Cheap energy, firm capacity, flexibility: which plant does which job?
Sort the technologies by the jobs they do
The idea in one table: three jobs, and who does each well
| Job | The question | Does it well |
|---|---|---|
| Cheap energy | Many MWh a year at a low running cost? | Wind, solar, nuclear, hydro, geothermal |
| Firm capacity | There in the riskiest hours, whatever the weather? | Nuclear, gas plants, geothermal, demand response |
| Flexibility | Up or down quickly on request? | Gas peakers, batteries, reservoir hydro |
With JavaScript on, you work through this step by step: nine technology cards, a chart of two numbers, a sort and a contested question.
Capacity factors: U.S. 2025 averages, average output ÷ size, from EIA's Electric Power Monthly, Tables 6.07.A and 6.07.B (released 23 October 2026).
Firm shares: PJM's ELCC class ratings for the 2028/2029 auction, one market and one auction. They rate technology classes; a single plant's accredited MW can differ.
Also: battery round trip, EIA (2019 data); demand response, FERC (2024 data).
What you just did, as diagrams
- Cheap energy
- Fill the year with many MWh at a low running cost; capacity factor measures the MWh per MW, running cost (lesson 7) the rest.
- Firm capacity
- Be there in the riskiest hours; firm share measures it.
- Flexibility
- Change output quickly on request; ramp rate and start-up time measure it.
- Two different numbers
- Gas peakers averaged 13.6% of their size in 2025, yet PJM counts 67% of their MW at the peak; solar averaged 24.0% but counts 7%.
- No single winner
- No technology does all three jobs best, so grids combine them.
Go deeper · AnalystHow a grid measures firm share
A firm share, which PJM calls an ELCC class rating, asks: how much steady, always-available capacity would keep the grid exactly as reliable as this plant does?
Planners answer it by simulating many weather years with random plant outages.
- It depends on the grid. The numbers here are PJM’s for its 2028/29 auction; another region, or another year, gives others.
- It falls as more of the same is added. The tenth solar farm shifts the riskiest hours into the evening, so it counts for less than the first, though not always.
- Longer storage counts more. PJM counts 59% of a 4-hour battery’s MW, 68% at 6 hours, 71% at 8 and 78% at 10.
- Classes, not plants. These are ratings for technology classes; PJM can accredit a single plant differently.
Analyst lesson A5 runs this simulation for the Plains grid.
Go deeper · ModelerThe three jobs as three kinds of rule in a planning model
A planning model turns the three jobs into rules. In words:
- Decisions. How many MW of each technology to build, and how much each one makes every hour.
- Goal. The lowest total cost of building and running them.
- Rule 1, energy. Supply equals demand in every hour.
- Rule 2, firm capacity. The firm MW add up to the year’s peak demand plus a safety margin.
- Rule 3, flexibility. No plant changes its output faster than its ramp limit.
Rule 2 in symbols, with \(C_g\) the MW of technology \(g\), \(f_g\) its firm share and \(m\) the margin (for example 0.15):
\[ \textstyle\sum_g f_g\, C_g \;\ge\; (1 + m)\, D_\text{peak} \]
Lesson 10 teaches how to write rules like these down; lesson 13 asks a solver for Prairie Town’s least-cost mix.
Go deeperWhat this lesson leaves out
- Costs: what each job costs to buy is lesson 7.
- Wires: a firm plant behind a full power line cannot help the town that needs it.
- Rules: markets reward the three jobs in different ways (lesson 8, and Analyst lesson A3).
- Averages: every number is a fleet average or one market's rating; single plants differ.
Check yourself
Myth or fact? “A plant's capacity factor tells you how much it counts at the peak.”
- Myth
- Fact
Answer
Myth. Solar's capacity factor was 24.0% in 2025, but PJM counts 7% of its MW at the peak; gas peakers averaged 13.6% but count 67%.
Myth or fact? “A battery is a source of energy.”
- Myth
- Fact
Answer
Myth. It moves energy in time and loses some on the way: the U.S. battery fleet gave back 82% of what it stored (2019 data).
Myth or fact? “A gas peaker with a capacity factor of only 13.6% is not doing a useful job.”
- Myth
- Fact
Answer
Myth. A gas peaker's job is to be there at the peak and to move fast; whether that is the best way to do the job is a question of cost (lesson 7).
Which number tells you most about a plant's firm capacity?
- Its capacity factor
- Its firm share at the riskiest hours
- Its size in MW
- Its running cost
Answer
Its firm share: the part of its MW a grid can count on in the riskiest hours. Size alone overstates it; capacity factor measures energy per MW.
Which job does a 4-hour battery do best?
- Cheap energy
- Firm capacity, for as long as any peak lasts
- Flexibility
Answer
Flexibility: it switches between charging and discharging almost at once. It helps with firm capacity only for as long as its stored energy lasts (PJM's 4-hour rating: 59%).
Numbers, sources and what’s next
Numbers to remember
| Number | What it means | Source |
|---|---|---|
| 58.7% | Average capacity factor of U.S. gas combined-cycle plants in 2025. | EIA, Electric Power Monthly, Table 6.07.A · 2025 data (final), released October 2026 |
| 13.6% | Average capacity factor of U.S. gas turbines (peakers) in 2025. | EIA, Electric Power Monthly, Table 6.07.A · 2025 data (final), released October 2026 |
| 7% | PJM's class rating for fixed-tilt solar: the share of its MW counted toward peak needs (gas peaker 67%, 4-hour battery 59%). | PJM, ELCC Class Ratings for the 2028/2029 Base Residual Auction · read November 2026 |
| 82% | Share of stored electricity the U.S. battery fleet gave back. | EIA, Today in Energy · data for 2019 |
| 33,272 MW | Demand response enrolled in the seven U.S. wholesale markets, about 6.5% of their peak demand. | FERC staff, 2025 Assessment of Demand Response · data for 2024 |
Next on the starter path
In the real world
Regions reward these services differently: through energy prices, capacity payments or planning requirements, and payments for reserves and fast response. Which technology wins each depends on the region's rules and weather.
Sources and further reading
- EIA. Electric Power Monthly, Table 6.07.A and Table 6.07.B: capacity factors of U.S. utility-scale generators. Data for August 2026, released 23 October 2026.
- PJM (2026). ELCC Class Ratings for the 2028/2029 Base Residual Auction (PDF).
- EIA (2021). Utility-scale batteries and pumped storage return about 80% of the electricity they store. Today in Energy.
- NREL (2019). Grid-Friendly Renewable Energy: Solar and Wind Participation in Automatic Generation Control Systems: wind and solar held below their available output can raise it on request.
- FERC staff (2025). 2025 Assessment of Demand Response and Advanced Metering (PDF).
- EIA. Hourly Electric Grid Monitor (Form EIA-930): the Southwest Power Pool's 2025 wind data behind the calm spell.
Spotted an error, or stuck? Write to hlli@ksu.edu. Corrections are made in place and dated here.
Text licensed CC BY 4.0, code MIT: teachers and students may reuse and adapt it, with credit to Honglin Li.