Africa AI LedgerThe record of artificial intelligence across Africa

Africa AI Ledger · News

Ethio Telecom Trials AI-Enabled Mobile Network

A pre-commercial radio-network deployment with ZTE reportedly increased 4G capacity by 11% and peak-hour throughput by 40%, though the results remain vendor reported.

By Africa AI Ledger Desk · 4 November 2026 · Event: 4 November 2026 · 2 min read
AI-directed radio beams distributing capacity across mobile towers and Addis Ababa at dusk

Background generated in ChatGPT; Africa AI Ledger editorial template

Ethio Telecom and ZTE have completed a pre-commercial trial of an artificial-intelligence-enabled radio access network designed to improve the performance of Ethiopia's existing 4G infrastructure.

The deployment uses ZTE's AIREngine system and intelligent beam scheduling to allocate radio resources in response to changing traffic. In its 4 November announcement, ZTE reported an 11 per cent increase in 4G capacity and a 40 per cent improvement in peak-hour user throughput compared with a conventional configuration.

Those figures are promising, but they are vendor-reported trial results. The announcement does not identify the number of sites, duration of testing, traffic conditions or an independent evaluator. It also does not establish whether the same improvement would persist across rural networks, dense city districts and different equipment configurations.

Radio access networks connect phones and other devices to the wider telecommunications system. Operators traditionally plan capacity through engineering rules, historical demand and manual optimisation. An AI-enabled system can adjust beams and resources more dynamically when demand shifts across time and location.

For Ethiopia, improving existing infrastructure may be more practical than relying only on new towers or spectrum. Ethio Telecom serves more than 90 million subscribers, according to the partners. Even modest improvements can affect a large number of users if they are deployed broadly and do not increase operating costs disproportionately.

The trial should be judged through measures beyond peak speed. Consistency for users at the edge of coverage, call reliability, energy consumption, congestion during unusual events and performance on lower-cost devices all matter. An optimisation system can improve average throughput while still leaving some communities poorly served.

AI in network control also introduces governance questions. Engineers need visibility into why the system reallocates resources, a safe method for overriding it and clear responsibility when optimisation creates instability. Models trained on historical traffic can under-prioritise emerging neighbourhoods or rural demand that is not well represented in past data.

The strongest deployment case would compare the AI configuration with conventional engineering over a longer period and publish results across several site types. Energy and maintenance costs should be included, because extra computing may reduce the financial value of capacity gains.

Ethio Telecom links the project to its Next Horizon: Digital & Beyond 2028 strategy and Ethiopia's Digital Ethiopia 2030 agenda. Those strategies create a policy frame, but commercial rollout will depend on procurement, compatibility with existing equipment and evidence that the improvement is durable.

This is a useful infrastructure experiment because it applies AI to a physical constraint that affects everyday digital access. It is not yet proof of a nationwide transformation. The next milestone is an independently documented expansion showing where the system works, what it costs and whether customers experience a measurable improvement.

Source: ZTE and Ethio Telecom.

Source & attribution

Source: zte.com.cn

Related coverage