India’s ambition to become a global artificial intelligence powerhouse, backed by reports of nearly $300 billion in AI-linked investment commitments, is setting off a second, less discussed race — one for electricity, water and transmission infrastructure.
The country’s installed data centre capacity has already expanded from about 520 MW in 2020 to 1.5 GW by mid-2025, and is projected to reach 4.5–9 GW by 2030, according to estimates by the Council on Energy, Environment and Water (CEEW). That translates into an additional 3–7.5 GW of IT load within five years — a scale that will test cluster-level grids in Mumbai, Chennai, Hyderabad, NCR and Bengaluru.
Data centres today account for roughly 0.8-1% of India’s total power consumption, translating to about 18-22 terawatt-hours (TWh) annually, as installed capacity has expanded sharply over the past two years.
Demand is projected to rise to 35-45 TWh
By 2030, this demand is projected to rise to 35-45 TWh, driven by hyperscale cloud expansion, AI workloads, edge deployments and higher utilisation rates — potentially doubling India’s data centre power footprint within this decade.
“The real challenge is not aggregate power supply but cluster-level grid readiness and firm power architecture,” said Prateek Aggarwal, Programme Lead at CEEW. “Renewables combined with storage can meet 70–80% of annual consumption, but scaling sustainably will require transmission upgrades, flexible generation and market reforms.”
Labanya Prakash Jena, Director at Climate and Sustainability Initiative (CSI), offered a sharper warning. “Data centres today consume less than 1% of electricity, but this can rise to nearly 10% over the next five to ten years,” he said. “Since they require 24/7 supply, coal will remain part of the grid. LNG-based electricity is expensive in India and cannot substitute coal or renewables.”
Stressed urban clusters
The geography of growth compounds the challenge. Data centre expansion is concentrated in already stressed urban clusters.
Water use is emerging as the more immediate risk. In 2025, India’s data centre industry consumed an estimated 150 billion litres of water annually, and that figure could more than double by 2030 as capacity scales.
“It is very serious that data centres are coming to cities already facing water scarcity and high temperatures,” Jena said. “Today it may not be visible, but tomorrow policymakers may have to choose between water for communities or data centres.”
CEEW’s research finds water governance gaps across most state data centre policies. Of the 15 states promoting data centres, only five include explicit sustainability provisions. Rajasthan’s 2025 policy stands out for mandating treated wastewater use, zero liquid discharge norms and groundwater recharge.
Aggarwal said India still has a policy window. “Cooling technologies such as liquid immersion and direct-to-chip can significantly reduce water use, but adoption is limited. There is an opportunity to embed water disclosure norms and basin-linked siting guidelines before risks become structural.”
Transmission remains another bottleneck. As renewable capacity rises, inadequate grid upgrades could slow project execution. Jena noted, “India is not investing enough in transmission lines, especially when too much renewable energy is added.”
Speaking on the development, Adani Group executives — who are aware of the development — said large operators are preparing for this scale challenge. “Our data centre designs prioritise integrated power sourcing, combining renewable energy with storage and grid reliability mechanisms,” a senior Adani executive said. “Energy cost is critical for hyperscale facilities, and we are optimising procurement to ensure 24/7 supply without compromising decarbonisation goals.”
On water, the official said the company is planning to invest in advanced cooling systems. “We are planning to adopt treated wastewater use, closed-loop cooling and zero liquid discharge wherever feasible,” the person added.
Another executive emphasised efficiency metrics. “Power usage effectiveness and water usage effectiveness — are core to our engineering design. As AI workloads intensify, cooling innovation becomes central to cost and sustainability.”
India’s AI build-out is anchored in favourable open-access renewable frameworks and competitive clean energy tariffs, making it attractive to hyperscalers. Yet environmental governance remains uneven across states.
As corporate net-zero commitments converge with AI expansion, the next phase of India’s digital infrastructure story will hinge less on capital inflows and more on how effectively the country can power, cool and integrate a rapidly expanding fleet of energy-intensive facilities — without triggering a parallel resource crisis.