Africa’s AI ambitions hinge on solving infrastructure crisis, Schneider Electric executive warns

The race to position Africa as a destination for artificial intelligence investment is accelerating, but the continent risks being left behind unless it confronts a fundamental constraint: it does not yet have the infrastructure to support the technology it is being asked to host.

That was the stark message delivered by Steven Santini, vice president for secure power at Schneider Electric Sub-Saharan Africa, speaking this week at the IDC CIO Summit 2026 in Johannesburg, one of the region’s most prominent gatherings of senior technology decision-makers.

“Global AI players increasingly view Africa as the next frontier the new gold rush, in many respects,” Santini told delegates at the Sandton Convention Centre. “We have the land, the resources, and the growth potential. Data centres are being developed across Kenya, Nigeria, South Africa, and other regions where investment is welcomed.”

Yet the enthusiasm, he cautioned, is outpacing the groundwork.

Power is the central problem. As AI workloads grow in computational intensity, so do the energy demands of the facilities that run them. Santini drew a stark comparison to illustrate the scale of what is coming. “Some of the projects we are involved with in the Middle East have power requirements comparable to entire cities,” he said. Against that backdrop, Africa’s persistent energy supply challenges take on new strategic weight.

The continent already struggles with generation deficits, aging grid infrastructure, and intermittent supply in several of its fastest-growing economies. For hyperscale data centre development, the kind that supports large AI model training and inference, reliable, high-capacity power is not optional. It is the foundation on which everything else is built on.

But Santini pushed back against a framing that reduces Africa’s AI infrastructure challenge to a single, monolithic problem. Not every AI application requires a warehouse-sized facility drawing gigawatts of power, he argued. That misunderstanding, he suggested, is itself an obstacle to progress.

“When people hear ‘AI’, they often picture massive hyperscale data centres,” he said. “But AI exists in many different forms. Your laptop can run AI workloads. A small ten-node server cluster deployed at an industrial site can support AI applications. AI does not always require enormous, high-density centralised environments.”

For sectors that form the backbone of many African economies, mining, agriculture, financial services, and government, distributed, smaller-scale deployments may in fact be the more practical path to adoption. Prefabricated systems, containerised data centres, and single-rack installations within existing facilities are already gaining traction, Santini noted, allowing organisations to extend AI capabilities without waiting for major infrastructure overhauls.

“This allows them to leverage existing cooling and power infrastructure while simplifying deployment,” he said.

Connectivity, he added, is equally non-negotiable. “A data centre without reliable network infrastructure is effectively just an expensive paperweight. If data cannot move efficiently in and out, the infrastructure cannot deliver value.”

The conversation around African AI infrastructure has tended to focus on physical assets — cables, cooling systems, generators, and server racks. Santini argued that an often-overlooked layer of software intelligence is just as critical, particularly in environments where energy is scarce and expensive.

“We live in a world where power is constrained, and nowhere is that reality felt more strongly than in Africa,” he said. “We need both the right physical infrastructure and the right software intelligence to maximise efficiency and performance.”

The practical implication is that organisations cannot simply replicate infrastructure models developed in Europe or North America and expect them to work at equivalent cost or reliability on the continent. Solutions need to be engineered for constrained environments from the outset.

For all the momentum building around AI on the continent, Santini closed with a note of discipline. Investment in AI infrastructure, he argued, must be tethered to clearly defined business outcomes, not driven by the fear of missing out on a technology wave.

“AI in Africa is not a future concept; it is already happening,” he said. “But success will depend on defining the right operational outcomes first, and then aligning the appropriate technologies, power, cooling, computing, storage, and networking around those goals.”

For Schneider Electric, which has deepened its footprint across African markets in recent years, the opportunity is less about selling hardware than about positioning itself as an end-to-end partner for what could become one of the most consequential infrastructure build-outs in the continent’s recent history.

“As Schneider Electric, we position ourselves as the energy technology partner helping organisations achieve those outcomes efficiently and sustainably,” Santini said.