Paris's Duplicate Image Problem: The Numbers Driving a Digital Overhaul Across the City's Archives

Paris's public digital infrastructure is drowning in copies of itself. A cross-institutional audit completed in spring 2026 by the Direction des Systèmes d'Information de la Ville de Paris identified more than 14 million duplicate image files spread across the city's administrative servers — redundant assets accumulated over two decades of uncoordinated digitisation drives, Olympic legacy archiving and post-pandemic remote-working migration. The finding has put data governance at the top of the agenda for bodies ranging from the Bibliothèque nationale de France to the Grand Paris Express metro authority.

The timing matters. Paris 2024's organisers transferred an estimated 2.3 petabytes of photographic and video content to the city's cultural stewardship bodies after the Games closed, flooding systems that were already struggling with legacy backlogs from the Seine-Saint-Denis urban regeneration programme. At the same time, the Grand Paris Express project — which spans 68 new stations and four new metro lines — has generated its own torrent of engineering schematics, drone survey images and public-consultation visuals, many filed in multiple formats by different contracting teams. When storage budgets tightened under the current Macron government's 2026 public-spending review, administrators could no longer afford to ignore the redundancy problem.

The Scale of the Problem in Hard Numbers

Storage is not cheap. Enterprise-grade archival space used by institutions on the city's shared infrastructure runs at roughly €180 per terabyte per year when maintenance and security contracts are included, according to publicly available procurement frameworks published by the Agence nationale de la sécurité des systèmes d'information. At that rate, a conservative estimate puts the annual carrying cost of the 14 million flagged duplicates — assuming an average uncompressed file weight of 8 megabytes — at somewhere north of €20 million citywide. That figure excludes the labour cost of cataloguers and archivists at institutions such as the Musée Carnavalet on the Rue de Sévigné, where staff have spent months manually reconciling image metadata after two separate digitisation contractors delivered overlapping batches of the same historical prints.

The Carnavalet is not alone. The Paris urbanism agency Apur, based near the Place du Panthéon, began its own internal deduplication project in January 2026 after discovering that nearly 40 percent of its urban-photography holdings existed in at least two identical or near-identical copies across different departmental drives. The Institute for Urban Planning research teams working on the Seine Rive Gauche regeneration corridor — stretching from the 13th arrondissement toward Ivry-sur-Seine — flagged the same pattern: aerial survey images commissioned at different project phases had been saved under different file names, fooling basic cataloguing software into treating them as distinct assets.

What Comes Next for Paris's Digital Housekeeping

The city's response is taking shape through a procurement process launched in March 2026 under the Direction des Affaires Culturelles. Three technology vendors have been shortlisted to deploy perceptual-hashing deduplication tools — software that identifies visually identical images even when file names, formats or metadata differ. A decision is expected before the end of the third quarter of 2026, with a phased rollout beginning at the Hôtel de Ville's central document management system before expanding to partner institutions.

For institutions on the Grand Paris Express corridor, the stakes are partly legal. Construction and environmental-impact documentation must meet specific retention standards under French administrative law, meaning files cannot simply be deleted without verification. The Société du Grand Paris, which oversees the metro expansion, confirmed in its 2025 annual report that it had begun a separate records-management exercise, though it gave no figure for the volume of duplicated assets identified.

Practically, archivists at smaller municipal bodies — including neighbourhood mairies across the 10th and 11th arrondissements, which digitised local planning records during the 2020–22 lockdown period — are being advised to freeze new uploads to shared drives until the deduplication framework is in place. The priority, officials have made clear, is to stop the duplication rate from compounding further before the audit tools are deployed. One thing is already settled: the era of treating server space as essentially free is over in Paris, and the reckoning is arriving file by file.