Can MENA connect AI data centres without straining power and water?
Only with site-level evidence and joined-up planning. A new World Economic Forum and SRMG Think report says the region can connect projects faster than mature markets, but its heat, water scarcity and volatile AI loads turn speed into a resilience test.
The short version
3 key points
- 1.The report says developers in parts of MENA can typically obtain grid access in two to five years, compared with lead times exceeding ten years in some mature markets, but it does not publish a country-by-country queue dataset.
- 2.It estimates current regional data-centre capacity at about 800 MW of IT load, equivalent to 1.44–1.84 GW of total energy capacity, while announced projects point towards gigawatt-scale expansion.
- 3.Its central proposal is to treat large data centres as dynamic critical loads and coordinate electricity, cooling, water, siting and operational data rather than approving each resource in isolation.
The answer: grid speed is an advantage only if resources are planned together
MENA can connect new AI data centres more quickly than some mature markets, but speed does not make the expansion sustainable. A new World Economic Forum report produced with SRMG Think argues that the region's shorter development timelines could become a liability unless utilities, water authorities, regulators and developers plan the facilities as variable critical infrastructure. That is a useful warning, not proof that the proposed safeguards already work.
The report says developers in the region typically secure grid access within two to five years, while comparable waits can exceed ten years in the United States and parts of Europe. It also estimates that MENA currently has about 800 megawatts of information-technology load, corresponding to 1.44–1.84 gigawatts of total energy capacity. Those figures describe the order of the challenge. They are drawn from cited market and industry material rather than a disclosed census of every operating site and connection queue.
For people who rely on the same grids and water systems, the question is not simply whether a campus can be built. It is whether new demand raises peak electricity costs, displaces other connections, deepens dependence on desalination or weakens reliability during extreme heat. The report's strongest contribution is therefore institutional: it asks public authorities to connect infrastructure approvals to resource evidence before gigawatt-scale projects lock in decades of demand.[1][2][3]
Why AI workloads change the infrastructure problem
The report distinguishes AI facilities from conventional web hosting. It says AI racks can draw 30–100 kilowatts, compared with 3–10 kilowatts for traditional web workloads, and training can create sharp demand changes over milliseconds. The precise numbers will vary by hardware and workload, but the operational point is credible: a utility that models a large AI campus as a steady industrial load can miss the ramping, reserve and power-quality requirements created by clustered accelerators.
The proposed response is to classify hyperscale sites as 'critical dynamic loads'. Developers would submit detailed load forecasts, while utilities and regulators would coordinate connection planning, behind-the-meter generation and emergency procedures. The report also recommends modular design, solar-plus-storage, digital twins and predictive controls, and it discusses small modular reactors and enhanced geothermal systems as medium-term possibilities rather than current answers.
That ordering matters. Firm grid capacity, transmission, storage, backup arrangements and demand flexibility are near-term engineering questions. Nuclear or geothermal projects may eventually add low-carbon supply, but neither should be counted as available capacity before licensing, construction, cost and delivery are resolved. A resilient plan needs a credible downside case for delays, not only a technology menu.[2]
Heat turns cooling into an energy and water decision
Cooling is the report's clearest link between digital policy and ordinary resource pressure. It says cooling can account for more than 40% of data-centre electricity use globally and may be higher in MENA's climate. In the report's regional framing, extreme temperatures reduce the efficiency of air cooling at the same time that electricity systems are serving high residential cooling demand. A new campus can therefore add load when the grid is already under greatest stress.
Water introduces another constraint. The report cites an upper estimate of 25.5 million litres a year for a one-megawatt facility using traditional cooling, and says conventional evaporative cooling can consume 1.5–2 litres per kilowatt-hour of cooling in temperate conditions, with potentially higher use in hot, dry locations. These are illustrative values, not measurements from a representative MENA sample. Actual withdrawals depend on climate, operating hours, cooling architecture, water quality and the boundary used for direct and electricity-related consumption.
The authors favour combinations of liquid and immersion cooling, closed-loop systems, seasonal hybrid operation and treated wastewater where local conditions support them. Each has trade-offs. Closed loops reduce ongoing water loss but still require heat rejection and electricity; wastewater needs pipelines, treatment quality and reliable contracts; liquid cooling changes equipment and maintenance requirements. The relevant comparison is whole-system water, energy, cost and reliability at a named site—not a generic claim that one cooling technology is sustainable everywhere.[2][3]
Desalination creates a feedback loop
In water-scarce parts of MENA, replacing freshwater demand with desalinated water can shift rather than remove the burden. The report says reverse-osmosis desalination commonly consumes 3–4 kilowatt-hours per cubic metre of potable water. Using its assumption of 1.8 cubic metres of cooling water per megawatt-hour, the authors estimate that producing that water could add 1–1.5% to total data-centre energy demand. This is a modelling example, and the result changes with water intensity, plant efficiency and distribution losses.
The policy implication is still important: electricity planners need to count the energy needed to produce and move cooling water, while water planners need the facility's hourly and seasonal demand rather than only its annual total. Site approval should also distinguish potable water, treated wastewater and seawater; direct withdrawals from indirect water used in electricity generation; and normal operations from extreme-weather conditions.
For communities, annual efficiency ratios can conceal the moment that matters most. A campus may look efficient over a year yet compete with households, hospitals, farms or industry during a heatwave or outage. Publishing peak demand, curtailment arrangements and emergency priorities would make the distributional choice visible instead of burying it in a regional growth forecast.[2]
The resource–intelligence matrix is a diagnostic, not a result
The report organises facilities along two dimensions: the sustainability of their energy supply and their use of AI to manage demand. It labels the least advanced position 'legacy infrastructure' and the target state a 'sustainable nexus', where cleaner diversified supply is combined with data-driven operational control and grid integration. It proposes power usage effectiveness below 1.3 and low water usage effectiveness as useful target characteristics.
The matrix can help a government ask consistent questions across projects, but it has not been validated as a predictor of lower system cost, fewer outages or reduced water stress. A facility could score well on an efficiency measure while increasing total resource use because it is much larger. Power usage effectiveness also measures the ratio between total facility energy and IT energy; it does not reveal whether the electricity is available at the right time, how carbon intensive it is, or whether the computing produces public value.
AI-based controls deserve the same scrutiny as the workload they manage. Digital twins and predictive cooling may reduce waste, but the report does not present controlled facility-level tests, error rates or verified savings. Operators need fallback procedures when sensors fail, models drift or cybersecurity controls interrupt automation. Human engineers remain responsible for safe operating envelopes and for challenging optimisation goals that overlook water, resilience or maintenance.[1][2]
What the report did—and did not—measure
This is a 35-page insight report developed by the World Economic Forum and SRMG Think with acknowledgements to government, technology, data-centre, consultancy and energy participants. It synthesises published sources, market estimates, examples and stakeholder expertise into four pathways: strategic principles, energy optimisation, water and cooling management, and institutional frameworks. It does not report a systematic search method, sampling frame, number of facilities studied or a common denominator for its regional comparisons.
The report is not peer reviewed, and no independent impact evaluator is identified. Several contributors represent organisations that build, finance, regulate or advise on the infrastructure under discussion. Their practical knowledge is relevant, but commercial and institutional interests make transparent operating evidence especially important. The report's figures should be treated as planning inputs to verify locally, not as an audited regional baseline.
MENA is also too diverse for one infrastructure trajectory. Gulf states with large investment vehicles, different fuel mixes and desalination capacity do not represent every North African or conflict-affected electricity system. Even within one country, coastal and inland sites face different heat, water, grid and network constraints. The two-to-five-year connection claim cannot substitute for public country-level queue and delivery data.[1][2][3]
What evidence would change the assessment
A stronger case would begin with a public project register covering requested and delivered megawatts, connection dates, local generation mix, transmission upgrades and who pays for them. Developers should report hourly load profiles, peak demand, curtailment performance and outage events. Site-level power usage effectiveness and water usage effectiveness should be published with climate, utilisation and measurement boundaries so comparisons are meaningful.
Water reporting should include source, annual and peak withdrawals, consumptive use, wastewater quality, discharge and drought contingency plans. Energy reporting should include marginal as well as annual carbon intensity, backup fuel, storage performance and the additional load from desalination. Independent assurance would help test whether contractual renewable claims match physical availability during the hours the data centre operates.
The assessment would become more positive if early MENA projects demonstrated reliable service through peak heat while keeping household and public-service supply secure, meeting independently verified water and carbon targets, and paying transparently for necessary grid upgrades. Evidence of rising peak prices, missed connections, emergency water use or repeated reliance on high-emission backup generation would push it the other way. Until those outcomes exist, rapid connection is an opportunity with conditions—not a sustainability achievement.[1][2][3]
What this means for people
- Residents, hospitals and businesses share electricity and water systems with new campuses, so peak-demand planning and emergency priority rules affect service reliability and costs.
- Utility engineers, cooling technicians and regulators will need workload forecasts, monitoring and authority to curtail or challenge unsafe operations rather than relying on annual efficiency claims.
- Communities near proposed sites need transparent information about grid upgrades, water sources, backup generation, pollution, construction and who bears the infrastructure cost.
Global context
The report concerns MENA, where extreme heat, water scarcity, desalination and relatively fast permitting create a distinctive resource problem. Its coordination principles are relevant elsewhere, but its connection timelines, energy mix and cooling assumptions cannot be transferred without local grid, climate and water evidence.
What the evidence does not yet show
- This is an industry and policy insight report produced by the World Economic Forum with SRMG Think, not a peer-reviewed study or independent audit.
- It does not disclose a systematic facility sample, common denominator or measured before-and-after outcomes for its recommendations.
- Regional capacity, grid-timing, cooling and water figures rely on cited external estimates and illustrative assumptions that need site-level verification.
- MENA countries and locations differ substantially in generation mix, grid strength, heat, water access, permitting and institutional capacity.
- Efficiency ratios do not by themselves show total resource use, peak-system effects, cost distribution or public benefit from the computing workload.
What to watch next
- Public country-level connection queues and delivered-capacity data for announced AI campuses
- Site-level hourly electricity demand, PUE, water use, cooling performance, outages and curtailment during peak heat
- Transparent allocation of grid, desalination, storage and backup-generation costs
- Independent verification of carbon, water and reliability claims after facilities enter ordinary operation
Living evidence record
Impact record IAI-0Q6DHY9
Evidence stage
Observed
Confidence
Supported
Reporting basis
Source analysis
Independent or research support
Not yet
Record status
Monitoring
Last checked
14 October 2026
Source trail
3 direct sources across 2 source types.
People impact
Documented in this record.
Uncertainty
Limits and next checks are explicit.
Stages describe the evidence available—not whether a technology is good or bad. See the public method.
Related-source reporting disclosure
This record analyses 3 linked source records around the same underlying development. The extra records add method, date or context, but they do not by themselves constitute independent replication of every performance claim or predicted outcome.
Evidence trail
Sources used for this report
Links checked 14 October 2026
Research question, topics and regions
Whether rapid AI data-centre expansion in MENA can be reconciled with grid reliability, extreme heat and water scarcity, and what evidence is still missing
Topics: Data centres · Electricity grids · Water · Cooling · Energy resilience · AI infrastructure · Climate adaptation
Regions: MENA · Gulf · Middle East · North Africa · Global
This report is labelled source analysis. We summarise and analyse source material in our own words; company statements remain attributed claims until independently supported. Translated summaries preserve the meaning of the original source and link back to it. Read our editorial standards.
The Impact Brief · Free
Keep following the stories that matter to you.
Get our weekday briefing on what changed and why it matters. Choose your topics below.
Continue the story
There’s more to this story
Climate & Energy
Can AI forecast Egypt’s groundwater decline?
A peer-reviewed study combines 21 years of satellite water-storage estimates with land-use mapping and XGBoost for Wadi El-Assiuti. The held-out fit is strong, but the study area is smaller than one GRACE cell and the 2030 path assumes recent conditions continue.
9 min · 1 source
Climate & Energy
Can Japan's planned 400MW AI data centre grow without locking in gas?
JERA, Dell and RHAELM plan a more than $15 billion AI-computing site beside a Chiba power station. The 400MW figure describes intended capacity—not measured electricity use—and the announcement provides no emissions, water or efficiency forecast.
6 min · 3 sources
Climate & Energy
IEA says ageing electricity grids are becoming an AI constraint
The International Energy Agency's 2026 grid report examines the investment, planning and digital tools needed as electricity demand rises from electrification, industry and data centres.
2 min · 1 source
Reader commentary
What does this mean where you are?
Tell us what you have seen at work, in your community or in everyday life. A question is welcome too. No account is needed; we review notes before publishing them.
Explore commentary across the portal →Published reader notes
0No published reader notes yet. You can start the evidence-led discussion above.
Prefer a private correction or response? Contact the newsroom.