The Scale of India’s Digital Ambition
India is witnessing an unprecedented surge in data center construction as global technology giants and domestic conglomerates race to establish a foothold in one of the world’s fastest-growing digital economies. Amazon Web Services has pledged US$12.7 billion to construct cloud computing facilities in Mumbai and Hyderabad, while Google announced a US$15 billion investment for an artificial intelligence hub in Visakhapatnam, marking the largest technology investment in the country to date. These massive financial commitments reflect a broader trend: India’s data center capacity is projected to explode from 1.5 gigawatts today to 13.2 gigawatts by 2032, according to estimates from the Ministry of Electronics and Information Technology.
This expansion is driven by a confluence of factors. With 800 million internet users and rapidly accelerating adoption of artificial intelligence applications, domestic demand for computing power is skyrocketing. Additionally, geopolitical tensions in the Middle East have prompted Gulf-based clients and global hyperscalers to view India as a strategic hedge, diversifying their data center exposure away from conflict zones. Sunil Gupta, chief executive of Yotta Data Services, reported a sharp rise in inquiries since March, with capacity requirements ranging between 200 and 500 megawatts from companies seeking to rework their infrastructure strategies to ensure business continuity.
Currently, India hosts approximately 270 data centers, accounting for roughly 0.5 percent of the nation’s electricity generation. By 2030, this share is expected to climb to 3 percent. While these percentages may appear modest at the national level, the concentration of power demand in specific geographic clusters creates localized strain that threatens to destabilize aging grid infrastructure in major metropolitan areas. The challenge is compounded by the nature of modern AI workloads, which require computational intensity far exceeding traditional cloud storage or web hosting services.
When AI Meets Infrastructure Reality
The energy appetite of artificial intelligence data centers is staggering. A single AI-focused facility consuming 100 megawatts requires as much electricity as approximately 100,000 American households or nearly 750,000 Indian households, given the lower average consumption patterns in India. These facilities operate continuously, drawing power 24 hours a day to maintain servers and specialized cooling systems. Unlike conventional industrial loads that may fluctuate with production cycles, data centers represent a steady, unrelenting baseload demand that offers no respite to grid operators.
According to the International Energy Agency, data centers in the United States consumed more than 4 percent of the country’s total electricity in 2024, a proportion that could more than double by 2030. However, the American grid generates approximately 4,500 terawatt hours annually, compared with India’s 1,900 terawatt hours. This disparity means that India’s grid has far less headroom to absorb new demand without impacting existing consumers. A parliamentary panel warned in March that highly concentrated power demand in data center clusters could strain the country’s transmission infrastructure while threatening clean power ambitions, as many regional grids still rely heavily on coal-fired generation to meet base demand.
Narendra Sen, founder of NeevCloud, an AI cloud service provider, explained that the computational requirements of artificial intelligence create a compounding effect on energy use.
“Data centres are very power hungry and consume more electricity than many of our cities. As more people are using AI now, more queries mean more analyses and that will translate into 20 to 30 times more power consumption in the future,” he told The Straits Times.
Mumbai’s Warning Shot
Mumbai, India’s busiest data center hub with 61 operational sites, exemplifies the looming crisis. These facilities provide services for global firms including Amazon, Google, and Microsoft while drawing power from the same electricity grid that supports 20 million commercial and residential users. By 2030, data centers alone could consume a third of the electricity produced in the city, according to current projections. This concentration creates a zero sum competition between digital infrastructure and basic urban electricity needs.
The strain is not merely theoretical. Across India, authorities struggle to maintain consistent power supply in many cities and towns, with regular power cuts affecting daily life. In April, the national grid faced night time power shortfalls of up to 5.4 gigawatts even before peak summer conditions arrived, as recorded by grid operators. Demand surged from 215,832 megawatts on April 1 to a record 256,117 megawatts by April 25, creating deficits during evening hours when solar generation drops but cooling needs remain high.
Data from the Indian Energy Exchange reveals the economic pressure of this mismatch. During daylight hours, abundant solar generation pushes prices down to 1.5 rupees per unit. After sunset, however, spot prices hit the regulatory ceiling of 10 rupees per unit, representing a nearly seven fold increase within hours. For data centers, which require uninterrupted power to prevent costly service disruptions, this volatility creates both operational and financial challenges.
The Transmission Bottleneck
India’s power grid is structurally unprepared for the round the clock, high load demand profile of modern data centers. The infrastructure was designed for traditional consumption patterns, not for the silent exits and rapid load ramps characteristic of computing clusters. Grid India, the national grid operator, has highlighted that sudden withdrawal of 1 to 2 gigawatts by inverter based data center systems could destabilize grid operations, particularly when these facilities represent large, dynamic loads requiring direct transmission level connectivity.
The physical infrastructure simply cannot keep pace with planned generation capacity. Despite efforts to upgrade transmission and distribution systems since 2024, implementation delays have plagued expansion projects. Only 42 percent of the targeted 15,253 circuit kilometers of transmission lines have been commissioned, including just 3,253 circuit kilometers of interstate transmission lines. This bottleneck prevents the movement of power from renewable energy generation hubs, such as Gujarat and Rajasthan, to demand centers like Visakhapatnam, where Google is building its massive AI hub.
The gap between renewable generation and grid absorption is already evident. In fiscal year 2025-26, India added over 50 gigawatts of renewable energy capacity, yet approximately 2.3 terawatt hours of solar power went unused due to curtailment. Coal plants, which provide baseload stability, cannot ramp down below 55 percent of their load during sunny afternoons, forcing grid operators to discard clean power. This solar cliff phenomenon, where supply collapses after sunset just as demand peaks, creates a structural mismatch that data center expansion will only worsen.
Geographic Diversification and Its Limits
Faced with congestion in metropolitan areas, operators are exploring locations beyond traditional hubs. Vibhuti Garg, director for South Asia at the Institute for Energy Economics and Financial Analysis, noted that industry discussions increasingly focus on moving new facilities 100 to 200 kilometers away from established centers like Hyderabad, Delhi-National Capital Region, Bengaluru, and Chennai. These cities were originally selected for their proximity to undersea fiber cables and availability of technical talent, but they now face severe power and water constraints.
“These locations were picked for their nearness to undersea fibre cables that transfer data, especially for offshore clients like hyperscalers. These cities also have the availability of good talent. But given that they are high demand areas, there is now a lot of discussion about moving the new data centres 100-200km away,” said Ms Vibhuti Garg.
NeevCloud has established AI computing centers in Indore and Raipur, central Indian cities located in low seismic zones with adequate land parcels. These facilities purchase renewable energy from nearby states at rates 30 percent cheaper than metropolitan grid power. Similarly, industry leaders are scouting inactive mines in central India for underground data centers, where natural cooling could reduce energy consumption. CapitaLand Investment, which is developing four data centers across Mumbai, Hyderabad, Chennai, and Bengaluru with plans to double capacity to 500 megawatts by 2030, is evaluating locations in adjacent corridors where fiber connectivity has matured.
However, non-metropolitan areas lack adequate transmission infrastructure to support gigawatt-scale computing clusters. State power distribution companies have struggled to keep up with increasing demand, and the shift of data centers to these regions risks simply moving the strain rather than resolving it. Corporate power purchase agreements, where businesses buy renewable electricity directly from generators, offer a partial solution but remain constrained by the same transmission limitations that affect grid supply.
Global Context and Local Stakes
India’s predicament mirrors a worldwide backlash against the resource demands of artificial intelligence infrastructure. In Mexico, residents near a Microsoft data center report increased power cuts and water outages lasting weeks, leading to school cancellations and health clinic operations conducted by flashlight. In Ireland, data centers consume more than 20 percent of the national electricity supply, while in Chile, aquifers face depletion from cooling demands. Similar concerns have surfaced in Brazil, Britain, Malaysia, the Netherlands, Singapore, and South Africa.
The common thread across these cases is the tension between technological advancement and basic resource availability for local populations. Researchers at the Pahle India Foundation highlighted this concern in a March 2026 analysis.
“The lights cannot afford to dim for a billion people so that the tech cloud stays bright for the few.”
This sentiment captures the ethical and political dimensions of infrastructure planning, where the benefits of AI accrue to global corporations and urban elites while the environmental and reliability costs often fall on ordinary citizens. Unlike the United States, where the grid can accommodate growth through massive investment and diverse geographic dispersal across multiple states and climate zones, India’s infrastructure operates closer to its physical and economic limits. The country’s per capita electricity consumption remains a fraction of developed nations, meaning that diversion of limited supply to energy intensive computing could directly impact residential and agricultural users who still lack reliable access to power for basic needs and irrigation.
Technical Solutions and Industry Responses
Industry players are deploying various strategies to mitigate their energy impact. Major corporations including Reliance Industries and AdaniConneX are moving toward establishing dedicated clean energy power plants for their own facilities, reducing reliance on the public grid. Iron Mountain Data Centers has partnered with CleanMax Enviro Energy Solutions to supply 32 million kilowatt hours of hybrid solar and wind power annually to its Mumbai, Pune, and Bengaluru locations, increasing renewable energy share to 75 percent under a group captive model.
Cooling technology represents another frontier for efficiency gains. Approximately 39 percent of a data center’s power consumption goes to cooling, while 41 percent runs the servers themselves. CapitaLand Investment has implemented direct-to-chip liquid cooling technology, which reduces water consumption by 30 to 40 percent while enabling rack density increases from 75 kilowatts per rack to 110-120 kilowatts. This approach supports AI led deployment while reducing dependency on traditional water cooling methods.
Energy storage and flexible generation are emerging as critical complements to renewable sourcing. India currently operates only 1,082 megawatt hours of battery storage capacity, though the government has approved 43 gigawatt hours under viability gap funding. For longer-duration flexibility, experts recommend modern gas-based reciprocating engines that can reach full output in under five minutes and operate efficiently during short balancing periods. Unlike traditional coal or combined cycle gas plants, these engines can run on sustainable fuels such as green hydrogen, offering a transition path to zero-carbon generation.
The Policy Imperative
Despite these corporate initiatives, experts argue that voluntary measures are insufficient to prevent grid destabilization. Krishna V. Giri and Payal Seth, fellows at the Pahle India Foundation, have called for mandatory regulation requiring energy self-sufficiency for data centers above defined capacity thresholds. They advocate for pre-planning dedicated data center zones with coordinated transmission upgrades, rather than allowing ad hoc expansion that leads to inefficiencies and rising consumer tariffs.
Deloitte’s research on Asia Pacific data center growth emphasizes a power first approach to planning, where energy sourcing considerations shape location and design decisions from the earliest stages. This strategy involves combining onsite renewables, utility green tariffs, long-term power purchase agreements, and advanced load-shifting systems that move workloads to periods of abundant renewable generation. The consulting firm projects that data center electricity consumption across Asia Pacific could grow more than fivefold from under 200 terawatt hours in 2025 to over 1,000 terawatt hours by the mid 2030s, making proactive planning essential.
The Indian government has offered a 20-year tax holiday until 2047 for foreign cloud service providers, alongside discounted power tariffs and capital subsidies from states like Tamil Nadu and Andhra Pradesh. However, these incentives may prove counterproductive if they accelerate construction faster than grid infrastructure can adapt. As Surajit Chatterjee, CapitaLand Investment Managing Director and Head of Data Centre India, stated, sustainability is becoming a mandate for both environmental and operational reasons, requiring close coordination between industry bodies and government agencies to ensure data center needs are integrated into evolving renewable energy policies.
Key Points
- India’s data center capacity is projected to grow from 1.5 GW to 13.2 GW by 2032, with electricity consumption rising from 0.5% to 3% of national generation
- Mumbai’s 61 data centers could consume one-third of the city’s electricity by 2030, straining a grid shared with 20 million commercial and residential users
- A single 100 MW AI data center uses as much power as 750,000 Indian households, creating concentrated demand that threatens grid stability
- Transmission infrastructure is lagging behind generation capacity, with only 42% of targeted power lines commissioned and 2.3 TWh of solar energy curtailed due to grid constraints
- Industry is moving toward tier-2 cities and corporate renewable energy agreements, though transmission bottlenecks persist in these locations
- Experts recommend mandatory energy self-sufficiency regulations and power first planning to prevent competition for limited electricity between data centers and residential users
