India’s Data Centre Investment Potential
Ashwini Vaishnaw, India’s Minister of Information Technology, has articulated that the nation stands poised to attract investments exceeding $200 billion in data centres over the coming years.
During a press conference at the 3DExperience World event in Houston, Jensen Huang, CEO of Nvidia, underscored the potential of these infrastructures to create substantial upstream and downstream employment opportunities in India, similar to the transformative impact of the internet.
Yet, the success of this endeavor hinges critically on effective execution, which is impeded by a myriad of challenges, encompassing land availability, energy supply, water resources, and skilled personnel.
Capacity Expansion
India’s data-centre capacity has experienced remarkable growth, jumping from 350 megawatts (MW) in 2019 to an anticipated 1.3 gigawatts (GW), or 1,000 MW, by 2025, as estimated by India Ratings and Research.
This capacity is projected to further revolutionize the sector, reaching 1.7 GW in 2026 and exceeding 2 GW by 2027. The introduction of a tax holiday may catalyze this upward trajectory.
The market is currently dominated by three primary categories of players: global hyperscalers such as Amazon Web Services, Microsoft, Google, and Meta, which are increasingly colocating AI-oriented facilities; Indian operators and joint ventures encompassing Yotta, Nxtra, AdaniConneX, Sify, and CtrlS, which cater to enterprise and cloud clientele; and telecom-linked conglomerates like Reliance Industries, integrating cloud, connectivity, and data centre capabilities for both domestic and regional markets.
All three segments are witnessing growth stimulated by surging digital demand. The announcement aims to further invigorate these entities while safeguarding domestic ownership and taxation.
For instance, foreign companies seeking tax exemptions are mandated to utilize data centres owned and operated by Indian entities. Likewise, services offered to Indian users must be routed through an Indian reseller entity.
The AI Imperative
Artificial intelligence acts as the primary catalyst for data-centre proliferation in India, paralleling global trends. By 2025, AI workloads are anticipated to constitute 23% of data-centre capacity, a substantial increase from 5% in 2020, with projections suggesting a rise to 50% by 2030, according to JLL, a real estate services firm.
Despite India generating nearly 20% of global data, a substantial portion has historically been stored or processed abroad. This trend is now shifting, motivated by the pressing demand for local storage solutions and low-latency access for AI systems.
Major cloud providers, including Amazon Web Services (AWS), Microsoft Azure, and Google Cloud, are increasingly relocating AI computational operations to India.
India’s overarching ambition lies in constructing the requisite infrastructure to facilitate domestic AI development, thereby reducing reliance on external resources for computational power.
Policy initiatives, such as long-term tax stability, are designed to anchor early-stage AI training, model deployment, and enterprise applications within the country.
This aligns with India’s expressed goal of cultivating autonomous AI capabilities and enhancing its role in the global AI ecosystem.
Should the tax holiday prove effective in luring investments, it could further expand local capacity. This expansion may lead to reduced costs and improved efficiency, thereby lessening dependence on overseas facilities, a scenario that may greatly benefit Indian startups.
Shifting Centres of Operations
Data-centre capacity in India is heavily concentrated in specific hubs, with Mumbai dominating at 53% of the market share, followed by Chennai at approximately 20%. Other regions, including Delhi, Bengaluru, Pune, Hyderabad, and Kolkata, account for the remainder, as reported by CBRE.
These clusters have emerged due to reliable energy supply, robust fibre routes, subsea cable access, and availability of suitable industrial land.
As the situations in Mumbai and Chennai evolve, operators are migrating towards locations that offer more expansive land and enhanced grid capabilities, propelling growth into Navi Mumbai, Greater Noida, Visakhapatnam, and Kochi.
This transition is actively reinforced by AI workloads, which impose demands for larger premises, fortified substations, and enhanced cooling systems—capabilities that several tier-II regions can adequately provide.
New AI-centric projects, particularly those in Visakhapatnam, manifest this trend. Supportive measures anticipated in the upcoming Budget 2026 will further facilitate this dispersion.
To qualify for the tax holiday, data centres must be Indian-owned and certified by MeitY, promoting expansion into tier-II markets where land costs are considerably lower.
The extended tax holiday through 2047 mitigates future tax exposure anxieties, making sites beyond the traditional Mumbai-Chennai corridor more attractive to global cloud operators.
Resource Strain
The push for geographical dispersion reflects the inherent constraints confronting data centres, which notoriously consume vast amounts of energy, water, and land. Power demand is expected to escalate to 57 terawatt-hours (TWh) by 2030, up from 13 TWh in 2024, according to S&P Global.
This surge is predominantly driven by AI workloads, necessitating between 30 and 50 kW per rack. In stark contrast, traditional server racks require only 6 to 12 kW.
Furthermore, robust water supply is crucial for cooling; a single 100 MW facility operates with water usage comparable to that of an entire town of 85,000 residents annually.
Such demands exacerbate friction in water-scarce locales like Mumbai, Chennai, and the National Capital Region, where groundwater levels have plummeted sharply adjacent to data-centre clusters.
S&P Global predicts that between 60% and 80% of India’s data centres will experience high water stress within this decade.
The need for land is equally pressing, as a 10 MW facility typically requires around 10 acres. Hyperscale campuses require considerably more, thereby intensifying the competition for prime locations near cable landing stations and depleting valuable urban spaces.
The government has classified data centres as infrastructure, granting them state-level incentives, including waivers on electricity duties. Furthermore, ambitious renewable energy targets of 500 GW by 2030 are bolstered by advancements in nuclear power and liquid-cooling technologies.
A Global Contest
India is actively participating in a global race to attract investments in data centres and AI computation. Jensen Huang of Nvidia has characterized this movement as “the largest infrastructure build-out in human history.”
Global data-centre capacity is projected to swell to 200 GW by 2030 from 103 GW in 2025, according to JLL.
India enjoys specific competitive advantages. Development costs hover around $5.4 to $6.8 million per MW, making it the second most economical option following China and significantly cheaper than markets such as the US, UK, or Japan, as indicated by Cushman & Wakefield.
Additionally, industrial power tariffs in critical Indian states are lower than those in developed markets including the UK and Germany.
Strategic geographical positioning places India at the convergence of cable routes linking Europe, the Middle East, and Southeast Asia, with Mumbai and Chennai augmenting subsea cable capacity that rivals Singapore’s current offerings, even as it faces land and power limitations.
Nonetheless, India must navigate the risk of “data dumping,” where foreign companies transfer resource-inefficient facilities rejected by more stringent jurisdictions. An illustrative case is Google’s halted project in Chile due to concerns over aquifer depletion.
Furthermore, India is confronted with a talent gap. While the nation boasts a vast IT workforce, there exists a significant deficiency in specialized technicians proficient in supporting data centres, an area where advanced economies exhibit greater operational maturity.
Source link: Livemint.com.
