Enzi President Calls for Lower AI Inference Costs Across Africa

Reducing the cost of artificial intelligence inference is essential if businesses and individuals across Africa are to use the technology consistently, Enzi president Oji Udezue has said.

Speaking to Moonshot by TechCabal in Lagos, Udezue said Africa’s ability to benefit from AI would depend on lowering the cost of using it. He highlighted the gap between the expense of running AI systems and what many Africans earn.

“The future of AI really goes through inference,” Udezue said.

Recurring costs affect AI adoption

Inference is the process through which an AI model generates responses or performs tasks. Unlike a one-time purchase, these costs recur whenever a system handles activities such as answering customer enquiries or supporting employees.

Udezue’s argument is that the repeated expense can influence whether an AI application remains practical beyond occasional use. A service that is affordable for a limited number of queries may become too costly when used continuously in a workplace or to serve a large customer base.

He also warned that access to an AI tool has limited value if people cannot afford to use it regularly. Lower inference costs, he said, are therefore necessary to make AI more accessible to businesses and individuals throughout the continent.

“Our goal is to make AI cheaper for Africa and America,” Udezue said, describing Enzi’s objective.

Infrastructure and local relevance remain important

Lower prices alone may not be enough to support broad AI adoption. The World Bank’s Digital Progress and Trends Report 2025: Strengthening AI Foundations identifies connectivity, computing resources, locally relevant data and digital skills as foundations for the use of AI.

Udezue also emphasised practical demonstrations as a way to show developers what they could achieve with the technology. The source of his remarks did not provide details of the demonstrations.

For African users, affordability is also linked to whether AI systems are suited to their circumstances. Cheaper services would have limited value if models do not work in local languages or provide responses that fit users’ needs.

The discussion places recurring inference costs alongside infrastructure, skills and local-language capability as key considerations for expanding access to AI across Africa.