Big Tech's Data Centre Expansion in India: Opportunity, infrastructure and environmental trade-offs

India is fast emerging as one of the world’s most attractive destinations for hyperscale data centres, driven by a potent mix of cloud adoption, AI workloads, data localisation rules and the country’s exploding digital consumption. Global technology giants including Google, Amazon Web Services (AWS), Microsoft and Meta are committing billions of dollars to large-scale campuses, while domestic players such as Adani Group, L&T and CtrlS are racing to build gigawatt-scale capacity.

But as India positions itself as the next global data centre hub, industry experts warn that the pace and scale of this expansion are putting unprecedented pressure on three already-strained resources: electricity, water and carbon headroom.

Connectivity, climate, cost: Why South India has emerged as data centre magnet for big tech

India’s installed data centre IT load stands at about 1.4–1.5 GW in 2025, with another 1.4 GW under construction, according to industry estimates. In an optimistic scenario, capacity could surge to 9–10 GW by 2030, making India the second-largest data centre market in Asia-Pacific, overtaking Japan and Australia in electricity demand from data centres.

“This is not incremental growth. It is an order-of-magnitude expansion,” said S. Anjani Kumar, Partner at Deloitte India. “If sustainability is not embedded at the design and approval stage, environmental risks will scale faster than mitigation efforts.”

Meghna Bal, Director at Esya Centre, notes, India’s scale itself makes localisation inevitable. “India’s digital public infrastructure, streaming, fintech, e-commerce and now AI workloads are driving in-country compute needs. India is simply too large to serve only from Singapore or the Middle East,” she says.

The hidden cost of hyperscale: power, water and AI

The most immediate challenge is power.

Indian data centres currently consume around 13 terawatt-hours (TWh) of electricity annually, roughly 0.8% of national demand. By 2030, that figure is projected to rise to 57 TWh, or 2.6% of India’s total electricity consumption, according to estimates cited by industry experts.

This surge is being accelerated by AI.

Over the past two to three years, rack power densities have jumped sharply as GPU-driven workloads proliferate. AI-heavy hyperscale facilities require continuous, high-quality power, placing additional strain on already congested urban grids in Mumbai, Chennai, Hyderabad and the National Capital Region.

The carbon implications are equally stark.

More than 70% of India’s electricity still comes from coal, and despite growing renewable procurement by tech firms, grid power remains emissions-intensive. Data centres also rely heavily on diesel generators for backup power, an often under-reported source of local air pollution and carbon emissions.

Strong power infrastructure makes India preferred destination for data centres, says Piyush Goyal

Water, however, may prove to be the most politically sensitive constraint. A typical mid-sized data centre can consume around 416 million litres of water annually, equivalent to the needs of nearly 1,000 households. Large AI-driven giga-campuses are even more resource-intensive, drawing up to 19 million litres per day, according to the US-based Environmental and Energy Study Institute (EESI).

“With AI workloads, the pressure intensifies dramatically,” Kumar said. “Each GPU can require roughly 30,000 litres of water per year for evaporative cooling. Multiply that by thousands of GPUs, and the scale becomes enormous.”

Most hyperscale facilities are clustered in water-stressed metros such as Bengaluru, Chennai and Mumbai, where summer shortages are already routine.

Mumbai ranks second-cheapest city worldwide for data centre construction

Industry watchers warn that concentration risk—multiple giga-campuses tapping the same water catchment, could trigger operational disruptions, regulatory intervention or public backlash.

Big Tech’s green pledges: progress, but not a silver bullet

Global cloud majors have responded with ambitious sustainability commitments.

Google is pursuing 24/7 carbon-free energy, Microsoft has pledged to be carbon negative and water positive by 2030, and Amazon, the world’s largest corporate buyer of renewable energy, has committed to net-zero emissions by 2040. Meta says it has matched 100% of its electricity use with renewables since 2020.

In India, data centres already account for roughly 15% of corporate renewable power procurement, with long-term power purchase agreements (PPAs) signed for solar and wind capacity. Google’s renewable partnerships and Microsoft’s expanding cloud regions illustrate how central India has become to Big Tech’s AI infrastructure plans.

Yet experts caution that voluntary, global pledges often fail to capture local impacts.

“Net-zero targets are important, but they don’t automatically solve water stress in Bengaluru or grid congestion in Mumbai,” said Rishi Agrawal, CEO and co-founder of TeamLease RegTech.

“The environmental footprint of a data centre is highly location-specific. That’s where policy and regulation have to step in.”

Agrawal pointed out that without structural planning, the sector risks exacerbating India’s emissions trajectory, complicating both the country’s net-zero by 2070 commitment and its target of 500 GW of non-fossil capacity by 2030.

The case for tougher guardrails and transparency

A growing consensus is emerging that self-regulation alone will not be sufficient as India’s data centre footprint multiplies. Experts are calling for mandatory sustainability disclosures, including public reporting of metrics such as Power Usage Effectiveness (PUE) and Water Usage Effectiveness (WUE) at the site level.

“Future regulations should require water and grid impact assessments before approving any project,” Kumar said. “Projects that integrate renewables, water-efficient cooling and sustainable design will have a clear advantage—not just environmentally, but in financing, approvals and long-term operating costs.”

Policy reform is also being discussed around urban planning and land use. Treating data centres like conventional office buildings under existing building codes limits vertical densification and pushes sprawl.

AI data centres to drive global water usage 11-fold by 2028: Morgan Stanley

Industry experts argue that the National Building Code 2025 should classify data centres separately, allowing higher floor-space indices and excluding utility-heavy infrastructure such as cooling plants and substations from FSI calculations.

Decentralisation is another key lever. India is expected to need 45–50 million sq ft of additional data centre space over the next few years. Pushing capacity into Tier-2 and Tier-3 cities, supported by edge and micro data centres, could ease pressure on metro utilities while improving network resilience.

Some experts have proposed Data Centre Economic Zones (DCEZs) in non-metro regions, offering single-window clearances, higher FSI and fiscal incentives covering 20–25% of land, power and capex costs.

Can India turn sustainability into an advantage?

Despite the risks, several industry leaders believe India has a rare opportunity to define a new global template for responsible hyperscale infrastructure.

“India starts with a clean slate,” said Santosh Singh, Chief Strategy and AI Officer, arguing that renewables-first planning, efficient cooling and smarter siting could turn sustainability into a competitive advantage rather than a constraint. “Unchecked scale invites backlash. Responsible scale builds legacies.”

Technological innovation could also play a role. Vineet Mittal, Senior Vice President at Ziroh Labs, pointed to advances in energy-efficient AI runtimes, CPU-based AI processing and waterless cooling designs that could sharply reduce resource intensity. India’s long coastline also opens the door to deep seawater cooling, which can cut cooling energy demand by up to 79% for coastal campuses.

Amazon to invest over $35 billion in India’s cloud, AI infrastructure by 2030

Policy analysts add that sustainability must extend beyond operational emissions to include Scope 3 impacts, such as embodied carbon in concrete, steel and hardware supply chains, as well as diesel generator usage during outages—factors often glossed over in headline net-zero claims.

A narrow window to get it right

The stakes are high. Data centres are rapidly becoming core national infrastructure, underpinning everything from digital public services and fintech to streaming and generative AI. But without stronger governance, transparency and planning, the very infrastructure powering India’s digital future could strain trust, resources and social licence.

“The future belongs to those who scale with conscience,” Singh said. “India can lead the world in sustainable hyperscale—or learn the hard way.”

For policymakers, Big Tech and operators alike, the message is clear: how India builds its data centres may matter as much as how many it builds.