Econet InfraCo, the tower infrastructure arm of Econet Wireless Zimbabwe, has backed its push to use artificial intelligence to tackle rising diesel bills with numbers. Its first half-year results as a listed company report a 30% reduction in fuel consumption in its power business. For telecom operators and other heavy generator users, the path it took is worth studying.
The Problem Every Tower Operator Knows
Zimbabwe’s grid is unreliable, so tower sites lean on diesel generators to stay online. Middle East tensions have pushed up the local price of diesel, although the company says it has seen no supply disruptions. Fuel is one of the highest costs of keeping a tower estate running, so even small efficiency gains matter.
Start With How Each Generator Is Used
The company built its AI Fuel Mode tool with sister company Cassava Compute. The team studied how and when each generator was deployed, then used those patterns to build a fuel management model. The aim is to cut unnecessary runtime while keeping sites powered, which goes further than simply tracking litres burned.
Pair AI With Solar and Batteries
The 30% figure is not credited to AI alone. In its half-year statement, the company attributes the reduction to site solarisation, AI-enabled fuel management, advanced battery cycling and periods of better national grid supply. The lesson for other operators is that AI works best as one layer in a wider energy strategy, with solar and storage lowering the load that generators must carry.
Build on a Wider AI Toolkit
Fuel optimisation sits alongside other AI work at the company. It is using AI for predictive generator maintenance and is developing a remote monitoring system and a digital twin of its towers. These tools support better visibility and uptime, which matter as much as fuel savings when a site must never go dark.
Roll Out in Phases
InfraCo describes its AI Fuel Manager rollout as phased. Management said it is expected to deliver lower energy consumption and better site uptime as deployment widens. A staged approach lets an operator test results at a few sites before committing across a whole network.
What Others Can Learn, and What to Watch
Cassava Compute has said the platform will be offered to other operators and could be adapted for industries such as mining and manufacturing. The company also describes the tool as the first of its kind from a tower operator globally, a claim that comes from the company and that this report has not independently confirmed.
Readers should also note that the headline figure covers fuel consumption across several measures. The portion due to AI specifically has not been published. Operators copying the model should measure their own baseline and track each intervention separately.
Even so, the core approach is clear and repeatable. Understand generator behaviour with data, let AI optimise it, back it with solar and batteries, and expand step by step.
