On a January morning in 2025, Mukesh Ambani told investors that Jamnagar — a city historically associated with refinery infrastructure rather than digital connectivity — would be transformed into a gigawatt-scale AI data centre hub.
The announcement drew limited attention in the Western financial press, but within India it carried the weight of a strategic declaration. The country’s richest businessman was making a generational bet that the core infrastructure powering the AI era should be built, owned, and operated on Indian soil.
By the close of 2025, Google had committed $15 billion to AI data centre projects with Adani and Bharti Airtel. Microsoft pledged $17.5 billion. Amazon announced $35 billion. Between 2019 and the third quarter of 2025, India secured nearly $94 billion in total digital infrastructure investment commitments from global and domestic players. The scale of these investments is too big to ignore.
While India’s digital infrastructure is expanding at breakneck speed, leading urbanists and technologists view the nation’s preparedness with deep-seated ambivalence. The strain on the power grid is described as palpable.
The depletion of water tables in critical hubs like Bengaluru, Chennai, and Hyderabad has emerged as a critical vulnerability that corporate leaders often avoid addressing. And then there is the regulatory critique: India’s regulatory ambitions are grand, but they are currently undermined, the argument runs, by a lack of rigorous enforcement.
The core argument is direct: India must seize this opportunity immediately. This is not a task for the distant future or a project to be delayed until a new political term. The current opening — to secure regional data sovereignty and guarantee that the infrastructure of South Asia’s AI evolution remains on Indian soil and under Indian control — is too valuable to let slip.
Competitors like Thailand, Indonesia, and Malaysia are already moving swiftly. While they possess distinct strengths, India offers a unique combination of scale spanning population, data creation, technical expertise, and political determination. The critical uncertainty is whether India will strategically deploy this immense capital or allow it to dissipate through hesitation.
The architecture of a moment
To understand why now matters so acutely, we must first understand the structural forces that have converged simultaneously — a coincidence of tailwinds so unlikely it is difficult to model.
India’s digital base is extraordinary. The country has nearly 970 million internet connections. UPI processed over 100 billion transactions in 2024, a volume that exceeds the transaction throughput of most G20 economies.
Total data consumption has almost surpassed 25 exabytes per month by 2025. The Aadhaar biometric identity system has enrolled over a billion people. No country outside China has digitised at this speed or at this scale, and China’s digital infrastructure exists behind a wall that makes it functionally irrelevant to international operators seeking to serve the Indo-Pacific.
This demand base is the first tailwind in India’s case.
The second is more structural and more durable: data localisation. India’s Digital Personal Data Protection Act of 2023 mandates that certain categories of personal data be stored within Indian borders. The Reserve Bank of India already requires all payment system data to reside in Tier 4 data centres on Indian soil.
These are compliance obligations that global companies from Meta to Mastercard must satisfy if they wish to access Indian users. India has, in essence, created guaranteed domestic demand for its own infrastructure. HDFC Bank and ICICI Bank have both commissioned dedicated campuses in response.
The third tailwind is AI itself. The global race to build artificial intelligence infrastructure has redefined what a data centre is. A traditional data centre rack consumes between 5 and 10 kilowatts of power. An AI-focused rack demands 10 to 15 times that amount.
The world’s hyperscalers — AWS, Google, Microsoft — are scrambling to build GPU clusters at scale, and they need to build them where land is available, talent is concentrated, and, most importantly, regulation is navigable. India, with its enormous engineering graduate output and comparatively affordable real estate outside the primary metros, sits at an attractive intersection.
And then there is the infrastructure status windfall. In a decision that went largely underappreciated outside the sector, the Government of India classified data centres as infrastructure, placing them in the same legal category as highways and ports. This single administrative act has eased access to long-term, low-cost financing from development finance institutions.
Combined with the 2025–26 Union Budget’s offer of a 20-year tax holiday for foreign cloud operators serving global demand from Indian facilities, the policy architecture has become genuinely competitive with Singapore and Hong Kong. As one JLL report observed, “India’s data centre market stands at an inflection point where robust fundamentals, emerging technologies, and supportive regulatory frameworks converge to create exceptional growth opportunities.”
The scorecard: where India actually stands
Capacity
India’s operational data centre capacity reached approximately 1.53 gigawatts of IT load as of the third quarter of 2025 — a figure that, while impressive in absolute terms, represents roughly the IT load of a single large American hyperscale campus cluster. The vacancy rate, however, tells a more revealing story: just 4.3 per cent, down from historical averages, signalling that demand is outpacing supply faster than any prior forecast suggested.
The pipeline is where the story becomes genuinely interesting. According to S&P Global Market Intelligence’s 451 Research, another 1.4 gigawatts is currently under construction — meaning India’s capacity is set to double within two years. Looking further out, more than 5 gigawatts of additional IT load capacity is expected to be operational by 2030. IDC projects the Asia-Pacific (ex-Japan) data centre market will reach 142,600 MW by 2029, and India will be one of its primary engines.
Geography
Mumbai remains India’s undisputed data centre capital, commanding 54 per cent of total national capacity and serving as the terminus for several of the world’s most significant submarine cable systems. Chennai is the second major node, with particular strategic importance as a cable landing hub connecting India to Southeast Asia and the Gulf. Delhi-NCR and Bengaluru form the third and fourth clusters respectively, driven by enterprise demand from the financial sector and IT industry. Collectively, these four metros account for roughly 90 per cent of India’s current operational capacity.
But the story is beginning to decentralise. Hyderabad now accounts for 6 per cent of market share, buoyed by Andhra Pradesh’s Data Centre Policy 4.0 which explicitly targets an additional 200 megawatts of capacity and offers land allotment and fiscal support.
Google’s $6 billion data centre investment in Visakhapatnam, announced in 2025, is the clearest signal that major operators are willing to move beyond the traditional quad of metro hubs. RackBank’s $120 million AI-focused campus in Raipur, Chhattisgarh — a state rarely mentioned in technology conversations — suggests that the frontier is expanding faster than the policy maps can track.
Connectivity
India’s submarine cable ecosystem has been steadily maturing into one of the most strategically significant in the world. The arrival of 220 terabits per second of new subsea capacity landing at Chennai and Mumbai gives India the raw bandwidth to serve as a transit and hosting node for traffic flowing between Europe, the Middle East, and East Asia. The NTT-operated MIST cable, connecting Singapore with India and broader Southeast Asia, is one example of how India is being woven into the backbone of global data movement as a routing hub.
The competition: why settling for second would be a civilisational error
The analogy of the Green Revolution illuminates India’s current situation, though it may seem an unusual parallel for the data centre sector. In the 1970s, India confronted a pivotal decision: whether to invest aggressively in domestic agricultural productivity or remain tethered to food imports. Despite its controversies and shortcomings, the Green Revolution acted as a transformative engine that liberated the nation from agricultural dependency. Today, data centre expansion represents India’s Green Revolution for the AI era — as global competitors take decisive strides toward dominance.
The predicament of Singapore is rich and instructive in this regard. For two decades, it was the unchallenged nerve centre of Asia-Pacific digital infrastructure. Then its land ran out, its power grid hit capacity, and a moratorium froze construction for three years. The repercussions were predictable: operators fled to Johor, Batam, and Bangkok. Singapore is now recalibrating toward higher-value, lower-volume workloads — financial data processing, critical AI inference, government cloud. The mass-market hyperscale business is flowing elsewhere.
Forecasts from BMI Country Risk and Industry Research identify India, Malaysia, and Indonesia as the primary growth engines for the Asia-Pacific data centre market between 2025 and 2028. India’s structural advantages, however, remain distinct.
Malaysia’s boom in Johor is largely a spillover effect from Singapore’s constraints and lacks a comparable domestic demand base. Indonesia’s growth is tangible, but it remains geographically limited to Greater Jakarta and is hampered by power grid unreliability outside Java. India alone offers the unique confluence of a billion-plus consumer base, a vast continental footprint, a legitimate pipeline for renewable energy, and a government capable of executing policy at exceptional speed.
The IndiaAI Mission serves as the most prominent evidence of this execution capability. Initiated in 2024 with a $1.2 billion allocation for GPU acquisition through public-private collaborations, the mission has already demonstrated significant operational progress. By mid-2025, it had successfully allocated 34,371 GPUs through two rounds of tendering, with nearly 50 per cent of that capacity currently deployed. This initiative marks a milestone in state-supported AI infrastructure development that remains unparalleled by any Southeast Asian counterpart.
The green veto, and why India must reject it
Historically, every significant infrastructure initiative in India has faced a unified front of critics brandishing provocative data and moral outrage. Whether the grievance concerns highways penetrating forests, dams uprooting communities, or factories impacting waterways, a familiar pattern emerges.
Today, this opposition targets data centres, citing concerns over water consumption and coal-based energy use. While the underlying statistics presented by these objectors are often accurate, their conclusions regarding national policy are fundamentally misguided.
The ecological opposition to India’s data centre growth is organised around three primary arguments. Rather than receiving uncritical validation, these claims require rigorous refutation. Allowing such critiques to dictate national strategy would relegate India to a status of lasting infrastructural dependence — even as nations that established their own digital bases decades prior provide guidance from a state of established development.
Claim I — “Data centres will devour India’s grid and turbocharge its carbon footprint.”
Projections indicate that India’s data centre electricity requirements will surge from 13 TWh in 2024 to approximately 57 TWh by 2030, representing an increase from 0.8 per cent to 2.6 per cent of total national demand. While environmentalists frame this escalation as a looming emergency, it actually serves as a compelling reason to expand power infrastructure rather than stifle digital growth.
Historical precedents justify this path. The United States fuelled its industrial rise with coal without prioritising carbon metrics. Germany established its manufacturing dominance through inexpensive Russian gas. China has commissioned more coal capacity in the last ten years than India’s entire current base. The nations currently pressuring India to scrutinise its data centre energy consumption are the very ones whose past emissions precipitated the global climate crisis.
While home to 18 per cent of the global population, India is responsible for only about 7 per cent of world CO₂ emissions. Its per capita output remains significantly lower than that of Europe or America.
Demanding that India prematurely limit a sector poised to contribute hundreds of billions to its GDP in order to safeguard a climate damaged by wealthier nations defies both moral and economic reason. Such a stance is not true environmentalism; it is development gatekeeping masked by ecological rhetoric.
What the data actually shows
India’s renewable energy pipeline is already among the world’s most ambitious. Between 2022 and 2025, the country added more solar capacity annually than many European nations have installed in total.
The 2025–26 Union Budget allocated Rs 20,000 crore for the development of indigenous Small Modular Reactors by 2033, which falls within the output range required for large data centre campuses. Liquid cooling and AI-driven energy management systems, already being deployed commercially, can reduce electricity consumption by 30–40 per cent. The grid will expand alongside demand, as it always has. That is what development looks like in the true sense.
Claim II — “Data centres will drain India’s aquifers and steal water from ordinary families.”
This particular critique, laden with emotional resonance, demands a stringent analytical breakdown. Forecasts suggest water consumption by India’s data centres will rise from 150 billion litres in 2025 to 358 billion litres by 2030.
To stir public resentment, advocacy groups often use inflammatory comparisons — for instance, claiming a 20 MW facility’s daily water requirement equals that of 27,000 families. This narrative paints data centre operators as entities seizing a critical public asset from the needy to power infrastructure for the elite.
This perspective obscures a fundamental reality: India’s water challenges far precede the data centre boom. The crisis is predominantly driven by agriculture, which NITI Aayog identifies as responsible for nearly 89 per cent of India’s freshwater withdrawals.
Data centres do not significantly aggravate or alleviate this situation. Even at its projected 2030 peak, the 358 billion litres consumed by data centres would constitute less than 0.05 per cent of the water utilised annually by agriculture. To suggest that curbing data centre development is a viable strategy for mitigating water stress is to use statistical ignorance as a policy tool.
What the data actually shows
Modern data centre operators are, by necessity, among the most water-efficient large-scale industrial users in India. Closed-loop cooling systems, liquid immersion technology, and AI-driven thermal management can reduce water intensity by 30–40 per cent compared to legacy infrastructure.
Amazon has invested in lake restoration projects near Bengaluru and Hyderabad that are expected to replenish more than 570 million litres of water annually — exceeding the consumption levels of many such facilities. The primary financial motivation for the sector is to minimise water usage rather than increase it: cooling represents a significant operational expense rather than a luxury.
Claim III — “Data centres are being built without environmental impact assessments, and this is lawless.”
Procedural objections often become the final refuge for critics who have lost the argument on substance. India’s current regulatory framework requires environmental clearances for data centres under building construction norms for projects exceeding 20,000 square metres — a system some critics argue is inadequate because it focuses more on construction-related effects than long-term operational impact.
But if one believes data centres are ultimately a net positive for the economy, the logical response is to build a more specialised regulatory framework, not to use procedural gaps as grounds for indefinite delay. Nearly every major infrastructure sector in India — from power generation and telecom towers to national highways — expanded before a fully evolved regulatory architecture was in place.
The frameworks need to evolve alongside the industry, and to evolve, the industry must experience its boom without the guardrails of irrational scrutiny.
What the data actually shows
India’s regulatory framework for data centres is already evolving alongside the sector’s growth. In August 2025, the Supreme Court clarified that large construction projects between 20,000 and 1,50,000 square metres could proceed through State Environment Impact Assessment Authority (SEIAA) approvals rather than requiring central-level clearance, effectively decentralising and accelerating environmental processing for large infrastructure projects.
The judgment also reaffirmed that environmental scrutiny would continue through the EIA framework instead of blanket exemptions.
What seizing the moment actually requires
When political determination takes hold, India has demonstrated a remarkable ability to deliver infrastructure projects once thought unattainable.
The implementation of GST, which consolidated 29 distinct state tax regimes, was written off by many as a bureaucratic impossibility. UPI transformed the commercial landscape for 1.4 billion citizens. Aadhaar achieved a billion enrolments in just a few years. Such achievements were far from minor improvements; they were significant leaps for society, realised even under the pressures of democratic timelines.
While the technical hurdles of the data centre sector surpass these prior examples, the core governance issues remain recognisable. There are vital interventions that can be realistically completed within one or two electoral terms.
The National Data Centre Policy must be bolstered by first acknowledging that the 2025 draft serves as an enabling roadmap for a sector of strategic national importance, rather than just a tool for environmental oversight. Key objectives should include enhancing regulatory transparency, streamlining approval timelines, and ensuring the dependability of infrastructure.
For hyperscale sites exceeding a 20 MW IT load, the policy can implement uniform reporting standards for Power Usage Effectiveness (PUE), Water Usage Effectiveness (WUE), the sourcing of green energy, and backup power setups — while avoiding the creation of an overly complex, multi-tiered licensing system.
Instead of prescriptive operational mandates, the framework should encourage market-led efficiency through open-access renewable power procurement, banking provisions for intermittent solar and wind generation, time-of-day tariff flexibility, captive energy integration, and faster grid interconnection approvals.
Liquid cooling systems, high-density rack optimisation, AI-driven thermal load balancing, and closed-loop water systems are already reducing resource intensity across modern facilities, because efficiency directly improves operating margins and lowers total cost of ownership.
India’s national grid must continue balancing household consumption, manufacturing expansion, transport electrification, and industrial growth alongside the rapid scaling of AI and cloud infrastructure. Accelerated approvals for open-access renewable procurement, captive generation, energy storage integration, and dedicated transmission connectivity can help hyperscale operators build long-term energy reliability without depending excessively on subsidised public power infrastructure.
India demonstrated during the vaccine rollout that large-scale administrative coordination and fast-track clearances are possible when strategic capacity creation becomes a national priority. The same execution model can be applied to digital infrastructure and power integration over the next five years.
India’s concentration of data centre infrastructure in hubs such as Mumbai, Delhi-NCR, Chennai, and Bengaluru should be viewed as a strategic strength rather than a structural weakness. Digital infrastructure clusters naturally emerge around dense fibre connectivity, subsea cable landing stations, financial ecosystems, enterprise demand, and reliable power networks.
Mumbai alone handles the majority of India’s international internet traffic through its subsea cable infrastructure, while Chennai’s coastal connectivity and Bengaluru’s enterprise ecosystem make them globally competitive locations for hyperscale deployments.
At the same time, the expansion of these hubs does not prevent the rise of secondary markets. Tier-2 cities such as Visakhapatnam, Coimbatore, Jaipur, Lucknow, and Bhubaneswar are already benefiting from spillover investment as operators seek lower latency distribution, edge computing capacity, and lower-cost land banks.
Andhra Pradesh and Rajasthan have demonstrated that targeted incentive structures, pre-cleared industrial zones, renewable power integration, and transmission support can gradually expand the digital infrastructure footprint beyond legacy metros without weakening the efficiency advantages of existing clusters.
Infrastructure concentration is a normal feature of industrial development. Financial centres, manufacturing corridors, ports, and technology ecosystems globally tend to cluster before diffusing outward over time. India’s leading data centre hubs are creating the scale, connectivity density, and operational expertise that will eventually support broader nationwide expansion.
The reckoning
History rarely presents itself as an obvious crossroads. But sometimes the underlying trends become clear enough that ambiguity is no longer intellectually honest. According to S&P Global Commodity Insights, India is expected to become the second-largest data centre electricity demand market in the Asia-Pacific region within two years, overtaking both Japan and Australia.
IDC projects that the APAC data centre market will grow at a 22 per cent CAGR through 2029, while CBRE estimates that investment commitments in India’s digital infrastructure sector have already crossed $94 billion.
Mordor Intelligence projects that India’s total installed data centre capacity, including facilities under development, will rise from 4,480 MW in 2025 to more than 15,210 MW by 2031 — a 240 per cent expansion within six years, on a measure that captures both operational IT load and projects in the pipeline.
These are structural forecasts derived from accelerating enterprise digitisation, AI compute demand, cloud adoption, financial data localisation requirements, and the rapid expansion of India’s digital economy. The scale of projected growth suggests that data infrastructure is increasingly becoming a foundational layer of national economic capacity.
Many critics of data centre expansion frame their objections as environmental caution, but the practical effect is often developmental paralysis. No major industrial power — whether the United States during the rise of the internet economy, Germany during its manufacturing expansion, South Korea during its shipbuilding surge, or China throughout its infrastructure boom — developed by freezing strategic capacity creation until every environmental trade-off was fully resolved.
The countries that built the foundations of the digital economy in the 1990s and 2000s did not hesitate over whether server infrastructure would consume electricity or water. They built scale first, accumulated economic leverage, and only later acquired the fiscal space to optimise and regulate at higher levels of sophistication.
India does not have the luxury of delaying that sequence. Data generation is already accelerating — estimated at nearly 25 exabytes per month — while AI workloads, cloud infrastructure demand, and global data flows continue expanding regardless of domestic hesitation.
GPUs will be installed somewhere. Submarine cable traffic will be routed somewhere. Hyperscale infrastructure will be built somewhere. The strategic question is whether that infrastructure capacity is created in India or diverted to competing hubs such as Johor in Malaysia, Singapore’s data parks, or large server clusters in Virginia and the Gulf region.
The shift of investment to other regions does not eliminate the environmental impact. Energy consumption, water use, and emissions will persist elsewhere — but they will be accompanied by the benefits of job creation, tax income, supply-chain development, and technology advancement. In that scenario, India would remain burdened by its hesitancy while losing out on all the economic benefits.
India should build. India should build fast. And India should do so with the confidence of a nation that understands, better than any Western think-tank ever will, the true environmental catastrophe it is trying to escape — the catastrophe of poverty.
