France's grid queue is the data center's first asset
With transmission capacity saturated in key regions, a position in RTE's connection queue determines whether a French data center gets built — and how fast.
In the saturated reaches of France's transmission network, a position in RTE's connection queue is what separates a data center that gets built from a site that never will. RTE operates the public electricity transmission system and runs the queue where those decisions are made, and a legal briefing by Orrick, Herrington & Sutcliffe published by Data Center Dynamics maps how the framework works in practice. The authors — Geoffroy Berthon, a partner in Orrick's public law practice in Paris, and associate Charlotte Berrat — say they concentrate on the issues that surface most often.
The standard route sits under RTE's grid connection request procedure and requires four agreements: a technical and financial proposal called the PTF, a grid connection agreement, an operation agreement, and a network access agreement. Each carries two layers — general conditions that RTE publishes and the CRE must approve, and specific conditions particular to the project. Because the CRE signs off on the general conditions, the commercial terms of connection carry a regulatory overlay; the project-specific conditions are where the negotiation happens.
Under the current legal framework, the decisive act is acceptance of the PTF, which places the project in RTE's connection queue and sets the priority order in which capacity is allocated. In saturated areas, securing and maintaining that position is a key prerequisite — in practical terms, a lost queue position is a lost project, whatever leases have been signed and whatever equipment has been ordered. That puts the legal work on the critical path: the connection contracts set the schedule, and a lapse in the queue unwinds the financing, the permitting, and the construction plan. The queue position has to be treated as an asset, because it is what converts a site into a project.
RTE has built an alternative path for the largest consumers in the form of the HTB3 initiative, which connects large-scale consumption facilities to the 400kV very high voltage network within three to four years for power demand typically between 400MW and 1GW — hyperscale territory.
A state-chosen fast lane
The trade-off is that sites come preselected by a State process that identifies suitable locations using three criteria: availability of industrial land, proximity to the 400kV network, and the capacity to offer high power rapidly, including through redispatching. The developer's site choice is therefore limited to a state-drawn list, and the selection itself becomes a form of planning consent.
Under the Fast Track, the central document is the grid connection commitment — the functional equivalent of the PTF — and it sets the technical and financial conditions of the connection, with a component for network extension works and a capacity component reflecting the additional cost of accelerated power delivery. The two-part price makes speed an explicit, priced option rather than a scheduling hope.
The French system is a clean expression of what this publication has argued: grid access is the scarce asset, and power positions are priced as contingent claims before leases are signed. The standard queue rations capacity by seniority; the Fast Track rations it by state selection. Both are rationing mechanisms, and the difference is who decides.
The standard queue rations capacity by seniority; the Fast Track rations it by state selection.
The Fast Track does not solve scarcity; it relocates the rationing decision, shifting the developer's binding constraint from waiting for a queue to winning a place on a state list — a different project risk. The diligence question for investors changes accordingly.
For investors underwriting French data centers, the first question is where the project sits in the queue. If the Fast Track is the route, the question is whether the site is on the State's list and what the capacity component actually costs. Those answers determine whether the project returns capital more than the renderings do.