The Hidden Water Cost of India’s Race to Become an AI Superpower

India is in the midst of the most ambitious digital infrastructure expansion any developing nation has ever attempted. Governments are offering tax holidays. Conglomerates are committing billions. Prime Ministers are posing for billboards at AI summits. What is not on the billboards, or in the press releases, is the water bill.

Beneath the gleaming announcements of gigawatt-scale data centres lies an increasingly urgent question: in a country that holds 18% of the world’s population but just 4% of its freshwater, who pays the environmental price for the AI boom?

The Scale of What’s Coming

Between March 2025 and April 2026, operators announced roughly 30 large projects across India, collectively adding about 3.5 gigawatts of planned capacity. Andhra Pradesh and Telangana alone account for over 2 GW, driven by AI-centric mega-campuses in Visakhapatnam and Hyderabad.

The investment roster reads like a who’s who of global and domestic capital. Google announced plans to establish an AI and data centre hub in Visakhapatnam, Andhra Pradesh, its first AI-focused facility in India, as part of a five-year, $15 billion investment plan from 2026 to 2030, developed in partnership with AdaniConneX and Airtel and expected to deliver “gigawatt-scale computing power.” Reliance Industries is building a 3 GW AI-ready data centre in Jamnagar, Gujarat, which Bloomberg has reported could become one of the largest such facilities in the world by capacity. Larsen & Toubro has partnered with NVIDIA to establish a sovereign AI factory as part of the IndiaAI Mission, including a new 40 MW AI-ready data centre in Mumbai.

India’s built data centre capacity is projected to nearly triple from 1.6 gigawatts in 2025 to about 5 gigawatts by 2030, growing at a compound annual rate of 26% over five years. The economic logic is hard to argue with: the digital economy is expected to contribute 20% of India’s national income by 2030, and the Union Budget 2026–27 announced a tax holiday until 2047 for eligible foreign cloud service providers operating through India-based data centre infrastructure.

What is not factored prominently into those projections is water.

The Cooling Problem

Every AI query, every cloud-stored document, every model training run generates heat. That heat must go somewhere. In India’s climate, the answer is almost invariably water.

A typical 100 MW hyperscale facility using water-based evaporative cooling can consume around 800,000 litres of water per day for on-site cooling alone, according to the Council on Energy, Environment and Water. The rise of generative AI has sharply increased these cooling demands. Widely-cited estimates suggest that even a short AI chatbot exchange may consume on the order of half a litre of water, though researchers caution that such per-query figures vary enormously by model, location, and cooling method, and should be treated as rough illustrations rather than precise measurements.

The problem is not just volume; it is permanence. Data centres typically evaporate about 80% of the water they consume. Most facilities in India rely on evaporative cooling due to high ambient temperatures, and the water used to dissipate heat is simply lost to the atmosphere. By one Deccan Chronicle estimate, a large 100 MW facility can draw close to two million litres a day, comparable to the daily needs of several thousand households.

There are compounding, less visible costs too. AI’s enormous electricity demands draw power from plants (often coal or nuclear in India) that themselves require large volumes of water to cool their turbines. This off-site water footprint sometimes exceeds what the data centre consumes directly.

A Nation Already Under Stress

India supports 18% of the global population but has access to only 4% of global freshwater resources. A 2025 analysis by UK-based non-profit Planet Tracker found that as many as 50 of India’s data centres are already located in “extremely high” water-stressed regions.

The hotspots are also, not coincidentally, India’s primary data centre hubs. In Mumbai, declining lake levels recently forced a 10% water cut across the city. In Hyderabad, groundwater levels in IT hubs like Gachibowli dropped nearly a metre in just three months in early 2026. Chennai is already caught in an intense competition between drinking water and industrial cooling water.

Visakhapatnam, where Google’s flagship AI hub is now planned, presents a particularly sharp irony. Visakhapatnam district has the lowest levels of groundwater available for domestic, agricultural, or industrial use in Andhra Pradesh, at just 2.12 TMC as of April 2026. The state government’s data centre policy proposes seawater cooling as a partial answer, but analysts have flagged unresolved questions around coastal zone regulations and potential impacts on biodiversity and fishing communities.

Looking forward, the picture worsens. Industry estimates, drawn largely from consultancy and market-research projections rather than official government data, suggest data centre water consumption in India could rise from about 150 billion litres in 2025 to 358 billion litres by 2030, more than doubling in five years. An S&P Global study forecasts that 60–80% of India’s data centres could face high water stress within this decade.

The Transparency Gap

Perhaps the most troubling dimension of this story is not the consumption itself, but how little anyone including regulators, communities and journalists actually know about it.

India’s data centre industry faces mounting scrutiny over water usage transparency. Despite claims of “water-neutral” facilities, companies like AdaniConneX and Sify provide scant evidence to support these assertions, and experts warn that without independent verification, corporate sustainability claims risk becoming digital greenwash.

AdaniConneX has no standalone ESG report. Its parent company Adani Enterprises Ltd reported total water withdrawals across all operations (mining, airports, solar manufacturing, data centres, and more) at 4,390 million litres in 2025, a 59% increase from 2022, without disaggregating how much was attributable to data centre operations specifically.

Globally, only about half of data centre operators tracked their water usage in 2020, and just 10% did so across all their facilities. Google’s global water consumption increased by 17% in 2023 alone.

When India’s Internet Freedom Foundation filed an RTI with the Telecom Regulatory Authority of India seeking data on environmental impacts of data centres, the Ministry of Electronics and Information Technology responded that it had zero information about data centres: the very ministry that created India’s draft data centre policy in 2020.

A Governance Vacuum

India has no national policy framework to guide data centre development or mandate transparency on water and energy use, even as both are expected to more than double by 2030. Regulations are left to state governments, where the incentive is to attract investment rather than constrain it. Only five of the 15 Indian states with data centre policies mention sustainability-related issues in their guidelines at all.

In Maharashtra and Telangana, data centres are classified as “essential services,” giving them priority access to water and electricity. In times of scarcity, they would be prioritised over the drinking water and sanitation needs of residents.

The contrast with peer economies is stark. China’s Green Data Centre Standards set caps on water-to-energy ratios. Singapore’s Green Data Centre Roadmap establishes rigorous monitoring and reporting standards. Malaysia has introduced national guidelines on power, water, and carbon efficiency, increasingly linked to project approvals. India has no equivalent.

What Responsible Growth Would Require

Solutions are available, though they require policy commitment rather than just corporate pledges.

Closed-loop cooling systems could reduce freshwater use by 70% or more. Treated wastewater, renewable-powered facilities, and more deliberate site selection in lower-stress locations are all viable alternatives. Mandatory reporting requirements and water usage limits, similar to environmental disclosure requirements in other industries, would be a baseline minimum to ensure rapid digital expansion does not create disproportionate environmental burdens.

The Council on Energy, Environment and Water has put it plainly: decisions being taken now on siting, power sourcing, and cooling technologies will lock in land, energy, and water impacts for decades. Without integrated planning, the sector risks becoming a systemic liability rather than a national asset.

The Question India Has Not Asked

India’s AI ambitions are legitimate. The desire for digital sovereignty is understandable. But the current trajectory reflects a choice, not an inevitability.

When a single data centre consumes the daily water needs of thousands of households, and when those households live in villages already watching their borewells deepen year by year, the question is no longer purely technical. It is a question of whose needs count, and whose water gets to be called infrastructure.

The servers will run around the clock. The cooling towers will not stop. And across the country, in the aquifers beneath village fields and urban neighbourhoods alike, the water table has nowhere to go but down.

References:

https://www.marketsandmarkets.com/blog/ICT/Future-of-the-Data-Center-and-AI-Infrastructure-Ecosystem-in-India

https://www.datacenterknowledge.com/data-center-construction/google-to-build-15b-ai-data-center-hub-in-india

https://www.outlookbusiness.com/deeptech/ai-expansion-india-data-centre-investments-report

https://www.crnasia.com/india/news/2026/india-s-ai-data-centre-boom-is-real-but-execution-not-announcements-will-decide-outcomes

https://www.outlookindia.com/national/ai-impact-summit-2026-water-stress-from-data-centres-a-cause-for-concern

https://www.ceew.in/blogs/why-is-water-based-cooling-a-big-issue-for-ai-data-centres-in-india

https://www.ceew.in/publications/how-is-data-centre-infrastructure-in-india-shaping-power-and-water-use

https://www.reccessary.com/en/news/india-ai-data-center-water-crisis

https://www.bloomberg.com/graphics/2025-ai-impacts-data-centers-water-data

https://www.computeforecast.com/Opinion/india-data-center-water-challenge-shapes-infrastructure-growth

https://www.deccanchronicle.com/nation/current-affairs/indias-ai-data-centres-and-the-invisible-groundwater-cost-1959255

https://www.downtoearth.org.in/science-technology/indias-digital-thirst-what-data-centre-giants-arent-saying-about-their-water-use

https://www.downtoearth.org.in/science-technology/indias-digital-thirst-data-centres-are-rising-in-water-scarce-regions-and-locals-are-paying-the-price

14. Earth Journalism Network: India’s Tech Boom Collides With Deepening Water Crisis in Bengaluru

15. Earth Journalism Network: What Data Center Giants Aren’t Saying About Their Water Use in India

16. Mongabay India: India Bets on Data Centres Even as Water, Energy-Use Concerns Mount

17. CBC News: India Is Going All-In on AI Data Centres. The Environmental Costs Will Have to Wait

18. Outlook Business: Cooling the Cloud, Draining the Cities? India’s Data Centre Dilemma

19. ResearchGate / PUIRP: Data Centers and Water Crisis in India: Why Digital Infrastructure Could Drain Our Wells Dry by 2030

Banner Image: Canva

Manjori Borkotoky
Manjori Borkotoky
Articles: 212