India's Data Centre Demand Expected to Surge by 26.3 GW by 2031-32

Artificial Intelligence (AI) data centres will put an additional load of 26.3 gigawatts (GW) by 2031-32, the Ministry of Power informed the parliament on Monday. This additional load is expected to be integrated into the grid and will primarily be served by renewable energy capacity, the ministry added. The fresh projection is almost double the government’s earlier projection in March this year when it estimated that electricity demand from data centres will reach 13.56 GW by 2031-32. The present estimate is based on projects received by different states to set up data centres as the ministry is also holding consultation with stakeholders such as state transmission utilities, distribution utilities and data centre developers, the ministry officials said. Discussions are going on regarding reliable power supply, requirement of transmission and grid balancing as power load from data centers could be very spiky and have sharp ramps. Policy-level discussions have also been underway on locating data centres away from urban centres to avoid transmission congestion, The Indian Express has learned. Notably, most data centres tend to be concentrated near IT hubs, which could lead to transmission congestion if more capacity is added to meet the data centre demand. There is also some opinion that data centres should come up near green energy sources. This could be beneficial in avoiding transmission congestion while enabling the electricity demand from these data centres to be met through green energy sources, people aware of the discussions said. Data centre expansion These discussions came as the country is expected to witness a sharp increase in data centre capacity in the coming years. From a mere 375 megawatt (MW) data centre capacity in 2020, India’s current capacity stood at around 1.5 GW by 2025. According to a report by Wood Mackenzie, a global energy consultancy firm, India’s operational data centre capacity will increase to 12 GW by 2030, at a compound annual growth rate (CAGR) of approximately 40%. “AI-dedicated capacity is expected to expand nearly 24-fold over the same period, rising from 275 MW to 6,546 MW, as hyperscale cloud providers, enterprise digitalisation and artificial intelligence accelerate demand for computer infrastructure,” it said in a statement on Monday. The report claimed that India’s expanding digital economy is underpinning unprecedented demand for data centre capacity. “Valued at INR 32 trillion in 2025 and contributing approximately 12% of GDP, the country’s digital ecosystem serves more than 1.03 billion active internet users and processes around 22 billion UPI transactions each month. Meanwhile, India’s domestic AI market is projected to reach INR 11.7 trillion by 2032, further increasing demand for high-performance computing infrastructure,” it said. Data centre electricity demand is forecast to grow 20-fold by 2040 from 10 TWh in 2025 and would account for 7% of the total electricity demand, it added. Risks to grid stability While AI runs on algorithms, it also requires electricity in large volumes. As a result, sector experts had expressed concerns over grid risks associated with these huge data centre loads. In a discussion at the India AI Impact Summit in February, Samir Chandra Saxena, the Chairman and Managing Director of Grid Controller of India Ltd (Grid India), had expressed concern over sharp variability of data centre loads posing risk to grid stability. “It is variable, it is spiky, it has sharp ramps,” he had said, adding that they can also make “silent exits” from the grid. “Anything that happens on the grid side, the data centres prefer to quietly isolate themselves.” Since most are inverter-based, he had warned that a sudden withdrawal of 1-1.5 GW — or even a couple of gigawatts — could “simply walk out of the system quietly and create a disturbance kind of situation for the grid.” According to him, this needs to be planned for and handled in a more planned manner rather than in random fashion. In the same discussion, Abhishek Ranjan, CEO of BSES Rajdhani Power Ltd, had warned that rapid hyperscale expansion could strain the grid if not carefully mapped.