Image Generation

DALL-E 3 Environmental Impact

HeavyEstimated

OpenAI's image generation model

Architecture
Diffusion Transformer
Provider
OpenAI
3.0 Wh
Energy per query
1.2 g
CO₂ per query
15 mL
Water per query
10x more than
vs Google search

Energy per query

3.0 Wh

10x more than a Google search (0.3 Wh)

CO2 per query

1.2 g

US East (Virginia) grid (450 gCO₂/kWh)

Water per query

15 mL

~67 queries to fill 1 litre

Processing location

Azure US East / Sweden

Provider

OpenAI

Category

Image Generation

Grid carbon intensity

450 g CO2/kWh (25% renewable)

How does DALL-E 3 compare?

Ranked #94 of 152 models by energy per query

0 Wh2 Wh4 Wh6 Wh8 WhImagen 3Stable DiffusionXLDALL-E 3MidjourneyGoogle search (0.3 Wh)

Detailed Breakdown

Energy Consumption

DALL-E 3 consumes approximately 3-11 Wh per image depending on resolution and complexity. We use 3 Wh as a conservative baseline for standard 1024x1024 output. Higher resolutions, inpainting, and variations may consume significantly more. Image generation is inherently more compute-intensive than text because diffusion models perform dozens of iterative denoising steps across a high-dimensional pixel space, with each step requiring a full forward pass through the neural network.

Power Source & Carbon

DALL-E 3 runs on Microsoft Azure, sharing infrastructure with other OpenAI models in US East (Virginia) and Sweden. The Virginia grid's relatively high carbon intensity means each AI-generated image carries a larger carbon footprint than if the same computation happened in a hydro- or nuclear-powered region.

Water Usage

At approximately 15 mL per image, DALL-E 3's water use comes from cooling the GPU clusters during the multi-step diffusion process. Generating 100 images would consume about 1.5 liters of water.

About DALL-E 3

Every AI-generated image has a physical cost that most users never see. DALL-E 3 uses roughly 3.4 Wh per image — enough energy to fully charge a smartphone. That places it in the middle of the image generation pack: more efficient than Midjourney's estimated 4.2 Wh but far more intensive than any text query. The jump from text to image represents one of the sharpest efficiency cliffs in AI: generating a single picture consumes 8-10x the energy of an entire ChatGPT conversation.

These figures are estimates derived from hardware specifications and API benchmarks — OpenAI has not published official energy data for DALL-E 3. Actual consumption may vary significantly depending on batching, quantisation, and infrastructure optimisations that we cannot observe from outside.

DALL-E 3 in Context

300 MWh
estimated daily

At global scale

With an estimated 10M+ daily users averaging 10 queries each, DALL-E 3 consumes roughly 300 MWh of electricity per day — enough to power 10000 homes.

27.4 kWh
per year

Your yearly DALL-E 3 footprint

At 25 queries per day, your annual DALL-E 3 usage consumes 27.4 kWh — roughly what a fridge uses in a month. That produces 10.9 kg of CO₂.

10x
vs a text query

Image vs text: the energy gap

Generating one image with DALL-E 3 uses 10x more energy than a text conversation with ChatGPT. The diffusion process that creates images requires many iterative passes through the model, each refining the image from noise — far more compute per output than generating text one token at a time.

What does your DALL-E 3 usage cost the planet?

Use our calculator to estimate your personal environmental footprint based on how often you use DALL-E 3.

Calculate My Compute

Frequently Asked Questions

How much energy does DALL-E 3 use per query?

Each DALL-E 3 query consumes approximately 3.0 Wh of energy. This is 10x more than a traditional Google search (~0.3 Wh).

What is DALL-E 3's carbon footprint?

Based on the carbon intensity of Azure US East / Sweden, each query produces approximately 1.2 g of CO2. The grid in this region has a carbon intensity of 450 g CO2/kWh with 25% renewable energy.

How much water does DALL-E 3 use?

Each query consumes approximately 15 mL of water, primarily used for cooling the data centers that process the request.

How does DALL-E 3 compare to a Google search?

A DALL-E 3 query uses 10x more than a Google search in terms of energy. A Google search uses approximately 0.3 Wh, while DALL-E 3 uses 3.0 Wh.

Technical Details

Architecture

Diffusion Transformer

Release date

2023-10-03

Open source

No