Comparing without cheating

Power, energy, land, cost and emissions put to work for an honest comparison.

Two headlines, two traps

The same word can mean different things

"This plant uses less land." "This electricity is cheaper." Both statements can be literally true and still not be enough to make a well-founded decision, because the words "land" and "cheaper" hide very different definitions depending on who uses them and with what intent. The land occupied by a facility can be measured as the exclusive plot actually sealed under concrete or panels, as the project's total area including the spacing needed between turbines or rows of panels, or as the surface that in practice remains compatible with other uses — grazing under elevated solar panels, farming between wind turbines — while hosting the facility. The price of electricity, in turn, can refer to the cost of building a completely new plant from scratch, the cost of operating a plant already paid off years ago, the wholesale price of a specific hour in the electricity market, or the final bill a household pays, which also includes grid costs, taxes and commercial margins that have nothing to do with the generation cost itself. And greenhouse gas emissions can be counted as only what comes out of the smokestack during operation, or as the full lifecycle, including equipment manufacturing, fuel extraction and final decommissioning.

Levelized cost of energy: useful, but incomplete by design

The levelized cost of energy (LCOE) is the most cited metric for comparing electricity-generation technologies against each other: it summarizes, in a single cost-per-megawatt-hour figure, the initial investment, fuel, maintenance and decommissioning of a plant over its entire service life, divided by all the energy it is expected to produce. It is a genuinely useful tool for comparing technologies on equal footing, but it has an important structural limitation: it says nothing about when that energy is produced or about what the electricity system needs at that specific moment. A variable generation source — solar, wind — delivers energy when the natural resource is available, not necessarily when demand needs it; a dispatchable source — gas, reservoir hydropower, nuclear — can instead respond according to its own design and fuel availability, with more or less flexibility depending on the specific technology. That is why an isolated euros-per-megawatt-hour figure, however well calculated, never on its own decides how a complete electricity system should be designed: that system also needs transmission networks, storage, interconnection with other regions, forecasting capacity for both demand and renewable resource, and flexible-demand mechanisms that adapt to real generation availability.

Compare equivalent services, not just isolated technologies

Producing electricity usefully is not simply a matter of adding up the installed capacity of solar panels or wind turbines: the electricity grid has to balance supply and demand at every moment, second by second, and that balance is itself a service that has to be provided, not an automatic side effect of having enough installed capacity on paper. The capacity factor — the fraction of real time during which a plant actually produces electricity at its rated maximum power — varies enormously between technologies, and confusing installed capacity with energy actually produced is one of the most common comparison mistakes, and one of the easiest to make without noticing.

Checklist before sharing a comparison

  • Is it talking about installed capacity or energy actually produced?
  • Does it include construction, fuel, maintenance, decommissioning and waste management, or only operation?
  • Does it compare new plants with existing, already paid-off plants?
  • Does it account for grid, backup, storage and production constraints when they matter for the whole system?
  • Is the land counted exclusive, or does it allow farming, grazing or already-built rooftops?
  • Is the figure a global average, or does it describe a specific place and climate?
  • What year, what financial discount rate and what equipment service life were assumed in the calculation?

A practical criterion

Start by cutting avoidable demand; then choose low-impact technologies suited to the locally available resource; preserve technological diversity and system flexibility; and measure real outcomes, not just projections. The best combination of technologies for an isolated island, a dense city and a heavy-industry region does not have to be the same in any of these cases.

The numbers, with their source

Levelized cost varies by a factor of five or more within the same technology

According to Lazard's 2025 report, the levelized-cost range for utility-scale solar photovoltaic runs from 38to38 to 217 per megawatt-hour, and onshore wind from 37to37 to 86 per megawatt-hour — a range so wide within the same technology that any single figure quoted without further context is almost certainly misleading. For the other technologies, the 2025 range is: combined-cycle gas 48−109/MWh,coal48-109/MWh, coal 71-173/MWh, nuclear 141−220/MWh,andpeakinggasturbines141-220/MWh, and peaking gas turbines 149-251/MWh. Solar and onshore wind have, according to Lazard, been the cheapest new-build generation source for ten consecutive years, though in 2025 wind costs rose sharply — 55% for onshore and 23% for offshore compared with 2024 — due to more expensive materials and supply chains.

Source: Lazard — Levelized Cost of Energy+, June 2025.

Capacity factor separates installed power from real energy

In the United States, in recent years, average capacity factors were approximately: nuclear 92-93%, geothermal 65%, offshore wind 41%, onshore wind 34-36%, solar photovoltaic 23-25%, and coal 50-80% depending on the source and year. Natural gas, often used to cover demand peaks, has a highly variable capacity factor, between 10% and 90% depending on the role that specific plant plays in the system. A nuclear plant with a 93% capacity factor produces, per megawatt of installed capacity, almost four times more real energy per year than a solar plant with a 24% capacity factor.

Sources: US Energy Information Administration (EIA); PNAS — Capacity factors for electrical power generation.

Land use can differ by several orders of magnitude

Counting the total land area occupied per terawatt-hour of electricity produced annually, nuclear uses on the order of 7 hectares per TWh/year, the most compact of all sources studied; wind uses 130 ha/TWh/year if only the turbines' physical base is counted, but 12,000 ha/TWh/year if all the spacing between turbines is included — land that, in practice, remains compatible with farming or grazing; ground-mounted solar photovoltaic uses about 2,000 ha/TWh/year; natural gas is around 410 ha/TWh/year and coal around 1,000 ha/TWh/year, counting fuel extraction.

Source: land-use intensity study published in PLOS One, Land-use intensity of electricity production.

What this table is missing

None of these figures are valid without context: levelized cost depends on location, year and the assumed financing rate; capacity factor depends on the natural resource available at each specific site; and land use depends on whether exclusive land or land shared with other uses is counted. Comparing without specifying these conditions is precisely the trap this leaf tries to teach how to spot.

Sources

We do not set a universal table: costs, capacity factors and electricity mixes change over time and by location. Each leaf in this block explains which variables to ask for before trusting a number.