Expert SpeakYoung Voices
Published on Jun 30, 2026
Without reliable power, India’s AI ambitions risk becoming dependent on infrastructure beyond its borders
Image Source: ChatGPT
During the industrial era, access to coal and oil shaped national power. In the artificial intelligence (AI) era, a structurally similar logic is taking shape. Most discussions about the global AI race focus on semiconductors, frontier models, and engineering talent. Yet these visible elements of AI development rest on a less discussed but equally essential foundation: electricity. Every large language model, data centre, and inference cluster ultimately depends on a steady supply of power. As digital technologies become progressively central to economic growth, commercial development, and national security, the infrastructure that powers them acquires strategic significance. For India, which has set bold AI targets while contending with significant energy infrastructure constraints, the availability of reliable electricity will determine not just the pace of AI adoption but its geography, and with it, its sovereignty.
The Physical Scale of AI’s Energy Demand
The numbers are now well established. According to the International Energy Agency (IEA)’s Energy and AI report, global data centre electricity consumption stood at approximately 415 terawatt-hours (TWh) in 2024, which is around 1.5 percent of global electricity demand. The IEA projects this to nearly double to 945 TWh by 2030, growing at 15 percent per year, more than four times faster than all other sectors combined. The trend is already evident. A follow-up report confirmed that data centre electricity demand surged 17 percent in 2025 alone, significantly outpacing overall global electricity demand growth, which stood at three percent. What makes this strategically significant is not merely the volume but the character of the demand. AI data centres require 24/7 uninterrupted, high-quality power, and the power density of AI servers has increased eleven-fold since 2020. As a result, access to reliable electricity increasingly influences where AI can be deployed at scale and, by extension, which countries are best positioned to compete in the emerging digital economy.
AI, Energy, and Geopolitics
The major powers have already embedded this understanding into their strategic behaviour. The United States has designated federal nuclear sites at Savannah River and Oak Ridge for AI data centres, with Energy Secretary Chris Wright describing the effort as the next “Manhattan Project for AI leadership.” Technology companies including Microsoft, Google, and Amazon have contracted over 10 gigawatts of new nuclear capacity to secure firm baseload power for round-the-clock AI operations
The Gulf states have taken a different but equally deliberate path. The United Arab Emirates (UAE)’s 5-gigawatt UAE–US AI Campus in Abu Dhabi (the largest AI campus outside the United States) and Saudi Arabia’s Humain entity, whose CEO has estimated theproject cost at US$77 billion, illustrate how energy resources are being leveraged to build strategic computing capacity. China is pursuing a parallel approach. Having accounted for 25 percent of global data centre electricity consumption in 2024, China added more than 430 gigawatts of new renewable capacity in 2025, marking one of the largest single-year expansions of renewable power in modern history. Across these cases, energy infrastructure is no longer viewed simply as a supporting input for economic activity. It is increasingly being treated as a source of strategic advantage.
India’s AI Ambitions and the Energy Challenge
India’s AI ambitions are significant. The IndiaAI Mission, backed by an outlay of INR 10,372 crore, is designed to expand domestic compute capacity, support AI startups, and establish India as a global AI hub. Progress on that front is already visible. The country now hosts more than 268 major data centres across Mumbai, Hyderabad, Bengaluru, Chennai, and Delhi-NCR, with investment from global hyperscalers continuing to accelerate
India is one of the world’s fastest-growing electricity markets, with ambitious clean energy targets and major solar and wind additions underway. However, the critical bottleneck is not generation in the aggregate; it is reliability, transmission, and the specific demands of AI-grade power consumption
According to the Institute for Energy Economics and Financial Analysis (IEEFA), India’s data centre capacity is projected to grow from approximately 1.4 gigawatts today to 9 gigawatts by 2030. Reflecting this rapid expansion, it is estimated that an additional 40–45 TWh of electricity would be required by 2030 to meet projected AI and data centre demand. In the first quarter of 2026 alone, India lost 300 gigawatt-hours of already-generated renewable electricity to transmission curtailment, a direct consequence of grid infrastructure that cannot deliver power where it is needed.
The concentration of data centres in a few metros, particularly Mumbai and Chennai, amplifies localised grid pressure. India’s hot climate further increases cooling requirements: one-third of India’s data centres are located in regions where average annual temperatures exceed the recommended 27°C operating threshold, raising power consumption and costs
If India cannot provide electricity for AI operations at scale, its AI ecosystem may become increasingly dependent on overseas data centres, many of which are owned and operated by non-Indian entities. This creates vulnerabilities across data governance, national security, and the terms on which Indian institutions and public services access AI capabilities
India’s Nuclear Energy Mission, which targets 100 gigawatts of capacity by 2047 and is supported by the Sustainable Harnessing and Advancement of Nuclear Energy for Transforming India (SHANTI) Act 2025, reflects recognition of this challenge. However, nuclear power at a meaningful scale remains a medium- to long-term prospect, and any rapid expansion will bring social costs and regulatory challenges that require careful policy attention. The more pressing challenge, however, lies in the near term.
Taken together, these constraints do more than slow India’s AI ambitions. They could gradually push them beyond its borders. When reliable energy is difficult to secure, demand naturally shifts toward foreign-hosted compute infrastructure.If India cannot provide electricity for AI operations at scale, its AI ecosystem may become increasingly dependent on overseas data centres, many of which are owned and operated by non-Indian entities. This creates vulnerabilities across data governance, national security, and the terms on which Indian institutions and public services access AI capabilities. Energy adequacy is, in this sense, a foundational precondition for AI sovereignty, not an afterthought.
What Should India Prioritise?
Three priorities require immediate attention. First, AI demand projections must be incorporated into the National Electricity Plan. Currently, they are absent, which means India’s grid is being designed for a world that no longer exists. A working precedent already exists: the United Kingdom’s National Energy System Operator (NESO) has formally integrated data centre and AI demand growth into its Future Energy Scenarios and Strategic Spatial Energy Plan, a model India’s Central Electricity Authority could directly adapt.
Second, transmission infrastructure must expand faster than data centre capacity, or India will keep generating power it cannot deliver, as the 300 gigawatt-hour curtailment in early 2026 already demonstrates
Over the medium term, two parallel tracks are necessary: renewable energy paired with long-duration storage to address intermittency, and accelerated nuclear deployment under the SHANTI Act to secure firm baseload. Energy and AI cannot be planned in separate rooms
Third, while several Indian states like Maharashtra, Tamil Nadu, and Telangana already offer dedicated power infrastructure and earmarked zones for data centres under their state-level policies, these remain fragmented commercial incentives rather than a nationally coordinated framework. What is missing is integration between these state-level zones and the Central Electricity Authority’s grid planning, so that power commitments made to data centre developers are reflected in national transmission investment. Over the medium term, two parallel tracks are necessary: renewable energy paired with long-duration storage to address intermittency, and accelerated nuclear deployment under the SHANTI Act to secure firm baseload. Energy and AI cannot be planned in separate rooms.
Conclusion
India’s AI strategy will achieve its stated ambitions only if policymakers recognise that data centre capacity and energy infrastructure are not parallel policy domains; they are the same challenge. Success depends as much on grid modernisation and firm-power deployment as on compute quality or model ambition
If AI is to become a pillar of India’s economic and strategic future, its energy requirements must be incorporated into electricity planning today, not retrofitted later. For India, the path to AI leadership runs through its power grid
The timing is particularly significant. The IndiaAI Mission and the Ministry of Power are currently planning India’s AI and energy futures in separate rooms. The Draft National Electricity Policy 2026 that will shape the country’s power sector through 2047 is currently out for public consultation. Yet the draft framework contains no dedicated assessment of AI-driven electricity demand despite the rapid expansion of data centres and compute infrastructure. If AI is to become a pillar of India’s economic and strategic future, its energy requirements must be incorporated into electricity planning today, not retrofitted later. For India, the path to AI leadership runs through its power grid.
Sairah Zahooris a Research Intern at the Observer Research Foundation
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.
PREVNEXT
Author
Sairah Zahoor
Sairah Zahoor is a Research Intern at the Observer Research Foundation
Read More +
Related Search Terms
Publications
The Energy Behind Intelligence: Linking AI Ambitions and Energy Security
Energy Security | Artificial Intelligence
Jun 30, 2026
