Can wind and solar power reliably meet electricity demand?
Short answer. Not on their own. Wind- and solar-only systems sized to match annual electricity demand meet roughly 72-91% of hourly demand in most countries, and 83-94% with twelve hours of storage — leaving hundreds of hours of unmet demand each year even in the best cases. The residual gap is set by the geophysics of the atmosphere, so closing it requires storage, demand flexibility, long-distance transmission, or firm generation rather than more wind and solar capacity alone.
Why the question matters
Debates about wind and solar power often conflate two different questions: whether the resource is large enough, and whether it arrives when it is needed. The resource question is settled — the wind and sunlight reaching the Earth's surface exceed human energy demand by orders of magnitude. The timing question is not.
Framing reliability as a geophysical property rather than an engineering shortcoming changes what counts as a solution. If the gap between supply and demand is imposed by the atmosphere, then cheaper solar panels do not close it; only storage, flexibility, geographic aggregation, or complementary firm generation can.
It also changes what counts as a fair comparison between places. Two countries with identical average wind resources can face very different reliability problems, because what matters is not the mean but the depth and duration of the shortfalls.
What our research finds
- Where a system is planned on a short weather record and then operated across decades of real variability, spending on additional storage and additional generation together improved resource adequacy more than spending the same amount on either alone — wind-only additions reached adequacy standards whatever storage was available, while solar-only additions often could not (Reich et al., 2026).
- The regions least exposed to wind droughts combine high wind power density with low seasonal and low weather variability — the American Midwest, Australia, the Sahara, Argentina, Central Asia, and Southern Africa (Antonini et al., 2024).
- Northwestern Europe has high wind power density but suffers more frequent and more prolonged wind droughts than its resource alone would suggest, because its weather is more variable (Antonini et al., 2024).
- There is little evidence of strong trends in wind droughts over recent decades in most places, and in many regions the most severe droughts on record predate any substantial wind power on the system (Antonini et al., 2024).
- A wind, solar, and battery system planned on one year of weather loses load when operated on weather it has not seen; reaching zero lost load across a decade takes nearly 40 years of planning data, or 15 years if dispatchable generation may supply 5 per cent of demand (Ruggles et al., 2024).
- On Oahu, a least-cost wind, solar, battery, and hydrogen system meeting 100% of hourly demand across 14 years of weather would cost $0.1673 per kilowatt-hour, below the $0.2126 to $0.2987 per kilowatt-hour that the island's petroleum-dominated system cost from October 2022 to September 2023 (Covelli et al., 2024).
- The emissions-reduction level at which firm low-carbon generation enters a least-cost system is set mainly by wind quality — countries with annual mean wind capacity factors near 0.4 need far deeper decarbonization before it competes than countries at or below 0.25 (Duan et al., 2022).
- Across 42 countries, the most reliable wind- and solar-only systems, sized so that annual generation equals annual demand and with no storage, met 72-91% of hourly electricity demand; twelve hours of storage raised this to 83-94% (Tong et al., 2021).
- Even in systems meeting more than 90% of demand, hundreds of hours of unmet demand still occurred in a typical year, concentrated in particular periods rather than spread evenly (Tong et al., 2021).
- In the United States, the daily and seasonal variability of solar resources, wind resources, and electricity demand means that deep decarbonization at high reliability requires substantial dispatchable generating capacity (Shaner et al., 2018).
Evidence and methods
Both analyses are built on multi-decadal hourly weather reanalysis data rather than on model scenarios — 39 years covering 1980 to 2018 in the case of the 42-country reliability analysis. Using observed weather means the variability being characterised is the variability that actually occurred.
The reliability analysis deliberately assumes idealised conditions — perfect transmission within each country, and annual generation set equal to annual demand — so that the remaining shortfall is attributable to the geophysical mismatch between resource and demand rather than to any particular system design.
Wind droughts are measured with an energy-deficit metric that integrates both the depth and the duration of each drought, rather than counting low-wind hours. A long shallow drought and a short deep one are not equivalent problems for an electricity system.