Not All Electricity is the Same
Imagine two identical data centers running identical workloads. One is in Iceland, powered almost entirely by geothermal and hydroelectric energy. The other is in a coal-heavy region of the American Midwest. They consume the same number of kilowatt-hours, but their carbon footprints are radically different. This is the central insight of GreenOps: the carbon cost of electricity depends entirely on where and when it is generated.
The Carbon Intensity of electricity is measured in grams of CO₂ equivalent per kilowatt-hour (gCO₂eq/kWh). It represents how much carbon dioxide (and equivalent greenhouse gases) are emitted to produce one unit of electrical energy.
- Low carbon intensity: Wind, solar, hydro, nuclear — typically 10–50 gCO₂eq/kWh.
- High carbon intensity: Coal — typically 800–1,000 gCO₂eq/kWh. Natural gas — typically 400–500 gCO₂eq/kWh.
The Grid is Dynamic — Carbon Intensity Changes Constantly
This is the crucial insight that makes carbon-aware computing possible and necessary: carbon intensity is not a fixed number. It fluctuates hour by hour.
Time dimension: At 2:00 PM on a sunny, windy afternoon, solar farms and wind turbines may be generating so much power that the grid is very clean. By 7:00 PM, when millions of people arrive home and turn on appliances, demand spikes. The grid operator must bring additional "peaker" plants online — often the dirtiest, least efficient gas or coal plants — to meet that demand. The same data center, running the same workload, has a dramatically different carbon footprint at 2:00 PM versus 7:00 PM.
Location dimension: A data center in the Pacific Northwest (Oregon, Washington) draws from a grid dominated by hydroelectric dams — one of the cleanest grids in North America. A data center in the US Midwest may draw from a grid still heavily dependent on coal. The carbon intensity difference between these two regions can be a factor of 10 or more.
Why This Matters for Engineers
Carbon intensity data is publicly available through APIs. The Electricity Maps API (formerly ElectricityMap) and the National Grid ESO Carbon Intensity API (for the UK) provide real-time and forecast carbon intensity data by region. This means your software can query this data and make intelligent scheduling decisions — the same way it might query a pricing API to make cost-aware decisions.