A new solar farm near Dubbo publishes a nameplate of 333 megawatts. A gas peaker at Kurri Kurri is rated around 660 megawatts. On paper the gas plant looks like it does twice the work. In practice, over a full year, the solar farm will probably generate more electricity than a peaker that mostly sits idle waiting for a hot Tuesday in February. That gap between what a plant could produce running flat out and what it actually delivers is the capacity factor, and it’s the single most misunderstood number in the Australian energy debate.
I’ve lost count of how many press releases cross my desk that quote nameplate capacity like it settles an argument. It doesn’t. Nameplate tells you the size of the asset. Capacity factor tells you how hard that asset actually works across a year, a season, or a five-minute dispatch interval. Mix the two up and you’ll misprice a project, misjudge a reliability forecast, or write a headline that doesn’t survive contact with AEMO’s generation data.
The basic definition, and the arithmetic behind it
Capacity factor is a ratio: the energy a generator actually produces over a period, divided by the energy it would have produced if it ran at full nameplate output for that entire period. It’s expressed as a percentage, and it’s unitless in the sense that it strips away the megawatt-hour totals and leaves you with something comparable across technologies of wildly different sizes.
Work it through with real numbers. Take a 100 megawatt wind farm. Run flat out for a year: 8,760 hours, and it would produce 876,000 megawatt-hours. Say AEMO’s generation data shows it actually produced 320,000 megawatt-hours over that year. Divide the two: 320,000 ÷ 876,000, and you get roughly 37%. That’s a solid, fairly typical capacity factor for an onshore wind farm in a good Australian location: better than you’d get in much of Europe, for what it’s worth, which is part of why this country’s wind resource gets talked up the way it does.
Compare that with a black coal unit running as baseload, which might sit anywhere from 60% to 85% depending on age, maintenance cycles and how often the market calls on it. A gas peaker built specifically to cover the top few percent of demand on the worst afternoons of the year might run at a capacity factor under 10%. None of these numbers describes whether the plant is “good”. They describe the job it’s doing.
Power versus energy, again
I keep hammering this distinction because confusing it is where most bad energy reporting starts. Megawatts measure the rate at which a plant can push electricity onto the grid at an instant – power. Megawatt-hours measure the actual electricity delivered over time – energy. Capacity factor is the bridge between the two, and it’s precisely because people quote a nameplate megawatt figure and assume it behaves like an energy figure that we get headlines claiming a new wind or solar project will “power” some improbably large number of homes.
Let’s be careful with that number whenever you see it. A 500 megawatt solar farm at a 25% capacity factor delivers the equivalent of a 125 megawatt plant running continuously. That’s the number you’d use to compare it with, say, a baseload gas plant or to work out how many average households its output actually covers across a full year, not the headline 500.
Why the ratio differs so much by technology
Three separate things drive capacity factor down from the theoretical 100%, and they don’t act the same way across technologies.
The first is resource availability. Solar panels only produce during daylight, and output tapers at the edges of the day and drops hard under cloud. Wind turbines need wind within a usable speed band – too little and the blades don’t turn, too much and the turbine has to feather out or shut down for safety. These are physical constraints no amount of better engineering removes; they’re why AEMO’s Integrated System Plan models solar and wind capacity factors by region, because a wind farm on the Robbins Island side of Tasmania behaves nothing like one in the Western Downs.
The second is economic dispatch. Gas peakers aren’t idle because something’s broken. They’re idle because the market doesn’t need them: batteries, hydro and cheaper generation cover most hours, and the peaker only gets called when wholesale prices spike on scarcity. A low capacity factor here is the plant doing exactly its job, not failing at it. I’d argue this is the bit that gets lost whenever someone complains that gas plants are “underutilised” – underutilisation is the business model for a peaking asset, not a defect in it.
The third is maintenance and forced outages, which hit every technology but matters most for large thermal units where a boiler tube failure can take a unit offline for weeks. Eraring’s ageing units, for instance, have had their availability affected by exactly this kind of unplanned outage in recent years, part of the reason behind the arrangement that kept the plant running past its original retirement date – see our piece on the deal that bought NSW two more winters.
Why capacity factor decides the real cost of a project
Here’s where the ratio stops being trivia and starts deciding whether a project stacks up financially. Build a generator, and most of your cost is fixed: the turbines, the panels, the transformer, the civil works, the finance charges on all of it. That fixed cost gets spread across however many megawatt-hours the plant actually produces over its life. A plant with a high capacity factor spreads its fixed costs over more units of energy, which pulls down its cost per megawatt-hour even if its capital cost per megawatt was expensive to begin with.
This is the whole logic behind nuclear’s claimed advantage, and also its practical problem in Australia. Nuclear plants overseas commonly run at capacity factors above 90%, spreading an enormous upfront capital cost over a huge volume of generation. But that math assumes the plant actually gets built on budget and on schedule, which is precisely the contested part of the debate our colleagues have covered in detail – see does Australia actually need nuclear power? and the cost and timeline questions around the Coalition’s nuclear plan.
Run the same logic on batteries and it flips the question entirely. A big battery might have a capacity factor, by this strict energy definition, of well under 20%, because it’s designed to charge cheap and discharge expensive, not to run continuously. Judging a battery on capacity factor the way you’d judge a coal plant is close to meaningless, which is exactly the argument in our earlier piece on whether big batteries are being built for the wrong job. The right metric for storage is closer to cycling frequency and revenue per megawatt of power capacity, not energy throughput against nameplate.
A worked comparison across technologies
Put rough, representative Australian figures side by side and the pattern becomes obvious.
| Technology | Typical capacity factor | What drives it |
|---|---|---|
| Rooftop solar (NEM average) | roughly 15–18% | daylight hours, panel orientation, cloud |
| Utility-scale solar (single-axis tracking) | roughly 25–30% | better tracking, still daylight-bound |
| Onshore wind | roughly 35–45% | site wind resource, turbine hub height |
| Offshore wind (where built overseas) | roughly 45–55% | stronger, steadier offshore wind |
| Black coal baseload | roughly 55–85% | age, maintenance, market dispatch order |
| Gas peaker | often under 10% | dispatched only in scarcity hours |
| Pumped hydro / battery storage | highly variable, often under 20% | designed for arbitrage, not continuous output |
None of these numbers are AEMO’s exact published figures for any single asset. They vary site to site and year to year, and anyone quoting a precise single national average should be naming the document. AEMO’s Integrated System Plan and its Quarterly Energy Dynamics reports are the places to go for the real, current regional figures rather than a rule of thumb like mine.
Where the number gets abused in public argument
The most common misuse I see is stacking nameplate megawatts across a portfolio of renewable projects and announcing it equals some number of coal units retired. It doesn’t, unless you’ve adjusted for capacity factor and for the fact that wind and solar output doesn’t arrive on demand the way coal and gas dispatch does. The second most common misuse runs the other way – critics of renewables citing a low capacity factor as proof the technology is a failure, when a 30% capacity factor on a wind farm built at a fraction of the capital cost of a coal unit can still produce electricity more cheaply per megawatt-hour over its life. Capacity factor is one input into a cost comparison, not the whole comparison.
It also gets abused in reliability arguments, which is really a different question dressed up in the same clothes. Keeping the lights on during a tight winter evening isn’t about annual capacity factor at all. It’s about whether enough power is available in that specific half-hour, which is why firming capacity matters so much in the debate we’ve covered around gas peakers versus big batteries. A technology with a brilliant annual capacity factor that happens to produce nothing during the one hour the grid is stressed hasn’t solved the problem that actually matters that evening.
The practical reading for anyone sizing up a project
If you’re trying to judge whether a new generation announcement is a big deal, ask two questions before you look at the nameplate number at all: what capacity factor is this technology likely to achieve at this site, and what job is it actually meant to do – bulk energy, peak power, or arbitrage? A 1,000 megawatt battery announcement sounds enormous next to a 300 megawatt wind farm, but they’re not playing the same game, and comparing them on nameplate alone is like comparing a sprinter’s top speed to a marathoner’s: both numbers are real, neither tells you who’d win the other’s race.
I’ll admit I used to get mildly obsessive about this on long bike rides, working out capacity factors for imaginary plants just to pass the kilometres on the Monaro Highway. It’s a habit that’s served me well in this job, if nothing else: the number looks dry until you realise it’s quietly deciding which projects get financed and which ones get a polite rejection from a bank’s credit committee.
– Anjali Rao, Grid & Storage Correspondent
Photo by American Public Power Association on Unsplash