In the just concluded AI Summit at New Delhi, the government of India has majorly played up the issue of this country being “prepared” for Artificial Intelligence. The claim was reinforced weeks earlier in the Union Budget for the financial year 2026-27, which rolled out fiscal incentives and policy support aimed at making India a global hub for AI infrastructure, particularly large-scale data centres. But what does being “prepared” truly mean?
If preparedness means indigenous breakthrough in cutting edge AI research, globally competitive foundational models, semiconductor dominance, or deep public investment in AI laboratories, the claim appears rather ambitious at best. For, India does not yet possess the scale of capital, advanced chip ecosystems, or concentrated research clusters seen in the United States or China to even be considered a major player in AI. Nor has it demonstrated a sustained public funding that buttresses ground-breaking technological leaps. Instead, India appears to be settling for a modest ambition of hosting the physical infrastructure of AI by building data centres.
Also Read
No doubt, there is an economic logic behind this strategy. For, data centres bring investment, construction activity and ancillary services. Long term tax holidays, lasting about 21 years, and land allocations, as outlined in the Budget, are designed to attract precisely this capital. Yet building data centres is not the same as dominating the AI ecosystem.
Also Read: Enforced Reverence
The global experience, when it comes to investing in data centres, offers a sobering reality. Several technologically advanced nations have slowed or restricted data centre expansion for the simple reason that these facilities take a heavy toll on power grids, land, and water resources.
The Netherlands imposed a nine-month moratorium in 2022 on hyperscale data centres larger than 10 hectares or consuming over 70 MW, and local authorities continue to restrict expansion due to grid congestion and land-use conflicts. Likewise, Ireland maintained a de facto halt on new grid connections in the Dublin region for years. Under its 2026 framework, new facilities must now provide on-site generation or storage and source at least 80 per cent of their energy from additional renewables. Singapore, after a moratorium between 2019 and 2022, shifted to a tightly controlled “managed growth” model, permitting only best-in-class projects with strict energy efficiency benchmarks. Even China has moved to assert control, directing state-funded data centres to avoid foreign AI chips and rely on domestic hardware.
These technologically advanced economies have confronted one basic truth: data centres are enormous consumers of electricity and water, and their growth can destabilise local infrastructure if left unfettered.
India, a country that faces chronic power distribution bottlenecks, uneven renewable integration, and water stress across major urban clusters, ideally should not be investing heavily in data centres. To invite an unrestrained expansion of energy-guzzling data centres without corresponding upgrades in grid resilience, storage capacity, and water management could compound existing vulnerabilities.
Building large AI data centres can also impose serious ecological costs on India. These facilities consume colossal amounts of electricity, often sourced from coal-based power, increasing carbon emissions. Land acquisition for hyperscale campuses can disrupt local ecosystems and agriculture. In the absence of stringent renewable energy mandates, water-recycling norms, and efficiency standards, expanding data centres risk deepening India’s energy insecurity, urban water crisis, and environmental degradation.
If India aspires to be central to the global AI ecosystem, it must go for sustainable power planning, upskill its workforce, and invest heavily in research. Otherwise, it may end up as a host to AI servers and bear heavy environmental cost, while the real intellectual and economic value of AI is created elsewhere. This leaves policymakers with two choices — build merely the infrastructure of AI, or build the capability to shape its future. Being “prepared” means doing the latter.
