector press having covered RTE reform
