What is grid carbon intensity, and why it changes hour to hour
When you plug something in, the carbon footprint of that electricity is not a fixed number. It depends on which power plants happened to be running at that moment.
Grid carbon intensity, usually expressed in grams of CO₂ per kilowatt-hour (gCO₂/kWh), is the metric that captures this, and understanding it is increasingly important for anyone trying to reduce real emissions rather than just buy offsets.
Why intensity moves
A grid is a constantly rebalancing mix of generators. Overnight, when demand is low, a grid might be carried by nuclear, hydro, and wind, pushing carbon intensity down toward 50 gCO₂/kWh or lower. On a hot evening when demand peaks and renewables fade, the grid fires up natural gas (or, in some regions, coal) and intensity can climb past 400 or 500 gCO₂/kWh. Same grid, same day, a tenfold difference. This is why an annual average can be misleading: it hides the hours when your consumption is doing the most damage and the hours when it's nearly free of carbon.
How it's calculated
The method is conceptually simple. Take the share of electricity coming from each fuel at a given moment, multiply each share by that fuel's emissions factor, and add them up. Approximate lifecycle emissions factors:
| Fuel | Approx. gCO₂/kWh (lifecycle) |
|---|---|
| Coal | ~820 |
| Natural gas | ~490 |
| Solar | ~40 |
| Wind | ~11 |
| Hydro | ~24 |
| Nuclear | ~12 |
A grid that is 60% wind and 40% gas at a given hour will have a far lower intensity than the same grid running 60% gas and 40% coal later that evening. PowerGridIQ derives its carbon-intensity figures exactly this way, from the live or estimated fuel mix.
Marginal versus average intensity
There's a subtlety worth knowing. The average intensity describes the whole mix. The marginal intensity describes the specific plant that responds when you add one more unit of demand, often a gas peaker. For decisions about when to consume more (charging a fleet, scheduling a compute job), marginal intensity is arguably the more honest signal, because it reflects what your extra demand actually causes to run. Most public tools, including this one, report average intensity because the underlying marginal data is rarely published, but it's worth understanding the distinction.
How to use it
If you control when you consume (charging electric vehicles, running batch compute, pre-cooling a building), you can shift load toward low-intensity hours and cut emissions with no change in total energy used. In many grids the cleanest hours are also among the cheapest, so timing pays twice. If you're choosing where to operate, comparing the typical intensity curves of candidate grids tells you far more than comparing single snapshots. And if you report emissions under an hourly-matching framework, intensity data is the backbone of those calculations.
A caution on precision
Carbon-intensity figures are estimates, not laboratory measurements. Emissions factors vary by source and methodology, transmission losses complicate the picture, and imported electricity across interconnectors carries the carbon of a neighboring grid that may not be fully visible. Treat intensity as a strong directional signal (clean hour versus dirty hour, clean grid versus dirty grid) rather than an exact accounting figure. Used that way, it's one of the most actionable numbers in energy.