India’s AI data-centre boom runs into a power and water reality check

India’s rapidly expanding data-centre industry could face its biggest bottleneck from access to electricity, water and sustainable infrastructure, according to a new report that positions data centres as the next strategic layer of national infrastructure.

A new report by the Centre for Accelerating India’s Growth (CAIG), titled Beyond Infrastructure: India’s Data Centre Pathway to Digital Sovereignty, argues that India’s data-centre sector is entering a phase where electricity access, cooling infrastructure and sustainability planning could become bigger constraints than capital availability itself.

The report comes amid a sharp acceleration in AI-led infrastructure investments globally and in India. According to the study, India has attracted nearly $64 billion in data-centre investment commitments since 2019, with close to $30 billion announced during the January-September 2025 period alone.

India currently has about 1.2-1.4 GW of installed data-centre capacity, but the report estimates this could rise to between 4.1 GW and 8.3 GW by 2030 depending on AI adoption and hyperscale expansion. Industry estimates from recent KPMG India research similarly suggest India’s data-centre capacity could approach 9 GW by the end of the decade as AI workloads, sovereign cloud demand and enterprise digitisation accelerate.

The larger concern is that AI-era data centres consume far more resources than conventional enterprise server facilities.

Unlike traditional cloud infrastructure, generative AI systems rely on dense GPU clusters and advanced cooling systems that require significantly higher electricity and water usage. Recent academic research on AI workload energy modelling has pointed to “unprecedented computational demands” from large AI infrastructure deployments, with major implications for energy planning and sustainability.

India’s Economic Survey 2025-26 had also flagged AI data centres as resource-intensive infrastructure, warning that rising compute demand could significantly increase electricity and water consumption.

The CAIG report estimates that annual operational water consumption from Indian data centres could rise sharply by 2030 under high-density cooling scenarios, potentially crossing 1,900 million litres annually. The report warns that unmanaged expansion could intensify urban water stress and carbon emissions as AI adoption scales.

Energy demand is emerging as an equally important concern, said the report. Recent energy-sector studies examining hyperscale AI infrastructure have estimated that supporting future data-centre electricity loads could require substantial renewable-energy additions, transmission upgrades and energy-storage investments over the next decade. Globally, technology companies are increasingly tying AI data-centre expansion to dedicated renewable-energy procurement and power-generation partnerships as electricity demand from compute infrastructure surges.

The CAIG report argues that India’s data-centre sector is evolving from a real-estate-led business into “strategic national infrastructure” tied to digital sovereignty, AI competitiveness and long-term economic resilience.

That shift is already visible in the scale of recent infrastructure announcements.

Companies such as Yotta Data Services, Nxtra by Airtel, CtrlS Datacenters and AdaniConneX have expanded hyperscale and AI-ready campus plans amid rising demand for sovereign cloud, AI compute and GPU infrastructure.

The report also highlights large-scale investments by global hyperscalers. Amazon Web Services has reiterated plans to invest $12.7 billion in India’s cloud infrastructure by 2030, while Google has announced a gigawatt-scale AI data-centre campus in Visakhapatnam linked to renewable-energy infrastructure.

India’s data-centre footprint remains concentrated in Mumbai, Chennai, Delhi-NCR, Bengaluru, Hyderabad, Pune and Kolkata because of fibre connectivity and enterprise demand. But the report suggests future growth may increasingly move toward Tier-2 cities and dedicated digital infrastructure corridors where land, power and cooling resources can be planned more systematically.

The report recommends dedicated digital infrastructure zones, wastewater reuse mandates, renewable-energy-linked incentives and stricter sustainability standards for future approvals.