Industrial IoT in Uganda — one platform, from the sensor to the number you can act on.
addaNet is our own platform — hardware, firmware and software built in-house rather than assembled from three vendors' products. On a Ugandan site the practical consequence is that it reads what you already have: PLCs, energy meters, generator controllers, flow meters. New instruments go in only where a measurement is genuinely absent, and the whole record keeps running on battery when the supply doesn't.
Four places the platform earns its keep.
Interruptions, costed
UEDCL has told Parliament that outages almost doubled and restoration went from about 12 hours to nearly 20. Grid events timestamped at the intake, linked to what the line was doing, so the cost stops being an impression.
What interruptions cost →Process records that hold up
UNBS Q-Mark certification requires process monitoring records, cleaning records and calibration records from the first re-certification onwards. Continuous logging makes them exist without anyone filling in a book.
UNBS record keeping →Standby plant you can rely on
Run hours, fuel, litres per kWh, changeover timing and start failures found during exercise runs rather than during production.
Generator monitoring →Water, from borehole to discharge
Abstraction, tank levels, pump behaviour, zone metering for loss reduction, and effluent flow and quality proxies where a discharge permit applies.
Water telemetry →Engineered for Ugandan site conditions.
- Cheap power that stops without warning. Uganda has surplus generation — reliable capacity near 1,550 MW against a February 2026 peak of about 1,337 MW, with roughly 270 MW exported — and ERA's July–September 2026 industrial tariffs are UGX 308.1/kWh large industrial and 207.7 extra-large. The problem is reliability, not price, and reliability problems are invisible on an invoice. That makes measurement the whole game.
- Unstable voltage, not just outages. Around four in ten firms in EPRC's 1,152-firm Business Climate Index reported harm from the distribution handover, naming more outages, unstable voltage and degraded power quality. Voltage events damage drives and motors without ever stopping the line, so they have to be logged deliberately.
- The record has to survive the event. Nodes are battery-buffered, store locally and back-fill, so an interruption appears as a logged event rather than as missing data.
- Industry concentrated, but spread inside the park. Namanve now counts roughly 480 factories allocated and close to 280 operating across 2,400 hectares. Cellular where it is solid, LoRaWAN for sites spread beyond it, wired where the plant supports it.
- Boreholes and solar pumping are normal. Water infrastructure is often off-grid and remote, so nodes run from the same panel as the pump and report over whatever link exists.
What an addaNet deployment in Uganda looks like.
- Remote assessment. Site layout, panel and controller photographs, nameplates, and what can already be read electronically. Most plants have more available signal than they expect.
- Written scope and cost band before anything is ordered, separating what is read from existing equipment from what needs a new instrument.
- Hardware pre-configured in Johannesburg and shipped ready to commission, so the on-site work is installation rather than configuration.
- Installation by your technicians or a local partner under live remote guidance from our engineers during East Africa Time working hours.
- Dashboards and alerts built around the decisions people actually make — the screen a supervisor opens at shift start, and the alert that should reach a phone at night.
- Remote support and iteration from the team that built the platform.
First deployments are typically live within two to six weeks of the assessment call.
Ugandan sites, and the first signal
| Site type | First measurement | Who reads the record |
|---|---|---|
| Factory in Namanve (~480 firms, ~280 producing) | Counts + run state + mains presence | Plant management |
| Site under UEDCL distribution (restoration ~20 h) | Outage duration + restart time + genset state | Operations, finance |
| Q-Mark certified plant (Scheme B) | Process monitoring records + calibration records | UNBS auditor |
| NWSC zone (NRW 34%, Kampala 37–46%) | Bulk input + DMA night flow | NWSC, VEI programme |
| Agro-processor (coffee, grain, sugar) | Drier temperatures + mill vibration | Factory management |
| Large user considering bulk supply (≥1,500 kVA) | Load profile, metered per feeder | ERA, the supplier |
Where this lands first in Uganda.
Food, beverage and dairy — the largest group on the UNBS scheme, where process records and cold chain both carry obligations · Grain milling and edible oils — continuous plant where energy per tonne and downtime decide the margin · Plastics, packaging and building materials — fast lines where a small stoppage rate compounds · Steel and metal fabrication — drive-heavy loads where sub-metering and power quality matter together · Pharmaceuticals — storage and process conditions with documented evidence · Water utilities, umbrella authorities and estates — zone metering, boreholes, tanks and prepaid verification.
Industrial IoT in Uganda — common questions
Can addaNet read our existing PLCs, meters and generator controllers?
That is the normal starting point. Modbus, OPC, pulse outputs and analogue signals are read directly, and generator controllers from DSE, ComAp and similar expose run state, hours, load and fuel the same way. On most plants a good share of what you want to see already exists somewhere and simply is not retained — which keeps a first project sized to the gap rather than to the site.
Does monitoring keep recording through an interruption?
Yes, and in Uganda that is the feature that decides whether the data is worth having. Nodes are battery-buffered and log locally, back-filling when the link returns. The interruption is captured as an event with a start, an end and whatever the plant was doing on either side of it — which is the record you need to cost it.
Cloud or on-premise?
Cloud for most sites, because it removes a server and its backups from a plant with enough to maintain. On-premise or hybrid where a client's policy requires data to stay on site, or where connectivity makes local-first sensible. Since we build the platform, both are configurations rather than different products.
Can it produce the records UNBS asks for?
That is one of its main uses in Uganda. UNBS scheme requirements from the first re-certification onwards include process monitoring records, cleaning schedules and records, and verification or calibration of measuring equipment. Continuous logging with batch context makes those records exist automatically and exportable for any period — see UNBS certification record keeping.
How does support work in Uganda?
Our engineering team is in Johannesburg, working the same hours as your plant in East Africa Time, with support on WhatsApp. Installation is by your technicians or a local partner under live remote guidance, and hardware ships pre-configured. First deployments typically go live within two to six weeks of the assessment call.
Where should a Ugandan plant start?
With the intake and one line. Grid events at the intake plus production and energy on a single line gives you a complete picture of one part of the operation, including what interruptions actually cost it. That is usually enough to decide whether the rest of the plant is worth instrumenting — and it is a far better first step than a site-wide rollout that produces dashboards nobody opens.
See your Ugandan site on one dashboard.
Tell an engineer what you run and what you can already read. You'll get an honest view of what to monitor first and what it would cost.