IFC projects access to computing as driver of next phase of economic growth

He identified reliable power supply as an indispensable requirement for sustaining the boom in data centres across the continent for hyperscale infrastructure deployment.

“Reliable electricity is the single most important constraint affecting data centre expansion in many emerging markets,” Isiadinso said.

According to him, the energy challenge is not unique to Africa, as mature digital hubs in Europe and North America are also facing grid pressures driven by rising AI-related power demand. However, the implications are especially significant for African markets seeking to attract cloud regions and large colocation campuses.

To fix this challenge, infrastructure developers, he said, are increasingly exploring hybrid energy models, including gas-to-power systems, renewable energy integration and private power purchase agreements to ensure a stable electricity supply.

Read Also: FG directs mobile network operators to compensate consumers

Beyond energy, financing large-scale data centre projects rely heavily on securing long-term customers, often referred to as ‘anchor tenants’. Contracts with hyperscale cloud providers, telecom operators or government institutions help provide predictable revenue streams that make projects bankable.

“Anchor tenants provide predictable revenue streams that reduce investment risk,” Isiadinso said. “Where those conditions are present, projects become significantly easier to finance.”

Government digitisation initiatives could therefore play a critical role in accelerating infrastructure deployment, particularly when public services migrate to cloud platforms or require local data hosting.

While Africa may not immediately compete with energy-rich regions hosting massive AI training clusters, the IFC sees significant opportunity in distributed AI inference infrastructure—systems designed to run AI applications closer to end users.

Inference workloads support applications such as language processing, financial automation, recommendation systems and real-time analytics, all of which are expanding rapidly across African markets as digital adoption grows.

According to Isiadinso, training facilities typically require extremely large compute clusters and high-power density, making them more suited to regions with abundant energy resources. Inference infrastructure, by contrast, is more distributed and aligns with Africa’s growing need for localised computing capacity.

“As digital services expand across the continent, demand for localised inference capacity will increase,” he said, noting that this creates opportunities for regional data centre operators and infrastructure investors.

The session also highlighted the limited participation of domestic institutional capital in Africa’s digital infrastructure sector. Pension funds across the continent are expanding but remain under-allocated to infrastructure assets, particularly digital infrastructure.

“With appropriate regulatory frameworks and risk mitigation mechanisms, these funds could become important sources of long-term capital,” Isiadinso said, adding that improving investor awareness of digital infrastructure as an asset class would be key to unlocking local financing.

Globally, hyperscale investment increasingly flows to markets offering policy clarity across energy, connectivity and digital regulation. Countries that can provide reliable power, regulatory certainty and investment stability are more likely to attract the next wave of infrastructure deployment, he added.

Africa’s digital foundation is improving, supported by expanding subsea connectivity, growing enterprise adoption and rising investor interest.

 However, Isiadinso stressed that coordinated action across energy, policy and finance will determine whether the continent captures more value from the global digital economy.

“Infrastructure determines where digital value is created. Ensuring that a greater share of that value is created within Africa will depend on the investments made today,” he said, stressing that Africa’s opportunity lies in AI inference infrastructure.

While the world’s largest AI training clusters are likely to remain concentrated in energy-abundant regions, Africa has a strong opportunity in distributed inference infrastructure - systems designed to run AI applications closer to users.

“As digital services expand across the continent, demand for localized inference capacity will increase. This creates opportunities for regional data-center operators and infrastructure investors,” Isiadinso said.

Inference workloads support applications such as language processing, financial automation, and real-time analytics - use cases already expanding across African markets.

During the session, Isiadinso distinguished between AI training infrastructure and AI inference infrastructure, as they represent very different investment models and deployment geographies. According to him, training facilities require extremely large compute clusters and very high power densities, and are therefore typically concentrated in markets with abundant energy resources and advanced connectivity ecosystems.

Inference infrastructure, by contrast, is more distributed and deployed closer to users. It supports real-time applications such as language models, recommendation systems, automation platforms, and enterprise analytics. In this context, he noted that Africa’s most immediate opportunities in AI infrastructure are likely to emerge on the inference side, where localized compute capacity becomes increasingly important as digital services expand across the continent. As governments digitize public services, enterprises migrate workloads to the cloud, and platform-based applications scale, demand for regional inference capacity is expected to increase significantly.

“As digital services expand across the continent, demand for localized inference capacity will increase. This creates opportunities for regional data-center operators and infrastructure investors,” Isiadinso said.