India is on the verge of a digital infrastructure supercycle. A new report from Crisil Ratings projects a five-fold expansion in data centre capacity to 7.5 gigawatts (GW) by 2030, fuelled by the massive computing needs of artificial intelligence.
The AI-Powered Growth Engine
The
driving force behind this unprecedented expansion is the global explosion in generative artificial intelligence (GenAI). Unlike traditional computing, AI workloads require immense processing power, which in turn demands high-density, energy-intensive data centres. According to a recent Crisil report, this surge in demand from AI, coupled with ongoing enterprise digitalisation and rising data consumption, is set to attract investments of more than Rs 3 lakh crore over the next four financial years. This isn't just about storing more data; it's about building the computational backbone for India's AI-driven economy. The transition to AI-led workloads is fundamentally reshaping the design, scale, and economics of data centres across the country.
What 7.5 Gigawatts Looks Like
To put the 7.5 GW figure in perspective, it represents a five-fold increase from current levels. This growth is being underwritten by global hyperscalers—large cloud providers like Amazon Web Services, Google, and Microsoft—who are flocking to India. These companies are expected to lease 75-80% of India's data centre capacity by 2030. The demand is so strong that existing facilities are operating at around 90% utilisation, and over 95% of the capacity currently under construction is already pre-leased. This provides a high degree of revenue certainty for developers, who are now signing longer-term contracts of 15-20 years, up from 10-12 years previously.
The Investment Hotspots
While this growth is a national phenomenon, it is heavily concentrated in a few key hubs. Nearly three-fourths of the upcoming 6 GW of capacity is expected to be built in Visakhapatnam, Mumbai, and Hyderabad. Mumbai and Chennai are established leaders, benefiting from their status as financial centres and landing points for subsea data cables. However, emerging hubs like Hyderabad, Pune (as an AI compute hub), and even Kolkata are attracting significant investment from both domestic and international players. Major operators expanding their footprint include AdaniConneX, Yotta Data Services, Nxtra by Airtel, and CtrlS Datacenters.
Power, Water, and Land: The Big Challenges
This massive buildout is not without significant challenges. The most critical hurdle is infrastructure, particularly power. By 2030, the projected data centre capacity in Visakhapatnam could equal roughly 80% of the city's current peak electricity demand, with Mumbai at 60%. This places enormous strain on local electricity grids and makes timely connectivity a crucial factor for project success. Beyond power, data centres are incredibly water intensive, primarily for cooling their servers. In a water stressed country like India, this raises serious environmental and community concerns, with local groups in cities like Thane and Visakhapatnam already protesting new projects over resource allocation. Developers are now focusing on securing not just land, but locations with robust power and water availability that can be sustained without disrupting local communities.
Building a Sustainable Digital Future
To mitigate these challenges, the industry is increasingly focused on sustainability. India's ability to develop large campuses powered by renewable energy is a key part of its appeal to global hyperscalers, who have their own green energy commitments. However, experts argue that a comprehensive national policy framework is needed to enforce standards for energy efficiency, water usage, and grid impact. Currently, regulations are fragmented across different states. The future competitiveness of India's data centre market will depend not just on building capacity, but on doing so responsibly. As the country builds the critical infrastructure for its digital ambitions, ensuring this growth is sustainable will be paramount for long term success.
