India is racing to become a global AI data centre hub, offering tax breaks and attracting billions in foreign investment. These facilities consume vast amounts of water for cooling, raising concerns in water-stressed states like Maharashtra. Can India strike a balance between its AI ambition and its resource constraints, or will shortages play a spoilsport? Mint explores.
How did Mumbai become the data centre capital of India?
Mumbai has a 53% share of India’s data centre capacity. Mumbai’s strategic advantages – hosting multiple subsea cable landing stations, and its status as the financial capital with a robust technology ecosystem – have created strong demand for secure, high‑capacity infrastructure. Reliable power supply, government incentives and land availability in Navi Mumbai further boosted growth. Global cloud providers like AWS, Microsoft and Google chose Mumbai as their hub to support various tasks such as advanced AI workloads, ecommerce, banking, media and so on. Mumbai accounts for nearly half of India’s operational data centres, with new campuses expanding rapidly. Its connectivity and ecosystem make it indispensable, despite growing concerns about water scarcity.
What kind of water resources do data centres need?
Hyperscale centres require millions of litres daily to cool the facilities, making water availability critical to run them. Without sustainable sourcing, data centres risk competing with local communities and agriculture for limited supplies, raising environmental and social concerns that could undermine long‑term viability in water‑stressed regions like Maharashtra. This year, water supply from local reservoirs in the state have dropped to around 10% of their capacities, forcing the local civic body to take measures to curb wastage. Data centres remain unaffected as the state government has classified them as ‘essential services’. On 22 June, ratings agency Moody’s flagged water stress due to rising demand from data centres and water-intensive sectors.
How much investment is going into data centres in India?
India’s data centre sector is witnessing unprecedented investment. Tax holidays until 2047, announced in this year’s annual budget, rising AI demand and cloud adoption are driving a surge in investments. Global giants — Amazon, Microsoft, Google — have pledged multi‑billion‑dollar expansions, while leading domestic players like the Adani Group and Reliance are building hyperscale campuses. Mumbai, Chennai, Hyderabad and Pune are key beneficiaries. According to the Economic Survey 2026, India’s data centre capacity will expand from 1.4 GW in 2025 to 8 GW in 2030.
Could water play a spoilsport in India’s data centre expansion plans?
Yes. Water scarcity is a looming risk for India’s data centre boom. Maharashtra already faces deficits, with reservoirs running low and urban areas rationing supply. Rajasthan, Karnataka and Tamil Nadu also face shortages, compounding risks. Agriculture consumes nearly 80% of India’s freshwater, leaving limited scope for industrial use. As AI workloads drive water demand to run hyperscale centres, water stress could trigger community backlash, regulatory hurdles and operational disruptions. India must enforce water‑efficient technologies and promote alternative sourcing for sustainable growth, lest water constraints undermine its ambition to emerge as a global AI hub.
Are there technology solutions to scale data centres even with limited water?
Companies can mitigate water risks by adopting advanced cooling technologies. Closed‑loop liquid cooling recycles water with minimal loss, while immersion cooling submerges servers in non‑conductive fluids, reducing loss due to evaporation. A 23 June Nvidia post on X recommends 45°C liquid cooling in favourable climates where dry coolers can replace water-cooled, tower-based systems. This could reduce water use from about 2.6 million gallons per megawatt per year to near zero. Global companies are testing “water‑positive” operations, replenishing more than they consume. Regulators could mandate water audits, incentivize reuse, and prioritize new centres in less stressed regions.