- India's data centre footprint is set to exceed 101 million sq ft by 2030, up from 27 million sq ft in 2026
- Data centres require specific infrastructure like steady power, water, connectivity, renewable energy access
- Land for data centres competes with residential and commercial projects without dedicated zoning in India
India's data centre story is no longer just about servers and storage. It is quickly becoming a real estate story too.
The country's data centre footprint is expected to cross 101 million sq ft by 2030, up sharply from nearly 27 million sq ft in the first half of 2026, according to an Anarock Capital report. The sector has come a long way from just 0.6 million sq ft in 2007.
More than $300 billion in investment commitments are expected to support the next phase of expansion. But building a data centre is not as simple as finding a large piece of land and putting up a building. It needs power, water, connectivity, cooling. And, increasingly, access to renewable energy.
Data Centres Fighting For The Same Land?
For developers, a data centre is competing with more familiar assets such as homes, offices, malls and hotels. Binitha Dalal, Founder and Managing Director, Mt K Kapital, said India does not yet have dedicated data centre land in the way it has established land categories for other real estate assets.
"There are no dedicated data centre lands, in India it is still an upcoming asset class," Dalal told NDTV. The question, she said, ultimately comes down to how a developer values the land and its potential use.
A developer has to weigh the returns from a data centre against what the same land could generate as a residential or commercial project. The calculation becomes even more important because data centres need very specific infrastructure.
"They need good and steady connectivity of electricity and water," Dalal said. They also need locations with an established municipal water network and access to a reliable electrical grid.
Data centres cannot simply be pushed to the far outskirts of a city because land is cheaper. They also need skilled teams that can reach and maintain the facilities.
Dalal said that, in many cases, traditional real estate may still offer developers better margins than data centres. "As a pure play, data centres are not something where developers will make the margins they would normally make in any other asset class," she said.
But the picture changes in locations where data centres are already established and approved for hyperscalers. In such markets, land specifically suited for data centres can command a premium over agricultural, residential or commercial use, she added.
This is already creating a new layer to India's land market. The value of land is no longer determined only by what can be built on it. It is also increasingly linked to whether the land can support the infrastructure needed to run a digital economy.
The AI Effect
The biggest force behind this demand may not be traditional cloud computing. It is AI.
Anarock expects the rapid adoption of artificial intelligence to become a major driver of data centre demand. AI workloads need significantly more computing power, storage and energy than many conventional digital services.
But Gurleen Khurana, Co-Founder, SimplifyGenAI, believes there is an important part of the AI story that often gets missed. It is not just about training large AI models. It is about what happens after those models are built.
"We look at India's data centre growth from the consumption side," Khurana told NDTV. "Enterprise AI is moving from pilots to production, and production workloads never switch off," Khurana said.
The distinction between AI training and AI inference is important. Training requires huge amounts of computing power to build models. Inference happens every time an AI model actually responds to a user, creates an image, processes a conversation or makes a decision.
That could mean millions of small workloads running continuously. Khurana said this creates a different infrastructure requirement. Inference needs capacity closer to users. Latency matters. A voice AI system cannot afford to respond too slowly. Energy efficiency matters too because these workloads can run around the clock.
This is why the next phase of India's data centre expansion will not simply be about adding square footage. It will be about building the right kind of capacity in the right places.
Mumbai Leads, But The Map Is Changing
Mumbai-MMR remains India's biggest data centre market. Its financial ecosystem, enterprise demand and connectivity have helped it retain that position.
But the industry is spreading. Chennai, Bengaluru, Delhi-NCR and Hyderabad are emerging as important data centre clusters, supported by land availability, power, connectivity, renewable energy and proximity to large demand centres.
India's installed data centre capacity has already crossed 1.8 GW in the first half of 2026. The development pipeline could take total capacity beyond 6.7 GW by the end of the decade, according to Anarock.
The expansion will touch much more than real estate. It will create demand for power infrastructure, renewable energy, engineering and construction services, fibre connectivity and cooling systems. And AI could make that demand even more intense.
