Cost of electricity by source
Based on Wikipedia: Cost of electricity by source
The number on your electric bill is a fiction. It is a single, rounded figure that suggests the cost of electricity is a static commodity, like a gallon of milk or a pound of apples, but the reality behind the kilowatt-hour is a chaotic storm of capital investment, geological luck, political subsidies, and invisible societal debts. When we speak of the "cost" of energy, we are often speaking in riddles, using metrics that can be twisted to make a coal plant look cheap or a solar farm look expensive depending on which variables are hidden in the shadows. To understand where our power comes from and what it truly costs us, we must dismantle the accounting tricks used by utilities and governments and look at the raw, unvarnished math of how electricity is born, distributed, and consumed.
Different methods of generating electricity incur a dizzying array of expenses that generally fall into three distinct buckets: wholesale costs, retail costs, and external costs. The wholesale cost is the price tag paid by utilities to acquire and distribute power. This includes the massive upfront capital required to build a plant, the ongoing operations and maintenance (O&M) needed to keep it running, the transmission lines that carry the juice across hundreds of miles, and finally, the grim bill for decommissioning the facility when its life ends. Depending on the local regulatory environment, some or all of these wholesale costs are passed through to you, the consumer, as your retail cost. But there is a third category, often ignored in quarterly earnings reports: external costs, or externalities. These are the damages imposed on society—the smog that clogs lungs, the water heated and returned to rivers killing fish, the carbon dioxide driving climate change—that are not reflected in the price of the megawatt-hour but are paid for by everyone anyway.
Calculating these costs is a statistical minefield, yet it is the primary tool governments use to make energy policy decisions. The industry standard metric is the Levelized Cost of Electricity (LCOE). On paper, LCOE sounds like an objective truth: it attempts to compare different generation methods on a consistent playing field by calculating the net present value of all costs over the lifetime of an asset divided by the total discounted energy output. It represents the minimum constant price at which electricity must be sold for the project to break even.
But LCOE is not a photograph; it is a painting, and like any painting, it reflects the choices of the artist. The calculation requires assumptions about the value of non-financial costs—environmental impacts, local availability, grid reliability—and these assumptions are where the controversy lies. On average, the LCOE for utility-scale solar power and onshore wind has dropped so dramatically that it is now less than the cost of coal and gas-fired stations in many regions. Yet, this average is a dangerous abstraction. The cost varies wildly by location, sunlight hours, wind patterns, and local labor markets. A solar farm in the Sahara is not economically equivalent to one in Seattle.
Roughly calculated, LCOE is the net present value of all costs over the lifetime of the asset divided by an appropriately discounted total of the energy output from that asset.
This metric becomes even more problematic when we introduce storage. The Levelized Cost of Storage (LCOS) is analogous to LCOE but applied to batteries and other storage technologies. However, a fundamental error in many public debates is treating storage as a primary source of electricity. It is not. Storage is secondary; it cannot generate power on its own. It relies entirely on a primary source like wind or solar to fill the bucket before releasing the water later. Therefore, a true cost accounting demands that we include the costs of both the primary generator and the secondary storage when comparing them to fossil fuels that generate in real-time to meet demand.
There is a hidden tax in this equation: efficiency losses. Every time you put energy into a battery and take it out, you lose some of it as heat due to inherent inefficiencies. Furthermore, if the primary source charging that battery is not 100% carbon-free—say, natural gas peaker plants running while wind isn't blowing—the storage operation actually generates emissions. A comprehensive 2015 study in the United States found that net system CO2 emissions resulting from storage operations were nontrivial compared to real-time generation. Depending on the location and operational mode, these emissions ranged from 104 to 407 kg of CO2 per megawatt-hour of delivered energy. The battery is not a magic zero-emission box; it is part of a complex chain where the carbon intensity of the grid matters profoundly.
To address the shortcomings of LCOE, economists developed the Levelized Avoided Cost of Energy (LACE). This metric asks a different question: how much value does this specific source provide to the grid? It considers dispatchability and the existing energy mix in a region. In 2014, the US Energy Information Administration (EIA) recommended that non-dispatchable sources like wind and solar be compared not to the LCOE of fossil fuels, but to their LACE. The logic is sound: a wind farm that only blows at night might not avoid the capital costs of a gas plant needed during the day.
The ratio of LACE to LCOE is known as the value-cost ratio. If LACE (the value) is greater than LCOE (the cost), the ratio exceeds 1, and the project is deemed economically feasible. This introduces a layer of nuance that simple "cost per kilowatt" figures miss entirely. The International Energy Agency took this further with the Value-Adjusted Levelized Cost of Electricity (VALCOE). VALCOE includes both the cost of generation and its value to the system at specific times.
Consider the concept of the "capture rate." This is the volume-weighted average market price a source receives divided by the time-weighted average price for electricity over a period. A hydroelectric dam can be incredibly valuable because it can release water only when prices are highest during peak demand, potentially achieving a capture rate of 200%. Conversely, a wind farm without batteries generates power whenever the wind blows, often flooding the grid at night or on weekends when prices are lowest, resulting in a capture rate under 100%.
This creates a perverse economic feedback loop. The more of a single type of renewable energy built in a pricing area, the lower the capture rate becomes for that specific source. If you build too many wind farms in Great Britain, and they all generate simultaneously on a windy Tuesday afternoon, the price of electricity collapses. This is the "cannibalization" effect. In extreme cases, grid connectivity issues can lead to curtailment, where operators pay wind farms to stop producing because there are no wires to carry the power from Scotland to consumers in England. The capture rate then fails to reflect the true cost or value of the technology, distorting investment signals.
When we peel back the layers of these metrics, we find that internal cost factors vary wildly between technologies. Capital costs—the price to build—are low for gas and oil plants, moderate for onshore wind and solar PV, higher for coal, and highest still for waste-to-energy, wave, tidal, solar thermal, offshore wind, and nuclear. Fuel tells a different story. Fossil fuels and biomass have high fuel costs that fluctuate unpredictably due to geopolitics; nuclear has low fuel costs; renewables have zero fuel costs. But the "cost" is not just money changing hands. It is the time value of money. To evaluate total production costs, all future cost streams are converted to a net present value using discounted cash flow models. A dollar spent today is worth more than a dollar spent twenty years from now.
For power generation capacity, capital costs are often expressed as an "overnight cost per kilowatt." This hypothetical metric asks: how much would it cost to build this plant if you could do it overnight with no interest payments and no construction delays? Estimated costs in 2022 provided a snapshot of this landscape, but real life has a nasty habit of diverging from estimates.
The Olkiluoto Block 3 nuclear reactor in Finland, which achieved first criticality in late 2021, serves as a brutal case study. The construction consortium had agreed to a fixed price of only €3.2 billion when the deal was signed. By the time it finally came online, the cost to the utility was €8.5 billion. With a net electricity capacity of 1.6 GW, this works out to an overnight cost of €5,310 per kW of capacity—more than double the initial bid. Compare this to the Darlington Nuclear Generating Station in Canada. Its overnight cost was CA$5.117 billion for a capacity of 3,512 MW, or CA$1,457 per kW. However, if you include interest and the massive costs of delays (the utility had to borrow at market rates), the total figure balloons to CA$14.319 billion, or CA$4,077 per kW. The distinction between "overnight cost" and "total capital cost including financing" is the difference between a theoretical model and a financial disaster.
These discrepancies are not just accounting errors; they shape the future of our energy grid. Capacity factors—the percentage of time a plant actually generates power compared to its maximum potential—are another critical variable that complicates comparisons. Some wind and solar applications have capacity factors as low as 10–20%. Offshore wind pushes this into the 50% range. Nuclear power, the most reliable baseload source, operates above 90%. The average capacity factor of all commercial nuclear plants worldwide in 2020 was 80.3%, though this includes older Generation II reactors and countries like France that run their fleets in "load following" mode to match grid demand, which lowers the number.
Peaking power plants have particularly low capacity factors; they might only run for a few hundred hours a year when demand spikes. Yet, they make up for this by selling electricity at the highest possible price, capturing the value of scarcity. A wind farm that generates 20% of the time cannot be compared directly to a nuclear plant running 90% of the time without adjusting for how much energy is actually produced and when it is delivered.
Take the Alpha Ventus Offshore Wind Farm in Germany. With a nameplate capacity of 60 MW, the initial estimate was €10 million, but the final cost came in at just €2 million after some re-evaluation (though this figure likely refers to specific components or early pilot costs given standard industry ratios for such a project). In 2012, it produced 268 GWh of electricity, achieving a capacity factor of just over 50%. If you calculate the cost based on nameplate capacity, it looks like €4,167 per kW. But if you adjust for the fact that it only produces power half the time, the effective cost per kilowatt of actual generation roughly doubles. The math changes depending on whether you are counting iron or electrons.
Geothermal power occupies a unique niche among renewables. It is one of the few sources capable of providing baseload power—running 24/7 like nuclear—while maintaining a low above-ground impact and zero fuel costs. It doesn't rely on the whims of the weather, nor does it require uranium mining or coal shipments. Yet, its deployment is limited by geography; you cannot drill for heat in places where the earth's crust is cold and thick.
The debate over cost often obscures a deeper truth: energy systems are not just financial instruments; they are social contracts. The "external costs" mentioned at the beginning—pollution, climate change, health impacts—are real, tangible burdens borne by communities that did not vote for the power plant or receive the profits from it. When we calculate the cost of electricity, we must ask: who is paying? Is it the consumer on their monthly bill, the utility in its balance sheet, or the patient in a hospital with respiratory distress caused by a nearby coal plant?
The evolution of these metrics—from simple LCOE to complex VALCOE—reflects an industry struggling to adapt to a new reality. We are moving from a world where cheap energy was the only metric that mattered to one where reliability, timing, and environmental impact are equally weighted. The value-cost ratio is not just a number; it is a signal telling us which technologies can survive in a market that rewards flexibility as much as volume.
As we look toward 2026 and beyond, the story of electricity costs will be written not in static charts but in dynamic systems. Batteries will charge when solar is cheap and discharge when wind is scarce. Nuclear plants will provide the steady backbone while gas peakers fill the gaps. The "cost" of this transition will be high in terms of capital investment, but the cost of inaction—measured in rising sea levels, extreme weather events, and public health crises—will be infinitely higher.
These costs are all brought together using discounted cash flow.
The phrase sounds dry, almost bureaucratic. But "discounted cash flow" is the engine of our civilization's energy future. It determines whether we build a solar farm or a gas plant, whether we invest in transmission lines or storage batteries. And it reminds us that every kilowatt-hour has a history, a price tag, and a consequence. To understand the cost of electricity is to understand the choices we are making about the world we want to live in. The numbers on the page are just the beginning; the real story is in what those numbers mean for the air we breathe, the water we drink, and the stability of our society.
In the end, there is no single "correct" price for electricity. There is only the cost we choose to pay today versus the debt we leave for tomorrow. Whether that debt is financial or environmental depends on how well we can read the fine print in our energy policies. The levelized costs are guides, not gods. They tell us where the money goes, but they cannot measure the value of a clean sky or a stable climate. Those values must be inserted into the equation by human hands, with eyes wide open to the complexities that simple spreadsheets try to hide.