India's artificial intelligence ambitions are entering a decisive phase. The Union Budget's 21-year tax holiday for data centres, followed by the India-US agreement to expand cooperation in advanced technologies such as graphics processing units (GPUs), sends a clear signal. India intends to host a larger share of the world's AI infrastructure, not merely consume it.
As the India AI Summit approaches, this moment invites a broader reflection. AI is often discussed in terms of talent, algorithms, and compute capacity. Increasingly, however, it is also an energy story. Watching the debate unfold from Washington, DC, it is striking how quickly AI has moved from being a technological conversation to one framed in the language of power plants, grid capacity, and fuel security. Beneath the excitement lies a simple truth: AI runs on electricity. It depends on power supply, cooling systems, and grid reliability, and the way countries plan these foundations will determine the sustainability and competitiveness of their AI ecosystems.
That shift is already visible globally. In several advanced economies, rapid growth in data centres is reshaping power system planning as electricity demand from the sector climbs sharply. The consequences are becoming clearer: rising costs for consumers, greater reliance on fossil fuels in some on-site deployments, and growing competition with other electrification priorities such as housing and transport.
In Ireland, where data centres account for roughly a fifth of national electricity use, regulators have tightened grid connection rules and required new facilities to provide on-site generation or storage. In parts of the United States, surging data-centre demand is driving gigawatts of new capacity, with some operators turning to integrated projects that combine data centres with dedicated gas plants to guarantee reliability. Singapore temporarily paused new data-centre approvals after the sector's rapid expansion, and the Netherlands has curtailed hyperscale projects amid concerns over grid constraints and renewable power allocation.
These developments are less about policy missteps and more about the sheer speed of AI-driven demand. GPU-intensive workloads are significantly more power-intensive than traditional cloud services and are designed to operate on uninterrupted, high-quality electricity with multiple layers of backup. Ensuring reliability at this scale is a genuine challenge, even for advanced grids. At the same time, rapid solutions adopted today can shape energy systems for decades.
It is against this backdrop that India's position stands out. Over the past decade, the country has undertaken one of the most ambitious energy transformations in the world. Electricity access has expanded to more than 230 million people. Installed power capacity has crossed 500 GW. India is now the world's fourth-largest renewable energy market, with renewable capacity more than tripling since 2014 and non-fossil sources accounting for over half of total installed capacity. Record annual additions and major investments in interstate transmission corridors are strengthening the system's ability to integrate large volumes of clean power.
These achievements have underpinned India's economic growth and digital expansion. They also offer a foundation for the next phase - one that will be more energy-intensive as AI adoption scales.
At the same time, AI-driven data centres introduce distinct system demands. Their electricity use is concentrated and continuous, often near urban centres. Cooling requirements intensify during periods of heat stress, when grids are already under pressure. What matters is not only annual generation capacity, but whether clean and reliable power is available at the precise hours it is needed.
This is where policy alignment becomes critical. The long-term tax incentive for data centres should be seen as a strategic enabler. It offers investors certainty while creating space for thoughtful system design. Aligning AI infrastructure growth with energy planning from the outset can allow India to avoid the carbon-intensive detours now visible elsewhere.
Energy need not be a constraint on India's AI ambitions. It can be a competitive advantage. Encouraging data-centre locations that reflect grid readiness and renewable availability can reduce system stress while improving reliability. Expanding frameworks for round-the-clock clean power supported by storage and flexible resources can ensure AI growth strengthens climate goals rather than complicates them. Continued investments in grid modernisation and advanced forecasting will further enhance resilience.
There is also a virtuous cycle to nurture. AI does not only consume energy; it can help manage it. From improving renewable integration and load forecasting to enabling predictive maintenance and smarter grid operations, AI applications can enhance the efficiency and reliability of power systems. Countries that build this dual capability - using AI to drive growth while strengthening the infrastructure that powers it - will shape the next generation of global standards.
India's choices, therefore, carry relevance beyond its borders. While some economies are responding to AI demand with fossil-backed solutions or slowing projects amid grid strain, India has the opportunity to demonstrate a balanced and forward-looking model - one that combines AI ambition with long-term energy responsibility.
If India succeeds, it will not merely host data centres. It will show that the infrastructure of the AI age can be built in a way that supports growth, resilience, and sustainability simultaneously. In a world navigating both AI acceleration and energy transition, that would be leadership of lasting consequence.
(Piyush Verma is Senior Fellow and Head of Energy and Climate Policy at ORF America based in Washington, D.C. The views expressed are personal.)
Disclaimer: These are the personal opinions of the author
