Marginal Cost by Technology
Prerequisites
The spread from cheapest to priciest generator is not 2x or 3x. It is more than 100x. That gap shapes every electricity price you have ever paid.
Marginal cost is what it costs to produce one additional megawatt-hour from a plant that is already running. It excludes capital already spent; it captures only fuel, variable operations, and emissions compliance. The spectrum, per Lazard's marginal-cost benchmarks: solar and wind at $0-5/MWh (fuel is free), nuclear at $10-20, coal at $25-50, gas combined-cycle at $30-60, gas peakers at $80-150, oil peakers at $150-300+.
When renewables are abundant, they suppress prices aggressively. When they are scarce (calm night, overcast week), the stack falls back on fossil fuels and prices can spike by an order of magnitude within a single day.
Tracing a price spike. A summer afternoon grid has 8 GW of solar, 4 GW of gas combined-cycle, and 1 GW of gas peakers. At 7 PM, solar drops to near zero as the sun sets. Demand stays high.
If the grid now needs peakers to balance supply, what happens to the clearing price?
The marginal unit shifts. From gas combined-cycle ($45/MWh) to gas peakers ($120/MWh). Every generator on the grid receives the higher price. Total grid spend roughly triples for that hour, even though 80% of running capacity is the same gas plants from the afternoon.
Knowing the cost stack tells you which plants are vulnerable to policy intervention. A carbon price raises coal and gas marginal costs; it leaves wind and nuclear untouched. A capacity market may keep expensive peakers alive even if they rarely run.
A grid adds 3 GW of new wind capacity. During windy afternoons, which outcome is most likely?
Wind's near-zero marginal cost pushes it to the front of the dispatch queue, displacing gas plants that previously set the clearing price. This is the merit order effect.
The answer is ALesson complete
Next: Zero Marginal Cost Problem→This unlocks