Substation and solar field viewed from above

The market context

Market context: growth constrained by deliverable power

AI infrastructure growth is increasingly constrained by deliverable power at specific locations — not by hardware alone. Energy, infrastructure and compute planning therefore need to be integrated.

950 TWh

Projected global data-centre electricity consumption in 2030 · IEA

GenerationStorageGridDemand
2025 · IEA485 TWh

Global data-centre electricity consumption

2030 · IEA~950 TWh

Projected global data-centre electricity consumption — roughly double 2025

Deliverable power

Hardware can arrive faster than infrastructure.

Because demand is concentrated at individual sites, grid access, transmission, cooling, fibre and permitting determine whether capacity is usable within the commercial timetable. Energy, infrastructure and compute therefore need to be planned as one system.

Where can power, infrastructure, workload and capital be assembled together?

Africa in the picture

Growth will not be uniform.

Power reliability, fibre routes, cooling conditions, regulation and financing depth differ materially between African markets. Opportunity concentrates where power, fibre, workload and capital can be assembled within one commercial timetable.

Africa's future AI infrastructure should therefore be considered alongside regional power, connectivity and sovereign capability — not as a separate hardware question.

  • Power
  • Fibre
  • Workload
  • Capital

Opportunity emerges where power, fibre, workload and capital can be assembled together.

Project fit is local

Source

Current stage · pre-development

Build where the constraint actually binds.