AI ignites India infra supercycle

Artificial intelligence is triggering a fresh investment boom in India’s data centre sector, with operators and investors rapidly scaling up capacity plans, deploying billions of dollars, to meet surging global demand for AI computing infrastructure.

Industry estimates suggest India’s operational data centre capacity could rise to nearly 4 GW by 2030 from around 1.5-1.7 GW in 2025, up sharply from about 375 MW in 2020, if current expansion plans materialise.

“The big change is that India’s data centre story is no longer being driven only by normal cloud migration,” said Varun Sharma, fund manager at Motilal Oswal Asset Management Company. “Demand is getting broader, the pipeline is getting larger, and AI is increasing the power density and value of future capacity.”


AI-led workloads are rapidly emerging as the key incremental growth driver, already influencing 25-35% of new demand visibility for hyperscale campuses, he said.

India’s cloud market is expected to grow to more than $24 billion by 2028 from about $8.3 billion in 2023, according to Motilal Oswal estimates.

Operators across the sector are accelerating capacity additions. Sunil Gupta, chief executive of Yotta Infrastructure, said the company is fast-tracking its Navi Mumbai campus towards nearly 2 GW over the next five years to meet rising demand for AI infrastructure.

“We are targeting around 90,000 graphic processing units (GPUs) within this financial year and a roadmap to 500,000 GPUs over five years,” Gupta said, adding that demand from global model builders and inference providers has resulted in bookings exceeding six months.

Sify Technologies is also expanding its footprint as enterprises increasingly shift from captive server rooms to colocation and hybrid models.

Sharad Agarwal, chief executive of Sify Infinit Spaces, said the company currently has a designed IT capacity of 192.36 MW across 14 operational data centres in six cities and is developing 12 new facilities focused on AI-ready infrastructure and edge deployments.

The colocation-to-captive capacity mix is expected to shift from 60:40 currently to 80:20 over the next five years as AI adoption increases the need for higher-density computing infrastructure and advanced cooling systems, Agarwal said.

Policy support and data localisation are further strengthening the investment case.

“Data localisation creates a structural, non-negotiable floor of domestic demand,” independent tech policy analyst Subimal Bhattacharjee said.

He said the government’s decision to grant infrastructure status to data centres has improved financing access and investor confidence, while states are increasingly competing to attract investments through incentives, land allocation and lower power tariffs.

Motilal Oswal estimates that India’s data centre sector could generate a hard-capex opportunity of Rs. 55,000-65,000 crore over the next two to three years across power infrastructure, cooling systems, networking and related digital infrastructure, as AI pushes the sector beyond a pure colocation and cloud-led growth story.