Africa Tech Festival: Time to cross the double digital divide

For two decades, African policy has revolved around closing the digital divide. Governments, regulators and operators built out fibre, extended mobile coverage and subsidised data to bring more people online. It was a campaign that transformed communication, commerce and education. Yet as the continent connects, another gap has emerged beneath the surface. Connectivity has brought access, but not capacity. The new challenge can be described as a double divide: the gulf between those who can connect and those who can compute.

The first divide is visible: cables, antennas and handsets. The second exists in the unseen engines of the digital economy. Africa generates vast amounts of data through mobile networks, online services and satellite imaging, but most of it is still processed and stored overseas. That means higher costs, slower services and weaker data protection. It also means that the real economic value of Africa’s data is often realised elsewhere.

The Ministerial Forum on Artificial Intelligence in Cape Town, held alongside the AfricaCom conference last week, placed this reality at the centre of its agenda. It drew together ministers, regulators and industry leaders who have long treated connectivity as an end in itself. The discussions shifted the focus from access to how Africa can gain the power to use and govern its own data.

“Artificial intelligence will only lift Africa if the continent builds the capacity to run it,” said Hover Gao, president of Huawei sub-Saharan Africa.

Gao pointed out that the global intelligent economy could reach a value of $18.8-trillion by 2030, growing almost three times faster than traditional sectors. Yet the continent still contributes only one percent of global computing capacity and three percent of AI talent. At least sixteen African countries have drafted national AI strategies, but most depend on cloud facilities abroad.

“To move from strategy to implementation, we must face the shortage of computing power, local data and AI skills,” he said.

His first recommendation was the creation of national computing power centres: shared infrastructure for research, government and enterprise. He drew on China’s “One Centre + Four Platforms” model, where a central hub supports public computing, innovation incubation, industrial development and talent training. Such an approach, he said, allows nations to localise and govern their own data and algorithms, safeguarding key sectors from dependency on external providers.

African policy makers have typically measured progress by kilometres of fibre or numbers of cell-phone subscribers. Gao suggested a new metric: processing power. Rather than a replacement for broadband, compute is its next layer.

South Africa’s Minister of Communications and Digital Technologies, Solly Malatsi, added a human dimension to the discussion.

“AI must work in our languages, in our contexts, and for our use cases,” he said. He reminded delegates that infrastructure alone cannot define intelligence. Algorithms trained solely on foreign data will reproduce foreign biases. African AI requires both local capacity and local content.

Gao expanded on this through a second proposal: the creation of African foundational AI models trained on regional languages, knowledge systems and social data. These models would allow African enterprises and governments to fine-tune AI tools for agriculture, healthcare and finance without starting from scratch, cutting costs and improving accuracy.

That ambition demands people as much as hardware. African Telecommunications Union secretary-general John Omo reminded participants that more than 70 percent of sub-Saharan Africans are under thirty, yet three out of four young people still lack the digital skills needed for the coming economy.

“If we do not act now, this demographic strength could become a lost opportunity,” he said. Training AI engineers, data scientists and policy specialists will determine whether the next generation shapes technology or merely uses it.

Clearly, the double divide is not a choice between networks and processors. The cables that carried Africa onto the first internet wave will carry the data that feeds its intelligent economy.

But compute power, data governance and skills development are now as critical as mobile coverage once was. Nations that invest in them will own more than their data, said Gao. They will own their digital futures.

A phrase that anchored his presentation also offered a challenge. “We firmly believe these exchanges will help African countries achieve leapfrog development in the AI era.”

 * Arthur Goldstuck is CEO of World Wide Worx, editor-in-chief of Gadget.co.za, and author of The Hitchhiker’s Guide to AI – The African Edge.