Power density: why land and space matter: Power density and the transition, with India in view

5.1 The transition reverses the spatial order of energy

For two centuries, industrial energy has flowed from points to areas. A few mines, oil fields and power stations, each occupying a small patch of land, supplied whole regions through rail lines, pipelines and grids. The energy industry was concentrated, and most people never saw it.

Renewable systems turn this around. Energy must be gathered from large areas, where sunlight, wind or crops are collected at a few watts per square metre or less, and then concentrated through transmission lines into cities and industrial zones using tens to thousands of watts per square metre. The energy system becomes visible everywhere: on roofs, across farmland, along highways and on the horizon.

This has practical consequences. Collection equipment covers far more land and requires far more materials per unit of power. Transmission becomes central rather than secondary, because the best solar and wind resources are rarely where demand is concentrated. In India the best solar land lies in Rajasthan and Gujarat and the best wind in a few western and southern states, while large demand centres lie across the Gangetic plain and in the cities of the south and west. Hence the importance of the 'AC transmission' and 'HVDC links' tiles, and of long, contested line corridors.

The biggest consequence concerns liquid fuels. Smil's chapter on biofuels argues that replacing oil with crop fuels at a few tenths of a watt per square metre is a land-hungry illusion. Replacing all US gasoline with corn ethanol would need more cropland than the country has. The much more efficient path is to electrify transport, so that the energy comes from solar panels and wind turbines at 1 to 10 W/m² rather than from crops at 0.1 to 0.4, and an electric motor uses that energy about three times as efficiently as an engine uses fuel.

Smil's conclusion is that renewables can and should grow, but low power densities make the transition slower, more land- and material-intensive and more dependent on transmission than its advocates usually admit.

• Smil, Power Density: A Key to Understanding Energy Sources and Uses (2015) · via Smil Mastery Course, M8 'Power density and the energy transition'
• Smil, Energy Myths and Realities: Bringing Science to the Energy Policy Debate (2010) · via Smil Mastery Course, M8 'Power density and the energy transition'
• Smil, Energy Transitions: History, Requirements, Prospects (2010; 2nd ed. 2017) · via Smil Mastery Course, M8 'Power density and the energy transition'

5.2 India's power growth needs thousands of square kilometres, a small share of its land

India is one of the most crowded large countries, with about 440 people per square kilometre (calculated) and over 40% of its land under crops. Land acquisition is often the slowest and most contested part of any infrastructure project. So it is worth working out what the power density numbers imply.

India used roughly 2,100 TWh of electricity in 2025 (Ember reports growth of 49 TWh, or 2.4%, which implies about 2,090 TWh; calculated), an average flow of about 240 GW. The IEA's Electricity 2026 report expects Indian demand to grow by about 6.4% a year, adding more than 570 TWh by 2030, which is an extra average flow of about 65 GW. If all of that growth came from ground-mounted solar at Lovering's median of 5.7 W/m², it would need about 11,400 km². At the 10 W/m² that good Indian sites can reach, it would need about 6,500 km². That is four to eight times the area of Delhi, but only 0.2 to 0.35% of India's 3.29 million km².

Scaling up to the whole of today's demand from solar alone would need roughly 24,000 to 42,000 km², about 0.7 to 1.3% of the country. For comparison, Lovering and colleagues estimate that the world's entire energy system today occupies about 0.4% of ice-free land, against 30 to 38% for agriculture. Land for solar is thus a real but manageable claim, comparable to the area of a medium-sized district or two per major growth step.

Wind is different. Supplying the same 65 GW from wind at about 1 W/m² would spread over some 65,000 km², about 2% of India, though nearly all of that land would stay in farming or grazing. The direct footprint would be under 1,000 km².

The harder constraints are local. Good solar land in Rajasthan and Gujarat overlaps with grazing commons and with the last habitat of the critically endangered great Indian bustard, a heavy bird easily killed by overhead wires. In April 2021 the Supreme Court restricted overhead lines across about 99,000 km² of the two states; in March 2024 it softened that order, and in December 2025 it accepted much smaller priority areas, about 14,750 km², where lines are to be buried or rerouted and large new plants barred. Developers have also turned to the salt desert of the Rann of Kutch, where the Khavda park is being built for about 30 GW, and to degraded land to avoid such conflicts, which is power density thinking applied in practice.

• IEA, Electricity 2026 (February 2026) · India demand growth to 2030, https://www.iea.org/reports/electricity-2026 (executive summary)
• Lovering, Swain, Blomqvist and Hernandez, Land-use intensity of electricity production and tomorrow's energy landscape, PLOS ONE 17(7): e0270155 (2022) · solar median and energy system share of ice-free land, https://doaj.org/article/58a75c2614134cd1b3372e1a6a930b2d
• Ember, Global Electricity Review 2026 (April 2026) · India 2025 demand, https://ember-energy.org/latest-insights/global-electricity-review-2026/2025-in-review/
• Bar and Bench, 'Underground power lines, ban on new wind turbines: Supreme Court accepts measures to protect Great Indian Bustard' (19 December 2025) · 2021 order, 2024 modification, revised priority areas, https://www.barandbench.com/news/underground-power-lines-ban-on-new-wind-turbines-supreme-court-accepts-measures-to-protect-great-indian-bustard
• IANS, 'India building world's largest 30 GW renewable park in Gujarat' (March 2026) · web

5.3 Built surfaces and shared land can carry much of the load

The cheapest land for solar is land already in use for something else. India has been experimenting with almost every variety.

Rooftops come first. Ember estimates that distributed solar, mostly on roofs, added about 22 TWh to India's generation in 2025. The PM Surya Ghar scheme, launched in February 2024, aims to equip ten million (one crore) homes by March 2027. Commercial and industrial roofs, warehouses, railway stations and airports add more. Rooftop panels claim no new land and generate where the power is used, avoiding transmission. Their limits are the ones described earlier: in dense cores, roof area is small relative to demand, and small systems cost more per watt than large parks.

Water surfaces come next. Gujarat pioneered panels mounted over irrigation canals more than a decade ago, which also reduces evaporation. Floating solar on reservoirs, such as the 600 MW park on the Omkareshwar reservoir in Madhya Pradesh, of which 278 MW was running by early 2025, uses water already impounded for hydro or irrigation and keeps panels cooler, slightly raising their efficiency.

Shared farmland is the third option. Agrivoltaic designs raise panels high enough, or space them widely enough, for crops or grazing to continue beneath. Yields of some shade-tolerant crops hold up well, and in hot, dry climates the shade can reduce water stress. Electricity output per hectare is lower than in a dense park, so it trades some power density for keeping land productive. Farmers can also lease land at the edges of fields, and India's scheme for solar irrigation pumps puts small arrays directly beside wells.

Finally, low-value land. Deserts, salt flats, degraded pasture and old mines have few competing uses. Restoring coal mine land with solar parks is especially neat: the land is already disturbed, often already connected to the grid, and close to the workforce of a declining industry.

None of these removes the basic arithmetic. They change which land carries the panels, and that is often what decides whether a project is welcomed or delayed for years.

• Ember, Global Electricity Review 2026 (April 2026) · India distributed solar, https://ember-energy.org/latest-insights/global-electricity-review-2026/2025-in-review/
• Smil, Power Density: A Key to Understanding Energy Sources and Uses (2015) · via Smil Mastery Course, M8 'Power density and the energy transition'
• Smil, How the World Really Works: A Scientist's Guide to Our Past, Present and Future (2022) · via Smil Mastery Course, M8 'Power density and the energy transition'
• The Tribune, 'Omkareshwar floating solar park showcases scale and ambition of India's clean energy goals' (4 January 2025) · 600 MW, 278 MW commissioned, https://www.tribuneindia.com/news/business/omkareshwar-floating-solar-park-showcases-scale-and-ambition-of-indias-clean-energy-goals-pralhad-joshi
• PIB, PM Surya Ghar: Muft Bijli Yojana factsheet (December 2024) · web

See it in the table

Check your understanding

Roughly what share of India's land would ground-mounted solar need to supply the IEA's expected demand growth to 2030?

Why does Smil favour electrifying transport over crop biofuels?

What is the main trade-off of agrivoltaics?