‘We can’t afford the electrical energy invoice’: Sridhar Vembu warns AI may overload India’s energy grid
What if the way forward for AI meant rolling blackouts in your house? Zoho founder Sridhar Vembu says that’s not science fiction, it’s a looming disaster as data-hungry AI fashions drive energy payments sky-high and stretch electrical grids to the brink.
In a stark warning on X, Vembu pointed to a 60% bounce in electrical energy prices since 2023 in Athens, Georgia—blaming the spike on AI-driven information facilities.
“Even when we may afford all of the GPUs (not!), we can’t afford the electrical energy invoice,” he wrote. “We can’t afford to harm households and factories.”
Vembu referred to as right this moment’s AI “terribly vitality inefficient” and mentioned India should rethink the way it builds digital infrastructure.
“We want vastly extra vitality environment friendly AI,” he mentioned, calling for a basic shift in how AI is computed.
India’s information heart capability is projected to balloon from 1.2 GW in 2024 to as a lot as 5 GW by 2030.
AI workloads may eat 40–50 terawatt-hours yearly by then—straining an influence system that’s already below strain to satisfy surging industrial and residential demand.
Whilst India ramps up its complete era capability to 777 GW, consultants warn that any shortfall in renewable vitality targets may create regional deficits, particularly in AI-heavy zones. If clear vitality progress slips, the nation may face a 6 TWh annual hole—with the danger of blackouts throughout heatwaves or excessive AI exercise.
Native grids might not be geared up to deal with the concentrated demand of knowledge facilities, elevating the specter of regional outages that hit properties and factories alike.
