Council on Energy, Environment and Water Integrated | International | Independent
Context: India’s data centre capacity is set to expand, driven by digitisation and AI adoption, with electricity demand from data centres projected to increase nearly fivefold by 2030.
Challenge: Concentrated, 24x7 power and cooling needs could strain local grids, slow power-sector decarbonisation, and intensify water stress around major urban hubs.
CEEW recommendation: Integrating data centres into energy, water, and land-use planning can help turn sectoral growth into a sustainable digital asset rather than a systemic risk.
Data centres are centralised facilities that house computing infrastructure to store, process, and transmit large volumes of digital information to distributed end users. Their use has grown in tandem with the rapid digitisation of the global economy, with the United States (US), Europe, and China collectively accounting for 85 per cent of global data centre electricity consumption in 2024.
The recent surge in demand for artificial intelligence (AI) software and services is driving the emergence of AI-focused data centres that require far greater processing power than conventional data centres. While traditional data centres mostly handle general computing, data storage, and retrieval, AI data centres use specialised hardware, such as numerous graphics processing units, for faster parallel processing, along with advanced cooling, power, and networking infrastructure. These facilities support both AI model training, which entails a high, short-term energy load, and inference (running queries), which creates a constant, growing power demand as applications and user bases expand. Conventional data centres typically consume 10–25 MW, while hyperscale AI facilities can exceed 100 MW, comparable to heavy industries such as aluminium smelters.
Despite having around 15 per cent of the world’s internet users, India hosts only three per cent of the world's data centres. However, this share is projected to rise due to India’s large and growing demand base, data localisation regulations, swift digitisation and an established IT services ecosystem that is increasingly adopting AI-driven processes. The Government of India has been active in strategising for and driving data centre growth through the recently announced IndiaAI Mission, and subsidies from the central and state governments. This policy push signals that data centres are being positioned as core digital infrastructure for India’s next phase of growth, which, through AI, could drive innovation and have a multiplier effect on the economy. As of April 2025, India's IT load capacity was approximately 1.4 GW. Based on under-construction and planned projects, the IT load is projected to increase to 6.5–17 GW by 2030.
There are strong imperatives for India to manage data centre growth sustainably, as unchecked deployment could significantly impact energy planning, decarbonisation, and water use. This blog reviews lessons from other regions that have experienced rapid data centre growth, examines potential issues in the Indian landscape that can be exacerbated by data centre demand, and identifies key research areas to support the sector’s sustainable development.
Data centre operations demand 24x7 uninterrupted power, effectively creating a continuous baseload demand. To ensure reliability, they depend on redundant grid connections, batteries, and backup generators. Therefore, countries with reliable and affordable power grids are best positioned to host such infrastructure. In both the US and China, most electricity consumed by data centres is generated from fossil fuels. In China, additional demand from data centres is largely met by increased generation from existing coal-fired power plants. In Ireland, the booming data centre sector has pushed electricity consumption to over 21 per cent of the national demand in 2023, prompting a moratorium on new data centre construction until 2028 to protect climate targets. In the US, where data centres have increased their share of total electricity consumption from 1.4 per cent in 2014 to 4.4 per cent in 2023, rapid deployment has delayed coal retirements and driven utilities to add gas-based power generation to meet high demand and reliability.
These challenges stem from a lack of preparedness for data centre demand growth. Sudden, concentrated demand creates competition for grid power, strains stability, and forces utilities to quickly augment power generation and supply capacity, thus raising prices for all grid users. It is estimated that electricity prices in parts of the US rose by 2.7 times compared to 2020 levels in areas close to new data centres. While data centres are expected to contribute less than 10 per cent of global power demand growth through 2030, the International Energy Agency alerts that their highly localised demand spikes can be far more challenging to integrate than distributed loads.
In India, data centre expansion is projected to increase electricity demand nearly fivefold by 2030, from ∼13 TWh in 2024 (0.8 per cent of total demand) to ∼57 TWh (2.6 per cent). We expect this demand to concentrate in regions around major urban load centres such as Mumbai, Bengaluru, Hyderabad, and Chennai. Meeting this surge through grid-scale renewable energy (RE) coupled with battery storage systems is a viable possibility. However, India already faces significant challenges across the RE value chain. These include grid integration constraints such as transmission capacity bottlenecks, as well as project development and market barriers, including land acquisition challenges and the lack of long-term offtake agreements.
Research by the Council on Energy, Environment, and Water (CEEW) indicates that grid-scale RE will increasingly face intermittency, higher land acquisition costs, land-use conflicts, and extreme climate events. Deploying sufficient RE capacity for data centre demand near major cities may hence prove difficult. Locating RE plants farther away would entail substantial investments in transmission infrastructure, adding pressure on the already strained finances of distribution companies (discoms).
In this context, local utilities could simply ramp up coal-based power generation to meet data centre demand, potentially slowing power-sector decarbonisation. For example, following the post-COVID economic recovery from 2021, India’s power-sector emission factor has increased despite record RE additions, with coal fulfilling more than 50 per cent of the additional annual generation.
Data centres tend to cluster in specific regions, creating localised stresses on power and land. This clustering is driven by proximity to fibre-optic networks, submarine cable landing stations, reliable electricity supply, and skilled workforce availability. Supportive policies, such as tax incentives, further reinforce the emergence of these hubs. Thus, while the national or global energy demand implications for data centre growth remain relatively modest at present, the local impacts are more pronounced.
Apart from power generation and supply challenges, data centres can also intensify local water stress due to their cooling needs. Although water use varies by cooling technology, data centre design, and local climate, 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. Given India’s high ambient temperatures and the extreme heat densities of modern AI workloads, liquid cooling is expected to become the dominant and most efficient technology for AI-focused data centres.
While liquid cooling in servers is typically closed-loop and uses very little water, data centres that rely on water-based heat-rejection systems, such as cooling towers, can have high overall water consumption, thereby exacerbating local water stress. Thus, more data centres mean more freshwater diversion for cooling, often coinciding with India’s recurrent water shortages and summer heat stress. In the US, it is reported that two-thirds of new data centres built since 2022 are in areas already facing water stress.
CEEW study shows that 57 per cent of districts are at high to very high risk from extreme heat. Many key data centre hubs are located in such districts. If data centres depend heavily on groundwater or municipal sources, they may end up competing with agriculture and households. Exposure to extreme heat could also raise cooling loads substantially during summer months or aggravate peak-load management issues that may reinforce reliance on fossil-based generation. Climate vulnerabilities compound these concerns: CEEW research also indicates that the same districts are also vulnerable to extreme climate events. Therefore, siting data centres in these districts could lead to potential data losses and service outages, resulting in financial losses especially as AI increasingly becomes essential in many areas.
As mentioned before, both the centre and state governments have rolled out policies to attract data centre investments. However, countries that have already witnessed or foresee significant data centre growth face data centre-driven environmental and energy challenges. They have responded with a mix of siting policies, zoning controls, minimum sustainability standards, and integrated infrastructure planning. At a regional level, the European Union’s (EU) latest Energy Efficiency Directive requires data centres with total energy consumption above 1 MW to utilise waste heat recovery where feasible. For siting incentives, South Korea offers a 50 per cent discount on the electricity facility levy for data centres built outside the Seoul metropolitan area. Singapore introduced a temporary moratorium on new data centres and later allowed only highly efficient facilities in designated zones under its Green Data Centre Roadmap. The Netherlands has implemented zoning restrictions and caps on hyperscale developments, while Ireland has linked grid-connection approvals to system constraints and flexibility requirements, effectively steering data centres towards locations with sufficient grid capacity. EU countries and China, among others, now have minimum standards for data centre water and energy efficiency.
In India, a comprehensive regulatory framework for data centres' environmental performance is still evolving. At present, there are no sector-wide water- or energy-based performance standards for data centres, nor any mandatory requirements to report water use, electricity consumption, or emissions. The Draft Data Centre Policy 2020 signals intent toward sustainability but does not yet translate it into enforceable provisions.
Guidance on energy efficiency for data centres was last issued by the Bureau of Energy Efficiency in 2010. Data centres are also not currently classified as ‘Designated Consumers’—a category that includes large, energy-intensive industries such as iron and steel, aluminium, and fertilisers, which are legally required to undertake periodic energy audits, appoint certified energy managers, report energy use, and comply with sector-specific energy-efficiency targets.
As data centre capacity expands, particularly with the growth of large hyperscale facilities, policymakers could consider the merits of extending the Designated Consumer framework, or elements of it, to this sector. Such an approach could enable systematic monitoring of energy use, strengthen accountability, and support efficiency improvements in line with India’s broader energy-governance objectives.
From a planning perspective, both central and state governments could take coordinated action to:
India’s digital ascendancy is inevitable, but its environmental cost is not. Policymakers and all stakeholders must deliberate on prioritising integrated energy, water, and climate planning to transform the data centre boom from a potential environmental and economic liability into a sustainable and strategic national asset.
Dr Aman Malik and Sabarish Elango are Programme Leads at the Council on Energy, Environment and Water (CEEW). Send your queries to [email protected].
