On 14 September 2026, three deputy mayors and a councillor in Étrechet, a small commune in the Indre département, announced their resignations. The next day, mayor Florence Laurent followed. They were protesting Google's planned data centre in the Ozans business zone near Châteauroux, and in particular the lack of guarantees on truck traffic, visual impact, water and property values (Korben). The episode is small in scale but large in meaning: it shows what happens when a national fast track is built for investors and the local bargain is left until last.
What is actually planned
The project would be Google's first wholly owned facility in France. It would cover about 195 hectares with eight to ten buildings, and the first building is expected around 2028-2029, with the others from 2031. Google plans a closed-loop cooling system without evaporation, but it would still draw on Châteauroux's drinking water network, and no consumption figures have been published (Korben). Investigative outlet Multinationales reports that Google's subsidiary Tricolore Computing would pay €58.5 million for the land and that demand would exceed 500 MW (Multinationales). The public consultation is only scheduled for November-December 2026, even though the metropolitan council approved the project in June 2025, which feeds the perception that the decision came first.
The strongest case for the fast track
The national approach deserves a fair hearing. France has real reasons to want compute on its own soil: sovereignty, jobs, and a low-carbon nuclear grid that few other large economies can offer. Grid connection is the genuine bottleneck, and slow, unpredictable permitting pushes investment to neighbours. A government that wants projects built by 2028-2029 cannot run each one through years of sequential procedures. Nor is the framework lawless.
On grid access, the energy regulator CRE approved on 7 May 2025 a "fast track" procedure that accelerates, to a 2028-2029 horizon, the connection of very large consumers (400 MW to 1 GW) to RTE's 400 kV network, on sites previously identified by the State and for projects showing credible prospects of completion (CRE). On planning, the 2026 simplification law lets a data centre of particular importance be designated a Major National Interest Project (PINM). In decision 2026-903 DC of 21 May 2026, the Constitutional Council upheld that designation, reasoning that the criteria are sufficiently defined and that the status streamlines planning-document compatibility without altering the substantive conditions for authorisation or bypassing environmental protections (Conseil constitutionnel). The same decision struck down a separate paragraph on derogations from land-use accounting as a legislative rider.
Where the fast track leaves a gap
The problem Étrechet exposes is not legality. It is sequencing and who captures the benefits. The Constitutional Council's reasoning stresses that the PINM label does not change the substantive authorisation conditions. That is reassuring, but it also means the commune's hardest questions, such as water draw, noise, truck routes and visual impact, are answered in the permit stage, which for this project is expected only in early 2027, after land, grid and political commitments have largely been made.
The Châteauroux metropolitan authority, not Étrechet, controls the project and presented it to the Étrechet council only on 9 September, five days before the first resignations. The commune that hosts the buildings and the lines held little leverage over the decision. Multinationales describes the consultation process as tightly controlled, with limited impact on final decisions according to experts, and reports that the grid reinforcement the site needs, doubling a 400 kV line at a cost of about €300 million, is not expected to start before 2032. If that is accurate, the timeline gives local officials a reason for doubt: a project promoted as imminent depends on infrastructure that may be years away.
That tension matters for innovation as much as for democracy. Investors want certainty, and certainty comes from consent that holds. A project whose host commune loses its mayor before the consultation opens is carrying political risk that no fast-track label removes.
A proportionate fix
The answer is not to slow infrastructure with blanket moratoria or open-ended vetoes, which would hurt a sector France is courting and would not serve its sovereignty goals. It is to move the local bargain earlier and make it concrete:
- Publish the numbers before the vote. Water consumption, peak power, heat reuse and truck movements should be public when the host commune is first briefed, not after.
- Give host communes standing. A binding role for the commune that bears the impacts, whether through a contract or a defined share of local revenue, is cheaper than a failed consultation.
- Tie speed to transparency. Fast-track access to grid capacity and PINM status should come with a requirement that impact data and community commitments are on the table before designation.
- Use the water-stress safeguard. The simplification law allows permits to be refused where water resources are under structural strain; regulators should say in advance how it will be applied.
None of this requires rolling back the national framework. The Constitutional Council has confirmed that the framework is constitutional, and CRE's grid procedure addresses a real constraint. What it needs is a front-loaded local process, so that the November-December consultation is a negotiation rather than a ratification.
The takeaway
Five resignations will not stop a project backed by a metropolitan authority and a global operator. But they are an early warning that speed without local consent is not durable. If France wants hyperscale investment to arrive on schedule, it should treat the commune as a partner at the start of the process rather than an objector at the end.