LCOE (Levelized Cost of Electricity) is a metric that takes a plant’s construction, fuel, operation and maintenance, and decommissioning costs and divides them by the electricity it produces over its lifetime, converting everything into a cost per kWh. It lets you compare capital-heavy sources such as renewables and nuclear with fuel-heavy thermal generation on the same basis.
| Source | Cost structure | Key variables |
|---|---|---|
| Solar PV | High share of upfront investment, no fuel cost | Irradiance, module prices, whether ESS is attached |
| Onshore wind | High share of upfront investment, no fuel cost | Wind resource, land acquisition cost |
| LNG combined cycle | Low upfront investment, high share of fuel cost | Gas price volatility |
| Nuclear | Very high upfront investment, low fuel cost | Construction period, regulatory and licensing costs |
A caution — LCOE compares only the cost of generation; it does not capture the cost of the ESS and reserves that back up intermittent resources, or the cost of connecting to the grid. Even when solar and wind come out with low LCOE, the gap can narrow once actual grid operating costs are added, so a single figure is not enough to declare one source superior to another.
Reading the cost structure from a grid-operations standpoint
The table above sorts sources by where the money goes. From the seat of someone who has to keep a power system running, the same table reads differently: it shows where each source’s risk sits and when that risk shows up on the bill.
Capital-heavy sources (solar, wind, nuclear) lock in most of their cost on the day construction finishes. After that, the cost per kWh depends almost entirely on how many kWh the plant actually delivers. Anything that cuts delivered energy, such as curtailment during low-demand hours, extended outages or a slipping construction schedule, raises the effective LCOE even though no one spent an extra won on equipment. For nuclear, the “construction period” variable in the table works exactly this way: every year of delay adds financing cost while producing no energy.
Fuel-heavy sources (LNG combined cycle) are the reverse. Their upfront commitment is small, but their cost per kWh moves with the fuel market every month they run. That makes them cheap to hold in reserve and expensive to run continuously when fuel prices spike. In system terms, this is why the same plant can be the right answer for peak and ramping duty and the wrong answer for round-the-clock output.
Variable sources add a third dimension that LCOE does not show at all: the cost of making their output usable. When output follows the weather rather than the load, something else has to fill the gaps, whether that is ESS, dispatchable reserves or additional transmission. Those costs land on the system, not on the plant’s own ledger, which is why the caution above matters.
What LCOE does not show: a qualitative comparison
The table below deliberately contains no cost figures. It lists the system-level characteristics that decide how much extra cost a source brings with it beyond its own LCOE.
| Source | Dispatchability | Output variability | Firming need | Main exposure |
|---|---|---|---|---|
| Solar PV | Low (follows irradiance) | High, daily and seasonal | High (ESS, reserves) | Curtailment, grid connection |
| Onshore wind | Low (follows wind) | High, less predictable | High (reserves, transmission) | Siting, curtailment |
| LNG combined cycle | High (fast ramping) | Low (controllable) | Low | Fuel price |
| Nuclear | Limited (baseload operation) | Low (steady output) | Low to moderate | Schedule, licensing |
The takeaway from this table is that a low LCOE and a low system cost are not the same thing. The sources with the lowest fuel exposure tend to carry the highest firming need, and the sources with the lowest firming need carry either fuel risk or schedule risk. The comparison that actually decides an investment is LCOE plus the firming and connection cost the source imposes on the grid it joins.
Where it wins, where it doesn’t
Capital-heavy sources (solar, wind, nuclear) — Current condition: a grid that absorbs the output without curtailment, and a plant completed on schedule. After applying it: the fixed upfront investment is spread over many kWh, LCOE stays low, and the source is free of fuel price swings. Linking logic: they win where the delivered kWh is assured. They lose where curtailment is frequent, where the construction schedule is uncertain, or where the firming cost of intermittency lands heavily on the grid.
Fuel-heavy sources (LNG combined cycle) — Current condition: a duty with short annual running hours, such as peak and ramping or reserve capacity. After applying it: the small upfront commitment keeps standby cost low, and fuel price exposure is limited to the hours actually run. Linking logic: it wins where it must run occasionally and quickly. It loses when assigned to continuous baseload while fuel prices are rising.
The same plant can move up or down the ranking when its duty changes from baseload to peak or the reverse. If the duty is not fixed before reading an LCOE table, the comparison ends up being between equipment doing different jobs.
Judgment order: using LCOE without being misled by it
- Fix the duty first — decide whether the capacity is meant for baseload energy, peak and ramping, or backup. LCOE is only comparable between sources that can serve the same duty.
- Check the assumed utilization — every LCOE figure assumes an annual output. Confirm that the assumption survives curtailment, maintenance outages and grid constraints at the actual site.
- Check the schedule and financing assumptions — for capital-heavy sources, confirm the construction period and cost of capital behind the figure; these move the result more than equipment prices do.
- Check fuel price exposure — for fuel-heavy sources, ask what happens to the cost per kWh under a sustained fuel price rise, not just under the base case.
- Add the system cost — for variable sources, add the ESS, reserve and connection cost needed to deliver the output the duty requires. Only then compare.
The order matters. Starting from the LCOE figure and working backward tends to hide the duty mismatch; starting from the duty makes the relevant comparison obvious and exposes which assumptions actually carry the result.
Boundary conditions
LCOE becomes a sound basis for choosing between sources when the sources being compared serve the same duty and the grid can absorb their output without significant firming or connection upgrades. In that situation, the plant-level cost is close to the system-level cost, and the lower LCOE genuinely means cheaper electricity.
The judgment flips when the share of variable generation on the grid is already high or the connection point is constrained. There, each additional kWh of low-LCOE variable output carries a growing firming and curtailment burden, and a source with a higher LCOE but controllable output can end up cheaper for the system. It also flips for capital-heavy sources when the schedule is uncertain, because a delay changes the LCOE after the decision is made.
The single selection criterion: compare LCOE only after adding the firming and connection cost the source imposes on its specific grid; where that cost is small, trust the LCOE, and where it is large, let the system cost decide.
