The artificial intelligence (AI) push is creating a rapidly expanding source of electricity demand, with data centres expected to play an important role in shaping the country’s energy infrastructure. As AI workloads grow and data-centre campuses expand, this growth is creating opportunities for investment in power generation, transmission, storage, and energy-efficient technologies, while also raising questions about grid reliability, cost, and sustainability.
Anchor For Power Investment
Power demand from data centres in India is expected to rise ninefold over the next decade, according to the assessment discussed during a session on AI opportunities and challenges for the country’s energy sector at BloombergNEF Summit 2026. Independent estimates underline the scale of the expansion, with S&P Global forecasting India’s data-centre electricity consumption to rise from around 13 TWh in 2024 to 57 TWh by 2030, taking its share of national electricity demand to about 2.6 per cent.
Wood Mackenzie has also projected India’s operational data-centre capacity to rise from 2.2 GW in 2025 to 12 GW by 2030, reflecting the rapid expansion expected as AI workloads grow. The firm has also forecast an increase in AI-dedicated capacity over the same period, underlining the growing link between the country’s AI ambitions and its electricity infrastructure.
Deepak Chhabria, President, RMZ Infrastructure, said the sector is currently “infra constrained” rather than demand-constrained, making the expansion of data centres an opportunity to build broader power infrastructure. “Unlike many other countries, India needs power infrastructure, not only for data centres, but to ensure that we have stable, continuous, uninterrupted power for several other applications,” he said.
Chhabria said the sector could create “anchor demand” for new power generation and transmission while strengthening the investment case for companies developing electricity infrastructure. The opportunity extends beyond data centres, with India also seeing rising electricity demand from air conditioning, electric vehicles and industrial electrification.
Transmission And Clean Power
While the rapid addition of generation capacity is important, transmission could become the more immediate constraint as large data-centre campuses come online. Chhabria said states are currently indicating that generation is not the primary problem, with the bigger challenge being the ability to transmit electricity to where demand is emerging.
“Today, the states that we talk to, they say, on the generation side, we have more electrons than we need. But the real bottleneck is on the transmission side,” Chhabria said. He added that data-centre operators and infrastructure developers would need greater predictability around regional demand to enable capacity additions across both generation and transmission.
The scale of individual facilities is also changing the nature of the infrastructure requirement. Manish Pant, Executive Vice President, Schneider Electric, said India’s data-centre capacity is currently around 1.5 GW. Estimates suggest that between 10 GW and 30 GW could be added over the next eight to nine years.
A 1 GW compute load also does not translate into a 1 GW requirement for the entire facility. Pant said additional electricity is required for cooling and other systems, putting the requirement at around 1.2 to 1.3 GW for every 1 GW of power used for computing, depending on the technology and operating conditions.
India’s relatively low-cost renewable power remains a key advantage for its data-centre ambitions, but the cost of delivering reliable clean electricity goes beyond generation. Srinath Iyer, Commercial Head, Asia-Pacific Clean Energy & Power, Google, said transmission, storage and grid management costs would also need to be factored into the overall cost of supplying power to data centres.
“While solar is cheap today, the true cost of a delivered energy solution actually lies if you lay down the full equation of the value chain,” Iyer said. He added that integrating more wind and solar would require greater use of storage and a broader approach to grid planning as electricity demand rises across the economy.
Efficiency Becomes Critical
The growth in data-centre electricity demand is being accompanied by a rapid push to improve the efficiency of computing, cooling and power systems. Higher compute density means operators need to deliver more computing performance without allowing energy and water requirements to rise at the same pace.
“Token per watt” is also expected to become an important measure as the industry focuses on improving AI output from each unit of electricity consumed. At the same time, rising compute density is changing data-centre power architecture, with some chipset providers moving towards systems capable of more than 400 kilowatts and potentially up to 600 kilowatts per rack, requiring new approaches to power distribution and cooling. Pant said the industry would need to improve efficiency across the full data-centre cycle. “We have had up to 30 per cent efficiency improvement across the full data centre from year to year,” he said.
According to the industry, water efficiency is becoming part of the same equation, particularly as data-centre operators expand in water-stressed locations. Water emerged as the biggest non-power constraint in an audience poll during the session. Technologies such as dry-air cooling can reduce water consumption but may require more electricity, creating a need for balancing the two resources.
The challenge for India will ultimately be to expand data-centre infrastructure without treating electricity supply as an isolated requirement. With AI demand rising alongside broader electrification, the ability to expand transmission, integrate storage and renewables, and improve energy efficiency will determine how effectively India can turn its data-centre growth into a long-term advantage for its AI economy.