Newton County in the American state of Georgia and Tusiana village near Greater Noida in Uttar Pradesh appear to have little in common. They differ in language, culture and geography. Yet they are connected by a troubling reality: both have experienced the growing pressure that large AI-driven data centres can place on local water resources.
In Newton County, residents reported water shortages and declining well levels following the construction of a major Meta data centre. In Tusiana, villagers living near the Yotta Data Centre Park have witnessed a significant fall in groundwater levels over the past two decades, forcing many households to dig substantially deeper wells to access water. These are not isolated local concerns. They are early warnings of a much larger challenge that could accompany the rapid global expansion of artificial intelligence infrastructure. The roots of that challenge are now taking firm hold in India.
India stands at the threshold of a new industrial revolution driven by data centres. Government figures indicate that the country's data-centre capacity increased from 0.4 gigawatts in 2020 to 1.5 gigawatts by the end of 2025—almost a fourfold rise in just five years. According to estimates by Deloitte, an additional 8–10 gigawatts of capacity could be added by 2030. A September 2025 report by market research firm Arizton noted that 132 data centres were operating in India, while another 84 projects were planned across 17 cities by 2029.
Prime Minister Narendra Modi, while inaugurating the AI Impact Summit, highlighted the employment opportunities that data centres could generate and invited the world to store its data in India. The Adani Group has announced plans to invest $100 billion in renewable-energy-powered hyperscale data centres by 2035, while technology giants Microsoft, Amazon and Google have each expressed readiness to invest billions of dollars in India's digital infrastructure.
At first glance, this appears to be a promising story of economic growth and technological advancement. Beneath that narrative, however, lies a critical question: where will the enormous quantities of water required to operate these facilities come from?
Data centres are essentially vast complexes filled with computers that process and store the digital activities of billions of people. Every online search, video stream, photograph upload and AI interaction relies on servers housed in such facilities. Because these servers operate continuously, they generate substantial heat. Cooling systems are therefore essential, and many of those systems depend heavily on water.
In India's hot climate, evaporation-based cooling remains the most common approach. Such systems consume large volumes of freshwater, with much of it lost permanently through evaporation. Unlike many industrial uses of water, a significant portion cannot be recovered or reused.
Karnataka's Information Technology Minister Priyank Kharge informed the state assembly in March 2026 that every megawatt of data-centre capacity requires approximately 2.5 crore litres of water annually. Applying that estimate to India's current capacity of 1.5 gigawatts suggests annual water consumption of nearly 37.5 billion litres. This figure could increase dramatically as the sector expands. A medium-sized data centre is estimated to consume around 11.35 lakh litres of water every day. According to a September 2025 report by Morgan Stanley, global water demand from AI data centres could rise elevenfold by 2028, reaching 1,068 billion litres annually.
The scale becomes even more striking when viewed against India's existing water challenges. Many major cities already face seasonal shortages. Reservoir levels fluctuate sharply during summer months, while groundwater depletion continues across large parts of the country. Water levels in key river basins such as the Krishna and Godavari have also experienced recurring stress. Yet many of the country's largest data-centre investments are concentrated in Maharashtra, Karnataka, Tamil Nadu and Telangana—states that already face significant water pressures.
Cities in southern India may be particularly vulnerable because groundwater reserves are already under strain. Himanshu Thakkar, coordinator of the South Asia Network on Dams, Rivers and People (SANDRP), has argued that the data-centre industry must take greater responsibility for water management. Bengaluru, for instance, depends on water pumped over long distances and significant elevations from the Cauvery River, making every additional demand on the system a matter of concern.
India already extracts more groundwater than any other country in the world, accounting for roughly a quarter of global groundwater withdrawals. In many regions, extraction far exceeds natural recharge rates. Unlike seasonal drought, this is a structural problem that accumulates year after year. Against this backdrop, the rapid growth of AI infrastructure risks intensifying existing tensions over water allocation.
Water disputes are already politically sensitive. The Cauvery conflict between Karnataka and Tamil Nadu has persisted for decades. As urban centres such as Bengaluru, Hyderabad and Chennai continue expanding their digital infrastructure, competition for limited water resources may become even sharper. Climate change adds another layer of uncertainty. El Niño conditions, associated with higher temperatures and weaker monsoon rainfall, can increase both water scarcity and cooling requirements for data centres, creating a reinforcing cycle of demand and stress.
Potential solutions do exist. Professor Arnab Bhattacharya of IIT Kanpur has advocated dual-pipeline systems that separate freshwater and recycled water supplies. Such systems are already being implemented in some institutions. Closed-loop cooling technologies can reduce freshwater consumption significantly, while treated wastewater can substitute for potable water in many cooling applications. Renewable-energy-powered cooling systems and careful site selection in regions with lower water stress also offer viable alternatives.
However, these measures require substantial investment, effective regulation and rigorous monitoring. Experts repeatedly point to a lack of transparency regarding water use by the AI industry. While some companies claim to employ efficient cooling systems and reduce consumption, independent verification mechanisms remain limited.
The contradiction is particularly visible in Maharashtra. Every summer, districts in Vidarbha and Marathwada depend heavily on water tankers. At the same time, large-scale approvals continue for hyperscale data centres in the Mumbai-Pune corridor. This raises difficult questions about policy priorities and resource allocation. There is currently no comprehensive mechanism to balance the water needs of rural communities, farmers and urban poor populations against the demands of large industrial investments.
Many water experts argue that mandatory public disclosure should be the first step. Data-centre operators should be required to regularly publish audited information on water consumption, recycling rates and sourcing practices. Similar reporting requirements already exist in parts of the United States and Europe. A comparable framework in India could improve public accountability and enable informed local decision-making. Regions facing severe groundwater depletion could also be designated as restricted zones for future data-centre development.
An often-overlooked aspect of the debate is its gender dimension. Across much of rural India, women continue to bear primary responsibility for collecting water. As groundwater levels decline and water sources move farther away, the burden on women increases. Longer journeys for water reduce opportunities for education, employment and healthcare access. Water scarcity is therefore not merely an environmental issue; it is also a question of social and gender justice.
Technology itself is neither inherently beneficial nor harmful. Its impact depends on how societies choose to develop and govern it. India is simultaneously one of the world's most ambitious adopters of digital technologies and one of its most water-stressed nations. The challenge lies in ensuring that technological progress does not deepen existing inequalities or environmental vulnerabilities.
Major technology companies have already disclosed substantial water consumption associated with their global operations. According to various estimates, AI-driven data centres could become among the most water-intensive components of the digital economy. Some projections suggest that by 2030, global AI infrastructure may consume water volumes equivalent to the needs of more than a billion people.
Mumbai is rapidly emerging as India's largest data-centre hub, with Maharashtra expected to account for nearly 45 percent of the country's planned data-centre power capacity. Bengaluru, Hyderabad, Chennai, Pune and Noida are also witnessing rapid expansion. Ironically, many of these locations are already grappling with recurring droughts and declining groundwater tables.
Some analysts contend that the real issue is not water availability but water management. India receives substantial annual rainfall, and its per-capita water consumption remains lower than that of many developed countries. Better storage, distribution and recycling systems could potentially accommodate both human needs and industrial growth. While this argument has merit, it cannot ignore the reality that improvements in water infrastructure have been delayed for decades. Data-centre expansion, by contrast, is proceeding at remarkable speed.
Ultimately, India's AI infrastructure boom is not simply a technological story. It is also a story about the allocation of a scarce and essential resource. The costs of water scarcity are rarely borne by those who benefit most from technological growth. Instead, they fall disproportionately on tribal communities, small farmers, rural women and urban slum dwellers—people whose daily lives already revolve around securing access to water.
A widely cited 2023 study from the University of California estimated that a series of interactions with AI systems can indirectly consume measurable quantities of water through cooling and electricity generation processes. Most users never see this hidden cost. Yet behind every digital interaction lies a physical infrastructure that depends on land, energy and water.
The challenge is not to reject AI or technological progress. Rather, it is to ensure that innovation proceeds within ecological limits and with democratic accountability. Before extending incentives and subsidies to data-centre developers, governments must establish robust water-accountability standards, mandatory disclosure requirements and meaningful consultations with affected communities.
The water crisis associated with AI is not yet fully visible. It remains a crisis of tomorrow. But unless action is taken today, tomorrow's warning signs may arrive in the form of dry wells, depleted aquifers and empty taps. The question is not whether India should embrace AI. The question is whether it can do so without sacrificing the water security of millions.
Comments
Post a Comment
NOTE: Hateful, abusive comments won't be published. -- Editor