MTN Group plans to transform its mobile tower network across Africa into a large-scale artificial intelligence infrastructure platform by installing GPU-powered computing systems at telecom tower sites.
The plan was revealed by Charles Molapisi during an event hosted by law firm Bowmans in Johannesburg.
According to Molapisi, MTN wants to replace traditional baseband units at cellular towers with open GPU infrastructure capable of handling both mobile network operations and AI inference workloads at the same time.
AI inference refers to the process where trained artificial intelligence models analyse data and generate responses or decisions in real time.
Today, telecom towers mainly use dedicated hardware designed only for managing radio and mobile network traffic. MTN’s new approach would allow the same infrastructure to process AI tasks much closer to users instead of sending workloads back to distant data centres.
The company described the idea as a “distributed AI grid” spread across its tower network.
Molapisi said one major benefit would be lower latency, meaning faster response times for digital services. He gave the example of children playing online games such as PlayStation in a housing estate. With edge computing installed at a nearby tower, processing could happen locally instead of routing data through distant data centres and back again.
MTN believes this approach could improve network efficiency while reducing pressure on central infrastructure.
The tower-based AI network forms part of MTN’s wider artificial intelligence strategy. Earlier this year, the company confirmed plans to build two AI-enabled data centres in South Africa and Nigeria.
Molapisi said MTN’s AI ambitions now cover several areas, including semiconductor procurement, cloud infrastructure, data centre development, AI model management, and application partnerships.
The company is also expanding terrestrial fibre networks across Africa, including in countries where it does not operate mobile services. MTN says the goal is to strengthen the continent’s digital infrastructure and improve connectivity for future AI-driven services.
The AI infrastructure push is part of MTN’s broader “Ambition 2030” strategy, which focuses on connectivity, fintech, and digital infrastructure businesses.
MTN has also been increasing partnerships within the global AI ecosystem. In March, MTN invested in US networking startup ORAN Development Company alongside major technology companies including Nvidia, Cisco, Nokia, AT&T, and Telecom Italia.
At the time, Mazen Mroué said the investment supported the idea of “sovereign AI”, where African countries host and manage their own AI infrastructure instead of depending entirely on overseas systems.
Molapisi warned that Africa risks repeating the same economic pattern seen in natural resources if it fails to build local AI infrastructure. He said the continent currently accounts for only around 1% of global computing power, raising concerns that African data could be processed abroad before being sold back to the continent through expensive digital services.
He also acknowledged the challenge of rapidly changing AI chip technology, noting that companies must carefully balance investments in training and inference hardware to avoid costly infrastructure mistakes.
