The rapid expansion of AI data centres has intensified concerns about electricity consumption, water scarcity, environmental degradation and the rights of local communities. Read here to learn more.
Opposition to large-scale data-centre development has emerged in several countries, including the United States, Finland and Kenya, reflecting growing tensions between digital infrastructure investment and local environmental interests.
In India, petitions before the National Green Tribunal (NGT) have reportedly challenged aspects of the proposed USD 15 billion Google-Adani AI data-centre development in Visakhapatnam, Andhra Pradesh, particularly its environmental implications and land allocation.
The debate highlights an emerging policy challenge: How can India build the computing infrastructure its digital economy needs without compromising water security, energy affordability, ecological sustainability, and community welfare?
What Is a Data Centre?
A data centre is a physical facility that houses computing and networking infrastructure used to store, process, transmit and manage digital information.
- It typically contains servers, storage devices, networking equipment, cooling systems, power infrastructure and cybersecurity systems.
- Data centres support cloud computing, digital payments, e-governance, online communication and artificial intelligence.
How are AI Data Centres Different from Conventional Data Centres?
AI data centres are designed to handle computationally intensive workloads, particularly training and inferring machine-learning models.
Feature |
Conventional data centres |
AI data centres |
Primary workload |
Cloud services, storage, databases and enterprise applications |
AI model training and inference |
Computing infrastructure |
Primarily CPU-based servers |
Dense clusters of GPUs and specialised AI accelerators |
Power requirements |
Vary according to workload and facility size |
Often substantially higher per rack |
Cooling requirements |
Air cooling and conventional liquid-cooling systems |
Increasing use of advanced liquid cooling |
Network architecture |
General-purpose networking |
High-bandwidth, low-latency interconnections between accelerators |
Resource implications |
Electricity, water and land requirements |
Potentially greater power density and cooling complexity |
AI data centres are particularly energy-intensive because training and operating large models can require thousands of processors working simultaneously.
However, resource consumption varies considerably with model architecture, hardware efficiency, workload and cooling technology.
Why Does India Need to Expand its Data-Centre Infrastructure?
- Rapid Growth of the Digital Economy
India’s expanding digital ecosystem requires reliable computing and storage infrastructure.
Data centres support:
- Unified Payments Interface (UPI)
- Digital banking and financial services
- E-governance platforms
- Cloud-based business applications
- Telecommunications and digital entertainment
- Artificial intelligence and machine learning
The growing adoption of digital services is increasing demand for domestic computing capacity.
- Bridging India’s Data-Hosting Gap
- According to the Economic Survey 2025-26, India generates approximately 20% of global data but hosts only around 3% of the world’s data centres.
- The Survey estimated India’s data-centre capacity at approximately 1.4 GW in mid-2025, with projections of substantial expansion by 2030.
- This gap presents an opportunity to strengthen domestic digital infrastructure and attract investment.
- Strengthening Data Security and Digital Sovereignty
- Domestic data centres can improve operational control over critical digital infrastructure and reduce dependence on overseas computing facilities.
- They can also support compliance with applicable sector-specific data-localisation and cybersecurity requirements.
- However, the Digital Personal Data Protection Act, 2023, does not impose a universal requirement that all personal data be stored in India.
- Domestic hosting should therefore not be confused with a blanket legal localisation mandate.
- Supporting India’s AI Ecosystem
The IndiaAI Mission seeks to expand access to computing resources, promote indigenous AI development and strengthen India’s artificial intelligence ecosystem.
Domestic data centres can support:
- Training of Indian-language AI models
- Development of sector-specific AI applications
- Research and innovation
- AI adoption by startups and MSMEs
- Public-sector digital services
- Attracting Investment and Technological Capabilities
- Data-centre development can attract investment in computing infrastructure, fibre-optic networks, electricity systems and related services.
- The inclusion of eligible data centres in the Harmonised Master List of Infrastructure Sub-sectors has also supported their recognition as infrastructure assets.
- The Union Budget 2026-27 proposed tax incentives extending to 2047 for qualifying foreign cloud-service providers operating through India-based data-centre arrangements, subject to prescribed conditions.
What are the Environmental and Social Concerns Associated with AI Data Centres?
- Rising Electricity Demand and Grid Pressure
AI data centres require continuous, reliable electricity to operate computing equipment and cooling systems.
In July 2026, the Ministry of Power projected that AI data centres could add approximately 26.3 GW of electricity demand by 2031-32.
Such demand creates challenges involving:
- Electricity generation capacity
- Transmission infrastructure
- Grid stability
- Power quality and reliability
- Energy affordability
- Integration of renewable electricity
Large data-centre clusters can also concentrate electricity demand in particular regions, increasing the need for local grid upgrades.
- High Cooling Requirements and the Thermal Challenge
- Almost all electricity consumed by computing equipment is eventually converted into heat.
- Consequently, large AI facilities require continuous thermal management to maintain safe operating temperatures.
- Modern high-density AI racks can exceed 100 kW of power demand, making conventional air cooling increasingly difficult in some configurations.
Advanced cooling technologies include:
- Direct-to-chip liquid cooling
- Rear-door heat exchangers
- Immersion cooling
- Hybrid cooling systems
- Heat-recovery technologies
These systems can improve cooling efficiency, although their water and energy implications depend on the overall facility design.
- Water Consumption and Groundwater Stress
- Water is used in certain data-centre cooling systems, particularly evaporative cooling towers.
- High water demand can intensify competition among data centres, households, agriculture and industry.
- This is particularly concerning in regions already experiencing groundwater depletion or seasonal water scarcity.
- According to estimates cited in the policy discussion, India’s data-centre water consumption could increase substantially by 2030.
- However, water-use projections depend on the assumed cooling technologies, facility utilisation and accounting methods.
- Direct-to-chip liquid cooling does not necessarily eliminate water consumption. The cooling loop may be closed, while the facility’s external heat-rejection system can still consume water.
- Carbon Emissions and Fossil-Fuel Dependence
The environmental impact of data centres depends heavily on the electricity used to operate them.
Where electricity comes from coal-based generation, expanding computing demand can increase greenhouse gas emissions.
Other emission sources include:
- Backup diesel generators
- Construction materials
- Equipment manufacturing
- Refrigerant leakage
- Hardware replacement and disposal
Without corresponding growth in low-carbon electricity, AI infrastructure expansion could complicate India’s long-term Net Zero 2070 objective.
- Waste Heat and Local Environmental Impacts
- Data centres release substantial waste heat through cooling and ventilation systems.
- Large clusters may contribute to local thermal burdens, particularly in densely developed areas.
- However, their contribution to urban heat islands depends on facility design, heat-discharge mechanisms, local climate and surrounding land use.
- It should not be assumed that all data centres produce the same degree of local warming.
- Noise Pollution
- Cooling fans, chillers, transformers and backup generators can generate continuous mechanical noise.
- Potential impacts include disturbance to nearby residential communities and increased environmental noise exposure.
- The severity depends on equipment design, distance, acoustic insulation and operating conditions.
- Land Use and Ecological Disturbance
Large data-centre campuses require land for buildings, electricity infrastructure, cooling facilities and associated utilities.
Poorly planned development may lead to:
- Loss of vegetation
- Changes in drainage patterns
- Disturbance of natural habitats
- Increased pressure on nearby water bodies
- Land-use conflicts with local communities
Environmental risks become particularly significant when projects are located near reservoirs, wetlands, forests or protected areas.
- Limited Permanent Employment and Unequal Local Benefits
- Large data-centre projects can generate substantial employment during construction.
- However, permanent operations are often highly automated and may require relatively fewer workers.
- This creates a potential imbalance between local environmental costs and the distribution of economic benefits.
- Communities may bear the burden of land conversion, electricity infrastructure, water use and noise while many digital benefits accrue to users elsewhere.
- Such concerns can contribute to local opposition to data-centre expansion.
What does the Visakhapatnam Data-Centre Dispute Highlight?
- The proposed Google-Adani AI infrastructure development in Visakhapatnam has become an important example of the environmental governance challenges associated with hyperscale data centres.
- Petitions before the National Green Tribunal have challenged environmental clearances for a proposed 512 MW data-centre project, involving concerns about its location and environmental appraisal.
Major Issues Raised
- Environmental clearance: Petitioners have questioned whether the project was appropriately assessed under the building and construction category of the EIA Notification, 2006.
- Ecological sensitivity: Concerns have been raised about the project’s proximity to the Kambalakonda Wildlife Sanctuary and its Eco-Sensitive Zone.
- Water security: Petitioners have alleged potential impacts on the Mudasarlova reservoir catchment and associated drainage systems.
- Resource assessment: Questions have been raised regarding electricity demand, water requirements, waste generation and heat emissions.
In October 2026, the NGT identified gaps in the environmental appraisal framework and sought a response from the Ministry of Environment, Forest and Climate Change.
These proceedings raise questions about regulatory adequacy. The petitioners’ allegations should not be treated as established findings of environmental damage.
What are the Regulatory Challenges in India?
- Absence of a Dedicated National Data-Centre Regulatory Framework
- India has pursued data-centre development through infrastructure recognition, investment incentives and State-level policies.
- However, a comprehensive, binding national framework addressing the distinctive environmental impacts of hyperscale AI data centres remains necessary.
- Gaps in Environmental Impact Assessment
The EIA Notification, 2006, does not contain a dedicated category specifically designed for AI data centres.
Depending on their characteristics, projects may be assessed under existing categories such as building and construction projects.
This creates concerns that conventional appraisal methods may not fully account for:
- Continuous electricity demand
- Cooling-water consumption
- Waste heat
- Electronic and hazardous waste
- Cumulative environmental impacts
- Local grid and water-resource constraints
- Fragmented Institutional Responsibilities
Data-centre regulation involves multiple institutions, including:
- Ministry of Electronics and Information Technology
- Ministry of Environment, Forest and Climate Change
- Ministry of Power
- Central and State Pollution Control Boards
- State governments
- Local planning authorities
Without effective coordination, environmental safeguards and infrastructure planning may remain fragmented.
- Competition Among States
- States compete to attract investment by offering land, electricity incentives, tax concessions and simplified approvals.
- Although such competition can promote infrastructure development, it may also create pressure to prioritise investment commitments over environmental assessment and community consultation.
What Measures are Needed for Sustainable AI Data-Centre Development?
- Establish a National Green Data-Centre Framework
India should develop a coordinated national framework linking project approvals and incentives to environmental and infrastructure readiness.
The framework should establish standards for:
- Electricity demand and grid capacity
- Water sourcing and consumption
- Cooling-system efficiency
- Renewable-energy procurement
- Waste management
- Site selection
- Environmental disclosure
- Community consultation
A risk-based approach would allow regulatory requirements to reflect facility size, location and environmental sensitivity.
- Promote Renewable and Low-Carbon Electricity
Data-centre operators should be encouraged to procure additional renewable electricity and support investments in grid infrastructure.
Measures could include:
- Renewable-energy power purchase agreements
- Energy storage
- Demand flexibility where technically feasible
- Grid upgrades
- Energy-efficient computing equipment
- Low-carbon backup power technologies
Renewable-energy claims should distinguish annual certificate-based matching from actual hourly availability of low-carbon electricity.
- Improve Cooling and Water Efficiency
Water-efficient cooling technologies should be promoted according to local climatic and hydrological conditions.
Possible measures include:
- Closed-loop liquid cooling
- Air-cooled or hybrid heat rejection
- Use of treated wastewater
- Water recycling
- Leak detection and monitoring
- Heat recovery where practical
Water-intensive facilities should face stronger scrutiny in groundwater-stressed regions.
- Promote Frugal and Decentralised AI
India should encourage computing approaches that reduce unnecessary resource consumption.
These include:
- Smaller, task-specific AI models
- Model compression and quantisation
- Efficient inference
- Shared computing infrastructure
- Workload optimisation
- Decentralised and edge computing where appropriate
However, decentralisation is not automatically more energy-efficient. Its benefits depend on hardware utilisation, network requirements and the nature of the workload.
- Introduce Mandatory Resource Disclosure
Large data centres should disclose standardised environmental performance indicators.
Indicator |
Meaning |
Purpose |
Power Usage Effectiveness (PUE) |
Total facility energy divided by IT equipment energy |
Measures facility energy overhead |
Water Usage Effectiveness (WUE) |
Water consumed per unit of IT equipment energy |
Measures water efficiency |
Carbon Usage Effectiveness (CUE) |
Operational carbon emissions per unit of IT equipment energy |
Measures operational carbon intensity |
Renewable electricity share |
Proportion of electricity supplied or matched by renewable sources |
Assesses clean-energy procurement |
Waste generation |
Quantity and type of electronic and hazardous waste |
Supports waste-management accountability |
These indicators should be independently verified and interpreted alongside local water stress, electricity sources and workload characteristics.
- Strengthen Environmental Impact Assessment
Large data-centre projects should undergo proportionate environmental assessment that accounts for their full operational characteristics.
This should include:
- Cumulative electricity demand
- Water availability and competing users
- Impacts on nearby ecosystems
- Heat and noise emissions
- Backup power and fuel storage
- Electronic waste and battery disposal
- Public consultation where warranted
- Ensure Community Benefit-Sharing
Host communities should receive tangible benefits from data-centre development.
These may include:
- Fair land compensation
- Local skill development
- Employment and training opportunities
- Improved electricity infrastructure
- Water conservation and recharge projects
- Transparent grievance redressal
- Community participation in planning
Benefit-sharing mechanisms can improve public trust and reduce conflicts over resource allocation.
- Strengthen E-Waste and Hazardous Waste Management
- Data centres generate waste from servers, batteries, cooling systems and electronic equipment.
- Operators should comply with applicable e-waste, battery-waste and hazardous-waste regulations.
- Circular-economy practices should include equipment refurbishment, responsible recycling, recovery of critical materials and safe disposal of cooling fluids.
Conclusion
AI data centres are becoming essential infrastructure for India’s digital economy, technological innovation and strategic competitiveness.
However, their rapid expansion creates substantial challenges involving electricity demand, water consumption, environmental protection and community welfare.
India must therefore adopt a resource-efficient, environmentally responsible and socially inclusive data-centre development model.
A comprehensive national framework integrating green energy, sustainable cooling, transparent resource accounting, environmental assessment and community participation can help reconcile AI-led economic growth with ecological sustainability.
India’s emergence as a global AI hub should be measured not only by computing capacity and investment attracted, but also by the efficiency, sustainability and social value of the infrastructure supporting it.
Possible Mains Question
The expansion of AI data centres presents both opportunities for India’s digital economy and challenges for environmental sustainability. Examine the major concerns and suggest a regulatory framework for sustainable data-centre development. (250 words)





Leave a Reply