Compare footprints
Compare the estimated carbon footprint of food, deliveries and digital activities to put your AI use into perspective. Find travel context in the side note and full table below.
Different activities, explicitly stated quantities. The digital batches are illustrative amounts, not measured averages or scientifically established norms.
See all the numbersThe carbon comparison
One linear scale · from zeroFood, deliveries and digital activities, with one beef burger as the largest comparator at 7 kg CO₂e. Central estimates for the quantities shown; select any activity to see its full table row.
↔ Scroll horizontally if needed to explore the full scale. Keyboard: focus the chart and use the arrow keys.
- Beef burger1 burger · 113 g patty + bun — view table row 7 kg CO₂e
- Chicken1 portion · 150 g — view table row 1.5 kg CO₂e
- Cheese1 portion · 50 g — view table row 1.2 kg CO₂e
- Large agent task1 task · ~1 hour · 100+ tool calls — view table row 210 g CO₂e
- Next-day parcel1 parcel · NL · ~150 km — view table row 150 g CO₂e
- Computer work8 hours · laptop + monitor — view table row 112 g CO₂e
- Video streaming2 hours · HD — view table row 100 g CO₂e
- Teams + recording1 hour · 4 people · kept 1 year — view table row 60 g CO₂e
- Google search100 queries · no AI — view table row 20 g CO₂e
- Medium agent task1 task · ~10 min · ~20 tool calls — view table row 18 g CO₂e
- Music streaming8 hours · smartphone — view table row 12 g CO₂e
- Chat queries100 short prompts + answers — view table row 12 g CO₂e
- GitHub Actions CI10 × 5-minute runs · Linux CPU — view table row 6 g CO₂e
Read the scale, not just the rank. All bars start at zero on the same kg CO₂e scale, using fixed quantities. The endpoint rings locate small values; they do not add to bar length. Tiny bars are genuinely tiny at this scale.
These are central estimates with low–high ranges in the table below. Only differences of about 10× or more are meaningful; differences of 2× can be within uncertainty. This compares different activities, not interchangeable services.
The detail behind the bars
All the numbers, in context.
The original nine report columns are per report unit, including energy and chat-query equivalents. The last two columns apply only to charted activities and describe their quantities and scaled carbon totals. Transport is retained for context, with original full-trip values, but is not charted.
↔ Scroll to see all 11 columns. Keyboard: focus the table region and use the arrow keys.
| Item | Functional unit | CO₂e central | CO₂e range | Final energy | Primary energy | Order of magnitude (g / Wh primary) | 🤖-equiv. | Conf. | Chart quantity (report units) | Chart CO₂e total |
|---|---|---|---|---|---|---|---|---|---|---|
| ⛽️🚗 Petrol car | 1 car, ~1,000 km AMS→GVA, 1 occupant | 170 kg | 130–210 kg | 620 kWh | 740 kWh | 10⁵ / 10⁶ | ~1,400,000 | H | Not charted | — |
| ✈️ Flight | 1 pax, AMS→GVA one way (incl. RF) | 110 kg (60 kg excl. RF) | 70–150 kg | 230 kWh | 280 kWh | 10⁵ / 10⁵ | ~900,000 | H | Not charted | — |
| 🔋🚙 EV | 1 car, ~1,000 km, NL grid | 54 kg | 45–70 kg | 180 kWh | 340 kWh | 10⁴ / 10⁵ | ~450,000 | H | Not charted | — |
| 🍔 Beef burger | 1 burger (113 g patty + bun) | 7 kg | 4–12 kg | — | 1.5 kWh (CED) | 10⁴ / 10³ | ~58,000 | M | 1 × burger | 7 kg |
| 🍗 Chicken | 1 portion (150 g) | 1.5 kg | 0.9–2.2 kg | — | 1.0 kWh (CED) | 10³ / 10³ | ~12,500 | M | 1 × portion | 1.5 kg |
| 🧀 Cheese | 1 portion (50 g) | 1.2 kg | 0.5–1.5 kg | — | 0.7 kWh (CED) | 10³ / 10³ | ~10,000 | M | 1 × portion | 1.2 kg |
| 🦾🦾 Agent task, large | ~1 h, 100+ tool calls, ~10 M tokens | 210 g | 70–600 g | 545 Wh | 1,040 Wh | 10² / 10³ | ~1,750 | L | 1 × task | 210 g |
| 📦 Parcel, next-day | 1 parcel, NL, ~150 km | 150 g | 80–350 g | 780 Wh | 1,000 Wh | 10² / 10³ | ~1,250 | M | 1 × parcel | 150 g |
| 🧑🧑🧒🧒 Teams + recording | 1 meeting-hour, 4 people, recording kept 1 yr | 60 g (≈15 g/person) | 40–400 g | 200 Wh | 380 Wh | 10¹–10² / 10² | ~500 | L | 1 × meeting-hour | 60 g |
| 📺 Video streaming | 1 hour, HD | 50 g | 20–100 g | 100 Wh | 190 Wh | 10¹ / 10² | ~420 | M | 2 × hour | 100 g |
| 🦾 Agent task, medium | ~10 min, ~20 tool calls, ~600k tokens | 18 g | 6–50 g | 48 Wh | 90 Wh | 10¹ / 10² | ~150 | L | 1 × task | 18 g |
| 🧑💻 Computer work | 1 hour, laptop + monitor | 14 g | 8–27 g | 45 Wh | 86 Wh | 10¹ / 10² | ~115 | H | 8 × hour | 112 g |
| 🎧 Music streaming | 1 hour, smartphone | 1.5 g | 0.5–8 g | 5 Wh | 10 Wh | 10⁰ / 10¹ | ~12 | L | 8 × hour | 12 g |
| 🗄️ GitHub Actions CI | 5 min, Linux runner, CPU only | 0.6 g | 0.2–1.6 g | 1.5 Wh | 2.9 Wh | 10⁰ / 10⁰ | ~5 | M | 10 × run (5 min each) | 6 g |
| 🔍 Google search | 1 query, no AI | 0.2 g | 0.05–0.5 g | 0.3 Wh | 0.6 Wh | 10⁻¹ / 10⁰ | ~2 | L (2009 data) | 100 × query | 20 g |
| 🤖 Chat query | 1 short prompt + answer | 0.12 g | 0.03–0.4 g | 0.3 Wh | 0.6 Wh | 10⁻¹ / 10⁰ | 1 | M | 100 × query | 12 g |
Three assumptions worth knowing
- Grid: NL 300 g CO₂e/kWh for home, office, EV and warehouses; data centres 400 g CO₂e/kWh (global average, location-based).
- Primary energy: electricity × 1.9; fuels × 1.2 (well-to-tank uplift). Food energy is cumulative energy demand (CED), which is already primary.
- Flight: non-CO₂ effects (radiative forcing, RF) are included in the table (DESNZ method, ~1.9×); the value without RF is in brackets.