What does it cost to run one AI prompt?

The cost depends on prompt length and compute region. A typical prompt uses 0.001-0.01 Wh of electricity, 0.02-0.2 mL of water, and 0.001-0.01 g of CO₂. Use the calculator to see exact costs for your prompt.

How is environmental impact calculated?

We estimate token count from your prompt, then multiply by peer-reviewed baseline metrics: 0.00045 Wh electricity per token, 0.004 mL water per token, and 0.0002 g CO₂e per token. Regional multipliers adjust for local grid carbon intensity.

Is this based on my location?

No, you select the compute region (US East, US West, Europe, etc.) in the settings. We then apply regional multipliers to reflect local grid composition and carbon intensity. The default is Global Grid Average.

What is the Global Real-Time Context clock?

It's a separate estimate of worldwide AI prompt resource use (~2.5 billion daily). It updates in real-time to show aggregate impact, independent of your prompt input. It's provided for context about the scale of global AI usage.

Calculate the Environmental Cost of Your AI Prompt

Enter any AI prompt and see instant electricity, water, and carbon impact.

What's your prompt?

Enter any AI prompt and instantly see its environmental impact.

Estimated Compute Weight: 0 tokens

Where is this processed?

Global standard operational baseline.

How often do you use AI?

5 prompts per day

Estimate how many AI prompts you send per day to project annual impact.

Your Prompt's Cost

See how this prompt translates into electricity, water, and carbon through real-world equivalents.

Electricity

0.0 iPhone 17 charges

(0.00 Wh)

Water

0.0 Water bottles (500mL)

(0.00 mL)

Carbon

0.0 Miles driven (sedan)

(0.00 g CO₂e)

If You Use This Daily...

Here's how the impact scales over one year if you use this prompt every day.

Electricity

0.0 iPhone 17 charges

(0.00 kWh)

Water

0.0 Water bottles (500mL)

(0.00 L)

Carbon

0.0 Miles driven (sedan)

(0.00 kg CO₂e)

Methodology & References

Environmental Baseline Constants

Metric	Value per Token	Source
Electricity	0.00045 Wh	Luccioni, Jernite & Strubell, 2023 — Power Hungry Processing (arXiv:2311.16863)
Water	0.004 mL	Li, Yang, Islam & Ren, 2023 — Making AI Less "Thirsty" (arXiv:2304.03271)
Carbon	0.0002 g CO₂e	IEA — Data Centres & Data Transmission Networks IEA — Energy and AI

Regional Grid Multipliers

Region	Electricity Multiplier	Water Multiplier	Carbon Multiplier
Global Grid Average	1.00x	1.00x	1.00x
US East (Northern Virginia)	1.25x	1.20x	1.25x
US West (Oregon)	0.50x	0.80x	0.50x
Europe (France)	0.15x	1.10x	0.15x
Europe (Ireland)	0.85x	0.90x	0.85x

Human-Scale Equivalence Metrics

Equivalence	Baseline
iPhone 17 Full Battery Charge	14.5 Wh
Standard Water Bottle Capacity	500 mL
Average Gasoline Sedan Emissions	~200 g CO₂e per mile
Global Daily AI Prompts (Baseline)	2.5 billion

Methodology Notes

This calculator uses averaged peer-reviewed baselines and regional grid profiles. Actual prompt impact depends on model architecture, server utilization, and local data center practices.

Values are standardized per token and use averaged regional multipliers rather than live telemetry.
Water estimates reflect blended direct and upstream effects, not a single facility’s cooling system.
These results are intended as an order-of-magnitude estimate and conversation starter, not a precise carbon ledger.

If you have more granular infrastructure data or model-specific benchmarks, open an issue on GitHub to help refine the model.

Global Real-Time Context

For reference: ~2.5 billion AI prompts run daily worldwide. Here's the aggregate environmental cost in real-time — independent of the prompt you entered.

Global Prompts Generated

Grid Electricity Drawn

0.00 kWh

Water Evaporated

0.00 liters

Carbon Emitted

0.00 kg CO₂e