Getting close to 24/365 solar generation is cost-competitive in sunny regions

We used hourly solar radiation data from 12 cities across the world with diverse geographical conditions to evaluate how close they can get to 24/365 solar generation with the same solar-plus-battery configuration.

3.1 Sunny cities can get very close to 24/365 solar generation

In sunny cities like Las Vegas or Muscat, 24/365 solar generation is within reach, with cloudy days only occasionally providing small shortfalls. The analysis explores the limits of what solar and batteries can do, independent of the grid or other back-up power.

3.1.1 Las Vegas can get 97% of the way to 24/365 solar generation

Ember’s modelling shows 24/365 solar generation is possible in a sunny city like Las Vegas. 6 GW of solar panels paired with 17 GWh of battery capacity can deliver 1GW of continuous solar electricity for 97% of the year. This model assumes a utility-scale project, moving from the kW scale explored until now to GW scale. It also assumes more solar capacity to reach higher reliability (6 GW for 1 GW target supply vs 5 kW for 1 kW average supply previously).

Some years are sunnier than others, leading to varying reliability. The average reliability across 2005-2023 was 97%, with the lowest year (2010) reaching 95% and the best year (2007) achieving 99%. The graphic below shows the days of 2023 which had an average of 96%.

In 2023, there were 301 days when the system delivered a full 1 GW every hour; only four days had a shortfall of more than 12 hours. The lowest performing day, January 4, saw just 0.13 GW of electricity delivered across the entire day – equivalent to three hours of full 1 GW power over a 24-hour period.

Batteries help by carrying excess electricity from the previous day. However, this buffer can help bridge only small gaps, usually a couple of cloudy days, before the battery reaches its minimum state of charge. Once that point is reached, a shortfall for several hours can occur. To maintain 100% reliability during extended cloudy periods, additional long-duration storage, a back-up generator or grid electricity is needed, without building more solar and battery.

3.1.2 The sunniest cities can get more than 90% of the way to 24/365 solar generation

We tested the same solar and battery configuration of 6 GW solar and 17 GWh battery in 12 cities across the world.

Most cities analysed showed some seasonality with the biggest shortfalls happening during stretches of consecutive cloudy days in winter. Even Las Vegas gets only 9 hours and 21 minutes of daylight on its shortest day of the year – 2 hours 39 minutes less than the average 12-hour day.

In Hyderabad, supply shortfalls were driven not by winter, but by cloudy monsoon days. Madrid struggled in November and December, due to shorter, cloudier days in winter. Meanwhile in Wuhan, cloud cover was evenly scattered throughout the year – but on many days, it was dense enough to cut solar generation to very low levels. Birmingham barely had two good months of reliable sunshine. Even near the equator, like in Abuja, solar is not guaranteed year-round. Extended cloudy periods still push 100% reliability just out of reach.

Overall across the year, Muscat in Oman comes closest to 24/365 solar generation, at 99% of the way – beating Las Vegas’s 97%. Six out of the 12 cities we analysed achieved over 90%: Mexico City in Mexico (96%), Johannesburg in South Africa (95%), Manila in the Philippines (92%) and Abuja in Nigeria (92%). Four other cities achieved between 80-90% reliability: Hyderabad in India (89%), Madrid in Spain (88%), Buenos Aires in Argentina (81%) and Washington D.C in the US (81%). Only two cities fell short — Wuhan in China with persistent, dense cloud cover (74%) and Birmingham with short, dark winter days (62%).

The level of reliability achieved in the sunniest cities, like 97% in Las Vegas, is remarkably high and it is possible to imagine industrial users taking 97% without even needing a grid connection – relying just on solar and batteries and managing electricity use during the few cloudy days, or supplementing with small back-up generators.

Even Birmingham in the United Kingdom reached 62% of the way to 24/365 solar generation. While it is not imaginable for solar and batteries alone to power users in this city independently, it is high enough to play a substantive role in local generation alongside wind and other electricity sources.

3.2 How close to 24/365 solar generation is optimal?

If solar and battery capacity is sufficiently oversized, then it is of course possible to reach 100% of the way to 24/365 solar generation. But doing so is likely to be uneconomic — especially due to the high cost of batteries needed to store electricity across multiple days. The cheapest solar electricity comes without any battery at all, but that only provides daytime electricity – a long way from 24/365 solar generation.

In this section, we explore the trade-offs between different levels of solar and battery capacity and how these impact the Levelised Cost of Electricity (LCOE). We used global average prices (see Methodology) to reflect a typical international case and compare what price different cities can achieve based on their local solar resources.

In a sunny city like Las Vegas, standalone solar electricity costs $41/MWh (see mark 1 on graphic). While that is a very low price, it only delivers electricity for only 21% of the hours needed for full 24/365 solar generation. If low cost is the goal, and grid electricity is available for the rest of the time, then a battery is not optimal.

However, by adding some batteries and some more solar panels, it is possible to quickly get to 60% of the way to 1 GW of 24/365 electricity. This set-up based on 3 GW solar and 7 GWh of storage can generate enough electricity to cover much of the high value evening and morning hours and will increase the LCOE cost to $75 /MWh.

To reach 97% of the way to 24/365 solar generation in Las Vegas (as detailed in the previous chapter), 6 GW of solar panels and 17 GWh of battery capacity is needed. This takes solar generation way past high-value hours only, providing the same 1 GW power every hour of every day. The LCOE cost rises to $104 per MWh — $63/MWh more than standalone solar — but delivers nearly five times more reliability, moving from 21% to 97%.

Taken to the extreme, it is possible to get even further — reaching 99.4% of the way to 24/365 solar generation by continuing to overbuild solar and battery capacity (see mark 3 on graphic). This would need 7 GW of solar and 35 GWh of battery storage, raising the LCOE to $167/MWh. Compared to 97% configuration, the additional 2.4% of the year’s electricity would cost $2700/MWh more — making it clear why 97% is likely to be more optimal than 99.4%.

This shows that ultra-sunny places can economically come very close to 24/365 solar electricity. While the LCOE costs above are based on solar radiation in Las Vegas, the results are very similar for cities like Muscat, Mexico City and Johannesburg.

In moderately sunny places, solar and batteries can economically deliver 60-90% of the way to constant electricity every hour of every day. In sunny cities like Madrid and Abuja, extended cloudy periods mean storing electricity across multiple days becomes necessary to go above 90% of the way to 24/365 solar generation, something which isn’t currently cost-effective.

The answer to “what is optimal” depends on the location and value of the shortfall generation and the cost of the alternative solution to fill this shortfall. This could come from load shifting or load curtailment, grid electricity or onsite generation. For off-grid use cases, it could be gas generation – which may be grid or onsite, and maybe from existing gas power plants or new ones. While aiming for nearly 100% 24/365 solar generation may be feasible and desirable in the sunniest regions, in most sunny areas, the optimal and practical role for solar-plus-battery is to likely deliver 60-90% of the way to 24/365 electricity.

3.3 The LCOE of near 24/365 solar generation are falling fast – 22% in the last year alone

It has not always been economical to get close to 100% of the way to 24/365 generation. But that has changed rapidly. In the last year alone, LCOE prices have fallen by 22%, driven by a 40% fall in battery prices.

Getting 97% of the way at $104/MWh in Las Vegas is based on 2024 global average solar module and battery prices, a 22% fall from 2023.

In 2024, average global battery prices dropped to $165/MWh, down 40% from $275/MWh in 2023. The impact on overall system costs is striking: in 2023, battery capex accounted for $61/MWh, almost half (46%) of the total LCOE. Just a year later, in 2024, that contribution dropped to $37/MWh or 36% of the total.

While $165/MWh reflects the 2024 global average, actual battery prices can be far lower. In early 2025, tenders for large-scale battery storage projects in Tabuk and Hail, Saudi Arabia, reported prices as low as $72/kWh — less than half the global average. These record-low prices highlight the growing potential for even more affordable clean electricity.

3.4 And now 97% of the way to 24/365 solar generation is cheaper coal and nuclear in sunny places

The most recent cost declines and longer battery lifetimes have dramatically shifted the economics of solar generation with batteries compared to other sources of electricity.

As of 2024, combining solar and battery storage to meet 97% of round-the-clock electricity demand in sunny places ($104/MWh) is cheaper than new coal ($118/MWh) and nuclear ($182/MWh), according to Lazard’s latest LCOE data. This marks a dramatic reversal from 2019, when 97% uninterrupted solar was still more expensive than fossil and nuclear alternatives. In some countries – notably China and India – new coal can be significantly cheaper than the US-focused price of Lazard.

Even natural gas is being challenged. While the LCOE of new combined cycle gas stands at $76/MWh in 2024, 60% of the way to 24/365 solar electricity (using 3GW of solar and 7GWh of battery) costs $75 – a whisker less than new-build gas. This gas LCOE assumes relatively cheap fuel price, as Lazard’s figures are based on US-focused data. In other regions dependent on imported gas, the LCOE of gas could be more than double the US level.