Clouds Over Dry Wells: India’s Data Centre Boom and the Cost of Going Digital

When the Andhra Pradesh government recently accelerated its push to become India’s next major data-centre hub, the move was presented as a symbol of technological progress and economic ambition. Just days ago, foundation work began for Google’s proposed $15 billion AI and data-centre project near Visakhapatnam, expected to become one of the largest facilities of its kind in Asia. At the same time, the state introduced a policy allowing strategic data centres to secure dedicated power-distribution licences, reflecting the intense competition among Indian states to attract digital infrastructure investment.

India is witnessing an unprecedented expansion of data centres. Capacity is expected to nearly double to around 2 GW by 2026 from less than 1 GW a few years ago. This growth is being driven by cloud computing, AI expansion, 5G rollout, fintech growth, e-commerce, and data localisation policies. India generates nearly 20% of global mobile data traffic but still accounts for only a small share of global data-centre capacity, creating strong pressure for rapid expansion.

For policymakers and industry leaders, data centres are now treated as essential infrastructure for the digital economy. Banking systems, AI platforms, cloud services, digital payments, OTT streaming, e-governance systems, and e-commerce all depend upon uninterrupted data storage and processing. Global technology companies are expanding aggressively in India, while domestic conglomerates are investing heavily in AI and digital infrastructure ecosystems.

The economic argument appears compelling. Data centres attract investment, strengthen telecom and fibre infrastructure, support cloud adoption, and align with India’s ambition of becoming a global digital power. Governments also increasingly view domestic data storage as a matter of sovereignty and strategic autonomy. In the emerging AI-driven global order, computing infrastructure is acquiring the same geopolitical importance that oil infrastructure once held during the industrial age.

Yet the celebration around the digital economy often ignores a fundamental reality: the cloud is not virtual. It is physical infrastructure built upon land, electricity, water, and minerals. Behind every seamless digital service lies an industrial system consuming enormous quantities of resources.

The most serious concern is water.

Data centres generate immense heat and require continuous cooling systems. In India’s climatic conditions, cooling requirements become especially intensive. Estimates suggest that data-centre water consumption in India could rise from nearly 150 billion litres in 2025 to over 350 billion litres by 2030.

The geography of expansion makes the problem more alarming. Major hubs such as Bengaluru, Chennai, and Hyderabad are already struggling with groundwater depletion, shrinking reservoirs, and periodic water crises.

Southern India, in particular, presents a deeper contradiction. States such as Karnataka, Tamil Nadu, Andhra Pradesh, Telangana, and Kerala remain heavily dependent on fragile and politically contested river systems such as the Cauvery and Krishna basins. These disputes have repeatedly triggered litigation, protests, political stand-offs, and social tensions over drinking and irrigation water.

In such a situation, the rapid expansion of highly water-intensive data centres raises uncomfortable questions. States already fighting over river water are simultaneously competing to host server farms that may consume millions of litres daily for cooling systems. A single large hyperscale facility can use water equivalent to thousands of households every day. In a country possessing barely 4% of global freshwater resources despite supporting nearly 18% of the world’s population, the long-term sustainability of such expansion cannot be ignored.

Electricity demand presents another major challenge.

Large data centres operate continuously and require uninterrupted power backed by extensive redundancy systems and diesel generators. Andhra Pradesh’s decision to permit dedicated power-distribution licences for large facilities shows how energy policy itself is beginning to adapt around the needs of this industry.

The difficulty is that India’s electricity system remains heavily dependent on coal. As data-centre expansion accelerates, so does the risk of rising carbon emissions at a time when India is attempting to meet ambitious climate targets. Some projections suggest data centres could account for more than 2% of national electricity demand by 2030, with AI workloads likely to push demand even higher.

The debate is no longer confined to environmental groups. Similar concerns are emerging globally. In the United States, citizens in states such as Texas, Virginia, Ohio, and Wisconsin have protested against large-scale data-centre projects due to fears over excessive water consumption, rising electricity demand, noise pollution, and pressure on local infrastructure. Several projects have faced lawsuits, delays, and organised public opposition. Environmental groups there have demanded stronger transparency regarding the ecological impact of AI-driven infrastructure expansion.

India may eventually face similar tensions if expansion continues without adequate safeguards or public consultation.

The employment argument surrounding data centres also deserves a more realistic assessment. Construction activity undoubtedly generates temporary jobs and related economic activity. However, operational data centres are highly automated facilities that generate relatively limited long-term employment, largely confined to specialised technical roles. Meanwhile, local communities often bear the environmental costs through groundwater extraction, land-use changes, pressure on electricity systems, and urban heat effects.

This creates an uncomfortable imbalance. The benefits of the digital economy are concentrated among corporations, investors, and urban consumers, while ecological costs are transferred to already resource-stressed regions and communities.

The policy framework governing data-centre expansion in India remains insufficiently developed for the scale of proposed growth. States are competing aggressively through tax incentives, subsidised land, electricity concessions, and expedited approvals, but sustainability regulations remain fragmented and weakly enforced.

India still lacks comprehensive mandatory standards regarding water-use efficiency, renewable-energy sourcing, wastewater recycling, waste-heat management, and public disclosure of actual resource consumption. Environmental clearances for several large projects have already attracted criticism from activists and local groups alleging inadequate scrutiny and consultation.

India cannot realistically avoid building digital infrastructure. Data centres are now essential to economic activity, governance systems, AI development, and technological competitiveness. But the debate is not about whether India should build them. The real question is whether the country can expand digital infrastructure without deepening existing ecological and social stresses.

That requires a shift from viewing data centres merely as investment destinations toward treating them as major industrial infrastructure with significant environmental implications. Renewable-energy integration, strict water-efficiency norms, wastewater recycling, transparent reporting standards, decentralised infrastructure planning, and careful site selection in less water-stressed regions must become central to policy rather than optional commitments.

The digital economy cannot be sustained on ecological depletion. If India pursues technological expansion while ignoring groundwater stress, energy vulnerability, and climate realities, the country risks repeating an old developmental mistake in a new technological form.

Data centres may power the digital future, but they will ultimately depend on rivers, electricity grids, aquifers, and ecosystems that are already under severe strain. Ignoring that contradiction today may produce a far more expensive crisis tomorrow.