PIB TrackerIndia’s data centres draw 1.5 GW, up from 375 MW in 2020
NITI Aayog’s new framework asks States to appraise power, water, siting, economic value and governance before they compete for the next gigawatt of capacity.
What happened
- NITI Aayog’s Frontier Tech Hub released a Data Centre Evaluation Framework on 14 October 2026, built with MeitY, the Ministry of Power and DPIIT.
- India’s data-centre capacity is put at about 375 MW in 2020 and approximately 1.5 GW now.
- Five pillars: power and energy; water and cooling; siting and environmental sustainability; economic and strategic value; policy, governance and implementation.
- Recommendations include grid readiness and net grid draw, storage-backed renewables, basin-level water stress assessment, rainwater harvesting and lifecycle plans for e-waste and batteries.
- On value: identify data-centre zones, separate colocation from active compute, tie State incentives to measurable outcomes, and keep time-bound approvals with monitoring - explicitly not a new regulatory layer.
For Prelims
- What a data centre consumes: electricity for compute and for cooling, and water for evaporative cooling. The two are substitutes - air cooling raises power demand, water cooling lowers it, which is why a framework treats power and water as one problem.
- Net grid draw: the load a facility actually places on the grid after its own generation and storage. A plant with captive solar and batteries can have large connected capacity and a smaller net draw, which is the number a State utility needs.
- Colocation versus active compute: a colocation provider rents space, power and cooling to clients who own the servers; an active compute or cloud operator runs the machines. Employment, value added and tax treatment differ sharply between them.
- PUE: Power Usage Effectiveness, total facility energy divided by IT equipment energy - the standard efficiency metric, where 1.0 is the unreachable ideal and modern hyperscale facilities claim 1.1-1.3.
- Basin-level water stress: assessing demand against the river basin’s availability rather than the site’s own borewell. It is the same logic that underlies the Central Ground Water Authority categorisation of over-exploited blocks.
- E-waste and batteries: handled under the E-Waste (Management) Rules, 2022 and the Battery Waste Management Rules, 2022, both built on extended producer responsibility - a data centre refreshes servers every few years, so this is a recurring stream, not a decommissioning event.
- Why States compete: data centres are capital-intensive and land-light, and several State policies offer stamp-duty waivers, power tariff subsidies and infrastructure status. The framework’s answer is readiness, not richer subsidy.
- Frontier Tech Hub: NITI Aayog’s unit on emerging technologies, which also works on AI, quantum and advanced computing - the institutional home of this framework, and an advisory body, not a regulator.
For UPSC: Data centres are the physical layer of the AI economy, and this gives you the resource arithmetic to argue about them rather than the slogans. Deploy it on digital infrastructure, on the energy-water nexus, on Centre-State industrial policy and the subsidy race between States, and on e-waste and extended producer responsibility.
What it is NOT: No figure for how much electricity or water the existing 1.5 GW actually consumes, no projection of capacity to 2030, and no threshold values - what counts as adequate grid readiness, or as unacceptable basin stress, is left to States. The release does not name the provision granting the tax holiday to 2047, nor who qualifies as an eligible foreign cloud service provider. Nothing on where the existing capacity is concentrated, which is the relevant fact for a framework about siting, and nothing on data localisation or on the land acquisition that siting decisions turn on.
For Mains
Syllabus: GS3.13 · GS3.9 · Linkage L1
Anchor
NITI Aayog has published a Data Centre Evaluation Framework for State governments, structured on power, water, siting, economic value and governance, as India’s data-centre capacity reaches about 1.5 GW from about 375 MW in 2020.
Substantiation (data)
Capacity has roughly quadrupled in six years. The framework asks States to assess grid readiness and net grid draw, promote storage-backed renewables, evaluate basin-level water stress and the impact on public supply, require rainwater harvesting and on-site treatment, and demand lifecycle plans for e-waste and batteries. It also asks them to identify data-centre zones, separate colocation from active compute, and link incentives to measurable outcomes under time-bound approvals.
Position
The useful move here is treating a data centre as an infrastructure load rather than as an investment announcement. A gigawatt of compute is a power-sector planning problem, a basin-level water problem and a waste-stream problem before it is an IT success, and States that sanction capacity faster than their distribution companies and water utilities can serve it will discover the constraint after the subsidy has been given.
Counterpoint
A framework from an advisory body binds nobody. Land, power tariffs, water allocation and industrial incentives are State subjects or State decisions, and the competitive pressure that produces the subsidy race is exactly what a voluntary checklist cannot restrain. NITI explicitly rules out creating a regulatory layer, which means the framework works only where a State wanted to be disciplined anyway.
Way forward
The test is whether any State adopts the evaluation as a condition of its incentive approvals, and whether the power and water numbers for sanctioned capacity begin to be published. Reporting net grid draw and basin stress for each approved project would make the framework checkable by anyone.
Conclusion
A sober document about an exuberant sector: the capacity number is the story, and the framework’s contribution is to ask what the capacity will drink and burn before asking what it will earn.
Deploys into: Digital infrastructure and the AI build-out · Energy-water nexus · Centre-State industrial incentives · E-waste and extended producer responsibility
NITI Aayog · 2026-10-15 · PRID 2323840 · PIB source ↗