Which Energy Should Nepal Use?

You met the sources of energy in Grade 9: renewable ones like sunlight, water, wind and biomass, and non-renewable ones like coal, oil and gas Grade 9: Grade 9: Sources of energy. You also learned how a hydropower station works Grade 9: Grade 9: Hydroelectricity, and the advantages of each source, one at a time.

Here is how much of each Nepal really uses. Count all the energy used in the country in a year, for cooking, heating, transport, factories and electricity, and in 2022/23 it came from:

Source	Share of all the energy Nepal used (2022/23)
Firewood, crop waste and dung (traditional biomass)	64%
Imported petrol, diesel, LPG, kerosene and coal	26%
Electricity from the national grid	7%
Solar, biogas and micro hydro off the grid	3%

Hydropower is Nepal’s pride, but grid electricity was only about one fourteenth of all the energy used. Most families in villages still cook on wood, and almost all transport runs on imported oil. So the question is not only “how should Nepal make electricity?” but also “what should run on electricity instead of wood and oil?”

When a family chooses a stove, it thinks about more than one thing: what the stove costs, what the fuel costs each month, whether it smokes, whether the fuel can be bought when it is needed. A country choosing its energy does the same, on a bigger scale. The things it cares about are called its criteria. For energy they usually include:

• Money: what it costs to build, and what it costs to run, for each unit of energy.
• Pollution and climate: smoke in the air people breathe, and the greenhouse gases that warm the planet.
• Land and people: the land it covers, the forests cut, the families who have to move.
• Risk: what can go wrong, and how badly: a flood, a landslide, a fire.
• Reliability: whether it gives energy when it is needed: in winter, in the evening, during a border blockade.
• Self-reliance: whether the country controls it, or depends on another country.

No source is best at all of them. The cheapest may be the dirtiest; the cleanest may fail in winter. Giving up some of one thing to get more of another is a trade-off. Setting out every option’s costs and benefits, as far as possible in numbers, and comparing them, is called a cost–benefit analysis.

Engineers also separate two kinds of requirement. Some are constraints, which a solution must meet or it is rejected: “the lights must stay on on a winter evening”. Others are criteria to do as well as possible on: “as cheap as possible”. A diesel generator could keep every light on in winter, but at a price so high that no one would choose it.

Electricity is sold by the unit, which is one kilowatt-hour (kWh): the energy a 1,000 W heater uses in one hour. (1 kWh = 1,000 W × 3,600 s = 3.6 million joules.) To compare a hydropower station with a solar farm fairly, you need the cost of each unit each one makes, not the cost of building it.

Two things decide it. The first is what the station costs, spread over the years: to build it, a company or the government borrows money and pays it back with interest. A loan at about 10% interest, paid back over 25 to 50 years, costs roughly a tenth of the amount borrowed every year. On top of that come the running costs: staff, repairs, fuel if it burns any.

The second is how much electricity it really makes. A 1 MW station working at full power all year would make 1 MW × 8,760 hours = 8,760 MWh. None does. A run-of-river station makes less when the river is low; a solar panel makes nothing at night. The share of that maximum a station really makes is its capacity factor:

$$\text{capacity factor} = \frac{\text{energy made in a year}}{\text{rated power} \times 8{,}760\ \text{h}}$$

Then

$$\text{cost per unit} = \frac{\text{yearly cost (loan + running)}}{\text{units made in a year}}$$

This cost per unit, worked out over a station’s whole life, is what engineers call the levelized cost of electricity.

Now compare a diesel generator, which is what a factory, a hotel or a mobile-phone tower runs when the power is off. A generator makes about 3.5 kWh from each liter of diesel, and diesel cost about Rs 195 a liter in Kathmandu in mid-2026. The fuel alone costs Rs 195 ÷ 3.5 = Rs 56 a unit, more than ten times the cost from a river or the sun, before the generator is paid for.

Two stations can make electricity at the same price and still be very different choices. Here are the other costs, measured.

Carbon dioxide

Hydropower and solar panels burn nothing, but making them does give out greenhouse gases: the cement for a dam, the steel and glass for panels, the trucks. The fair comparison counts everything over a station’s whole life, from mining its materials to taking it apart: its life-cycle emissions.

Coal gives out about 35 times as much as hydropower for each unit. So when Nepal sells hydroelectricity to India in the monsoon, it can replace coal power there; when it buys from India in winter, it takes in electricity with about 30 times the carbon of its own.

Land and people

Every source needs land. A run-of-river plant needs little: a small dam to turn the water into a tunnel, a powerhouse, a road. A storage plant needs a reservoir, a valley filled with water. The biggest planned in Nepal, the Budhi Gandaki project on the border of Gorkha and Dhading, would make 1,200 MW and hold back a lake of about 50 km². Its reservoir would cover the homes of about 3,560 households completely and affect more than 8,000 (as of 2026 it has not been built: most of the land has been paid for, and the government plans to start building in 2028, which would take about eight years).

Solar farms need land too, about 2 hectares (20,000 m²) for each MW. Figure 43.2 compares the two for the same electricity in a year.

Surprisingly, they are about the same size. But the land is of a different kind. The reservoir floods a river valley: its farms, villages, forest and temples. Solar panels can go on dry, stony land, along canals, or on land already used, like the roofs of houses, schools and offices, where they need no new land at all.

Risk

Nepal’s rivers are steep, young and dangerous. On the night of 8 July 2025, a lake on a glacier in Tibet burst and sent a flood down the Bhotekoshi and Trishuli rivers in Rasuwa. It swept away the Friendship Bridge to China at Rasuwagadhi, destroyed the headworks of the 111 MW Rasuwagadhi hydropower plant, and damaged several others: about 200 MW of power was lost in one night. Thirteen months later, on 26 August 2026, an even bigger flood from Tibet came down the same valley. It killed hundreds of people, and the Nepal Electricity Authority reported about 430 MW of power stations cut off from the grid, among them Rasuwagadhi again, swept away completely. A hydropower plant in a Himalayan valley has to be designed for floods, landslides and earthquakes, and that adds to its cost.

Every source has risks of its own. A solar farm’s risks are small: a fire, a hailstorm. A dam that fails can drown the towns below it, so big dams are built to far stricter standards. Imported fuel has a different risk: that the supply stops, or the price jumps, for reasons outside the country, as in 2015.

Nepal’s rivers are fed by the monsoon and by melting snow. From June to October they are full; by February they have shrunk to a fraction of that. Most of Nepal’s hydropower is run-of-river: it has no big reservoir, so it can only use the water flowing past each day. In the dry season, run-of-river plants often make less than 40% of their rated power, while people use as much electricity as ever, more on cold evenings.

You can see what that is worth in the prices. For years the Nepal Electricity Authority has paid run-of-river plants Rs 8.40 a unit in the dry season and Rs 4.80 in the wet season. A unit in winter is worth almost twice as much, because it is scarce.

That is why Nepal both sells and buys electricity. In the year 2025/26 it sold electricity worth about Rs 29.3 billion to India and Bangladesh, nearly all in the monsoon, and bought electricity worth about Rs 10.6 billion from India, nearly all in the dry season.

Spring in Nepal is sunny and dry, and solar panels make most just when the rivers make least. So solar and run-of-river hydropower suit each other. But look at what the model can’t show: these are monthly averages. Nepal uses most electricity on winter evenings, between about 6 and 9 pm, after the sun has set. Solar panels cover the day; the evening needs water stored in a reservoir, batteries charged by day, or electricity from India. Batteries are getting cheaper every year, so solar plus batteries is becoming another way to cover the evening.
The model also keeps Nepal’s use fixed at 2025’s. In fact Nepal uses more electricity every year, as more people cook on induction and drive electric cars. A dam like Budhi Gandaki, if building starts in 2028, would not be finished until the mid-2030s, and by then the dry-season gap it was meant to fill will have grown. One dam alone will not close it.

The same reasoning works for a single family. Which is cheaper for cooking: LPG or an induction stove? The trick is to compare the energy that reaches the pot. An LPG flame heats the air around the pot as well, and only about half of the gas’s energy gets into the food. An induction stove heats the pot itself, and about 85% of the electricity’s energy gets in. That share is the stove’s efficiency.

An induction stove costs about Rs 5,000 to Rs 6,000, and the family may need new flat-bottomed steel pots, because induction only heats pots that a magnet sticks to. Is it worth it? Two simple measures help:

• The payback time is how long the savings take to pay back the extra cost: extra cost ÷ saving per year.
• The benefit–cost ratio is the total benefit over some years divided by the total cost. If it is more than 1, the benefits are bigger than the costs.

Now put the evidence together. Here is what the earlier sections found, for four ways Nepal could get more electricity:

	Run-of-river hydro	Storage dam	Solar farms	Buying from India
Cost per unit	about Rs 4–8	about Rs 13 (Question a)	about Rs 4–6	the market price in India, dearer at peak hours
CO₂ (g per kWh)	about 24	about 24, more if the reservoir floods forest that rots	about 48	about 730 (mostly coal)
Land and people	little land; the river below the dam is left with less water	floods a valley: about 50 km² and 3,560 homes for Budhi Gandaki	about 2 ha per MW; none on roofs	none in Nepal
Winter and evenings	falls to about a third in winter	any month, any hour	sunny winter days; nothing at night	if India has power to spare and the lines are free
Self-reliance	Nepal’s own	Nepal’s own	Nepal’s sunshine; the panels are imported	depends on another country

How do you choose? One way engineers use is a decision matrix. Give each option a score for each criterion, from the evidence. Then decide how much each criterion matters, its weight, and add up the weighted scores.

The scores are where science comes in: measured costs, emissions, areas, river flows. The weights are not science. They are values: how much do you care about the families in the valley compared with the lights on a winter evening? Different people will honestly choose different weights, and should be able to see what each choice of weights leads to.

Try it. With every criterion weighted equally, run-of-river hydro comes first and solar a close second. If winter and evenings matter most, the storage dam wins. Only if using no land in Nepal matters far more than anything else does buying from India come first. Notice which options stay near the top whatever you choose: they are the safe choices. And notice that the real answer is usually not one option but a mix: run-of-river and solar for cheap clean units, some storage for winter evenings, and trade with neighbors for the rest.

Every power station, road and house is built of materials dug from the ground. The questions are the same as for energy: which source, at what cost, and can it last?

Sand and gravel

Concrete is about seven-tenths sand and gravel, by volume. The world now uses about 50 billion tonnes of sand and gravel a year, about 17 kg for every person every day, more than any other solid material we take from the Earth. In Nepal nearly all of it comes from three places:

• Riverbeds. Rivers bring sand and gravel down from the mountains, already washed and sorted by size. It is the cheapest source: dig it out, load it on a truck.
• The Chure hills (the Siwalik range), the lowest hills above the Terai. They are made of young, weak rock and loose gravel, easy to dig and easy to wash away. Their slopes soak up the rain that feeds the ground water of the Terai below them.
• Quarries in hard rock, crushed to size in crushers. The rock lasts longer as a resource, but it costs more to blast, crush and carry.

Is river sand renewable? Partly. Each monsoon, floods bring new sand and gravel down into every stretch of a river. That is the replenishment. If people take out no more than that, the river refills the pits and the riverbed stays where it was. If they take out more, the bed sinks a little more every year.

A sinking riverbed does harm far from the pit. The water table beside the river sinks with it, so wells and springs dry up. The fast water digs under the foundations of bridges and embankments. In Kathmandu, sand was dug from the Bagmati until, in the late 1980s, the bridge at Thapathali between Kathmandu and Lalitpur collapsed; the government then banned digging near bridges. Nepal’s rules today ban digging in the Chure hills and in the monsoon months, and keep it away from bridges, highways and settlements, but illegal digging is common.

Metals: use them again

Aluminium, copper and steel are made from ores dug from the ground: Nepal has very little of any that is worth mining (you met Nepal’s minerals in Grade 8: Minerals in Nepal). But a metal is not used up when a pot or a wire wears out. It can be melted and made into something new, again and again, and that takes far less energy than making it from ore:

Metal	Energy saved by making it from scrap instead of ore (roughly)
Aluminium	95%
Copper	85%
Steel	60–75%

Recycling also leaves the ore in the ground for later, and keeps mines, waste heaps and smelter smoke away. In Nepal, the kabadi collector who buys old cans, wires and pots by the kilo is a recycling industry: the scrap goes on to furnaces that melt it down.

The cheapest and cleanest unit of energy, and kilogram of metal, is often the one you don’t need. A 10 W LED gives as much light as a 60 W filament bulb. A well-insulated house needs less heating. Using less is part of every resource plan.

• How Much Electricity Can a Small Stream Make? Work out the power of a stream from its drop and its flow, and design a micro-hydro scheme.
• Does the Fix Actually Work? How to test whether a solution to an environmental problem really does what it promises.
• What Will the Climate Be Like When You Are Grown Up? What the carbon from the world’s energy choices will do to the climate, and to Nepal’s glaciers and rivers.