- By 2050, 1 out of every 4 people on the planet will be African
- But Africa is woefully behind in terms of compute power
- An MTN executive proposes the idea of "data embassies"
DTW IGNITE 2026, COPENHAGEN, DENMARK – Europe is scrambling to plot out its digital future after the U.S. recently revoked global access to certain Anthropic AI models, confirming fears around the existence of a virtual “kill switch.” Africa is facing what is arguably a worse digital sovereignty crisis, but MTN Group’s CTIO Charles Molapisi said he has a plan to help solve it.
The United Nations Economic Commission for Africa has forecast that by 2050, 1 out of every 4 people on the planet will be African. And yet, Molapisi noted that Africa today represents less than 1% of total global compute power. That lack of local infrastructure could cripple the continent if operators don’t step in.
“Let’s say we don’t invest. What will happen is that we will export our data, our raw data, and we’ll buy intelligence back into the continent because intelligence will be a requirement for every nation,” he told Fierce. “We cannot find ourselves in that situation.”
MTN, which serves 300 million customers across 19 African countries, is perfectly placed to help solve the problem, Molapisi said.
Sovereignty of a different sort
The operator is investing in compute infrastructure on the continent to mitigate this risk, building one data center each in Nigeria and South Africa.
The idea is to create what Molapisi called “data embassies,” essentially regional sovereign hubs. Basically, nearby countries and companies can put their own servers in a centrally located data center and have those servers be subject to the laws of their home country. The way an embassy is a sort of independent outpost in a foreign country, so would these servers be little data islands.
“In the context of the laws, it is actually a sovereign area for the country. They can service their data with their own engineers. They can rip and replace it. They can pull the equipment and the local country cannot touch it,” Molapisi said.
Molapisi pitched data embassies as a way for Africa to make a start toward achieving its own AI and data sovereignty. But he also argued the country-by-country approach taken by European governments is “a little bit too extreme.”
He pointed out that “you cannot be 100% sovereign…Data? Yes, but not computing.” GPUs running in data centers will not be domestically made. Frontier models will likely remain centralized and even if countries or companies try to use open source options to make their own models, they’ll never own those entirely. So, he argued, data embassies are the happy medium that let folks retain control over their key asset: Data.
Big AI ambitions
The data embassy pitch is just one part of MTN’s Ambition 2030 initiative, which outlines how it plans to play in connectivity, financial services and digital infrastructure to grow revenue over the coming years. Molapisi said AI will apply across all three of those pillars and infuse all elements of the business, including everything from revenue assurance and network applications to IT, finance and energy management.
All told, MTN is aiming to achieve $1.8 billion (30 billion Rands) in revenue growth via this strategy. Around half of that number will come from AI applied inside the organization and the rest will come from its B2B and B2C businesses.
“AI inside will create a number of assets that become more relevant as you’re going to B2C. For example, the issue of hyperpersonalization, we are generating more data about customers than ever before. Fiber sensing, that we’re putting on the network. Energy management. Digital twins,” Molapisi said. “All of these insights are creating a whole new profile that enables us to build new products and services at the B2C level.”
But while Molapisi is bullish about MTN’s plans, he’s clear-eyed about some of the risks involved, especially those relating to token economics and the costs both for MTN and consumers.
Asked to name an uncomfortable truth about applying AI in telecom networks, he noted he’s concerned about whether MTN can produce AI for consumers “at the right economic profile.”
“I’m concerned about the token economics. I’m concerned about the energy profile, whether we survive the cost implication of AI and computing and energy. My view is that the operators or anybody who solves the token-to-watt ratio will become the winner. And I’m hoping we are up there with the rest,” he concluded.