While infrastructure, manufacturing and agriculture dominated the Union Budget narrative, one of its most consequential bets lay elsewhere. The government has offered a tax holiday until 2047 for foreign firms providing global cloud services from India-based data centres. The objective is explicit. India wants to position itself as a global compute hub rather than remain merely a large consumer of digital services.
The incentive is not marginal. Alongside the tax holiday, cloud infrastructure has been reclassified as core infrastructure, and compute capacity is now treated as capital. This places digital infrastructure on the same policy plane as highways, ports and power plants. The shift matters. It signals long-term fiscal certainty to hyperscalers deciding where to locate the next generation of data centres and AI workloads.
READ | AI data centres: The new backbone of national competitiveness
Industry response has been predictably positive. Global cloud providers point to India’s engineering talent, fast-growing domestic data demand and geopolitical positioning as reasons to scale capacity locally. Estimates suggest the policy could attract up to $200 billion in long-term investment over time. Whether that materialises depends less on tax treatment than on physical feasibility.
Capacity growth and its geography
India’s operational data centre capacity has already crossed roughly 1.5 gigawatts. Projections for the end of the decade range from 8 to 12 GW, depending on the pace of AI adoption and hyperscale expansion. Annual growth exceeds 20%, making data centres one of the fastest-growing infrastructure segments in the economy.
But this growth is highly concentrated. Mumbai–Navi Mumbai, Chennai, Hyderabad, Bengaluru and the Delhi region account for nearly all large-scale capacity. The logic is familiar: proximity to enterprise demand, financial markets, skilled labour, fibre backbones and subsea cable landing stations. The vulnerability is equally clear. These cities are also among India’s most water-stressed urban regions.
Concentration has also been reinforced by land economics. Hyperscale facilities require large, contiguous parcels close to substations and fibre routes. In major metros, land prices, fragmented ownership and restrictive zoning norms have already stretched project timelines. Land conversion approvals, floor-area regulations and right-of-way permissions sit largely with urban local bodies and state development authorities. These frictions rarely feature in national policy announcements, but they increasingly shape execution.
Water as a binding constraint
Data centres require continuous cooling to prevent outages. Most still rely on water-intensive cooling systems. India’s data centre sector already consumes an estimated 150 billion litres of water annually. If current expansion plans proceed without major efficiency gains, that figure could more than double by 2030.
The macro context is unforgiving. India uses over 80% of its available freshwater resources each year and ranks among the most water-stressed countries globally. Rising compute demand and shrinking water availability are now on a collision course. In key urban clusters, water may become the binding constraint on future capacity additions, regardless of fiscal incentives.
Some operators have begun investing in treated wastewater reuse, rainwater harvesting and closed-loop cooling. These efforts remain uneven. Disclosure standards are limited, and sustainability practices vary sharply across facilities. Without clearer norms, water stress risks translating into regulatory friction rather than technological adaptation.
READ | AI data centres risk derailing global climate action
The electricity challenge
Electricity poses a parallel constraint. Data centres currently account for less than 1% of India’s power consumption. That share could rise to 2.5–3% by 2030 as hyperscale facilities and AI workloads expand. Unlike many industrial users, data centres require uninterrupted, high-quality power.
Meeting this demand implies several gigawatts of new continuous supply, along with grid upgrades, transmission investment and firming capacity for renewables. Many states already struggle with peak summer shortages. Compute-intensive infrastructure adds a new layer of planning stress to an already tight system. The challenge is not only generation, but predictability.
Network dependence and concentration risk
The case for geographic diversification is often framed around water and power. Network infrastructure is an equally important constraint. India’s international bandwidth depends heavily on a small number of coastal subsea cable landing points. Inland data centre clusters rely on long-haul fibre redundancy that remains uneven across states.
This dependence reinforces concentration near coastal metros and limits the practical dispersal of hyperscale capacity. Without parallel investment in fibre resilience and inland connectivity, diversification beyond established corridors will remain policy intent rather than commercial reality.
READ | Global capability centres driving India’s knowledge economy
Localisation still matters
The policy push is not misplaced. India is the world’s fastest-growing data consumption market. Low mobile tariffs, expanding broadband, video streaming, digital payments and AI workloads are driving sustained growth in per-user data consumption. India generates roughly a fifth of global data but hosts only a small fraction of global data centre capacity.
Local compute reduces latency, improves resilience and anchors higher-value digital activity domestically. While data centres are not labour-intensive, their multiplier effects are substantial. They underpin AI development, fintech, e-commerce and enterprise digitalisation. The Budget’s attempt to shift India from a digital market to a digital production base is strategically sound.
Incentives versus constraints
Yet infrastructure policy fails when physical and regulatory constraints are ignored. Beyond land, water and power, regulatory predictability also shapes hyperscaler decisions. India’s data governance framework continues to evolve, particularly around cross-border data flows, compliance obligations and sectoral exemptions. Long-horizon investments assess rule stability alongside fiscal certainty.
If water availability tightens further in core clusters, and regulatory clarity lags, new projects will face delays, rising operating costs and potential pushback from local authorities. Incentives can accelerate interest. They cannot override urban capacity limits or institutional uncertainty.
Where the next phase must shift
The more relevant question is not whether India should build data centres, but where and how. Geographic diversification will be unavoidable. Encouraging new clusters in water-surplus regions, tier-2 cities or coastal zones with desalination access can ease pressure on stressed metros, but only if land aggregation, fibre connectivity and power planning move in tandem.
Policy must therefore move beyond incentives to mandates. Minimum recycling thresholds, compulsory use of non-potable water, renewable energy procurement norms and clearer disclosure standards should become baseline requirements. Accelerating the adoption of liquid immersion cooling, air-cooled systems and zero-water designs—already deployed globally—will determine whether India’s compute ambition scales without colliding with ecological and urban limits.
The Budget has made its bet. Execution will depend on whether digital infrastructure policy is aligned not just with fiscal ambition, but with physical and institutional reality.