Revenue model
Purpose
The revenue model estimates merchant energy revenue per MW of capacity. It uses a conservative generation assumption and historical prices, so that you can compare assets on the same basis.
Inputs
| Input | Description |
|---|---|
| P75 generation | The annual output that is met or exceeded with 75% probability. It comes from the location’s weather records for 2012–2024. |
| Local power price | The 2010–2024 average wholesale price from nearby pricing nodes. Where data are sparse, the model uses nearby locations in the same market or balancing area. |
| Curtailment adjustment | A reduction based on how often prices were negative in 2025. |
| Market-value factor | A regional adjustment for the value of solar or wind output, compared to average power prices. Source: Berkeley Lab. |
Method
Risked Revenue = P75 generation × local price × curtailment adjustment × market-value factor
Example with illustrative inputs:
2,000 MWh/MW-year × $40/MWh × 95% × 0.90 = $68,400/MW-year (shown as about $68K)
Outputs
- Risked Revenue ($/MW-year): the headline revenue estimate.
- Revenue Risk: one of the five InfraRisk Index families. It uses the year-to-year variation in the matched generation and price histories. It is different from Risked Revenue. A high Revenue Risk score shows variation to investigate. It does not explain the cause.
See the Metrics dictionary.
Assumptions
- Contracted energy: if you enter a PPA price on the asset’s Profile, the model uses that price for contracted energy instead of market prices. Check that the result matches the contract that you want to analyze.
- Curtailment: the curtailment adjustment is an estimate from negative-price hours. It is not a measurement of the plant’s curtailed energy.
Limitations
This is not a forward price forecast
The calculation combines different historical periods and a conservative generation input. It does not give a P75 revenue distribution or the amount that a lender will lend.
For all limitations, see Limitations.
Validation
Version
Applies to methodology version 1.0, September 2026.