India's Pumped Storage Is Routing Around the Himalaya

The Central Electricity Authority has identified 345.38 GW of pumped storage potential in India, across 320 project sites. The regional split is the part worth reading twice.

Maharashtra alone accounts for 67.36 GW, Gujarat 33.36 GW and Odisha 41.53 GW (The Tribune (opens in a new tab), Drishti IAS (opens in a new tab)).

Add the Northern and North Eastern regions together and you get 64.50 GW, which is 18.7% of the national total. The Western Region on its own holds 37.6%. India is preparing to build its largest grid-storage asset class, and the mountains that hold the country's hydraulic head are not where most of it is going.

That is a surprising result in a country whose hydropower story has always been a Himalayan story. It has a straightforward technical cause.

Off-stream changed the geography

Pumped storage works by moving water between two reservoirs at different elevations, pumping when power is cheap and generating when it is dear. The older on-stream design puts those reservoirs on a river. The newer off-stream, or closed-loop, design does not: it needs two basins, a height difference between them and enough water to fill the system once, with top-ups for evaporation and seepage.

Of the 345.38 GW identified, 288.71 GW across 258 sites is off-stream. That is 83.6% of the total. Only 56.67 GW across 62 sites is on-stream.

Once the design no longer requires a river, the siting question changes completely. A closed-loop project wants a reliable height difference, stable ground, a grid connection and proximity to the load it is balancing. The Deccan plateau, the Western Ghats and the Eastern Ghats supply all four. They have escarpments, abandoned quarries and existing reservoirs, they sit next to India's industrial demand, and they sit next to the solar capacity whose middle-of-the-day surplus is the thing being stored.

The Himalaya has far more elevation than any of them. Elevation stopped being the scarce input.

What the build-out actually looks like now

India had roughly 7.18 GW of pumped storage installed at the end of 2025, fourth in the world behind China at close to 66 GW, Japan at around 22 GW and the United States at approximately 19 GW, against a global total of 159.8 GW (IRENA, reported by The Tribune (opens in a new tab)). The CEA put operational capacity at 7.43 GW across ten projects as of August 2026, with 15.87 GW across eleven projects under construction or implementation.

The target is a different order of magnitude. The CEA envisages 100.81 GW of additions between 2025-26 and 2035-36, taking cumulative capacity to about 105.56 GW. For the nearer horizon it projects a total storage requirement of 73.93 GW by 2031-32, of which 26.69 GW is pumped storage and 47.24 GW is batteries.

Two things follow. The first is that pumped storage is not competing with batteries so much as dividing the work with them: batteries take the short, fast duty, pumped storage takes the long duty that batteries price badly. The second is that a fourteen-fold increase in a decade will be built where it can be permitted and financed quickly, and that is a bias towards sites with short clearance paths and short transmission runs.

A policy lever is already pointing the same way. Pumped storage projects awarded for construction by 30 June 2028 receive a full waiver of interstate transmission charges for 25 years from commercial operation. A waiver on a deadline rewards whoever can reach award soonest, which again favours the site with the fewer unknowns.

The caveat that makes the gap wider, not narrower

One correction is owed to the table above. The Northern Region is a grid region, not a mountain one. It contains Uttar Pradesh, Rajasthan, Haryana, Punjab and Delhi alongside Himachal Pradesh, Uttarakhand, Jammu and Kashmir and Ladakh. Uttar Pradesh and Himachal Pradesh are reported as its two leaders.

So the genuinely Himalayan share of the 50.32 GW is smaller than the regional figure suggests, and the honest reading of the national picture is that the Indian Himalaya holds rather less than the 18.7% that a region-level count implies. The direction of the finding holds either way, and the caveat strengthens it.

What the Himalaya still has

None of this makes the Himalayan opportunity small. It makes it specific.

Three things remain true of the mountains and are not true of the alternatives. Head is one: a closed-loop project with 600 metres of drop moves far more energy per cubic metre than one with 200, which cuts reservoir size and therefore land take for the same storage. Existing infrastructure is another: Himachal Pradesh and Uttarakhand already have operating hydropower cascades, access roads, substations and evacuation capacity, and an off-stream scheme built alongside an existing reservoir inherits a great deal of that. The third is the one the national grid will eventually have to price. India's solar build-out is concentrated in the west and south, and a grid that stores its surplus only where it generates it has bought correlation rather than diversity.

There is also a defensible argument that the Himalayan sites being counted are the wrong ones. The identified potential reflects where surveys have looked, and survey effort in the Indian Himalaya has historically followed rivers, because that is where on-stream hydropower was. A closed-loop survey of high-altitude plateau terrain, disused quarries and existing reservoir pairs has not been done at the same density. The 14.18 GW figure for the North East in particular reads like a measure of survey coverage as much as of geology.

The condition nobody has priced yet

The reason capital is cautious about Himalayan storage is not elevation and not distance. It is that the hazard record on these valleys has become impossible to ignore, and nobody has yet turned it into a number a lender can use.

HEF's Himalayan Hydropower Hazard Tracker records seven events since 2013 that damaged hydropower in the Indian and Nepali Himalaya, and fourteen plants on a single Nepali corridor, seven of which were damaged twice in fourteen months. A closed-loop scheme is materially less exposed than a run-of-river plant, because it does not sit in the main channel and does not depend on river flow to operate. That distinction is real and it is not currently reflected in how the risk is discussed.

This is where the opportunity sits. If an off-stream project above the flood path can be shown, with mapped upstream hazard, to carry a different risk profile from the run-of-river plants that have been damaged, then it should attract different terms. Nobody has done that mapping at the scale that would let an insurer or a lender act on it. Until somebody does, every Himalayan storage proposal is priced against the worst headline from the valley it sits in.

What would move the share

A larger Himalayan share of the 100 GW does not require the rest of India to build less. It requires four things that are inside the reach of state governments and the CEA between now and the 2028 award deadline.

A closed-loop survey of the Indian Himalaya on its own terms, looking for reservoir pairs and plateau basins rather than river sites, would establish whether the current figure reflects geology or survey history. Pairing new upper reservoirs with existing lower ones at operating hydropower stations would shorten both the clearance path and the transmission run, which are the two things the award deadline rewards. Publishing upstream hazard assessments per site, in the form a lender can underwrite, would let low-exposure sites separate themselves from high-exposure ones instead of sharing a reputation. And a transmission plan that treats Himalayan storage as a diversification asset for a solar fleet concentrated two thousand kilometres away would value it for what it does for the national grid rather than for what it does for its own state.

The 345 GW figure is not a verdict on the mountains. It is a snapshot of where India has looked, with a technology whose siting rules changed faster than the survey did. The useful question for the next decade is not why the Himalayan share is small. It is how much of that share is geology and how much is a map that has not been redrawn.