0 LWater saved this month
0.0 gEmissions saved this month
0.0% of your monthly budget used · weighted by model footprint · we offset over 200%
Free tool

AI Carbon Footprint Calculator

Estimate the CO₂, energy and water behind your AI usage, ChatGPT, Claude, Gemini and more. Adjust the inputs to see your footprint per day, month and year, with real-world equivalents.

1.4
kg CO₂ per month
3.4
kWh per month
11.7
liters water per month
≈ 281 smartphone charges
≈ 3.4 miles driven in an average car
≈ 23 water bottles (500 ml)
≈ 0.78 trees needed for a year to absorb this
≈ $0.03 to offset at ~$20/tonne CO₂

Estimates use published energy-per-token figures by model class, PUE ~1.56, US-average grid (0.395 kg CO₂/kWh) and water (~3.4 L/kWh). Your real footprint varies by model, data center and region.

Beyond the estimate

Don't just estimate it, track and offset it

This calculator approximates your impact. Ecoia measures the real footprint of every request and offsets over 200% of it.

Measured per request

Every token and image is metered for carbon, water and energy, no annual guesswork.

Offset >200%

Verified reforestation and clean-water offsets retired for more than double your usage.

Exportable reports

CSRD- and SB-253-ready carbon, water and energy data for compliance and ESG decks.

AI carbon footprint FAQ

How accurate is this AI carbon footprint calculator?

It produces a transparent estimate, not a lab measurement. We multiply your usage by published energy-per-token figures for each model class, add data-center overhead (PUE ~1.56), then apply US-average grid carbon (0.395 kg CO₂/kWh) and water-use factors (~3.4 L/kWh). Real impact varies with the exact model, data-center efficiency, and regional grid. See our How it Works page for the full methodology.

How much CO₂ does one ChatGPT prompt produce?

A single short exchange is small, often a few grams of CO₂ and a fraction of a liter of water. The footprint grows with longer prompts, larger or reasoning models, image generation, and especially with volume: billions of daily queries add up to a meaningful global total.

Does using a smaller model really reduce my footprint?

Yes. Energy scales with model size and token count. A small/fast model can use roughly a fifth of the energy per token of a frontier or reasoning model, so choosing the right-sized model for the task is one of the easiest ways to cut your AI carbon footprint.

How does Ecoia offset this footprint?

Ecoia measures the carbon, water and energy of every request you make and retires verified offsets for more than 200% of it, while 10% of revenue funds conservation. So your usage on Ecoia is carbon-negative rather than just neutral.

Is this calculator free to use?

Yes, completely free, with no sign-up required. If you want your real, per-request footprint tracked and offset automatically, that is what the Ecoia platform and API do.

Keep reading

Guide
How much electricity does AI use?
Guide
AI energy usage by model
Guide
How to reduce your AI carbon footprint
Carbon-negative AI

Use AI that gives back more than it takes

Ecoia.ai runs Claude, GPT & Gemini for chat, images and an API, and offsets over 200% of the water usage and carbon emissions your AI creates.

Get started with Ecoia.ai Try the chat free