India data centres to deploy 700,000 GPUs in five years

Avendus Capital has published a report estimating that India's data centres could deploy 650,000 to 700,000 GPUs over the next five years, creating a USD $23 billion investment opportunity.

Rising use of artificial intelligence, along with cloud and broader digital demand, could increase India's built data centre capacity from 1.6 GW in 2025 to about 5 GW by 2030, a compound annual growth rate of 26%.

Developers already have an active pipeline of more than 3 GW, including about 1 GW of AI data centre capacity. Avendus estimates this build-out will require nearly USD $25 billion of capital investment over the next five years.

Mumbai is expected to remain the country's largest data centre hub, accounting for nearly half of India's installed and upcoming capacity over that period.

AI demand

The forecast adds to evidence that AI workloads are reshaping infrastructure plans in India as companies seek more computing capacity for model training and deployment. Growth in AI demand is arriving alongside continued expansion in cloud services and digital applications, broadening demand for operators and investors.

Avendus identified GPU infrastructure as a particularly attractive part of the market. At current capital expenditure and pricing levels, large-scale GPU deployments can deliver equity internal rates of return of more than 28% on a hold-to-maturity basis, according to the report.

Those economics have helped draw investors into the sector. Dual demand from AI and cloud workloads has already generated USD $5 billion of transaction activity over the past three years, backed by global institutional investors, infrastructure funds, and strategic operators.

Private market activity in data centres has also remained strong globally, with transactions completed at EBITDA multiples of 20 to 30 times. The report adds that REITs and InvITs are receiving closer attention as structures for recycling capital because of the sector's long-term contracts and stable cash flows.

These projections are also underpinned by a broader AI build-out in India. Avendus expects the country's AI market to grow from USD $13 billion in 2025 to USD $131 billion by 2032, representing a compound annual growth rate of about 39%.

Public markets

The report suggests listed markets could become a larger source of funding for expansion, with public markets and other strategic transactions likely to play a bigger role in financing India's data centre growth.

Vaibhav Garg, Director, Infrastructure & Real Assets Investment Banking at Avendus Capital and chief author of the report, linked that funding outlook to the volume of demand forming in the market.

"AI adoption is emerging as a significant catalyst for next-generation infrastructure investments in data centers, alongside sustained demand from cloud and digital workloads. This dual demand trajectory has already translated into USD 5 bn of transaction activity over the last three years, with backing from global institutional investors, infrastructure funds, and strategic operators. Going forward, we also expect public markets and other strategic transactions to play a key role in funding India's data centre growth, with 3-4 IPOs expected in the next three years," said Vaibhav Garg, Director, Infrastructure & Real Assets Investment Banking at Avendus Capital.

The report also highlighted the role of government-backed demand in the early stages of AI infrastructure deployment. More than 38,000 GPUs have already been committed under the IndiaAI Mission.

Of those, more than 22,000 have been allocated to AI workloads and are due to be deployed in the near term. Avendus says that commitment has created substantial demand for high-density, liquid-cooled, and AI-ready data centre capacity across the country.

The figures underline how quickly technical requirements are shifting in the market. AI-focused facilities often need different cooling systems, higher power densities, and more specialised layouts than conventional data centres, raising both investment needs and the potential returns for operators able to supply that infrastructure.

For investors, the combination of a large development pipeline, concentrated capacity in major hubs such as Mumbai, and rising AI-specific workloads points to a market moving beyond traditional colocation demand toward more specialised computing infrastructure.