Blitz India Business
NEW DELHI: India’s nuclear fleet is 8.78 gigawatts. The target announced this morning is 100 by 2047 — and the demand case for it was made by two industries, not by the climate.
The Prime Minister set a target of 100 GW of nuclear generating capacity by 2047, and justified it against a specific source of demand: the electricity appetite of data centres and semiconductor fabrication plants. India’s installed nuclear capacity stood at 8.78 GW at the end of June, within a non-fossil fleet of 297.36 GW — which is to say nuclear is under 3 per cent of the country’s clean capacity today. Reaching 100 GW is therefore roughly an eleven-fold expansion over twenty-one years, and it is the intermediate step rather than the headline that will indicate whether the trajectory is real.
The commercial logic is about firmness rather than carbon. Solar has grown from about 2 GW to some 160 GW in twelve years and is now the cheapest electricity India has ever generated — but it is unavailable after sunset, precisely when Indian demand peaks. A semiconductor fab cannot tolerate an interruption: a wafer lot in process is scrapped if power fails mid-step, which is why fabs are located where grids do not blink. Data centres carry the same requirement written into commercial contracts. An economy that wants to host both must buy dispatchable capacity, and nuclear is the only large low-carbon source that supplies it around the clock.
Firmness is the product: a fab or a data centre buys guaranteed availability rather than cheap kilowatt-hours — the requirement that nuclear capacity is being built to meet.
Solar sells cheap electricity. A fab is not buying cheap electricity. It is buying electricity that never stops.
At a Glance
• Target: 100 GW of nuclear capacity by 2047
• Installed nuclear capacity, 30 June 2026: 8.78 GW
• Implied expansion: roughly eleven-fold over twenty-one years
• Non-fossil capacity, 30 June 2026: 297.36 GW
• Nuclear share of that: under 3 per cent
• Solar capacity: from about 2 GW to about 160 GW in twelve years
• Stated demand case: data centres and semiconductor fabrication
• Why firmness matters: a fab losing power scraps the wafers in process
• Related target: 500 GW of non-fossil capacity by 2030
An eleven-fold build over two decades implies a construction cadence India has never sustained, and the honest way to present that is as the central execution risk rather than as a reason for scepticism. Large conventional reactors take the better part of a decade from first concrete to first power, which means a 2047 target requires orders placed well inside the 2030s. The financing model matters as much as the engineering one: a capital structure that loads the entire construction risk onto a single public balance sheet limits how many units can be under way at once, which is the constraint that policy on private participation and on small modular designs is intended to relax.
The constructive path runs through three parallel tracks rather than one. Fuel security has to be settled alongside capacity, which is why the fast breeder programme milestone this year matters more commercially than it appears. Supply chain has to be developed early — heavy forgings, pressure vessels, control systems and the qualified welding capacity behind them are long-lead items that a domestic industrial base can supply if it is given visibility of an order pipeline. And transmission has to be planned to the load, since firm power stranded behind a congested corridor is not firm at the point of use. Get those three moving together and 100 GW becomes a schedule. Treat it as a single announcement and it remains an aspiration.