As India’s AI economy fires up demand for grid-based electricity, the Ministry of Power has nearly doubled its estimate of additional data-centre load to 26.3 GW by 2031-32 from an earlier 13.56 GW, prompting policy and energy experts on Wednesday to call for demand-response regulations and green tariffs that integrate these power-hungry facilities more closely with renewable energy and the grid.

About 17 GW of connectivity applications have already been filed with State Transmission Utilities, while the additional data-centre load could help push India’s peak electricity demand beyond 388 GW by 2031-32. Rather than meeting the emerging peak simply by adding generation capacity, experts want data centres to become more flexible power consumers as renewable generation expands.

TURNING AI’S POWER HUNGER INTO A GRID TOOL
“Data centres have significant potential to provide demand-response services to the electricity grid, particularly as the share of renewable energy increases and demand becomes more variable,” Vishal Tripathi, consultant at the Council on Energy, Environment & Water, said at a Policy Consensus Centre webinar.

“Instead of building additional generation capacity simply to meet peak demand, demand response can become an important part of the solution,” he said.

The idea effectively turns part of AI’s power problem on its head. Data centres would remain huge electricity consumers, but demand response could make their consumption more responsive to grid conditions, particularly as intermittent solar and wind power account for a greater share of electricity supply.

Tripathi said States need regulatory frameworks to unlock demand response. Distribution companies also need green tariffs aligned with evolving corporate renewable-energy procurement and Scope 2 accounting requirements.

Such tariffs could give data centres greater access to renewable electricity while creating an additional revenue opportunity for utilities.

A $30 BILLION COMPUTING BUILD-OUT
The stakes are rising quickly. India’s data-centre capacity is projected to reach 6-6.5 GW by 2030, while another industry estimate puts the development pipeline through the end of the decade at 3,860 MW, requiring nearly $29.9 billion in capital.

Cloud computing, digital services and increasingly artificial intelligence are driving that expansion, tying India’s computing ambitions ever more closely to its electricity and renewable-energy build-out.

The challenge extends beyond generating enough electricity. Where the next wave of facilities is built will determine the pressure it places on power, water, land and other infrastructure.

MAPPING WHERE AI CAPACITY GOES NEXT NASSCOM called for States to create integrated district- and cluster-level resource maps covering power, water, land, fibre, roads and other infrastructure, rather than keeping such information fragmented across government departments.

“We need comprehensive information integrated at the cluster and district level in the form of resource maps,’” said Sudipto Banerjee, Deputy Director, Public Policy, at NASSCOM.

He said the maps should be matched with a database tracking data-centre projects from announcement to commissioning, allowing investors, operators and governments to assess an area’s carrying capacity before deciding where facilities should be built.

The proposals bring together three infrastructure build-outs that have largely been viewed separately: AI computing, renewable-energy generation and the electricity grid.

Source: Business Line