India's data centre capacity could reach 15 GW in five years on AI demand: L&T Vyoma
As AI adoption expands globally and domestically, L&T Vyoma and Avendus Capital expect India's data centre ecosystem to become an increasingly important part of the country's digital infrastructure landscape.
India's data centre capacity could expand to between 10-gigawatt (GW) and 15 GW over the next five years as artificial intelligence (AI) adoption accelerates and global companies seek alternative infrastructure locations, according to Prashant Jain, CEO & MD of L&T Vyoma.
Jain said demand for AI-focused infrastructure is rising rapidly, while constraints in overseas markets are creating opportunities for India. He expects AI adoption, data localisation requirements and growth in the domestic digital economy to support long-term capacity additions.
L&T Vyoma, the data centre arm of Larsen & Toubro, currently has a land bank capable of supporting about 5.5 GW of capacity and 250 MW of power-ready infrastructure. The company has operationalised a 30 MW data centre in Chennai, started construction of a 35 MW facility in Mumbai and recently announced an AI factory initiative in partnership with Nvidia.
Jain said the company aims to serve hyperscale customers through colocation services while also building AI computing infrastructure and offering AI token as a service as the business scales.
Vaibhav Garg, Director-Infrastructure & Real Assets Investment Banking at Avendus Capital, said India is positioned to benefit from both global AI demand and domestic adoption of AI technologies.
According to Garg, India's data centre infrastructure costs are almost 40 to 50% cheaper than some of the other global geographies, making the country attractive for AI-related investments. He added that more than 70% of Indian companies have already begun incorporating AI into day-to-day operations, a trend that continues to gain momentum.
Garg believes India will benefit from a dual opportunity: providing AI infrastructure to global enterprises while also becoming a major consumer of AI-enabled services.
Garg expects competition in the industry to remain broad-based despite increasing investments from global and domestic players.
"The size of the pie is large enough for everyone to grow," he said, noting that the market is likely to remain fragmented with 15 to 20 meaningful players over time rather than consolidating into a handful of dominant operators.
Jain, however, expects a distinction to emerge between operators of large AI facilities and conventional edge data centres. He said AI-focused facilities require substantial financial commitments and strong execution capabilities.
He said that demand for GPU-powered infrastructure is significantly outpacing supply. He added that many customers are seeking facilities that can be operational within the current year rather than waiting for the 18-24 months typically required to build new data centres.
According to Jain, developing the next wave of AI infrastructure could require investment of roughly $100 billion over the next five years. A 100 MW AI data centre alone may require investment of ₹4,000-5,000 crore, creating high barriers to entry for large-scale projects.
The executives also highlighted emerging opportunities outside traditional data centre hubs. While Mumbai currently accounts for a significant share of installed capacity, rising land costs are encouraging developers to explore alternative locations.
Jain identified Visakhapatnam as a city that could become a major data centre hub over the next five to seven years. He said the city offers suitable access to land, fibre connectivity, water and electricity, making it competitive with established markets.
For the full interview, watch the accompanying video
Catch all the latest updates from the stock market here
Jain said demand for AI-focused infrastructure is rising rapidly, while constraints in overseas markets are creating opportunities for India. He expects AI adoption, data localisation requirements and growth in the domestic digital economy to support long-term capacity additions.
He added that power availability constraints in the US and geopolitical developments are also contributing to demand shifts toward India.
L&T Vyoma, the data centre arm of Larsen & Toubro, currently has a land bank capable of supporting about 5.5 GW of capacity and 250 MW of power-ready infrastructure. The company has operationalised a 30 MW data centre in Chennai, started construction of a 35 MW facility in Mumbai and recently announced an AI factory initiative in partnership with Nvidia.
Jain said the company aims to serve hyperscale customers through colocation services while also building AI computing infrastructure and offering AI token as a service as the business scales.
Vaibhav Garg, Director-Infrastructure & Real Assets Investment Banking at Avendus Capital, said India is positioned to benefit from both global AI demand and domestic adoption of AI technologies.
According to Garg, India's data centre infrastructure costs are almost 40 to 50% cheaper than some of the other global geographies, making the country attractive for AI-related investments. He added that more than 70% of Indian companies have already begun incorporating AI into day-to-day operations, a trend that continues to gain momentum.
Garg believes India will benefit from a dual opportunity: providing AI infrastructure to global enterprises while also becoming a major consumer of AI-enabled services.
Garg expects competition in the industry to remain broad-based despite increasing investments from global and domestic players.
"The size of the pie is large enough for everyone to grow," he said, noting that the market is likely to remain fragmented with 15 to 20 meaningful players over time rather than consolidating into a handful of dominant operators.
Jain, however, expects a distinction to emerge between operators of large AI facilities and conventional edge data centres. He said AI-focused facilities require substantial financial commitments and strong execution capabilities.
He said that demand for GPU-powered infrastructure is significantly outpacing supply. He added that many customers are seeking facilities that can be operational within the current year rather than waiting for the 18-24 months typically required to build new data centres.
According to Jain, developing the next wave of AI infrastructure could require investment of roughly $100 billion over the next five years. A 100 MW AI data centre alone may require investment of ₹4,000-5,000 crore, creating high barriers to entry for large-scale projects.
The executives also highlighted emerging opportunities outside traditional data centre hubs. While Mumbai currently accounts for a significant share of installed capacity, rising land costs are encouraging developers to explore alternative locations.
Jain identified Visakhapatnam as a city that could become a major data centre hub over the next five to seven years. He said the city offers suitable access to land, fibre connectivity, water and electricity, making it competitive with established markets.
For the full interview, watch the accompanying video
Source link
