This article explores the World Economic Forum's AIGA, which argues for a multistakeholder compact to ensure that “AI for impact” does not become “AI versus the planet.”
The World Economic Forum’s AI Governance Alliance is designed to align AI’s growth with clean power and efficiency
The spread of artificial intelligence (AI) has been nothing short of revolutionary, transforming industries by streamlining processes and boosting efficiency. But the same boom driving innovation is also fuelling a surge in electricity demand, as the data centres powering AI expand at breakneck speed. Even as these tools promise to cut carbon from energy, industry and transport, their soaring appetite for power risks undermining those very gains.
In June 2023, the World Economic Forum (WEF) launched the AI Governance Alliance (AIGA), a platform designed to bring companies and policymakers together around a common playbook to align AI’s growth with clean power, efficiency and transparency.
AIGA argues for a multistakeholder compact to ensure that “AI for impact” does not become “AI versus the planet.”
The scale of the crunch
Data centres already consume roughly
Amount of electricity consumed by data centres globally in 2024
Amount of electricity expected to be consumed by data centres globally in 2030,
according to the International Energy Agency (IEA). AI-optimised facilities are expected to be the biggest driver of that growth, with AI’s electricity needs projected to more than quadruple by decade’s end. While still smaller than looming loads from EVs and air-conditioning, this growth is material and concentrated, straining grids in clusters where hyperscale centres are located.
Corporate climate pledges are already under pressure. Microsoft says its total emissions are 29.1% above its 2020 baseline, largely due to data centre build-out. Google reports its 2023 emissions were 48% higher than in 2019, as cloud and AI growth lifted both energy and water footprints. This is even though both firms are among the world’s biggest clean-power buyers.
Governments, too, are scrambling to keep pace. Ireland has effectively paused new Dublin grid connections for data centres until at least 2028, pending new large energy user rules, a sign of the strain AI demand can create. The Netherlands has also imposed targeted moratoria and tighter siting rules for hyperscale projects. More planning frictions seem inevitable unless AI growth is paired with clear benefits to local grids and communities.
The case for optimism
AI consumes power, but it can also drive decarbonisation. AI-enabled analytics cut energy use by 10–60% in factories and logistics. In the UK, Octopus Energy’s Kraken platform coordinated 1.5 million homes to shift demand, while DeepMind boosted wind power’s value by 20% through forecasting. Inside data centres, AI cooling has cut energy use by 40%. Deployed at scale across industry, buildings and grids, AI could save more energy than it consumes.
Understanding the impact
The World Economic Forum’s AIGA, which spans more than 600 members from 500 organisations across business, government, academia and civil society, exists to prevent a tragedy of the commons: everyone racing to build computing power, with too few coordinating on clean power, metrics and safety.
“AI is one of the biggest stories in the energy world today – but until now, policymakers and markets lacked the tools to fully understand the wide-ranging impacts.” - Fatih Birol, Executive Director, IEA
In its 2024 Energy Paradox brief, the WEF sketches a pragmatic agenda: scale AI where it abates emissions, curb AI’s own energy and water footprints, and create transparent metrics so boards and regulators can tell the difference. Early pilots cited by WEF show material savings from AI-driven energy management; the task now is to productise and govern those wins, sector by sector.
The narrative risk, and the prize
Left unmanaged, AI could trigger a backlash: delayed grid upgrades, planning pushback in places like Dublin and Amsterdam, and more. Managed well, AI becomes a system optimiser, firming renewables, fine-tuning demand, accelerating materials discovery and cutting waste across the real economy.
The technology is not fated to be either villain or saviour; it will be the sum of decisions made by chipmakers, cloud providers, utilities, policymakers and boards.
That is why the AIGA’s multistakeholder framing matters. Getting measurement, market design and procurement right may be less glamorous than training the next frontier model, but it is how we ensure the gains from AI, in productivity and decarbonisation, outweigh its costs.