India's nuclear power push: Can India really reach 100 GW by 2047? | Green Shift

ELI5 / TLDR

India wants to build a huge fleet of nuclear power plants — 100 gigawatts of them by 2047. The problem, says energy journalist M. Ramesh, is that the government’s own numbers show new nuclear electricity will cost roughly 9 rupees per unit, while clean solar-plus-storage can be had for half that. Nobody wants to buy the expensive stuff, the private companies the government hoped to attract are staying away, and so state-owned firms will end up doing all the work. Realistic outcome: maybe 20 gigawatts, not 100.

The Full Story

In April, India’s Central Electricity Authority held a workshop to walk industry through the new “Shanti Act” — a law that, for the first time, lets private companies into nuclear power. Big names showed up: the state nuclear monopoly NPCIL, the power giant NTPC, and private players like L&T, Tata Power, and Westinghouse. Ramesh read the minutes. His verdict is blunt.

The number that breaks the math

For decades, Indians thought of nuclear power as cheap — around 3 to 3.5 rupees per unit. But those were old, long-paid-off plants. The new ones tell a different story.

“Electricity from the PHWR will cost something like 7 rupees 77 paise. And if it’s a light water reactor, it will cost something like 7 rupees 88 paise.”

(A quick chaperone for the jargon: a reactor’s job is to boil water with a controlled nuclear reaction, and the steam spins a turbine. A pressurized heavy water reactor, or PHWR, is India’s homegrown workhorse design; a light water reactor is the more common global design. The difference for our purposes is just the price tag — both land near 7.8 rupees.)

And that 7.8 is today’s cost. Two things push it higher. The running costs creep up about 10 paise a year. And a nuclear plant takes 12 to 13 years to build, so by the time it switches on you’re adding another rupee-plus. Do the arithmetic and you land near 9 rupees per unit.

“Nuclear power is simply pricing itself out of the market. That’s my biggest fear.”

The killer comparison: firm, round-the-clock renewable energy — solar or wind paired with storage so it doesn’t cut out when clouds roll in (the industry calls this FDRE, firm dispatchable renewable energy) — already sells for about 4 to 4.5 rupees. So a company that wants clean power has a green option at roughly half the price. Who chooses the 9-rupee one?

So why is the government pushing it anyway?

Ramesh can only guess, because the government hasn’t explained itself. His best theory: a nuclear plant lasts 60 to 100 years, and its cost is almost all upfront capital — the fuel and running costs are small. So over the decades, as inflation drags every other electricity price upward, the nuclear tariff barely moves. Eventually it looks cheap by comparison. The catch is the first decade, when someone has to swallow the 9-rupee price with no payoff yet.

His other guess is bundling. NTPC runs cheap coal plants and is building lots of renewables. It could blend a slice of pricey nuclear into that cheaper mix and sell the average to state distribution companies — high enough to hurt, but not 9 rupees.

The thorium prize at the end of the tunnel

There’s a longer-game reason to bother at all. India sits on enormous reserves of thorium, a weakly radioactive metal that can’t fuel a reactor directly. You have to “breed” it — run conventional uranium reactors first, which produce plutonium, then use that plutonium to unlock the thorium cycle. Get there, Ramesh says, and the lights stay on for centuries.

“Once we get to the thorium cycle, thorium alone can energize the entire country for hundreds of years or even millennia.”

But reaching it means running a fleet of expensive reactors for 40, 50, maybe 100 years first. The high-cost plants of today are the toll booth on that road.

The quieter red flags

Ramesh flagged several softer concerns from the minutes:

Shrinking the exclusion zone. The government is mulling cutting the no-habitation buffer around plants from 1 km to 700 or even 500 metres. Ramesh thinks this is foolish — not because of a Chernobyl-scale disaster (at that scale the extra 300 metres is meaningless anyway) but because small radiation leaks do happen, and even a minor incident near homes could torch public trust and derail the whole program. “There is really no need… to shrink the exclusion zone and leave it open for criticism.”
Water. A twin 700 MW plant needs about 80 million litres — roughly double a comparable coal plant. Plants must sit near water, and water is where people already live. An unsolved tension.
FDI is likely coming. The Atomic Energy Commission has recommended letting foreign money in. The percentage is undecided, but Ramesh expects Russia’s Rosatom, already deep in India, to want equity stakes.
A strategic fuel reserve. The government wants to stockpile roughly 20% of a plant’s lifetime fuel need — uranium — the way India stockpiles oil, so a geopolitical shock can’t starve the reactors.
Sites on dead coal plants. Retired thermal stations, which already have grid connections and infrastructure, are the leading candidates for new reactors.

Small modular reactors: lots of buzz, higher cost

The other star of the workshop was the SMR — a small modular reactor. Think of it like the difference between building a house on-site brick by brick versus shipping in factory-made rooms and bolting them together. SMRs are built in a factory, trucked or railed to the site, and assembled fast. India’s wish-list is specific: road-or-rail transportable, grid-connected within 36 months, dry-cooling for water-scarce spots, and crucially, compatible with that thorium fuel cycle.

Ramesh tells a nice anecdote: at a Russian nuclear conference in Sochi a couple of years ago, he couldn’t even squeeze into the SMR session — packed wall to wall — while the other rooms sat empty. The interest is real and growing. The snag is the same as everything else here: SMRs are more expensive per unit than big plants, not less.

The honest bottom line

“We’ll probably add 7,000 or 4,000, maybe 15 gigawatts or 20 gigawatts… I am not convinced that the government will be anywhere near its target of meeting 100 gigawatts by 2047.”

The plants that will get built are the ones already in motion — NPCIL’s ten 700 MW PHWRs, the Kudankulam reactors with Rosatom, some of NTPC’s promised 30 GW. State companies do the heavy lifting; the private excitement the law was meant to spark hasn’t shown up. Even NTPC’s own chairman said as much in the workshop, noting “with concern that the anticipated excitement from the private sector has yet to fully materialize.” Brace for high-cost nuclear, Ramesh says, and don’t get hyped.

Key Takeaways

New Indian nuclear electricity is indicated at ~7.8 rupees/unit today; factoring 12–13 year build times and rising O&M, the real delivered cost lands near 9 rupees.
Firm round-the-clock renewables (FDRE) already cost ~4–4.5 rupees/unit — roughly half of nuclear.
Nuclear’s cost is almost entirely upfront capital; fuel and running costs are minor, so tariffs stay flat for the plant’s 60–100 year life. That’s the long-game bet.
The “Shanti Act” opened nuclear to private players, but private interest has not materialized — confirmed by NTPC’s own chairman.
State-owned NPCIL, NTPC, and NHPC (with Rosatom and other foreign partners) will do nearly all the building.
The strategic reason to persist: a uranium-reactor fleet breeds plutonium, the gateway to India’s vast thorium reserves — potentially centuries of fuel, but 40–100 years away.
Government is considering shrinking the plant exclusion zone (1 km → 700 m → 500 m); Ramesh warns this risks public trust for trivial land savings.
A twin-700 MW plant needs ~80 million litres of water, about double a comparable coal plant — a siting headache.
FDI in nuclear is likely to be approved (level undecided); Rosatom is the probable taker.
A strategic uranium reserve (~20% of lifetime need) is proposed, mirroring oil reserves, to buffer geopolitical supply shocks.
New plants will likely go on retired thermal-plant sites to reuse grid infrastructure.
SMRs (factory-built, fast-deploy reactors) drew huge interest but cost more per unit than large plants.
Realistic 2047 outcome per Ramesh: ~20 GW, not the targeted 100 GW.

Claude’s Take

This is a clean, numbers-first reality check, and its strength is that the pessimism rests entirely on the government’s own indicative tariffs, not the host’s editorializing. The central argument — that ~9-rupee nuclear can’t compete with ~4.5-rupee firm renewables, so demand collapses and only state firms will build — is sound and well-supported, including the rare candor of NTPC’s chairman admitting private interest hasn’t shown.

Where I’d apply the BS filter: the analysis is one expert’s read of meeting minutes, framed as near-certain. A few things get hand-waved. The thorium endgame is invoked as the strategic justification but treated almost mystically (“millennia”) — India’s thorium program has been “40 years away” for half a century, and that deserves more skepticism than it gets here. The bundling theory is explicitly a guess. And the cost story is real but not the whole story — nuclear’s value is firm baseload and grid stability, which a pure price-per-unit comparison undersells; the video brushes past this. The 9-rupee figure also leans on the assumption that India keeps taking 12–13 years to build, which is partly the problem being solved, not a fixed law.

Still, it’s an honest, jargon-light, genuinely informative segment that resists the usual booster narrative. A 7: trustworthy, well-grounded, and refreshingly unhyped, but it’s a single-source opinion piece that could have stress-tested its own optimistic counterarguments harder. Not an 8 because it doesn’t engage seriously with the strongest case for nuclear (firmness, decarbonization at scale, energy security beyond the fuel reserve).

India's nuclear power push: Can India really reach 100 GW by 2047? | Green Shift | M Ramesh