The World Economic Forum has called on governments across the Middle East and North Africa to treat AI data centres as a distinct category of high-demand electricity consumers and to revise their power infrastructure planning accordingly, according to a policy brief published by the organisation.
The document identified power supply constraints, limited water availability, and periods of intense heat as tangible obstacles for the region's drive to establish itself as a centre for AI infrastructure development. Projections outlined in the document point to a 40 per cent expansion of electricity generation capacity across the MENA region before 2030, while forecasts separately indicate a 15 per cent rise in power supply — a figure the report noted falls short of what energy-intensive AI infrastructure is expected to demand.
The brief stated that the debate has shifted away from whether digital infrastructure should be expanded toward the question of how that expansion can be carried out at scale with efficiency, sustainability, and resilience in mind. The WEF document urged governments to treat AI data centres as a distinct regulatory class, separate from conventional industrial and commercial energy consumers, given the facilities' disproportionate power demands and volatile load behaviour.
Among its recommendations, the document outlined a push for compulsory projections of electricity demand alongside tighter collaboration between data centre developers, utilities, grid operators, and the bodies responsible for oversight. Under the same proposal, any project whose IT load exceeds 10 megawatts would be required to provide detailed power-demand projections at the point of seeking a grid connection.
The report noted that growing demand from AI-driven computing tasks is putting measurable strain on power grids and energy infrastructure. Power draw figures in the report place AI server racks at a range of 30 to 100 kilowatts per unit, a demand the document described as running roughly ten times higher than the 3 to 10 kilowatts typically drawn by conventional web workloads.
The brief also warned that volatile power consumption at this scale puts electricity network reliability at risk, particularly when multiple data centres go offline or come back online in quick succession. Cooling infrastructure — the systems tasked with preventing server hardware from overheating — accounts for more than two-fifths of total energy use at data centres throughout the MENA region, according to the report.
Rising heat and shrinking water supplies stand to push up day-to-day running costs while placing additional strain on resources that are already under pressure, the document stated. A proposal from the WEF centred on integrating power generation, cooling systems, and water stewardship into a unified operational framework.
Four priority areas were identified in the document: infrastructure design, the integration of energy systems, approaches to cooling and water management, and a regulatory framework built to anticipate future needs. Among the proposals put forward are pairing solar generation with battery storage, building electricity production capacity on-site, deploying AI-driven tools for managing energy consumption, and transitioning to closed-loop cooling systems.
The document also proposed channelling treated wastewater into supply chains as a way to ease demand on freshwater reserves. Small modular reactors and enhanced geothermal systems were identified as longer-range additions to the power mix, though the brief noted that questions around whether either technology can achieve commercial scale have yet to be resolved.
According to a WEF assessment, on-site power generation — situated behind the utility meter — could reduce the timeline for bringing new data centres online by between one and three years, chiefly by lowering reliance on external grid infrastructure. The document also recommended that MENA governments bring their regulatory frameworks into closer alignment and deepen cross-border cooperation mechanisms.
Tighter alignment across the sector could lift operational efficiency and strengthen investor confidence as appetite for AI computing capacity expands, according to the WEF. The organisation described its recommendations as an operational blueprint covering the next five years, a period when choices around energy supply, thermal management, and governance frameworks will carry growing weight in determining how the region positions itself within the broader global contest for AI infrastructure, the document stated.
