Grade by Grade
D. Topics in Computer Science · Topics in Computer Science: Sustainability, Ethics, Emerging Technologies and Careers

The Full Footprint of Computing: Embodied Carbon, Data Centres and the Grid

D1.1About 70 minutes
Learning goals
  • I can explain the difference between embodied and operational emissions of a computing device.
  • I can calculate a device's yearly footprint from its embodied emissions, power use and grid intensity.
  • I can explain power usage effectiveness (PUE) and interpret current, dated data-centre energy figures.
  • I can rank strategies for reducing the environmental impact of computers using evidence.

Words to know

CO2e (carbon dioxide equivalent)
A unit that converts all greenhouse gases into the amount of carbon dioxide with the same warming effect.
embodied emissions
Greenhouse gases released while a product's materials are mined, it is manufactured and it is shipped.
operational emissions
Greenhouse gases released to generate the electricity a device uses while it is working.
grid emission intensity
The average grams of CO2e released for each kilowatt-hour of electricity generated on a grid.
annualize
Spread a one-time amount evenly over the years something is used.
data centre
A building full of servers, storage and network equipment that runs cloud services.
power usage effectiveness (PUE)
A data centre's total facility energy divided by the energy used by its IT equipment; 1.0 is ideal.

Where does a laptop's carbon really come from?

In Grade 11 (ICS3U, The Environmental Cost of Computing) you learned the life cycle of a device, what e-waste is, and how to find energy with energy = power × time. This year we go one step further and do what sustainability analysts do: put all of a device's impact into one unit, so you can compare strategies and decide which ones actually matter.

The surprising answer for many devices used in Ontario: most of the carbon was released before you ever turned it on.

Two kinds of emissions

Scientists add up greenhouse gases in kilograms of CO2 equivalent (kg CO2e). Other gases, such as methane, are converted to the amount of carbon dioxide that would warm the planet by the same amount.

  • Embodied emissions come from mining materials, manufacturing chips, screens and batteries, and shipping. They are "locked in" the moment the device is made.
  • Operational emissions come from the electricity the device uses during its life.

Operational emissions depend on the grid emission intensity: how many grams of CO2e are released, on average, for each kilowatt-hour (kWh) generated.

operational emissions (g CO2e) = energy used (kWh) × grid intensity (g CO2e per kWh)

To compare devices fairly, analysts often annualize the embodied part: spread it evenly over the years the device is used.

yearly footprint = embodied ÷ years of use + operational emissions per year

Ontario's grid

The Government of Ontario's Ontario energy snapshot (last updated January 8, 2026) reports that in 2024 Ontario's electricity came from:

  • nuclear: 50.9%
  • hydroelectric: 24.1%
  • natural gas: 15.7%
  • wind: 8.5%
  • solar and bioenergy: each under 1%

It lists Ontario's grid emission intensity as 53 g CO2e per kWh in 2023. Grids that burn a lot of coal or gas have intensities many times higher. That is why the same computer has a different operational footprint in different places, and why an Ontario user's footprint is dominated by the embodied part.

Worked example

Example 1: Annualizing a laptop's footprint

A laptop's manufacturer reports (example value) 300 kg CO2e of embodied emissions. It draws an average of 20 W and is used 6 hours a day, every day. Find its yearly footprint in Ontario (use 53 g/kWh) if it is kept 3 years and if it is kept 6 years.

  1. Energy per year: 20 W × 6 h × 365 = 43 800 Wh = 43.8 kWh.
  2. Operational per year: 43.8 kWh × 53 g/kWh = 2321.4 g ≈ 2.3 kg CO2e.
  3. Kept 3 years: 300 ÷ 3 = 100 kg per year of embodied emissions; total ≈ 100 + 2.3 = 102.3 kg CO2e per year.
  4. Kept 6 years: 300 ÷ 6 = 50 kg per year; total ≈ 50 + 2.3 = 52.3 kg CO2e per year.
  5. Keeping the laptop twice as long almost halves its yearly footprint. Turning it off more would only change the small 2.3 kg part.

About 102.3 kg CO2e per year if kept 3 years and about 52.3 kg CO2e per year if kept 6 years. In Ontario, using devices longer matters far more than saving electricity.

Yearly footprint of one laptop (example values, kg CO2e per year)Ontario grid, kept 3 years102.3Ontario grid, kept 6 years52.3500 g/kWh grid, kept 3 years121.9500 g/kWh grid, kept 6 years71.9embodied (making the laptop)operational (electricity)
Each bar is one laptop's yearly footprint (example values from Example 1). The embodied part (yellow) shrinks when you keep the device longer; the operational part (blue) only grows large on a high-carbon grid.

Let Python compare the strategies

The program below turns the formula into a function. Each line of output is one scenario. Change the numbers to test your own device.

def yearly_footprint(embodied_kg, watts, hours_per_day, years, grid_g_per_kwh):
kwh_per_year = watts * hours_per_day * 365 / 1000
operational_kg = kwh_per_year * grid_g_per_kwh / 1000
return embodied_kg / years + operational_kg
scenarios = [("Ontario, 3 years", 3, 53), ("Ontario, 6 years", 6, 53),
("500 g/kWh grid, 3 years", 3, 500), ("500 g/kWh grid, 6 years", 6, 500)]
for name, years, grid in scenarios:
kg = yearly_footprint(300, 20, 6, years, grid)
print(f"{name:26}{kg:7.1f} kg CO2e per year")

The f-string format {kg:7.1f} prints the number in a 7-character column with one decimal place.

Ontario, 3 years 102.3 kg CO2e per year
Ontario, 6 years 52.3 kg CO2e per year
500 g/kWh grid, 3 years 121.9 kg CO2e per year
500 g/kWh grid, 6 years 71.9 kg CO2e per year

Output of the program above (what Python prints).

Data centres: the computers you never see

Streaming, cloud storage and AI chatbots run in data centres. The International Energy Agency (IEA) estimates:

  • about 415 TWh of electricity used by data centres worldwide in 2024, about 1.5% of global electricity (Energy and AI, April 2025);
  • about 485 TWh in 2025, a 17% increase, with electricity used by AI-focused data centres rising about 50% (Key Questions on Energy and AI, 2026);
  • about 950 TWh by 2030 in its projections, roughly double the 2025 level and about 3% of global electricity demand.

(1 TWh = 1 billion kWh.) These are estimates and projections, and the IEA updates them as new data arrive, so always check the date of any figure you quote.

Data centres measure their efficiency with power usage effectiveness (PUE): total facility energy divided by the energy that reaches the IT equipment. Many also report water use for cooling, because evaporative cooling can use a lot of water.

Power usage effectiveness (PUE) = total facility energy ÷ IT energyTotal facility energy: 2800 kWh per dayIT equipment (servers, storage, network): 2000 kWhOverhead: 800 kWhPUE = 2800 ÷ 2000 = 1.4Overhead is mostly cooling, power conversion losses and lighting.A perfect facility would have PUE = 1.0 (every kWh goes to computing).
PUE compares the whole building's energy use with the part that actually does computing. Lower is better; 1.0 is the ideal.
Worked example

Example 2: Reading a PUE

A data centre's IT equipment uses 2000 kWh a day and the whole facility uses 2800 kWh a day. (a) Find the PUE. (b) The operator improves cooling so the PUE becomes 1.2 with the same IT load. How much energy is saved each day?

  1. (a) PUE = 2800 ÷ 2000 = 1.4.
  2. (b) New total = 1.2 × 2000 = 2400 kWh per day.
  3. Saving = 2800 − 2400 = 400 kWh per day, or 146 000 kWh in a year (400 × 365).

(a) 1.4 (b) 400 kWh per day

Strategies that reduce the impact (D1.1)

The curriculum asks you to outline strategies for the environment and for human health. Ranked roughly by effect for an Ontario household or school:

  1. Reduce: buy fewer new devices; choose one device that does several jobs; don't upgrade just because a new model came out.
  2. Reuse and repair: keep devices longer, replace batteries, add memory, fix cracked screens, buy refurbished.
  3. Recycle properly at end of life, including printer cartridge and battery programs, so metals are recovered and hazardous substances are kept out of landfills.
  4. Power habits: turn computers and monitors off at the end of the day, use sleep settings, unplug chargers.
  5. Digital habits: delete files you don't need from the cloud, stream at the resolution you actually need, print only when necessary.

Health strategies (ergonomic standards, breaks and screen habits) are the topic of the lesson Healthy Computing: Ergonomic Standards and Workplace Rules in this unit.

Just for fun

A popular sticker in computing offices says: "There is no cloud. It's just someone else's computer."

It's a joke, but it is also exactly why data-centre electricity counts in your footprint.

Hands-on activity

Household device footprint audit

You need: a notebook or spreadsheet, the devices in your home (phones, laptops, tablets, game consoles, TVs, routers), and Python.

  1. List every device, its approximate age, and how many hours a day it is used.
  2. Find each device's power in watts. Look on the label or charger (it shows the maximum; typical use is lower, so use about half of a laptop charger's rating as an estimate) or search the model's specifications. A grown-up should handle anything behind a TV or near outlets.
  3. Use yearly_footprint with an embodied value. Many manufacturers publish "product environmental reports" with an embodied figure; if you can't find one, use the example value 300 kg for a laptop and say it is an estimate.
  4. Make a table of the yearly footprint for each device if kept 2 more years and if kept 4 more years.
  5. Write three recommendations for your household, ranked by impact, and justify the ranking with your numbers.

What to look for: correct energy arithmetic (W × h ÷ 1000 = kWh), estimates clearly labelled, and recommendations that put device lifespan ahead of small electricity savings when the numbers say so.

Your turn

1

What is the difference between embodied and operational emissions?

2

A monitor draws 25 W and is on 8 hours a day for 250 school days. How many kWh does it use in a year?

3

Using Ontario's 2023 grid intensity of 53 g/kWh, what are the monitor's yearly operational emissions?

4

A phone has 60 kg CO2e embodied emissions. What is its annualized embodied footprint if kept 2 years? 5 years?

5

Why is the operational footprint of the same laptop much smaller in Ontario than on a coal-heavy grid?

6

A data centre's facility uses 9000 MWh in a year and its IT equipment uses 6000 MWh. What is its PUE?

7

Give two strategies from this lesson that reduce environmental impact and say which one matters more for an Ontario laptop.

More practice: warm-up

1

What does CO2e stand for?

2

Convert 3500 Wh to kWh.

3

What does a PUE of 1.0 mean?

4

Name the two largest sources of Ontario's electricity in 2024.

5

Name one thing that should go to a printer cartridge recycling program instead of the garbage.

More practice: on your own

1

A gaming PC draws 300 W for 3 hours a day, every day. Find its yearly energy and its operational emissions on a grid of 53 g/kWh and on a grid of 500 g/kWh.

2

For the gaming PC above in Ontario, embodied emissions are 600 kg (example value). Is it better to cut daily use by one hour, or to keep the PC 5 years instead of 4?

3

The IEA estimated 415 TWh for data centres in 2024 and 485 TWh in 2025. Check the 17% growth figure.

4

A data centre's IT load is 5000 kWh per day. Its PUE drops from 1.6 to 1.3. How much energy does it save in a 365-day year?

5

Why should every data-centre or AI-energy number you quote include its date and source?

Challenge

1

Using yearly_footprint(300, 20, 6, years, 53), after how many whole years of use does the annualized embodied part first fall below 10 times the operational part?

2

A school replaces 200 laptops every 4 years. Each has 300 kg embodied emissions (example value). A repair program would make them last 5 years. How many kg CO2e per year does the school avoid?

3

A company says, "Our new AI service is green because our data centre has a PUE of 1.1." Explain what this claim leaves out. Give at least two missing pieces of information.

Check your understanding

Quiz

1.

Which formula gives operational emissions?

2.

For a laptop used in Ontario, which change usually reduces its yearly footprint the most?

3.

A facility uses 3000 kWh a day and its IT equipment uses 2500 kWh. What is its PUE?

4.

According to the IEA (2026), about how much electricity did data centres use worldwide in 2025?

5.

What does Ontario's grid emission intensity (53 g CO2e/kWh in 2023) describe?

6.

Which is an example of embodied emissions?

For parents and teachers

Big idea: Expectation D1.1 asks students to outline strategies that reduce the impact of computers on the environment. At Grade 12 they should back strategies with numbers. This lesson shows that, on Ontario's low-carbon grid, keeping and repairing devices usually matters far more than small electricity savings.

How to help: Do the household audit together. Let your teen read device labels (you handle anything behind furniture or near outlets) and ask them to defend their top recommendation with their calculation. Ask: "Would the answer be the same if we lived somewhere with a coal-heavy grid?"

Common difficulties:

  • Unit slips: watts × hours gives watt-hours; divide by 1000 for kWh, and grams by 1000 for kg.
  • Forgetting to annualize embodied emissions before adding them to yearly operational emissions.
  • Treating a low PUE as proof that a data centre is "green". PUE measures overhead only.
  • Quoting energy figures without a date. Data-centre figures change quickly; the ones here are IEA estimates published in 2025 and 2026.