Levelised cost of electricity: what LCOE leaves out

Levelised cost of electricity gives you a cost per unit of generated electricity. That makes two sites easier to compare. It does not tell you whether either site can export when the market pays, whether the connection date fits the land option, or whether the capital can tolerate the delay.

A lower LCOE is a useful result. It is not a build decision.

What is the levelised cost of energy?

Levelised cost of electricity, usually shortened to LCOE, discounts the expected costs of building and operating an asset and divides them by the discounted electricity it is expected to generate. The result depends on the inputs: capital cost, operating cost, financing assumption, useful life, production forecast and degradation assumption.

That definition matters because LCOE is a model of the asset. It is not the price the project receives. It does not know the shape of the market revenue unless the analyst adds a separate revenue model.

For a solar farm, the production forecast carries more weight than the nameplate capacity. A site with a cheaper build cost per MW can still produce less valuable electricity if its resource is weaker, its export is constrained or its output arrives in low-price hours.

Why can two sites with the same LCOE have different investment cases?

LCOE compresses the project into one average cost. Investment decisions depend on when electricity is produced, when cash leaves the project and what has to happen before construction can start.

Two sites can show the same LCOE while differing in several ways:

• One has a firm connection offer. The other has a proposed point of connection with reinforcement still to be settled.
• One exports most of its forecast output into stronger-price hours. The other produces into hours where solar output is already abundant.
• One has land control and a credible planning route. The other has an option that expires before the grid programme reaches a firm date.
• One carries a fixed-price equipment contract. The other still relies on an estimate for the solar farm cost per MW.

LCOE does not rank those differences. The investment case has to carry them separately.

Which project risks does LCOE leave out?

Connection and export risk

LCOE does not answer whether the network can accept the requested export at the point of connection. It also does not price reinforcement exposure, a queue delay or an export limit unless those effects are explicitly included in the model.

This is where a low LCOE can mislead. A cheaper generating asset with no firm route to market is worth less than a more expensive asset that can export on a defensible date. Curtailment removes output from the hours that can matter most to revenue, and redispatch explains who gives the instruction and who carries the resulting cost.

A preliminary screening report can flag these risks from the site information and documents supplied. It does not replace the network operator's official study.

Revenue shape and capture price

LCOE treats generated electricity as an output total. Revenue depends on the value of each interval's output, the offtake structure and the project's exposure to imbalance or market prices.

A solar site often produces most heavily in the same part of the day as other solar sites. Wind has a different production shape. Storage changes the timing again, but it brings its own operating and degradation assumptions. A single annual energy figure cannot show those differences.

The same issue appears across borders. A project in Germany and a project in the United Kingdom can have the same LCOE and different revenue because their market prices, grid conditions and contract terms differ. An interconnector can change the price a project sells at, but LCOE does not make that price comparison for you.

Development and delivery risk

LCOE usually starts with a project that exists in the model. Early-stage projects still have to secure land, planning consent, environmental evidence, a connection route, finance and a construction programme. Delay changes the financing cost and the date when revenue begins. Cancellation leaves development spend with no generation to divide it across.

Those risks belong in the investment case as probabilities, contingencies or explicit decision gates. Folding an unsupported allowance into LCOE gives the number a false precision.

What does Noda's screening record show about project scale?

Noda's screening record shows why a single cost comparison needs the site behind it. Between 5 July and 10 August 2026, the record covered 77 runs. Solar PV accounted for 1,373 MW across 28 runs. Ten of the 77 sites carried no technology yet and sit in the other bucket with no capacity attached.

That figure describes sites saved in the screening record, not a market-wide sample. It also does not say that the solar sites had the same export arrangement, development stage or revenue model. Capacity is the denominator for a cost comparison. It is not the answer to the connection question.

Should LCOE be high or low?

For otherwise comparable projects, lower LCOE is preferable. The comparison stops being useful when the inputs describe different risks or different products.

Before accepting a ranking, hold the following questions beside the LCOE model:

• Does the production forecast use the same treatment of degradation, availability and losses for both sites?
• Does the cost model include the same boundary, especially grid works, development costs and replacement assumptions?
• Is the discount rate consistent with the financing risk and timing of each project?
• Does a separate revenue model value output by time rather than multiplying annual generation by one price?
• What happens to the result if export is reduced, the connection date moves or curtailment increases?

The last question is the one a headline LCOE cannot answer. Run the change through the project model before the number reaches an investment committee.

What should you check before comparing two sites?

The developer should make the comparison auditable. Keep the LCOE inputs, the production forecast, the point of connection, the export request, the connection documents and the development milestones together. Record which values are source-backed, modelled, estimated or missing.

The investor should compare the risk-adjusted case, not the lowest displayed LCOE. The investor decision before money moves starts with the evidence behind the number and the conditions that could change it.

The grid review should remain separate from the LCOE calculation. Check the operator's current information and commission the formal study required for the project. LCOE can show that the generating asset is efficient. It cannot grant the asset a connection.

A project is worth advancing when its cost, export route, revenue shape and delivery assumptions still hold after the missing evidence is made explicit. That is the comparison the LCOE number cannot do on its own.

This article was written automatically from Noda's own screening data and checked against the official sources it cites. Editorial responsibility rests with Noda.