Estimate the emissions of a cloud estate from vCPU hours, PUE and grid carbon intensity, and what a greener region saves.
Server power is not proportional to load: a vCPU draws a substantial idle baseline and rises toward a loaded ceiling, so utilisation is interpolated between the two rather than multiplied. The resulting kilowatt-hours are scaled by PUE to account for cooling and distribution losses, then converted with the grid intensity of the specific region the workload runs in. Region choice usually dominates every other sustainability lever — the same workload can emit eight times more on a coal-heavy grid than on a hydro-backed one, and moving it costs nothing but latency. These are order-of-magnitude planning estimates; use your provider's own carbon reporting for anything you disclose.
Cloud Carbon Footprint
watts per vCPU = idle + (loaded − idle) × utilisation; kWh = (compute watts + storage watts) × hours ÷ 1000 × PUE; kgCO2e = kWh × grid intensity ÷ 1000 × (1 + embodied uplift).
watts per vCPU = idle + (loaded − idle) × utilisation; kWh = (compute watts + storage watts) × hours ÷ 1000 × PUE; kgCO2e = kWh × grid intensity ÷ 1000 × (1 + embodied uplift). Server power is not proportional to load: a vCPU draws a substantial idle baseline and rises toward a loaded ceiling, so utilisation is interpolated between the two rather than multiplied. The resulting kilowatt-hours are scaled by PUE to account for cooling and distribution losses, then converted with the grid intensity of the specific region the workload runs in.
Region choice usually dominates every other sustainability lever — the same workload can emit eight times more on a coal-heavy grid than on a hydro-backed one, and moving it costs nothing but latency. These are order-of-magnitude planning estimates; use your provider's own carbon reporting for anything you disclose.
This calculator takes 12 inputs: Total vCPUs running, Average CPU utilisation, Idle power per vCPU, Fully loaded power per vCPU, Storage provisioned, Power per terabyte of storage, Hours in the month, Provider PUE, Grid carbon intensity of your region, Intensity of a low-carbon region, Embodied carbon uplift, Matched renewable energy. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.
Providers apply market-based accounting that nets off the renewable energy certificates they buy, and they use measured rather than modelled server power. This calculator is location-based and model-driven, which is the more conservative view and the one most useful for comparing regions.