Nairobi Agencies Face Massive Digital Records Cleanup Bill

Kenya's public sector has a data problem it can no longer ignore. Across government ministries, county offices, and the sprawling digital infrastructure supporting the Huduma Namba national identity programme, administrators are grappling with vast libraries of duplicate scanned images, ID photos, land registry scans, health records, that are clogging storage systems, slowing down services, and eating into already stretched IT budgets. The question now is who decides what gets deleted, what gets kept, and who pays for the clean-up.

The issue has sharpened this year because the William Ruto administration is under IMF-linked fiscal pressure to cut wasteful government expenditure. Every server rack running redundant data has a cost. The Kenya National Archives and Documentation Service, headquartered along Moi Avenue in the central business district, has been in discussions with the ICT Authority, based in Westlands, about a unified deduplication framework that would set binding standards for how agencies handle image data. Those talks, which began in earnest in the first quarter of 2026, have yet to produce a published policy.

The Scale of the Problem on the Ground

Storage costs are not abstract. Commercial cloud storage from providers operating data centres in Nairobi, including facilities along Mombasa Road and the Sameer Business Park corridor, runs at roughly Ksh 8 to Ksh 15 per gigabyte per month for enterprise contracts, depending on redundancy tiers. Government agencies holding tens of thousands of unprocessed scan batches can accumulate hundreds of terabytes of duplicate files over a single budget cycle. The arithmetic is punishing for a treasury already managing a constrained fiscal envelope.

The Nairobi Metropolitan Services land records digitisation push, which covers title documents for properties across Starehe, Embakasi, and Dagoretti sub-counties among others, is one programme where the duplication problem is acute. Field digitisation teams working under tight deadlines have historically scanned documents multiple times to guard against data loss, a sensible precaution that has now become a storage liability. Sources familiar with the programme, speaking on background because the review is ongoing, say some document batches contain three or four identical image files for the same physical page.

The Silicon Savannah startup community has noticed the opening. Firms operating out of Nairobi Garage on Ngong Road and the iHub in Kilimani have begun pitching AI-assisted deduplication tools to county governments. These systems use perceptual hashing, a technique that identifies visually identical images even when file names or metadata differ, to flag duplicates for human review before permanent deletion. The pitch is cost savings; the concern among data governance advocates is that automated deletion, without proper audit trails, could irreversibly destroy official records.

The Decisions That Cannot Wait

Three choices will define what the next six months look like. First, the ICT Authority needs to publish the deduplication framework it has been drafting, giving agencies a legal basis to act rather than simply accumulate. Without that framework, no ministry head is likely to authorise mass deletion of government image files, no matter how obviously redundant they are. The political risk of erasing something that later turns out to matter is simply too high in an environment where the Gen Z-era scrutiny of government conduct remains fierce.

Second, the National Treasury must decide whether deduplication work qualifies for funding under the ongoing digital infrastructure budget line or whether agencies must absorb the cost internally. The difference matters enormously for smaller county offices, Makadara and Mathare sub-county administrations, for instance, are unlikely to have the internal IT capacity to run deduplication projects without dedicated resourcing.

Third, and perhaps most consequentially, policymakers must determine whether citizens have any right to know when their biometric or document images are being reviewed, merged, or deleted. Kenya's Data Protection Act 2019 establishes principles around personal data handling, but guidance specific to bulk image deduplication in government systems remains thin. The Office of the Data Protection Commissioner, which operates under the Ministry of Information, will need to weigh in publicly before agencies can move with confidence. The clock is ticking on all three fronts.