Policy Brief

Smart Energy for Smart Technologies: G20 Solutions for AI-Era Challenges

AI is rapidly becoming a technology that consumes a lot of electricity. As AI grows, it is driving a major shift in energy use. How the G20 responds to this challenge will determine if digital transformation supports or slows down the move to clean energy. The scale of the challenge is already visible. Data centres consumed around 485 terawatt-hours (TWh) of electricity globally in 2025, with electricity demand from data centres rising 17% in a single year. The International Energy Agency (IEA) projects that global data-centre electricity consumption could reach approximately 950 TWh by 2030, roughly double the 2025 level. AI-focused data centres are expanding even faster.¹

The challenge is not just about producing more electricity, but about making it more reliable, affordable, and cleaner, and delivering it where and when it is needed. Because data centres are often located in specific areas, they can put extra pressure on local power grids. The IEA (2025) reports that almost 20% of planned data centre projects could be delayed because of grid limitations.²

Because the G20 is responsible for most of the world’s economic activity and a large share of energy use and emissions, it should treat AI infrastructure as part of its energy and industrial policy, not just as a technology issue. It is important to make sure that new electricity demand is increasingly supplied by clean energy. The United Nations has called for new electricity demand, especially the electricity that is generated for rapidly expanding fields such as AI and data centres, to be met by means of renewable and other clean sources; it has also stressed the importance of modern grids, energy storage, stronger international cooperation towards trade and investment, and more financing for developing economies.³

Policy Recommendation for G20

For G20 countries, this means moving past the current approach where tech companies choose to buy renewable energy on their own. Approvals for new data centre investments should be linked to solid plans for using clean energy, meeting energy efficiency goals, following water-use rules, and sharing clear reports on electricity and carbon use. Governments could set 'clean digital infrastructure' standards, requiring new large data centres to show how they will meet extra electricity needs without causing too many emissions or risking the reliability of local power grids.

Second, the G20 needs to modernize its electricity grids quickly. To use more renewable energy, countries need flexible transmission and distribution networks, storage, and advanced ways to manage demand. The IEA says that yearly investment in grids must rise a lot, and its latest analysis shows that over 2,500 GW of renewable energy, storage, and large projects are stuck in grid queues worldwide.⁴

AI can also help solve these challenges. AI-powered forecasting can predict electricity demand and renewable energy output. Using AI widely in power plant operations and maintenance could save a lot of money, and AI management could unlock more transmission capacity from existing infrastructure. The main goal should be to use AI to improve energy efficiency before dedicating more energy to AI itself. Every time AI infrastructure grows, it should be paired with investments that use AI to make the whole energy system smarter.

India shows both the difficulties and the possibilities in this domain. Demand for electricity is rising quickly because industrialisation, urbanisation, the need for cooling, the electrification of transport, and digitalisation are all accelerating. Meanwhile, the country is pursuing an ambitious clean-energy expansion programme and has set a target of 500 GW of non-fossil electricity capacity by 2030. Non-fossil energy sources accounted for 50% of total installed electricity capacity by June 2025.⁵

India could use AI to link rooftop solar, batteries, electric vehicles, and other distributed resources into virtual power plants. These systems can unite thousands of small energy sources so they can respond together to changes in supply and demand. This would ease pressure on transmission lines and help bring more decentralized renewable energy into the grid.

The third priority is for the G20 to work together on standards, supply chains, and finance. Using AI in the energy transition will raise demand for transformers, semiconductors, copper, key minerals, batteries, and other specialized equipment. If supply chains are too concentrated, it could create new risks for energy security. The IEA has already noted that critical minerals and grid parts are new weak points at the intersection of AI and energy.

G20 members should set up a G20 AI-Energy Partnership focused on four areas: shared standards for sustainable data centres, secure supply chains for grid and digital infrastructure, joint research on AI-powered electricity systems, and funding options for emerging and developing economies. Cybersecurity must also be included in this cooperation. More digitalisation makes electricity systems more efficient but can also make them more vulnerable to attacks. AI can help with anomaly detection, predictive maintenance, and cyber defense, but it can also give attackers new tools. So, the energy sector needs shared cybersecurity standards, ways to report incidents, and regular stress tests for AI systems.

Conclusion

The key question is not just about balancing AI growth with energy goals, but whether governments can make these two priorities support each other. The G20 should follow a four-point strategy: Clean Supply, Smart Grids, Secure Systems, and Shared Prosperity. Clean Supply means making sure that growing digital demand leads to more renewable and low-carbon energy. Smart Grids should use AI, storage, and digital tools to improve flexibility. Secure Systems should protect connected energy networks and key supply chains. Shared Prosperity means making sure developing countries get the finance, technology, and skills needed to join the transition. AI is causing a new rise in electricity demand, but it also provides tools to manage more complex energy systems. The G20 can make this a win-win situation. By matching investments in digital infrastructure with clean energy, grid upgrades, cybersecurity, and inclusive finance, the energy needed for AI could help build a stronger energy system instead of slowing down the shift to clean energy.

Note: