Why the numbers have collapsed — and why they're not going back up
In the case of solar panels, the drop in cost was about 10% per year over the last 40 years, with an overall 99% decline in costs over that period.
Here's what we'll cover:
And that changes everything
Let's start with the single fact that shapes this whole module: solar and battery costs have collapsed. Not gently. Not temporarily. They have fallen by a factor of ten in the last 15 years, and the economics of electricity generation will not be going back to where they were.
A solar module in 2010 cost around $2.00 per watt. Today the global average is around $0.28, with Chinese spot prices closer to $0.15. A reduction of more than 85 % in 14 years. A lithium-ion battery cell in 2010 cost around $1,200 per kWh. By 2025 a complete, fully built battery storage system was below $120 per kWh globally, and around $75 per kWh in China. Meanwhile, gas turbines cost about the same as they did in 1990, while lead times have stretched to 5–7 years on the back of data-centre demand. Fuel prices continue to cause major global energy crises. (That is the whole energy transition in one paragraph.)
This isn't a temporary dip driven by oversupply. It's the product of compounding manufacturing scale and process improvement. It's the kind of change that's very hard to reverse. Once the factories are built, the engineers trained, and the supply chains optimised, the cost won't go back up, it will keep getting lower.
Before we can compare a solar farm with a gas plant, we need a common yardstick. That yardstick is LCOELevelised Cost of Energy — the industry-standard yardstick. Enables you to compare a 30-year nuclear plant with a 25-year solar farm on a single $/MWh number.: the total lifetime cost of a power plant divided by the total energy it produces over its lifetime. Usually quoted in $/MWh.
LCOE answers a very simple question: if you paid for everything on day one and ran the plant for 25 years, what would each kWh of electricity actually cost?
It lets us compare apples with apples. For example, a solar farm with no fuel bills, against a gas plant on the same scale, with huge ones.
Take a look at the chart below, which plots the LCoE for 5 major sources of electricity.
Look at the solar line. In 2009, it was the most expensive technology on the chart by a wide margin. By 2015, it had crossed below gas. Since 2020, it's the cheapest line on the board, alongside onshore wind. Nuclear has gone the other way; gas is static. That single visual is the story of the last 15 years of electricity economics and it's the reason the rest of this module matters.
Solar power is now the cheapest form of electricity generation. Simple.
So how did solar get so cheap? The answer has two parts: a manufacturing improvement pattern called the learning curve, and the factories that turned that pattern into reality.
Every time the world has doubled the amount of solar PV it produces in a year, the price of a module has fallen by roughly 24%. That relationship, sometimes called Wright's LawNamed after Theodore Wright, who first observed the pattern in aircraft manufacturing in 1936. It applies across solar, batteries, semiconductors, wind turbines and more — the solar-specific version is sometimes called Swanson's Law., has held for over 40 years, through multiple technology generations, multiple dominant manufacturers, and multiple global shocks. It is one of the most consistent patterns in industrial history.
On a log-log chartA chart where both axes increase by a factor of ten per major gridline, rather than by a fixed amount. Useful for showing patterns that span many orders of magnitude., it shows up as a straight line. Use the slider below to see how the learning rate you choose sets the curve and how well it matches the data.
A straight line on a log-log chart over nearly 50 years is not a coincidence. It's what happens when scale, process improvement, materials innovation, and ruthless competition compound over decades. The line isn't abstract, it's a demonstration of manufacturing engineers getting very good at making a relatively simple product.
The learning curve is a pattern, not a cause. The thing that actually moves the curve is each decision, made thousands of times, to build another factory. The cumulative output axis only grows if we keep building new capacity.
In 2010, the world shipped about 18 GW of solar modules. In 2025, that figure was around 700 GW — nearly a 40-fold increase, and the line is still bending upwards. Installed global manufacturing capacity is now over 1.5 TW/year, well ahead of annual demand. That oversupply is uncomfortable for manufacturers but excellent for everyone else, as it keeps costs down.
LONGi Green Energy. Became the world's biggest module maker by industrialising a single innovation — monocrystalline silicon wafers cut with diamond wire saws. Today operates over 150 GW/year of integrated wafer, cell, and module capacity.
JinkoSolar. Built its scale on upgraded cell technology, which now dominate shipments. Ships well over 80 GW/year and has delivered modules to more than 160 countries.
Trina Solar. Pioneer of the 210 mm "ultra-large" wafer format that drove the latest step-change in module power output. Module shipments are now in the 65–75 GW/year range.
Together, these three firms produce roughly half the world's solar modules. All three are vertically integrated from polysilicon to finished panel. None of them were major players 20 years ago.
And it's not just modules. The price of polysiliconThe raw material feeding every solar module. collapsed from around $40/kg in late 2022 to under $7/kg by late 2024 as new Chinese production capacity came online. Polysilicon is roughly 15 % of module cost, so a 5× drop in its price is one of the biggest single reasons modules are now under $0.10/W at the factory gate.
Once a factory is built and running, the cost of the modules produced is affected by materials and power, not the cost of building the factory itself. Stopping production means not making a return on that factory investment. Manufacturers won't do that. They'll keep producing, and — if anything — keep expanding.
The factories are the ratchet. Every time a new one is built, the next doubling of capacity and reduction in costs gets closer.
Batteries are on exactly the same ride as solar, they just started later. The learning rate is slightly lower (~18–20 %), but the absolute decline is just as dramatic.
A lithium-ion cell in 2010 cost around $1,200 per kWh. By 2024 it was below $130 — a 90 % decline in 14 years. Costs fell another ~30 % in 2025 alone: a complete and installed battery storage system is now under $120/kWh globally and around $75/kWh in China, the lowest level on record.
The driver has been the electric vehicle industry: billions of dollars of investment in cell chemistry, gigafactory construction, and supply chains have dragged stationary battery costs down with them.
Wright's Law and the learning curve apply to batteries too. Once again, use the slider below to see how the learning rate you choose matches the data.
As with solar, the learning curve is being pushed along by an extraordinary build-out of capacity. In 2015 global lithium-ion cell production was around 50 GWh/year — roughly enough for half a million electric vehicles. In 2024 it was over 1,100 GWh/year. By 2030, announced capacity exceeds 3,000 GWh/year. That's a 60-fold growth in 15 years.
CATL (China). The world's largest lithium-ion cell producer by a wide margin — roughly one in three cells made globally comes from CATL. Pioneered low-cost, cobalt-free lithium-ion chemistry for stationary storage and EVs. Operates over 500 GWh/year of capacity.
BYD (China). Uniquely combines cell manufacturing with finished EVs and grid-scale BESS. Its "Blade Battery" brought safety and cycle-life to the mass market. Cell output is rapidly approaching CATL's scale.
Tesla (US). Built the "gigafactory" model — enormous single-site facilities combining cell production, pack assembly, and vehicle final assembly. Giga Nevada, Giga Shanghai, and Giga Berlin between them produce hundreds of GWh/year of cells.
Northvolt (Sweden). Europe's flagship attempt at a home-grown gigafactory supply chain. Hit serious financial and production-scaling problems in 2024–25 — a reminder that the learning curve rewards volume, and catching up is hard.
Chinese cell-makers account for roughly 80 % of global production. Cheaper, cobalt-free lithium-ion now dominates stationary storage.
Put everything together and two claims emerge — both of which would have sounded absurd ten years ago.
Solar is the lowest-cost variable source of electricity in most of the world today. Variable means when the sun is up. In sunny regions, new solar is cheaper per kWh than new onshore wind, new gas, new coal, and new nuclear.
Solar + BESS is the lowest-cost firm source for daily firming in those same regions, where capital is reasonably priced. Firm means around the clock. Pair a solar farm with a 4-hour battery and you can deliver electricity through the evening peak at a lower LCOE than a new gas peaker.
The LCOE chart below answers "what does it cost to generate the kWh?" For the dispatchable sources on the chart (gas, coal, and nuclear), that's also broadly the cost to deliver whenever the grid needs it, because the fossil fuel power stations ramp on demand. For the variable sources (solar PV alone, onshore wind, offshore wind) LCOE only counts the kWh produced when the resource is available. Firming them up to a round-the-clock profile means adding storage, and maybe building extra generation too. The extra costs is called the firming premiumThe cost gap between LCOE (just generating) and Firm LCOE (delivering on a flat round-the-clock profile). Falls as battery costs fall., and the total — generation plus firming — is the Firm LCOEProject-level cost of round-the-clock renewable supply. IRENA introduced it in 2026 as a standard benchmark for hybrid solar+storage projects.. The Solar + Storage bar is already a Firm LCOE: solar paired with a 4-hour battery, so it compares like-for-like with gas.
In 2025, firm solar at 95 % reliability ranged from around $30/MWh in the cheapest Chinese projects, to $54–82/MWh across other sunny regions (Brazil, India, Oman, South Africa, Australia), and up to $113/MWh in the United States. By 2030, the mid-range is projected to fall to $44–58/MWh. The firming premium narrows every year as battery costs fall.
In truth, there are places where the chart doesn't yet land this way — high-latitude winters, regions with weak sun, grids with very high peak demand in dark hours. We'll come back to those limits in Module 10. But for the bulk of the world's electricity demand, and for most new capacity being financed right now, the ranking above is what investors are pricing.
The bars above are unsubsidised, new-build LCOE for plants in good resource conditions, expressed in $/MWh and averaged over a 25-year operating life. They blend 2025 mid-points from three industry-standard sources.
Sources: Lazard Levelised Cost of Energy+ v18 (2025); BloombergNEF LCOE Update 2026; IRENA Renewable Power Generation Costs in 2024 (2025) and 24/7 Renewables: The economics of firm solar and wind (2026). All figures are global mid-points; ranges around them are wide.
Solar's installed cost is roughly the same everywhere. Modules and inverters trade on a global market. Installation costs vary a bit. What changes wildly is the cost of financing the project. In emerging markets, the cost of debt for a solar project can be 5–10× higher than in Germany or Australia, and required equity returns can exceed 30%. That single variable, the cost of capital, can double or triple the LCOE.
A solar farm in Lagos buys the same panels as one in Hamburg, but pays roughly 8× more to finance them. Module 9 returns to why.
Source: Climate Policy Initiative 2023, via GSC How to Finance Solar 2025.
A UK household installs a 10 kW solar system costing £8,000 plus a 10 kWh home battery costing £6,000 — a "firm solar" setup that can deliver electricity through the evening as well as during the day. The system has an expected lifetime of 25 years. The UK capacity factor is about 10 %.
Questions:
1. How much energy (in kWh) will the system produce over its lifetime?
2. What is the approximate cost per kWh of firm solar electricity (i.e. the firm-solar LCoE)?
3. UK grid electricity is about 28 p/kWh. Is the firm solar electricity cheaper?
1. 10 kW × 8,760 hours/yr × 10 % capacity factor × 25 years = 219,000 kWh.
2. Lifetime cost divided by lifetime energy:
(Simple version: it ignores battery replacement around year 15, round-trip losses, discounting, and O&M. Accounting for those would lift the figure by very roughly a third, so treat 6.4 p/kWh as a rough floor rather than a precise number.)
3. Yes — by a large margin. At 6.4 p/kWh the firm solar costs roughly one-quarter of grid electricity. The battery also lifts how much of that solar generation you actually self-consume (typically from ~30 % up to 60–70 %), which means more of your kWh displaces 28 p/kWh imports rather than being exported at a low feed-in rate.
Solar and batteries don't just cost less than fossil alternatives. Their cost is different.
A gas or coal plant has two major cost streams: the upfront cost of building it, and the fuel it burns for 25 years. The fuel cost is the bigger of the two, and it is impossible to forecast. Fossil fuel prices swing on geopolitics, weather, and competing demand. A solar or battery project has essentially one cost stream: what you pay on day one. That's it. The fuel is sunlight, and its price is zero.
We call this capex-ledBanking-friendly: the lifetime cost is locked in on the day the financing closes, with no fuel-price surprises in subsequent years. (capex is short for capital expenditure, the money spent upfront to build it), as opposed to fuel-ledLifetime cost is hostage to fuel markets. The 2022 European gas spike turned brand-new gas plants into financial headaches almost overnight.. The distinction sounds minor, but it isn't. Knowing exactly what you will pay for energy in the future removes a major risk for investors and economies. When that energy source is also the cheapest, it reshapes electricity prices for everyone.
Fuel-cost volatility isn't hypothetical. European gas prices in 2022 hit roughly ten times their 2019 level. The pattern returned in 2026, when shipping disruptions in the Strait of Hormuz sent European gas benchmarks back to multi-year highs.
Drag the slider below to see what different gas fuel prices do to a new gas plant's LCOE compared with new firm solar (a solar farm paired with a 4-hour battery, as priced in the previous lesson).
The red slab on the gas bar is fuel. It stays there for 25 years, and its height is set by gas market prices. The gold bar for firm solar is the whole thing, panels and battery, paid up front. Once you've built it, that's it.
Right now, solar is around the same cost as gas generation when fuel costs around $4.5 per MMbtu, but look what happens to the LCOE once the price of gas goes up.
Lenders, governments, and households all prefer costs they can forecast. Firstly, a 25-year fixed-cost solar project is easier to finance than a gas plant whose fuel bill could double. Secondly, and equally importantly, all other energy related investments are easier to finance when running costs are likely to be stable in the long run.
This is not a minor advantage. It rewrites energy economics.
A solar farm built in 2026 is locked at 2026 costs for 25 years. A gas plant built in 2026 will pay whatever the gas market charges in 2035, 2040, 2050 — and no one has a reliable forecast for those numbers.
If you had a choice between a 25-year mortgage with a fixed interest rate and one that reset every week based on geopolitics, which would you take? That's the choice energy market planners are making today, at scale.
Five quick questions. Don't worry if you're unsure — you can retake the quiz as many times as you like.
Solar and battery costs have collapsed by roughly tenfold since 2010. That isn't a cycle — it's the learning curve, pulled along by a huge build-out of Chinese (and increasingly global) manufacturing capacity.
LCOE — total lifetime cost divided by total lifetime energy — is the yardstick that makes these technologies comparable with fossil alternatives. Solar and batteries are capex-led: almost all cost is upfront, fuel is zero, lifetime cost is locked. Gas and coal are fuel-led — 25 years of exposure to prices no one controls.
In most regions today, solar is the lowest-cost variable source and solar + BESS the lowest-cost firm source of new electricity. The factories being built right now make further cost falls close to inevitable.
And modules are no longer the expensive part — install speed, not panel price, now decides who grows fastest. That's where Module 6 picks up.
Next up: Module 6 shifts from why it's cheap to where it's being deployed — the markets, scaling dynamics, and the install-speed pacing factor that decides global solar and storage growth.
You now know why solar and battery costs have collapsed, why that collapse is likely permanent, and why solar + BESS is rewriting the economics of new electricity.
