By Soumyashree Mohanty, ImpactLens Correspondent
Technology is developed to make our lives easier. As society progresses, new technologies continue to change the way we work, communicate and solve problems. But, as the saying goes, every coin has two sides. While technology brings speed, convenience and new possibilities, it also has an environmental cost that deserves attention. Artificial intelligence (AI) is one such technology. In this article, we explore how AI consumes water and, at the same time, how AI-enabled systems can help us manage and conserve water more effectively. The article also looks at India’s growing adoption of AI data centres and the water-related concerns associated with their expansion.,
Does Your AI Prompt Really Use Water?
Most of us are now familiar with tools such as ChatGPT, Gemini, DeepSeek and Claude. We use them to write, search for information, analyse data, create images and complete many other tasks within seconds. But behind this convenience is a large digital infrastructure that requires computing power, electricity and water.
Take a very simple question: you ask ChatGPT, “Where is India?” It may seem like such a small question that it would have almost no environmental cost. But even this simple query requires computing power. OpenAI CEO Sam Altman has said that an average ChatGPT query uses about 0.000085 gallons of water. That is a very small amount—about one-fifteenth of a teaspoon. However, water use can vary depending on the AI model, the complexity and length of the task, the hardware being used, and the location and cooling system of the data centre. An estimated 500 millilitres (0.5 litres) of water may be used for every 5–50 prompts in a ChatGPT conversation. Research comparing different large language models has also found significant differences in energy consumption, with more computationally demanding models generally requiring more energy. Higher energy use can be linked to a higher water footprint, although this depends on the infrastructure being used (Jegham et al., 2025).
A single question may require only a tiny amount of water. But imagine the number of questions being asked every day by millions of people around the world. At that scale, even a small water footprint can become significant.
So, Where Does the Water Come In?
AI does not directly consume water when we type a prompt. The data centres that run AI systems may use water to manage the heat produced by their servers. When you send a prompt, powerful computer chips inside servers process the request. These chips perform large numbers of calculations, generating heat in the process. If this heat is not removed, the equipment can become too hot, affecting its performance and potentially damaging the hardware.
Data centres therefore use different cooling systems to control temperature. Some rely mainly on air cooling, while others use liquid-based cooling systems. The amount of water used depends on the technology and design of the facility.
In water-based cooling systems, water or another cooling liquid helps carry heat away from the equipment. The heated liquid is then cooled through a heat-exchange or cooling system and returned for reuse. In cooling towers, some water can be lost through evaporation, while the remaining water may be reused or discharged, depending on the system.
Therefore, the water footprint of AI is not only about the water used to cool one computer. It is part of a much larger system involving servers, cooling equipment, electricity generation and the infrastructure required to operate data centres.
Water Behind Technology
The water footprint of AI does not end at the data centre. Water is also required to manufacture the semiconductor chips that power computers and AI systems. Semiconductor manufacturing involves several water-intensive processes, including cleaning and processing materials. This means that water is involved at different stages of the AI supply chain—from manufacturing the chips to operating the data centres that run AI applications. Water is essential for drinking, sanitation, agriculture, ecosystems and almost every form of life. At the same time, the world is rapidly increasing its dependence on AI, which requires more computing infrastructure and data centres.
The Other Side: Can AI Also Help Save Water?
AI-enabled technologies can support water management across different areas, including irrigation, leak detection, water-quality monitoring and urban water planning. By using information from sensors, satellites, weather systems and other sources, AI can help people understand water conditions, identify risks and make more informed decisions about how water is used.
Efficient Irrigation
By integrating AI-based models into irrigation systems, water can be used more efficiently. AI-enabled irrigation systems can use information from soil-moisture sensors, weather forecasts, satellite data, and crop conditions to determine when and how much water crops need. IBM’s Watson Decision Platform for Agriculture, for example, combines AI, machine learning, IoT and agricultural data, including soil moisture, weather forecasts, and crop-stress information, to support irrigation and farm-management decisions (Ministry of Agriculture & Farmers Welfare, 2019). Research also shows that deep-learning models can improve the estimation and forecasting of evapotranspiration, which can support better irrigation planning according to crop water requirements (Sarkar et al., 2025).
Early Leak Detection
AI can also support early leak detection in water pipelines. By processing information from satellites and other monitoring systems, AI-based tools can identify signs of leakage before they become visible. In London, Thames Water has partnered with Origin Tech to use AI-powered satellite technology to locate leaks in its water network. Following successful trials, the company expects the technology to help reduce leakage by more than 100 million litres per day (Thames Water, 2026). Such systems can help water utilities locate problem areas faster and prioritise repairs.
Water-Quality Monitoring
Another important application is water-quality monitoring. Conventional monitoring often involves collecting samples and carrying out periodic laboratory tests, which can take time and may not provide continuous information. IoT sensors can continuously measure parameters such as pH, turbidity, temperature, dissolved oxygen, electrical conductivity and total dissolved solids. Machine-learning models can then analyse these measurements to detect unusual changes, assess water-quality conditions and provide early warnings of possible contamination (Essamlali et al., 2024).
Water-Sensitive Urban Planning
At the city level, AI can be combined with Geographic Information Systems (GIS) to support water-sensitive urban planning. These tools can help planners understand how buildings, land use, green spaces and drainage systems influence water demand and rainwater flow. AI can identify areas vulnerable to excessive heat or flooding, helping planners decide where to introduce trees, green roofs, rain gardens, drainage systems and water-storage areas. It can also support the planning of greywater reuse, such as using treated water from showers and sinks for toilet flushing or gardening.
India’s AI Expansion
The global discussion around AI and water becomes particularly important for India as the country expands its AI and data-centre infrastructure. Data-centre projects in several countries are already facing concerns about water use, electricity demand and their impact on local communities. India, meanwhile, is witnessing a rapid influx of investment in AI infrastructure.
Microsoft announced a $17.5 billion investment in India from 2026 to 2029, its largest investment in Asia. The company said the investment will expand its cloud and AI infrastructure, hyperscale data-centre capacity, skilling programmes and operations across the country (Microsoft, 2025). Google has committed approximately $15 billion over five years, from 2026 to 2030, to establish an AI hub in Visakhapatnam, Andhra Pradesh. The project includes a gigawatt-scale data-centre campus, new energy infrastructure and an international subsea cable gateway. Google describes it as its largest investment in India to date (Google, 2025). Meta has also entered into an agreement with Reliance Industries to lease a 168 MW AI-enabled data centre in Jamnagar, Gujarat. Meta says the facility will be powered by renewable energy and cooled using desalinated seawater, with the option to expand its capacity (Meta, 2026).
The scale of these investments raises questions about the resources needed to operate this expanding digital infrastructure. Water is emerging as one of the key concerns. A 2026 analysis by WRI India found that more than half of India’s data centres are located in water-stressed regions. It also found that 75% of the country’s data centres are concentrated across Maharashtra, Tamil Nadu, Karnataka, Telangana and Uttar Pradesh, with many facilities located in areas classified as having high or extremely high-water stress (World Resources Institute India, 2026).
The concern is particularly significant because data centres require water for cooling, although the amount used depends on the cooling technology, climate and source of water. The growing concentration of data centres in water-stressed regions therefore raises an important question: how can India expand its digital infrastructure without adding pressure to already limited water resources?
Protests Against Data Centres in India
The debate has become particularly visible in Visakhapatnam, where Google’s $15 billion AI hub, now under construction, has faced opposition from environmental groups and activists. The city has a reported daily water shortfall of around 70 million litres, with activists citing a supply of about 410 million litres against an estimated requirement of 480 million litres. Critics have raised concerns about the project’s potential impact on water resources and the nearby Kambalakonda Wildlife Sanctuary, and legal challenges have been filed over these concerns (Reuters, 2026).
The opposition has also taken the form of public protests. Activists and children have participated in demonstrations in Visakhapatnam, including campaigns highlighting concerns over water use and the project’s environmental impact. The Andhra Pradesh government has rejected allegations that residential or rural drinking-water supplies will be diverted to the project. Google, meanwhile, has said it plans to use advanced air-cooling and sound-dampening technologies to reduce the facility’s environmental impact (Reuters, 2026).
In Thane, Maharashtra, a proposed 54-acre data centre in Balkum has also drawn political and public opposition. In June and July 2026, activists, residents and political representatives raised concerns about the project’s potential impact on the city’s green cover and already stretched water and electricity infrastructure, as well as questions about its broader environmental impact (The Times of India, 2026).
These developments point to a challenge India will increasingly have to address as AI infrastructure expands. The question is not whether India should develop AI or build data centres, but how it can do so while protecting water security and reducing pressure on resources shared by households, agriculture and other industries. AI may consume water, but the same technology can also help us manage water more efficiently. The challenge is to ensure that the growth of AI does not create new pressures on resources that are already under stress. As India builds the infrastructure for its AI future, water security must remain part of the conversation.
References:
- Ministry of Agriculture & Farmers Welfare. (2019, July 3). Agriculture Ministry signs a Statement of Intent with IBM for pilot study to utilize artificial intelligence & weather technology solutions in agriculture. Press Information Bureau, Government of India
- Essamlali, I., Nhaila, H., & El Khaili, M. (2024). Advances in machine learning and IoT for water quality monitoring: A comprehensive review. Heliyon, 10(6), e27920. https://doi.org/10.1016/j.heliyon.2024.e27920
- Google. (2025, October 14). Our first AI hub in India, powered by a $15 billion investment. Google India.
- Google. (2026, April 28). Google AI Hub: Scaling India’s AI infrastructure and community impact. Google India.
- Jegham, N., Abdelatti, M., Elmoubarki, L., & Hendawi, A. (2025). How hungry is AI? Benchmarking energy, water, and carbon footprint of LLM inference. arXiv. https://doi.org/10.48550/arXiv.2505.09598
- Meta. (2026, June 9). Meta partners with Reliance on AI-enabled data centre in India. Meta Newsroom.
- Microsoft. (2025, December 9). Microsoft invests US$17.5 billion in India to drive AI diffusion at population scale. Microsoft Source Asia.
- Ministry of Agriculture & Farmers Welfare. (2019, July 3). Agriculture Ministry signs a Statement of Intent with IBM for pilot study to utilize artificial intelligence & weather technology solutions in agriculture. Press Information Bureau, Government of India.
- Reuters. (2026, August 6). Google’s $15 billion India data centre project battles water, wildlife concerns. Reuters.
- Sarkar, S. S., Bedi, J., & Jain, S. (2025). A deep learning based framework for enhanced reference evapotranspiration estimation: Evaluating accuracy and forecasting strategies. Scientific Reports, 15, 15136. https://doi.org/10.1038/s41598-025-99713-2
- Thames Water. (2026, May 1). Thames Water signs landmark leak detection contract with Origin Tech to save over 100 million litres of water per day.
- The Times of India. (2026, June 20). NCP (SP) warns of water crisis, health risks from 54-acre data centre planned in Thane. The Times of India.
- World Resources Institute India. (2026, May 18). More than half of India’s data centres are in water-stressed regions. WRI India.
