Global Residential Electricity Prices: The Demand Side of Solar Economics

Price is the other half of solar economics

Every self-consumed kilowatt-hour a solar system produces is worth exactly the grid price it displaces. The annual value of an installed kilowatt of panels is therefore the product of two numbers: the specific yield it generates (kilowatt-hours per kilowatt-peak per year) and the retail price of the electricity it offsets (US dollars per kilowatt-hour). Yield is set by sunlight, module technology and temperature. Price is set by a national tariff, a network charge and a subsidy regime that have nothing to do with the sun. The two carry equal weight in the result, and Solar Analytica has until now held only the supply side.

This report is the demand side. It maps residential retail electricity prices across 92 cities and 70 countries from 2016 to 2025, then joins the 2025 prices to modelled specific yield for the markets the two datasets share, to state what a kilowatt of panels is worth per year in each. The finding that matters most for screening is that the ranking by value looks nothing like the ranking by sunlight, and that the lowest headline prices are the ones most likely to mislead.

The 2025 price map

Residential retail prices in 2025 span a factor of roughly sixty, from about 0.02 US dollars per kilowatt-hour in the most heavily subsidised markets to about 0.45 in the most heavily taxed. Aggregated to region, the order is wide and stable:

Solar Analytica compilation of the city-level retail prices in the sources below, 2025. Prices are household rates inclusive of taxes, levies and network charges, converted to US dollars at prevailing exchange rates. Ordered by regional average.

Western Europe carries the highest regional average at 0.319 US dollars per kilowatt-hour, with Dublin (0.451), Berlin (0.422) and London (0.419) at the top of the single-city table. South Asia (0.064) and MENA (0.078) sit at the bottom. The high end is set by taxes, network charges and, in 2022 and 2023, an energy-price shock. The low end is set almost entirely by subsidy, which is where the screening problem begins.

The value of a kilowatt: where price meets sunlight

Joining the 2025 prices to modelled specific yield inverts the picture. Across the 23 markets that appear in both this dataset and Solar Analytica's yield baseline, the annual grid-offset value of one kilowatt-peak of TOPCon modules, price multiplied by yield, spans roughly 41 to 545 US dollars. A selection across that range, ordered by value:

Modelled annual grid-offset value of 1 kWp = 2025 retail price × modelled TOPCon specific yield. Prices per the sources below; yield per Solar Analytica's modelled specific-yield baseline (NASA MERRA-2 1994-2024 and CERES EBAF). Values rounded to the nearest US dollar; Perth and Nairobi both sit at about 377.

The most useful line in the table is the pairing of Berlin and Cairo. Cairo collects more sunlight than Berlin, a modelled 1,769 against 1,058 kilowatt-hours per kilowatt-peak per year, and yet a kilowatt of panels is worth about 461 US dollars a year in Berlin and about 41 in Cairo, a ratio of roughly eleven to one. Nothing about the sun explains that gap. It is the tariff: Berlin's household electricity costs about 0.44 US dollars per kilowatt-hour, Cairo's about 0.02, held there by an Egyptian lifeline subsidy. A screen that ranked these two markets on irradiance would put them in the wrong order by an order of magnitude. The full price times yield join is explorable market by market in The Value of a Kilowatt.

Perth sits at about 377 US dollars per kilowatt per year. Its price, at 0.210 US dollars per kilowatt-hour, is mid-pack among Australian capitals, below Adelaide, Sydney and Brisbane. Its modelled yield, at 1,791 kilowatt-hours per kilowatt-peak, is near the top of the whole set. The regulated Synergy tariff and the Western Australian solar resource combine to place it among the higher-value markets in the table without either input being the highest on its own, which is the pattern the value join exists to surface.

The subsidy trap

The largest error available to a cross-market screen is to read a low US-dollar price as a weak solar market. Several of the lowest headline prices in the dataset, Cairo (0.022), Addis Ababa (0.006), Kuwait City (0.045), Riyadh (0.053), Tashkent (0.049), Budapest (0.128) and Kyiv (0.096), are held below cost-reflective levels by state subsidy, lifeline tiers or price caps rather than by any structural cheapness of supply. Two things follow. First, the low price is a policy choice, not a market equilibrium, and it can move sharply when the policy changes: Romania's price cap expired in mid-2025 and Chile unwound a price freeze across 2024 and 2025, both driving steep increases. Second, the same subsidy that depresses the retail price also depresses the modelled solar value, so a subsidised market can screen as uneconomic on the day the dataset is compiled and become economic the day the subsidy is withdrawn.

A second distortion runs alongside the first. In markets with sharply depreciating currencies, Nigeria (0.038), Egypt (0.022), Turkey (0.060), Ethiopia (0.006) and Argentina (0.107), the US-dollar price is compressed by exchange-rate movement even where the local-currency bill is rising quickly for households. Three markets, Caracas, Yangon and Kinshasa, carry no reliable local-currency rate at all, reflecting hyperinflation or non-publication. These figures are recorded and flagged in the dataset rather than smoothed, because the flag is the finding: a low US-dollar price in a devaluing currency is not a durable indication of the cost of electricity.

Tariff structure, not just the headline rate

Two markets at the same headline price can support very different storage economics depending on how the tariff is shaped. A flat rate can be offset but not arbitraged. A time-of-use, dynamic or spot-linked tariff lets a battery move energy from a cheap window into an expensive one and capture the spread on top of solar self-consumption. The dataset records tariff structure alongside price for exactly this reason. Nordic and continental spot markets (Copenhagen, Stockholm, Helsinki, Oslo), regulated time-of-use regimes (Madrid's PVPC, Paris's Tarif Bleu, Rome, Zurich) and the time-of-use-common utilities of the United States and Australia all reward storage in a way a single-rate tariff at the same average price does not. Structure should be read next to the headline number, not after it.

The decade of movement, 2016 to 2025

The continuous series in this dataset, the markets with a source running the full ten years, show two patterns. European retail prices stepped up sharply through the 2022 and 2023 energy shock and then partly retreated: Berlin's household rate ran from about 0.33 US dollars per kilowatt-hour in 2016 to a 2023 peak near 0.45 before easing to about 0.44 in 2025, and Rome, London and Brussels trace the same peak-and-retreat shape. Regulated tariffs outside Europe moved more slowly and in one direction: Perth's Synergy rate rose from 0.265 to 0.324 Australian dollars per kilowatt-hour across the decade, a steady climb rather than a spike. The distinction matters for payback modelling, because a market whose price spiked and retreated carries different forward risk to one on a slow regulated climb, even where the two sit at the same price today.

Provenance is uneven across the decade and the dataset says so. Prices for many developing markets in the 2016 to 2019 window rest on a World Bank commercial-connection proxy, flagged low confidence, with a gap through 2020 to 2024 and a fresh 2025 reading from GlobalPetrolPrices; those series are left discontinuous rather than interpolated. The OECD-market series (EIA for the United States, Eurostat for the European Union, the IEA and the Australian regulators, Synergy for Perth) are continuous and carry higher confidence. Of the 920 city-year cells in the dataset, 665 carry a US-dollar price. The remaining gaps are left empty rather than filled by assumption.

What this dataset is, and is not

This is a screening layer as of mid-2026, not a substitute for the current published tariff schedule of a named retailer. Three limits matter for anyone applying the numbers directly. First, national figures stand in for the city where no city-specific series exists (Shanghai, Osaka, Sao Paulo, Ho Chi Minh City and Wellington among them), and where a major city diverges materially from its national average, such as Con Edison in New York or Meralco in Manila, that is noted in the underlying data. Second, the North American cities use state-average residential rates as a proxy, and the Canadian rates are quoted at 1,000 kilowatt-hours per month excluding tax, so they understate the all-in delivered price. Third, the Australian figures are usage rates excluding the daily supply charge. Every figure traces to a named source, and where a value is distorted by subsidy or by currency, the dataset flags it rather than resolving it by assumption.