Near-zero marginal cost rewrites the merit order.
For decades, power plants were dispatched in "merit order" — cheapest first. Coal plants with $25/MWh fuel costs ran baseload. Gas peakers with $60/MWh fuel costs ran during peaks. Nuclear plants with very low variable costs ran near-continuously. The merit order was stable and predictable.
Wind and solar changed everything. Their marginal cost — the cost to generate one more megawatt-hour once the equipment is installed — is essentially zero. No fuel. No emissions. The only cost is amortizing the upfront capital investment, and that cost is fixed regardless of whether the plant runs or not. An operating wind farm will always accept any positive price, because running earns something while not running earns nothing.
As solar penetration grew in California and other states, a new operational challenge emerged: the duck curve. On a clear spring day, solar generation peaks around noon, pushing down net demand (total demand minus renewable generation) to very low levels in the middle of the day. But when the sun sets in the evening, solar output falls and conventional generators must ramp up very quickly to meet the evening demand peak.
The ramp rate required — gigawatts per hour — strains the capabilities of conventional generators, which are designed for slower, steadier operation. CAISO's real-time market sometimes sees ramping requirements of 10-15 GW in three hours, a challenge that would have been unimaginable in 2005.
More dramatically: during high-solar periods, LMPs can go negative. If there's more power being generated than the grid can absorb, the marginal generator would have to pay to inject its power — because adding more supply when you're already long hurts the system. Negative prices send a signal: store this energy, curtail it, or export it. They're an artifact of zero-marginal-cost resources in a grid not yet built to handle them.
The combination of RPS mandates, federal tax credits, and plummeting costs drove massive renewable buildout in the 2010s. As renewable generation grew, so did the number of hours during which renewable energy was sufficient to meet most or all of demand. Conventional coal and nuclear plants — which have high fixed costs and need to run many hours to be economical — found their operating hours shrinking.
Utilities began retiring coal plants ahead of schedule, not primarily due to regulation, but because they couldn't earn enough revenue to cover their costs when the merit order put them behind zero-marginal-cost wind and solar. Nuclear plants — clean, reliable, and enormously expensive to build — faced the same market pressure. Several closed in competitive markets, even as policymakers scrambled to design "zero-emissions credits" and other support mechanisms to preserve their output.
The renewable revolution was reshaping not just the generation mix, but the entire economics of the electricity sector.
Tucson, Arizona, May 2017. When Tucson Electric Power announced a solar power purchase agreement at 2.999 cents per kilowatt-hour — the first utility-scale solar contract in U.S. history below three cents — analysts paused to note what the number signified. In 1977, solar photovoltaic panels cost approximately $77 per watt. By 2017, utility-scale modules had dropped below $0.30 per watt — a 99-percent cost reduction in four decades, faster than any energy technology in history. The Tucson contract was not an outlier; it was the leading edge of a price collapse that made solar the cheapest source of new electricity generation across much of the country. The implications for electricity markets were immediate: generation resources with zero fuel cost and near-zero marginal operating cost do not bid into energy markets the way gas plants do. Grid operators began confronting a fundamental question: how does a market designed around marginal cost pricing function when the cheapest resources have no marginal cost?
Lazard — Levelized Cost of Energy Analysis NREL — Utility-Scale Solar Cost Trends