India's Data-Centre Capacity Quadruples to 1,575 MW on AI Demand
- India's data-centre capacity has risen more than four-fold, from ~375 MW (2020) to ~1,575 MW, on AI/HPC demand (Lok Sabha written reply, 5 August 2026).
- MoS MeitY Jitin Prasada attributed the growth to AI and the IndiaAI Mission's sovereign-compute push, with measures flagged for environmentally sustainable expansion.
- Principal AI corridors face acute water stress - Bengaluru and Hyderabad are CGWB over-exploited; Chennai has seen Day-Zero conditions.
- The test is whether new capacity concentrates in water-abundant or renewable-rich corridors, or amplifies stress in the metro clusters.
- India's data-centre capacity has risen more than four-fold, from ~375 MW (2020) to ~1,575 MW, on AI/HPC demand (Lok Sabha written reply, 5 August 2026).
- MoS MeitY Jitin Prasada attributed the growth to AI and the IndiaAI Mission's sovereign-compute push, with measures flagged for environmentally sustainable expansion.
- Principal AI corridors face acute water stress - Bengaluru and Hyderabad are CGWB over-exploited; Chennai has seen Day-Zero conditions.
- The test is whether new capacity concentrates in water-abundant or renewable-rich corridors, or amplifies stress in the metro clusters.
- ANI (05 Aug 2026): https://www.aninews.in/news/business/indias-data-centre-capacity-rises-over-four-fold-to-1575-mw-as-ai-demand-accelerates-govt20260805171543/
- The Tribune (05 Aug 2026): https://www.tribuneindia.com/news/business/indias-data-centre-capacity-rises-over-four-fold-to-1575-mw-as-ai-demand-accelerates-govt/
India's operational data-centre capacity has risen more than four-fold, from about 375 MW in 2020 to roughly 1,575 MW now, driven by artificial-intelligence and high-performance-computing demand. The figure was given by Minister of State for Electronics & IT Jitin Prasada in a written reply to the Lok Sabha on 5 August 2026, which also noted that the IndiaAI Mission is building a sovereign AI stack - compute, foundation models and indigenous applications - while the government puts in place measures for the environmentally sustainable expansion of AI infrastructure and data centres.
The strategic significance is twofold. First, AI workloads require dense compute, and India's earlier capacity constraints limited the scale of domestic model training and inference; a four-fold increase lets hyperscalers and research institutions keep more data and computation within Indian jurisdiction, reinforcing the sovereign-AI narrative. Second, the expansion concentrates additional electricity and water demand - a modern hyperscale data centre draws both continuously - in the same handful of corridors where India's digital infrastructure already clusters.
The dependency frame connects directly to the power-and-water constraint. Bengaluru and Hyderabad are both classified by the Central Ground Water Board as over-exploited, and Chennai has experienced Day-Zero conditions. If the build-out continues toward the multi-gigawatt scale projected for 2030 while remaining tied to conventional grid supply in these stressed corridors, AI infrastructure will increasingly compete with residential and agricultural users for the same water and power - the binding constraint is local, not national.
The test to watch is where the next tranche of capacity lands: in water-abundant or renewable-rich corridors such as Gujarat, Rajasthan or the north-eastern states, or in the default clusters around Bengaluru, Hyderabad, Mumbai and Chennai. The government's choice of which corridors receive AI-compute incentives will determine whether this surge alleviates or amplifies India's infrastructure stress.
Track the systems we watch
Signals, reports and briefings on India’s industrial transformation.