I published a contribution in Lyon Mag on the data-centre projects in Greater Lyon: their value for the area, their electricity footprint and the conditions that should come with them.

The starting point: two projects in the metropolis

In Dardilly, Phoenix Group and DC Max plan an 18 MW centre; in Vénissieux, UltraEdge is aiming for 8.4 MW on a site where an extension is already running. Local benefit cannot simply be asserted: it depends on location, actual consumption, jobs created, services provided to local players and whether the heat can be put to use.

The argument

• From announced megawatts to actual use. 18 MW is the equivalent of 9,000 two-kilowatt heaters switched on at once. But what exactly is being measured — the servers alone, or the whole building including cooling? According to industry sources, recently connected centres use on average only 20% of the capacity they requested, and RTE notes that a single project can file multiple connection studies without ever materialising, which complicates grid planning.
• Benefits worth scrutinising. The mayor of Dardilly announces ten on-site jobs; UltraEdge cites around a hundred direct or subcontracted jobs in France and a supervision centre based in Lyon. The question is what these facilities actually bring in return for the resources they use.
• A grid that needs preparing. RTE is keeping its plan to transform the Lyon grid through 2035, with each connection on its own timetable. As Éric Sellin (INSA Lyon) points out, even 10 MW may require reinforcement depending on how the local network is organised.
• Planning where they go. The Metropolis is preparing a dedicated data-centre plan. Not all sites are equal in terms of grid, land, recoverable heat and proximity to users: the capacity map published by Enedis could be cross-referenced with customer needs and outlets for heat.
• Heat is less recoverable than often assumed. The law requires centres of at least 1 MW to recover their heat, but it often comes out at low temperature: a heat pump is needed, which itself consumes electricity, along with an outlet prepared in advance with the heat-network operator.

What I propose

Move from announcements to verifiable commitments: a public sheet for each project (server capacity, connection capacity, timetable, actual consumption); local agreements setting out expected jobs, local purchasing and services to SMEs, followed by an annual review; on public land, a building lease that keeps ownership with the local authority and anticipates possible abandonment. Volunteer sites could also explore demand flexibility with their grid operator, along the lines of the DCFLEX programme in which RTE takes part.

Greater Lyon has good reasons to welcome these investments, but not on just any terms: that is how the megawatts behind the cloud can become a genuine local project.

→ Read the contribution on Lyon Mag