redzilla
All tools
Power

Energy Cost & PUE

Enter IT power, PUE and tariff and I compute the energy and annual cost, the cooling overhead (PUE−1), DCiE and CO₂ emissions.

kW

Actual electrical load of the IT gear (servers, storage, network).

PUE = total energy ÷ IT energy. Ideal 1.0; typical 1.42.0.

$/kWh

Energy price per kWh. Use any currency; the result comes out in the same one.

h

Continuous operation = 8760 h (365 × 24). Lower it if it isn't 24/7.

kg/kWh

Grid carbon intensity. Chile's grid ≈ 0.4 kg CO₂/kWh.

Examples
redzilla.cl — pue
 
Annual energy cost
Cooling overhead
energy above IT (PUE−1)
CO₂ emissions
per year
Cost/yr
Energy
DCiE
CO₂

Calculation breakdown

ItemValueNote
How it's computed · PUE, DCiE and CO₂

1. Annual IT energy: E_IT = P_IT × hours (kW × h = kWh).

2. Total facility energy: E_total = E_IT × PUE. PUE covers cooling, UPS, distribution and lighting.

3. Annual cost: Cost = E_total × tariff.

4. Cooling and infrastructure overhead: = (PUE − 1) × E_IT × tariff. It's what you pay on top of the IT energy.

5. Efficiency: DCiE = 1 ÷ PUE (as %). Emissions: CO₂ = E_total × factor.

! Simplified model: assumes constant IT power and an average PUE. Real PUE varies with load and climate, and grid intensity changes hour by hour.

Runs locally in your browser · no sign-up · nothing leaves your browser

How it works

The calculator estimates the annual energy cost of a datacenter or equipment room from five inputs: the IT power (the real electrical load of servers, storage and network gear, in kW), the PUE (Power Usage Effectiveness, the ratio of total facility energy to IT energy), the electricity rate per kWh, the operating hours per year and the CO₂ emission factor of the power grid.

The formulas follow the standard efficiency model defined by The Green Grid: E_IT = P_IT × hours, E_total = E_IT × PUE and cost = E_total × rate. It also breaks out the cooling and infrastructure overhead as (PUE − 1) × E_IT × rate, the DCiE efficiency (1 ÷ PUE, as a percentage) and the annual emissions CO₂ = E_total × factor. It is a simplified model: it assumes constant load and an annual average PUE.

Example: a 10 kW rack with PUE 1.6 at $0.15/kWh

  1. Annual IT energy in continuous operation: 10 kW × 8,760 h = 87,600 kWh.
  2. Total energy with PUE 1.6: 87,600 × 1.6 = 140,160 kWh, and the annual cost: 140,160 × 0.15 = $21,024.
  3. Of that amount, (1.6 − 1) × 87,600 × 0.15 = $7,884 is cooling and infrastructure, not compute.
  4. With a 0.4 kg/kWh factor, annual emissions are 140,160 × 0.4 ≈ 56,064 kg, about 56 tonnes of CO₂.

Frequently asked questions

What is a good PUE for a datacenter?
The theoretical ideal is 1.0 (all energy goes to IT equipment). A modern efficient datacenter runs between 1.2 and 1.4; the typical range for enterprise rooms is 1.4 to 2.0, and older or poorly optimized facilities exceed 2.0. Hyperscalers publish averages close to 1.1.
How much do I save by lowering PUE from 2.0 to 1.5?
With the same IT load, total energy drops in direct proportion: 25 %. For 10 kW of IT at $0.15/kWh the annual cost goes from $26,280 to $19,710, about $6,570 saved per year, without touching a single server. Typical measures: hot/cold aisle containment, raising temperature setpoints and free cooling.
What is the difference between PUE and DCiE?
They are the same metric expressed inversely: DCiE = 1 ÷ PUE, as a percentage. A PUE of 1.6 equals a DCiE of 62.5 %: of every 100 kWh entering the facility, 62.5 reach the IT equipment and the rest goes to cooling, UPS, distribution and lighting.
Why can the result differ from my actual bill?
The model assumes constant IT power and an average PUE, but in reality the load varies with usage, the PUE changes with weather and occupancy, and tariffs usually include demand and time-of-use charges beyond the kWh. Use it to compare scenarios and size orders of magnitude, not to replace the utility bill.
Was this tool useful?
Disclaimer We take great care to keep every tool accurate and review it thoroughly; even so, we can't guarantee it is free of errors or take responsibility for how the results are used. We recommend double-checking anything critical.
Found an error? Let us know →