MENA Must Classify AI Data Centres as Critical Loads to Avoid Power Crisis: WEF
The World Economic Forum warns MENA nations must treat AI data centres as critical loads to avoid power crises amid surging electricity demand.
The World Economic Forum (WEF) has issued a stark warning to nations across the Middle East and North Africa (MENA), urging them to urgently overhaul their energy planning. As the region strives to position itself as a global hub for artificial intelligence, the WEF suggests that AI data centres power demand must be classified as "critical dynamic loads" to prevent systemic failures.
The alarm comes as the region faces a perfect storm of surging electricity needs, chronic water scarcity, and extreme heat, all of which could jeopardize the ambitious digital infrastructure goals of MENA countries.
The Growing Energy Gap
According to the WEF report, the disparity between current power projections and actual needs is alarming. To sustain the rapid expansion of energy-intensive AI facilities, the region's electricity capacity needs to surge by 40 per cent by 2030. However, current projections show a growth of only 15 per cent.
The report emphasizes that the conversation has shifted. It is no longer just about whether to expand digital infrastructure, but rather how to do so in a way that is resilient, sustainable, and efficient at a massive scale.
Why AI is Different from Traditional Tech
The report highlights a massive jump in power consumption when comparing traditional web workloads to AI-driven ones. While traditional server racks typically consume between 3 and 10 kilowatts, AI server racks demand between 30 and 100 kilowatts.
This intensity creates a volatile environment for power grids. The WEF warns that sudden fluctuations in demand—such as when multiple facilities disconnect or reconnect at the same time—could potentially destabilize entire electricity grids.
Proposed Regulatory Shifts
To mitigate these risks, the WEF recommends several strategic regulatory changes:
- Differentiation: Governments should distinguish AI data centres from standard commercial or industrial electricity users.
- Mandatory Forecasting: The introduction of compulsory energy-demand forecasts.
- Strict Requirements: Any data centre project with an IT load exceeding 10 megawatts should be required to submit detailed electricity-demand forecasts during their grid-connection applications.
- Coordination: Establishing formal mechanisms for cooperation between regulators, grid operators, utilities, and developers.
The Cooling Crisis: Water and Heat
In the MENA region, the environment is a significant adversary. Extreme temperatures mean that cooling systems account for more than 40 per cent of a data centre's total electricity consumption. This creates a vicious cycle where the need to cool the hardware further increases the power demand and puts immense pressure on limited water resources.
The Roadmap to Resilience
The WEF has outlined four primary areas for action to ensure the region remains digitally competitive:
- Smarter Infrastructure: Designing facilities with efficiency at the core.
- Integrated Energy Systems: Utilizing solar power paired with battery storage and on-site electricity generation.
- Advanced Resource Management: Implementing closed-loop cooling systems, utilizing treated wastewater, and employing AI-based energy management.
- Forward-Looking Regulation: Harmonizing regulatory standards across borders to boost investment confidence.
The report also suggests that "behind-the-meter" electricity generation could be a game-changer, potentially accelerating the deployment of data centres by 12 to 36 months by reducing the reliance on utility grid connections.
Long-term Energy Horizons
While solar and batteries provide immediate relief, the WEF pointed toward more advanced solutions for the future. Enhanced geothermal systems and small modular nuclear reactors were mentioned as potential long-term power sources, though the report noted that their commercial viability remains a significant constraint for now.
These recommendations serve as a critical framework for the next five years, as the decisions made today regarding power supply and regulation will ultimately determine the region's standing in the global AI race.