Whose Water Is AI Drinking in India?

Why It Matters

The water demands of AI infrastructure threaten to deepen existing social and environmental inequities in India, jeopardizing basic access for millions. Without regulatory safeguards, the AI race could exacerbate caste‑based and class‑based water deprivation.

Key Takeaways

• $167.5 bn AI data‑centre investments target water‑stressed Indian cities
• Data centres used ~150 bn litres water in 2025; 358 bn by 2030
• Bengaluru, Hyderabad, Chennai host facilities despite severe local water crises
• Dalits, urban poor, and women face heightened water insecurity from AI expansion
• Corporate pledges lack enforceable water‑equity standards or mandatory disclosures

Pulse Analysis

The race to position India as a global AI hub has attracted unprecedented capital. Microsoft, Amazon, Google and the Adani Group together announced a first‑wave investment of roughly $167.5 billion for data‑centre construction, a sector whose electricity load has already quadrupled from 0.4 GW in 2020 to an expected 1.5 GW by 2025. Cooling these servers requires water at scale: a 100‑megawatt plant uses about two million litres per day, translating to an estimated 150 bn litres consumed nationwide in 2025 and a projected 358 bn litres by 2030.

Beyond the technical footprint, the water draw intensifies long‑standing hierarchies of access. In cities like Bengaluru and Hyderabad, private tanker networks already price water out of reach for peripheral slums, while Dalit and Adivasi communities in rural districts face systemic denial of wells and taps. Women, who traditionally shoulder water collection, confront longer treks and higher health risks when supplies shrink. As AI facilities sprout in these water‑stressed zones, the most vulnerable bear the hidden cost of a digital economy that promises growth but delivers scarcity.

Corporate mitigation promises—recycled wastewater, air‑cooling, water‑positive pledges—remain voluntary and lack enforcement. To prevent a resource‑injustice cascade, India needs a binding framework that mandates public disclosure of water use, sets enforceable consumption caps tied to regional hydrology, and integrates water‑equity impact assessments into every data‑centre permit. Such policy would align the nation’s AI ambitions with the United Nations’ water security goals and protect the rights of marginalized groups. Only by embedding equity into the infrastructure blueprint can India reap AI’s benefits without draining the lifeblood of its most disadvantaged citizens.