From Sugar to Servers: Normandy's Derelict Cagny Plant Eyed for an AI Data Center

A shuttered 33-hectare sugar plant in Cagny, near Caen, is slated to become an AI data center. Brownfields is in talks to buy and rebuild, with grid and permits setting the pace.

Published on: Feb 11, 2026
From Sugar to Servers: Normandy's Derelict Cagny Plant Eyed for an AI Data Center

Derelict Sugar Factory Near Caen Poised for AI Data Center Redevelopment

A 33-hectare sugar refinery in Cagny, east of Caen (Calvados), is being lined up for conversion into an AI data center. Real estate developer Brownfields is in talks to acquire the Saint Louis Sucre site, with plans to demolish and rebuild.

The plant operated from 1950 until its closure in 2020. The two parties have worked together before, including the conversion of a Saint Louis Sucre facility in Marseille into a film studio, suggesting alignment on process and expectations.

Why this site works for data center use

  • Scale: 33 hectares gives room for a multi-phase campus, separate utility yards, and logistics access.
  • Power: A 220 kV line runs nearby, and there's a direct link to the Flamanville nuclear power station, supporting low-carbon supply and capacity planning. For context, see Flamanville.
  • Existing infrastructure: On-site wastewater treatment (currently offline) could be refurbished for cooling discharge or industrial reuse, reducing new civil works.
  • Acquisition cost: Reported €15-€20 million for the land and assets creates headroom for grid, cooling, and envelope capex.

What developers and contractors should expect

Negotiations with Saint Louis Sucre are ongoing. Assuming a clean transaction and standard remediation, demolition and site prep could begin quickly, but utility delivery will drive the critical path.

  • Grid connection: Engage RTE early for 220 kV substation siting, protection studies, and phasing. Expect multi-year timelines without fast-track agreements.
  • Permitting: Environmental impact assessment, noise, heat rejection, water discharge, and traffic plans will be scrutinized, especially given proximity to existing towns.
  • Water and cooling: Validate options across air-cooled, adiabatic, or closed-loop systems. Assess refurb of the wastewater plant for non-potable uses or thermal discharge routing.
  • Demolition and remediation: Former sugar processing can mean soils affected by organics and byproducts. Budget for remediation, backfill, and compaction to data hall tolerances.
  • Phasing: Design for modular expansion (power blocks, chilled water or CDU skids, prefabricated electrical rooms) to align capex with AI demand ramps - see AI Design Courses for modular approaches and prefabrication strategies.

Market context

France is scaling AI compute capacity with a national push for large-footprint data center development. The government named 35 "turn key" sites for rapid builds, though Cagny isn't on that list, which may add time for approvals and grid studies but doesn't block development.

For investors, this is a classic brownfield play: low land basis, existing utility corridors, and strong power provenance via nuclear. For GCs and MEPs, the opportunity sits in fast-track design-build, standardized blocks, and early utility coordination.

Practical next steps for interested parties

  • Run a preliminary power study with RTE to confirm capacity, interconnection costs, and energization dates.
  • Commission environmental and geotechnical surveys focused on demolition sequencing, soil quality, and bearing capacity for slab and equipment pads.
  • Model water balance across cooling scenarios and price the reactivation or replacement of the wastewater facility.
  • Structure a phased masterplan (zoning, roads, yards, stormwater, security) that enables quick commissioning of the first halls.
  • Align incentives: explore local tax agreements and employment commitments to smooth permitting and community relations.

Numbers that will shape the pro forma

  • Land and legacy asset purchase: €15-€20m (reported).
  • Substation and HV works: substantial portion of phase-one capex; sensitivity-test multiple delivery dates.
  • Cooling strategy: capex and water costs swing widely; lock this early with the power profile.
  • Schedule risk: grid lead times and permit paths will define NPV more than build speed inside the fence.

If your team is preparing for AI-focused build programs and needs role-specific upskilling, explore curated options here: AI courses by job.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)