After 20 years of exponential growth, the carbon footprint of computing is now greater than that of Saudi Arabia (European Commission, 2025, and The World Bank, 2024). But increased public awareness of the IT sector’s emissions impact has facilitated new ways of thinking about how we use digital services: carbon-aware computing. These strategies aim to reduce the emissions impact of computing processes, by prioritising renewable energy consumption.
Carbon-aware computing can be implemented in three ways: time shifting, location shifting, and demand shifting. Time and location shifting strategies have been practiced by GreenOps specialists for several years, but recent research suggests that their effectiveness can be limited if not implemented without awareness of a new, third aspect of carbon-aware computing: electricity demand.
Time Shifting
What is it?
Scheduling computing loads during periods when the local power grid is primarily generating renewable energy.
How do I do it?
Identify which data centers your workloads are based in, then find the grid mix of the country in which they are located. Consult the country’s grid operator to identify time periods where the carbon intensity of the grid is lowest, and schedule time-specific workloads during these periods.
Example:
In Spain, an excess of solar power around midday means that the grid is greenest between 11am and 3pm.
Source: Electricity Maps, 2026
What do I need to consider?
If you have workloads with predictable durations that can run at any time of day, time-shifting is a simple strategy to reduce your overall emissions from computing. However, on a larger scale, time-shifting does not actually reduce overall carbon emissions at the national level: it simply shifts energy consumption from one period to another. Because grid loads are preplanned, moving your workload to a low carbon time period doesn’t change the overall emissions generated by your grid.
Location Shifting
What is it?
Migrating workloads to data centers in countries with a low carbon intensity grid.
How do I do it?
Identify which data centers your workloads are based in and assess the carbon intensity of the national or local grid mix that powers them. Then, migrate those workloads to data centers in a location where the local grid mix is powered by a higher percentage of renewable energy sources.
Example:
The variety in carbon intensities of European grid mixes in February 2026, ranging from 16 gCO2e/kWh in Sweden to 625 gCO2e/kWh in Poland.
Source: Electricity Maps, 2026
What do I need to consider?
Workloads that are not restricted to one location by cloud service provider licensing or data sovereignty requirements can usually be migrated to alternative data centers without significant reconfiguration. It is important to assess whether the new data center location has the required hardware to maintain your original workload, as not all cloud services or providers are available in every country. From a carbon emissions perspective, location shifting operates similarly to time shifting, in that it is very effective at reducing an organization’s cloud emissions, but the overall impact on global emissions may be limited. This is because location shifting can cause a spike in demand in low carbon countries, which is offset with fossil fuel reserves or imported power from fossil fuel sources.
Demand Shifting
What is it?
Running workloads when demand for electricity is low to use energy that would otherwise have been curtailed.
Energy curtailment is the deliberate reduction of energy generation. Power suppliers curtail their energy sources when public demand is low, because the subsequent energy cannot be stored. Curtailment is primarily applied to renewable sources, because they cannot be managed or stored in the same way as finite fossil sources.
By shifting compute workloads to periods of time when public power demand is low in locations where renewables make up a large portion of the grid mix, such as overnight wind power, you signal to power suppliers that there is a demand for renewable energy that would otherwise not be being generated.
How do I do it?
Identify which data centers your workloads are based in. Then, find locations that utilise more renewable energy during periods of low power demand. Migrate your workloads to these locations, during these time periods.
Example:
The UK’s grid mix between 5pm and 10am – most of the UK’s energy is sourced via wind power between midnight and 6am, at the same time as overall power consumption drops.
Source: Electricity Maps, 2026
What do I need to consider?
The overlap between high renewable usage and low power demand can be difficult to predict and is subject to daily fluctuations. Complex compute processes do not align well with such a spontaneous schedule, but an awareness of electricity demand in combination with location and time scheduling allows GreenOps practitioners is likely to reduce demand for fossil fuels overall – even if workload scheduling does not always overlap with the exact periods in which renewables would have been curtailed.
What Now?
The field of carbon-aware computing is ever developing. If you’re interested in reducing your organization’s carbon emissions from the cloud, get in touch with Tailpipe to discuss how location shifting can help you to utilize renewable energy for your cloud workloads. If you’d like to find out more about demand-shifting, visit Climate Action Tech’s grid-aware software github page (with thanks to Hannah Smith and Ismael Velasco for their work developing the concept of demand shifting).
