The Hidden Cost of Duplicate Images: What the Numbers Reveal About Zurich's Digital Storage Crisis

Zurich's universities, hospitals and municipal archives are collectively storing vast quantities of duplicate image files, a problem that has quietly inflated IT budgets across the city's public and private sectors. Internal assessments circulating among Swiss digital infrastructure managers suggest that between 30 and 40 percent of all images held on institutional servers are exact or near-exact duplicates—a figure that translates directly into unnecessary storage costs running into the millions of francs annually.

The issue has sharpened in 2026 partly because of the UBS integration of Credit Suisse assets, which forced a reckoning with legacy data architecture across dozens of merged business units. When two major banks collapse into one, their image libraries—product photographs, compliance screenshots, scanned documents—do not merge cleanly. They stack. IT teams at the combined entity's Zurich headquarters on Bahnhofstrasse have reportedly been grappling with precisely this kind of redundancy since the merger was confirmed in 2023, though the full deduplication exercise is still underway.

What the Data Actually Shows

Commercial cloud storage in Switzerland costs roughly CHF 0.023 per gigabyte per month on standard enterprise contracts, according to published rate cards from Swiss providers as of early 2026. A hospital the size of UniversitätsSpital Zürich, which handles imaging for tens of thousands of patients a year across its campus in Rämistrasse, can accumulate petabytes of medical imagery. Even a 20 percent duplication rate on a two-petabyte archive translates to roughly 400 terabytes of redundant data—a storage bill of around CHF 220,000 per year for files that carry zero clinical or administrative value.

ETH Zurich, ranked consistently among the world's top ten technical universities, published a working paper in March 2026 through its Computational Science and Engineering department examining hash-based image deduplication at scale. The paper found that perceptual hashing algorithms—which detect visually similar rather than byte-identical images—can reduce image libraries by up to 47 percent without any loss of unique content. That figure has circulated widely in Zurich's civic tech community, appearing in presentations at the Impact Hub on Sihlquai and in workshops organised by the city's Digital Zurich 2025 strategy office at Stadthaus.

The problem compounds when organisations operate across multiple departments with separate storage buckets. The City of Zurich's own digital archive, administered through the Stadtarchiv on Neumarkt, contains historical photograph collections that were digitised in multiple overlapping grant-funded projects between 2010 and 2022. Project managers have acknowledged publicly that coordination between digitisation rounds was imperfect, meaning thousands of images exist in two or three versions—different resolutions, different file formats, same underlying photograph. The exact count has not been publicly released, but comparable municipal archives in cities like Vienna and Amsterdam have reported duplication rates above 25 percent after similar multi-phase digitisation efforts.

The Financial and Environmental Arithmetic

Storage is not just a budget line. Swiss data centres consumed an estimated 1.5 terawatt-hours of electricity in 2024, according to figures from the Swiss Federal Office of Energy. Redundant data contributes directly to that load. Zurich's climate action framework, the Netto Null 2040 target adopted by the city council, specifically identifies data centre efficiency as a lever for reducing the municipal carbon footprint. Deleting or consolidating duplicate image files is among the lowest-cost interventions available—requiring processing time and algorithm deployment rather than physical infrastructure investment.

Several Zurich-based startups working out of the Technopark on Technoparkstrasse have built commercial products around exactly this problem, offering deduplication-as-a-service to mid-sized enterprises that lack in-house data engineering capacity. Pricing for such services typically runs between CHF 5,000 and CHF 40,000 for a one-time library audit, depending on volume, with ongoing monitoring contracts available on annual subscription.

Organisations sitting on unaudited image libraries should start with a perceptual hash scan of their largest storage buckets before the end of Q3 2026—the point at which many Swiss IT departments finalise their 2027 budget submissions. A clear duplication rate gives procurement teams a defensible number when arguing for either expanded storage or reduced spend, and it gives compliance officers evidence that data minimisation obligations under the revised Swiss Data Protection Act, which took full effect in September 2023, are being actively addressed.