São Paulo Battles the Duplicate Image Problem Swamping City Databases — and Other Megacities Are Watching

São Paulo's city government is sitting on tens of millions of duplicate photographs, scanned documents, and surveillance stills clogging the municipal data infrastructure managed by the Empresa de Tecnologia da Informação e Comunicação do Município de São Paulo, known as PRODAM. The problem, which officials at the Secretaria Municipal de Inovação e Tecnologia have been publicly flagging since early 2026, is not cosmetic. Storage redundancy is slowing permit-processing times in districts like Mooca and Lapa, where urban renewal programs generate hundreds of georeferenced images per city block every month.

The timing matters because the Prefeitura under Mayor Ricardo Nunes committed in its 2026 digital governance plan to cutting municipal IT operating costs by streamlining data architecture ahead of the city's 2028 infrastructure investment cycle. Duplicate image files — generated by everything from the CET traffic camera network on Avenida 23 de Maio to the digitised building-permit ledgers held at subprefeituras across the Zona Leste — represent one of the least glamorous but most costly drains on that effort. Tech consultants working on similar municipal projects put storage sprawl among the top three drivers of unnecessary cloud expenditure for cities above ten million residents.

What São Paulo Is Actually Doing

PRODAM has been piloting a deduplication protocol since March 2026, starting with image archives linked to the Nota Fiscal Paulistana tax program and the city's georeferenced zoning database, GeoSampa. The pilot targets files with identical hash values — a standard cryptographic fingerprint check — before moving to perceptual hashing tools that can catch near-duplicate images taken seconds apart by the same sensor. GeoSampa alone holds satellite and aerial photography layers dating to 2004, and internal assessments described in city council budget hearings in April suggested that between 18 and 22 percent of stored image files across several municipal systems were exact or near-exact duplicates.

The Secretaria Municipal de Inovação ran a public procurement process in the first quarter of 2026, publishing a call for vendors capable of processing at least five million image files per week through automated deduplication pipelines. The contract value, according to documents posted on the city's procurement portal, was set at a ceiling of R$4.2 million for a twelve-month engagement. That figure reflects the scale: São Paulo is not dealing with a startup's photo library, but with decades of layered municipal record-keeping across more than 96 subprefectures and districts.

How the City Compares Globally

Seoul's Smart City Division tackled a structurally similar problem in 2023, when its integrated CCTV network across 25 autonomous districts had generated an estimated 40 petabytes of partially overlapping footage and associated metadata thumbnails. The city invested in a purpose-built deduplication layer inside its public cloud contract with a domestic provider, reportedly cutting active storage use by roughly 30 percent within the first year, according to the Seoul Digital Foundation's 2024 annual report. Amsterdam's Datapunt initiative, which manages open datasets for the municipality, publicly documented a comparable exercise on its street-level imagery archives in 2022, using open-source perceptual hashing libraries before migrating to a commercial solution.

São Paulo's R$4.2 million contract ceiling is modest against Seoul's investment but proportionate given that the initial PRODAM pilot is scoped narrowly around three legacy systems rather than the entire municipal stack. Cities like Bogotá and Mexico City, managing comparable population densities and multi-decade analogue-to-digital migration backlogs, have not yet published comparable deduplication programs at the municipal level, according to reporting by tech policy researchers cited in regional urban governance forums held in Brasília in May 2026.

For residents and businesses who interact daily with municipal systems — applying for construction licences in Pinheiros, registering food-service permits in the Mercado Municipal area near Rua da Cantareira, or accessing property records online — the practical payoff would be faster query times and fewer errors caused by systems returning conflicting versions of the same image. PRODAM has indicated that a second phase, covering the CET camera archive and the SP Urbanismo real-estate database, is planned for the second half of 2026, assuming the first-phase audit results meet internal benchmarks. The city's ability to execute will be a closely watched reference point for other Latin American capitals planning their own digital-infrastructure audits before the decade closes.