The AI Boom's Unseen Water Footprint
Artificial intelligence feels digital and weightless, but it runs on physical infrastructure with a very real and heavy environmental footprint. Data centres, the sprawling homes for servers that train and run AI models, generate immense heat. The primary
method for cooling these facilities is through evaporative cooling, a process that consumes enormous quantities of water. A single 100-megawatt hyperscale data centre can use about 2 million litres of water daily. With India's data centre capacity projected to skyrocket, water consumption by the sector is expected to more than double by 2030. This surge is happening in a country where 600 million people already face high to extreme water stress and which holds only 4% of the world's freshwater reserves despite having 18% of its population. The conflict is stark: the demand for computational power is growing in the very regions—like Mumbai, Chennai, and Hyderabad—that are already facing significant water challenges.
The Law of the Land: CGWA Rules
The Indian government has made it clear that the rules of the game for water use have not changed for the new digital economy. The Central Ground Water Authority (CGWA), established under the Environment (Protection) Act, 1986, is the key regulator. Its mandate is to ensure sustainable groundwater management across the country. For any industrial or commercial project, including data centres, that intends to extract groundwater, a No Objection Certificate (NOC) from the CGWA is mandatory. This isn't a simple rubber stamp. The approval process requires detailed hydrogeological reports, plans for rainwater harvesting, and the installation of digital flow meters to monitor usage. Crucially, the regulations are strictest in areas classified as 'Over-Exploited', where new industries are generally not granted NOCs for groundwater extraction. The government has affirmed that these existing regulations fully apply to AI data centres, meaning they cannot draw water without oversight.
A Collision of National Priorities
This regulatory stance highlights a fundamental tension in India's development goals. On one hand, ambitious national programmes like 'Digital India' and the IndiaAI Mission are aggressively promoting the country as a global hub for technology and data processing. States like Uttar Pradesh and Gujarat are rolling out policies to attract massive investments in data centre capacity. On the other hand, the nation is grappling with the severe impacts of climate change and dwindling water tables. Local communities in places like Visakhapatnam and Thane have already protested against proposed data centre projects, fearing the strain on local water and electricity resources. The government's clarification that data centres must adhere to existing groundwater rules serves as a critical check, forcing the tech industry to reckon with the environmental cost of its expansion rather than assuming a special status.
Industry's Response and the Path to Sustainability
Faced with stringent regulations and growing water scarcity, the data centre industry is being pushed towards innovation. The most significant shift is the move away from traditional evaporative cooling towards more water-efficient technologies. Closed-loop liquid cooling, including direct-to-chip and immersion cooling, is becoming a preferred option for new AI-focused facilities. These systems use a sealed coolant loop that virtually eliminates the massive water loss associated with evaporation. While more expensive upfront, these technologies are seen as essential for long-term operational viability in water-stressed India. Additionally, policies are starting to mandate higher efficiency. The Ministry of Electronics and Information Technology (MeitY) has set stricter Power Usage Effectiveness (PUE) benchmarks, and state-level policies are increasingly including requirements for wastewater recycling and even zero liquid discharge.
