Harnessing the Winds

Introduction

Solar and batteries cover the daily cycle. Our earlier analysis, Harnessing the Sun, found they can already serve a flat load for most hours of the year in most places: with the atlas's standard builds, they reach 80% uptime for 99.9% of people. What they struggle with is the rest of the year, above all the dark winters of the high latitudes. That is where wind comes in.

We added wind to the Solar + Wind + Battery Atlas. The model divides the world, without Antarctica and Greenland, into 4,465 regions and asks, in each, what it takes to serve a flat 1 GW load from local resources alone: up to 20 GW of solar, up to 8 GW of onshore or offshore wind and up to 36 GWh of batteries per GW of load, run through every hour of ten weather years, 2015–2024. For each target uptime – the share of the load's energy the build delivers, averaged over the ten years – the model finds the cheapest build that gets there.

It is a sceptic's test: no grid imports, no demand flexibility, no hydro and no backup. Costs are the atlas's local costs at 2026 prices: each country's own capital costs for solar, batteries and wind, and its own cost of capital, a median of 7.5% weighted by demand. Finance counts as much as physics: the same build costs more where money costs more.

The comparison is like for like. Wind counts as the cheaper route only where solar and batteries can also reach the target and the build with wind costs less. Where the standard builds fall short, the model also tries much larger solar and battery builds – up to 200 GW of solar and 4,608 GWh of batteries per GW of load – with the ten-year record repeating end to end, so a battery starts the record with what it held at its end. Builds this oversized are a thought experiment: the comparison is about cost, not feasibility.

The map shows the capacity factor each region's onshore wind simulates, set by the windiest fifth of its eligible land. Grey regions have no eligible land for a wind farm.

Adding wind lowers the cost for places that use 71% of the world's electricity

At 90% uptime, adding wind lowers the cost of round-the-clock clean power for places that use 71% of the world's electricity, and by more than $15/MWh for 47%.¹ Weighted by people, the share is smaller: 38% of the world's people gain, 21% by more than $15/MWh, with a median saving of $17/MWh where wind helps. The gap lies mostly in the tropics, home to many people who use little electricity today.

Solar and batteries alone can reach 90% uptime everywhere once the model allows oversized builds, so this is a comparison of cost, not of feasibility.

Where both can reach 90%, wind cuts the demand-weighted median cost from $112 to $92/MWh, or 18%. The cut is 22% at 50% uptime, falls to about a sixth between 70% and 90% and rises to 28% at 99% uptime, where solar and batteries alone need ever more storage.

Latitude matters more. Between 0° and 15°, wind changes almost nothing. North of 55°, it cuts the cost of 90% uptime by 64%, from $299 to $109/MWh. At 99% uptime, solar and batteries alone cost $650/MWh there; with wind, $202.

For 62% of the world's people, solar and batteries alone are the cheapest route to 90% uptime. About a third of them – one in five people (22%) – live where even the windiest fifth of the land would run a turbine at under 15% of its capacity: the Indo-Gangetic plain from Pakistan to Bangladesh, Indonesia, the Sichuan basin, southern Nigeria, Uganda and the Congo basin. These are low-wind places, not gaps in the data. India's 29 such regions have a median wind speed of 4.7 m/s at 100 m, and their median patch of eligible land would run at 10% of its capacity.

Weighted by people, the like-for-like cut at 90% uptime is 10%.

Onshore wind does nearly all of the work. Without offshore wind, wind is still the cheaper route for 71% of demand against solar and batteries of any size.

On the map, dark green marks regions where wind saves little or nothing at 90% uptime, and lighter shades larger savings.

The wind resource is vast

On conservative assumptions – capacity factors scaled to those of new wind farms, 3 MW per km² of eligible land and only land that still reaches a 20% capacity factor – the world's technical wind potential is at least 19 times its electricity demand in 2025 (31,820 TWh): about 620 PWh a year, of which 554 PWh is onshore and 64 PWh offshore (fixed-bottom). On the atlas's default assumptions, 5 MW per km² and each region's good site, it is 38 times: about 1,220 PWh a year, of which 1,157 PWh is onshore and 68 PWh offshore. Floating offshore wind adds about 150 PWh on either set of assumptions.

The resource is widely spread. North of 45°N, wind potential exceeds local demand in regions home to 96% of people on the default assumptions; in the tropics, between 61% and 78%. Worldwide the share is 65% on the conservative assumptions and 78% on the default ones.

Sunlight is more abundant still: the atlas's technical solar potential is 2,647 to 3,848 PWh a year. Wind's case is not size. It is timing and latitude.

Technical potential is what could physically fit, not what will be built. The atlas's eligible land excludes no forest or farmland, and its offshore area excludes no shipping lanes, fishing grounds or sea ice.

The map shows each region's wind potential as a multiple of its own electricity demand. Grey marks the regions where it falls short.

In Europe, wind fills solar's winter gap

The atlas's published results are annual, so they cannot show the seasons. Observed output can. Ember's monthly data for national power fleets show solar and wind peaking in opposite halves of the year.

In Germany, solar delivered 12% of its annual output in the four months from November to February, and wind 44% (2023–2025 average). An even spread would put a third in those months. Across nine European countries, from Sweden to Spain, wind delivered 39–44% of its output in those four months and solar 7–19%. Month by month, Germany's solar and wind shares have a correlation of −0.95.

The pattern is common, not universal. Of the 62 countries with complete monthly data for both sources, 47 show them moving in opposite directions. But the sample leans on Europe (34 of the 62) and says little about the tropics. In India, wind is strongest in the wet season, when solar dips: the complement follows the monsoon, not the winter.

These are national fleets, and capacity added during a year tilts its shares towards later months. They point to why wind keeps costs down in the north.

The atlas's own wind sites tell the same story. Across Europe, the demand-weighted median region's onshore wind farm delivered 42% of its 2024 output in November–February, and regions that use 92% of the electricity in the map's view get more than a third.

The map shows the share of each region's simulated onshore wind output in November–February 2024, the only year with daily data in the atlas. Green marks more than a third; grey, less.

The north needs wind to keep costs down

North of 50°N, the standard solar and battery builds – up to 20 GW of solar and 36 GWh of batteries per GW of load – reach 90% uptime for 84% of people. The share falls to 52% north of 55°N and to 0% north of 65°N. Where the standard builds fall short, solar and batteries alone get there by oversizing solar and batteries far beyond the atlas's sliders. In 574 of its 1,043 regions, the cheapest solar and battery build that gets there lies beyond the sliders; the median such build has 60 MW of solar and 145 MWh of batteries per MW of load. Across the band, solar and batteries alone cost a median $180/MWh – about three-quarters more than the cost with wind ($103). The builds beyond the sliders also lean on a weather record that repeats.

Add wind and 90% uptime is within reach for everyone there, at a median $103/MWh. That is above the world median with wind, weighted the same way by people ($87).

At 80% uptime north of 45°N, wind cuts the median cost by 29%, from $126 to $90/MWh. The chart names a city from the atlas in each band, from Tromsø to Buenos Aires.

In the United Kingdom, Germany, Poland, the Netherlands and Sweden, wind cuts the cost of 80% uptime by 29–47% against solar and batteries alone. With wind, the median cost of 90% uptime is $96/MWh in the United Kingdom, $100 in Germany and $86 in Spain, against $94 in the United States. Onshore wind is enough: without offshore, the median cost with wind is unchanged in every latitude band.

Two cautions. On the atlas's global costs – the same capital costs and cost of capital everywhere – 90% uptime north of 50°N would cost a median $75/MWh with wind: much of the north's cost is finance and local prices, not weather. And ten weather years, 2015–2024, cannot test the longest winter wind droughts.

The map zooms to Europe: lighter shades mark larger savings from wind.

Round-the-clock clean power keeps getting cheaper

Wind turbines are manufactured, and their costs follow a learning curve. IRENA finds that onshore wind's installed cost falls 25% for every doubling of global capacity: 33% in Europe, 31% in the United States and 13% in China, which started from a low base. Solar's rate is higher still, at 34%.

The global average cost of onshore wind power fell 71% between 2010 and 2025, from $114 to $33/MWh; offshore wind fell 63%. IRENA expects onshore installed costs to fall a further 20% by 2035, though more slowly than before.

The atlas's local cost paths assume steeper falls still. Weighted by demand, the median onshore capital cost drops 27% by 2035, from $1,616/kW in 2026 to $1,173 in 2035 and $770 in 2050; solar's from $779/kW to $370, and batteries' from $128/kWh to $65. Each country keeps today's cost of capital.

As solar, batteries and wind get cheaper, so does power at high uptime. Weighted by demand, the cheapest build for 99% uptime falls from $133/MWh in 2026 to $67 in 2050, and for 90% uptime from $92 to $46: the cost of 90% uptime with wind halves by 2050. North of 55°, round-the-clock power falls from $202 to $106/MWh.

Wind keeps its place, though not everywhere. North of 30°, solar, batteries and wind together cost less than solar and batteries alone in every year to 2050; between 15° and 30°, the median gain shrinks to nothing by 2050 as solar and batteries get cheaper. Worldwide and weighted by demand, wind cuts the median cost by 9–18% depending on the year, where both can reach 90% uptime. Wind stays in the cheapest build for most demand to 2050: as the cheaper route for 61–71% of demand, against solar and batteries of any size. North of 55°, with wind, 90% uptime costs $58/MWh in 2050, against $145 without.

That is not every grid. The low-wind tropics stay solar and battery systems in every year. And wind's advantage narrows to 2035, because the atlas's local solar and battery costs fall faster than wind's until then. The paths are the atlas's assumptions, not forecasts.

Wind power itself gets cheaper everywhere. At each region's simulated site, on local costs, the demand-weighted median onshore wind farm produces power at $65/MWh in 2026 and $39 in 2050.

The map shows the cost of power from an onshore wind farm in each region, on local costs: move the slider from 2026 to 2050, or press Play.

Conclusion

Solar and batteries cover the daily cycle. Wind adds what they lack at high latitudes. Together they make round-the-clock clean power cheaper at 90% uptime for places that use 71% of the world's electricity. In the north, wind does the job at less than three-fifths of the cost of solar and batteries alone.

The low-wind tropics are the exception: there, solar and batteries do the job without help.

Elsewhere, planners who want clean power at high uptime should build wind beside solar and batteries, not instead of them. Wind is still getting cheaper.