Skip to content
Bundu Foundation

Infrastructure

· · 5 min read

Load-shedding is a design parameter, not a contingency

32% of primary and ~50% of secondary African schools are off-grid. Where there is grid, 4–16 hours of cuts is routine. Solar + LiFePO4 sizing, classroom power budgets, and the SLA adjustments that make support credible.

By Bryan Fawcett

Most school-IT plans drafted in the global north include a single line about "ensure UPS coverage for critical systems". That line is the polite, US-template version of a reality that on most of the African continent reads very differently: power is unreliable for most of the school year, in most of the schools we work with, in most of the countries we ship into.

Treating that reality as a contingency — something the SLA will gently apologise about — leaves you with a deployment that breaks on contact with the school's actual operating environment. The Bundu open standards group treats power as a design parameter equal to connectivity. Here is what that means in practice.

The base numbers

From the European Commission Joint Research Centre's 2025 mapping of African school electricity (Solar power offers a brighter future for African schools):

  • 32% of primary schools in Africa operate off-grid.
  • ~50% of secondary schools operate off-grid.
  • Beyond off-grid: load-shedding regularly removes 4–16 hours of grid power per day in South Africa, Zimbabwe, Nigeria, and others.

Zimbabwe's national grid suffered a total blackout in July 2025; South Africa's load-shedding returned to its highest level in a year in early 2025. These are not edge cases — they are the operating environment.

Power budget per classroom

A useful rule of thumb. A 30-learner classroom needs roughly:

Mode Peak load Average load
Phones / cheap tablets + one teacher projector ~300 W ~150 W
Mixed: tablets + 2–3 laptops ~800 W ~400 W
Full desktop computer lab ~2,000 W ~1,500 W

The phones-plus-projector classroom is the African default and the right design centre. A school IT plan that assumes a desktop lab is the unit of compute is sizing for 5× the power it needs — and 5× the solar and battery cost.

Solar + LiFePO4 sizing

Indicative ranges that the Foundation's standard worksheet uses:

  • Small primary (≤ 200 learners, no labs). 3 kW solar + 5 kWh lithium-iron-phosphate (LiFePO4) battery. Grid or genset failover. Roughly USD 4–7K depending on procurement and country.
  • Medium school (≤ 500 learners, one lab). 5 kW solar + 10 kWh LiFePO4. USD 7–12K.
  • Secondary with multiple labs. 6–10 kW solar + 15–20 kWh LiFePO4. USD 12–25K.

LiFePO4 is the standard chemistry for school-scale deployments; pack costs are now below USD 100 per kWh and falling. Rwanda's national programme aims to deploy solar to roughly 1,000 schools by end of 2025 (Africa Energy Portal, 2025).

The backup hierarchy

Cheapest, most resilient configuration we ship against:

  1. Solar + LiFePO4 battery — primary off-grid spine. Sized for daytime production + overnight reserve.
  2. Grid — top-up where reliable. Treat as bonus, not as primary.
  3. Genset — sized to LMS server + WiFi + a few charging stations, not the whole school. A 3–5 kVA inverter generator (~USD 800–2,500) is enough to keep digital learning alive through an extended cut.
  4. UPS — for the LMS server and main switch only, with ≥ 10 minutes runtime. A whole-school UPS is a budget trap.

Operating discipline

  • Charge during solar peak. Device-fleet charging happens between 10:00 and 14:00 when the panels are running hot — not overnight.
  • Daytime data bundles. Align bandwidth-heavy work to the windows when both power and connectivity are best.
  • Server consolidation. A single Raspberry Pi 5 running Kolibri at roughly 10 W can serve a 30-classroom WiFi with curated offline content for the entire school day. That is not a workaround — it is the right architecture for the conditions.

What this changes in the SLA

If an SLA promises 99.5% uptime in a context with 8 hours of daily load-shedding, the SLA is fiction. The K-12 Support Process framework treats SLAs as load-shedding-adjusted:

  • Define the school's documented load-shedding window (Eskom stage, ZESA schedule, Nigerian grid status).
  • Exclude that window from response-time guarantees; include it in resolution-time accounting.
  • Publish the assumption with the SLA so parents and ministries can read it.

An honest SLA outperforms a fictional one every time. A school IT lead can plan around 95% uptime in school hours; nobody can plan around a promise the grid does not deliver.

Why this is a Bundu open standard

The Foundation publishes Off-grid power resilience as a measurable benchmark with sizing tables, a procurement reference, and an integration-test checklist. School operators, ministries, and partners can audit against it without buying a closed accreditation. Built for Africa.

Sources cited in this note

Back to newsroom