---
title: "What is the carbon footprint of AI and ChatGPT? How much does it weigh in a Bilan Carbone®?"
description: "A ChatGPT query uses about 0.3 Wh, like a Google search: the old 2-3 Wh estimate was divided by 10. AI weighs under 1% of a service-sector SME's footprint."
canonical: https://projetcelsius.com/en/blog/chatgpt-carbon-footprint/
author: "Guillaume Pakula"
datePublished: 2026-02-21
dateModified: 2026-10-24
language: en
translationOf: https://projetcelsius.com/blog/empreinte-carbone-ia/
---

# What is the carbon footprint of AI and ChatGPT? How much does it weigh in a Bilan Carbone®?

**Author**: Guillaume Pakula
**Published**: 2026-02-21
**Updated**: 2026-10-24
**Section**: Explained
**Canonical URL**: https://projetcelsius.com/en/blog/chatgpt-carbon-footprint/
**French original**: https://projetcelsius.com/blog/empreinte-carbone-ia/
**Source**: Projet Celsius (https://projetcelsius.com/en/), a climate strategy consultancy of engineers and PhDs based in Paris and Marseille.

> A **ChatGPT** query uses **about 0.3 Wh**, the same order of magnitude as a Google search: the widely repeated "10 times more" rested on an outdated figure that recent measurements have divided by 10. For an SME with 200 employees, AI accounts for less than **1%** of its [Bilan Carbone®](https://projetcelsius.com/en/blog/bilan-carbone-method-explained/) (the French carbon accounting method). Within digital, the largest item is the **manufacturing of equipment**: extending the lifespan of a PC fleet from 4 to 6 years cuts the annualised manufacturing footprint by about **33%**.

## Key takeaways

- **A ChatGPT query uses about 0.3 Wh**, the same order of magnitude as a Google search.
- **AI weighs less than 1% of the carbon footprint** of an SME, even with intensive use.
- **The real digital lever** is making equipment last, which cuts the annualised manufacturing footprint by about 33%.
- **Hosting in France or Scandinavia** cuts emissions by a factor of 5 to 10.

[Google's 2026 environmental report shows emissions (in French)](https://projetcelsius.com/ia/graphiques/emissions-google-microsoft/) **81% higher than in 2019**, including an 18% rise in 2025 alone, driven by the construction of data centres for AI. This is the most visible effect of AI on the climate, along with [the water AI consumes](https://projetcelsius.com/en/blog/ai-water-footprint/) to cool these facilities. The question is what this means for a company that uses these tools.

In most cases, it matters far less than the headlines suggest, as long as you think in [orders of magnitude](https://projetcelsius.com/en/blog/company-carbon-footprint-main-emission-sources/) and compare AI with the other emission sources in the footprint.

## An AI query uses about as much energy as a Google search

For a long time, a query to an AI model was said to use **2 to 3 Wh**, or 10 to 15 times as much as a Google search. That figure is now out of date: recent analyses have **divided it by about 10**, and according to [Epoch AI](https://epoch.ai/gradient-updates/how-much-energy-does-chatgpt-use) the old estimate was about 10 times too high. For a model-by-model breakdown, the [ChatGPT, Claude and Gemini comparison](https://projetcelsius.com/en/blog/chatgpt-vs-claude-vs-gemini-environmental-impact/) ranks the published figures.

A [short text query](https://huggingface.co/blog/sasha/ai-environment-primer) to ChatGPT, Gemini or Claude uses about **0.3 Wh**. Google measured a median Gemini query at ~0.24 Wh ([arXiv study 2508.15734](https://arxiv.org/abs/2508.15734)) and Sam Altman puts ChatGPT at ~0.34 Wh. A benchmark of GPT-4o, a model withdrawn from ChatGPT in February 2026, gives ~0.43 Wh. An AI query and a Google search are therefore of the same order of magnitude, nowhere near a factor of 10. For more on this comparison, see [ChatGPT query vs Google search](https://projetcelsius.com/en/blog/chatgpt-vs-google-search-energy/).

- AI query (generative AI, short text): ~0.3 Wh, not the 2-3 Wh that used to be quoted
- Standard Google search: same order of magnitude, ~0.3 Wh
- [Life cycle of a large model (in French)](https://projetcelsius.com/ia/graphiques/acv-mistral-large-2/): 20,400 tCO2e for Mistral Large 2, from training to January 2025, including 18 months of use, according to the [life cycle assessment (LCA)](https://mistral.ai/news/our-contribution-to-a-global-environmental-standard-for-ai) that Mistral published with Carbone 4 and ADEME (the French Agency for Ecological Transition)
- Data centres worldwide: ~1.5% of global electricity consumption in 2024, projected to reach ~3% in 2030 (base case of the [IEA](https://www.iea.org/reports/energy-and-ai), the International Energy Agency)

## How much does AI contribute to a company's carbon footprint?

For an **SME with 200 employees** that uses AI tools intensively, at 100 queries per employee per day (which is already a lot), the annual footprint of that use stays **below 1 tonne of CO2**. That is not much next to the main emission sources:

- Purchases and subcontracting ([Scope 3](https://projetcelsius.com/en/blog/scope-3-emissions-guide/)): often 40-70% of the total footprint
- Building energy: often 10-20%
- Business travel: an item that grows quickly with air travel, counted in Scope 3 for all employees, [executives included](https://projetcelsius.com/en/blog/ultra-rich-climate-429-tonnes/)
- IT equipment (manufacturing): [70-80% of the digital carbon footprint](https://projetcelsius.com/en/blog/corporate-digital-carbon-footprint/)
- AI and cloud: less than 1% for most SMEs

> **Order of magnitude** For most SMEs, the footprint of AI amounts to **less than one Paris-New York flight** a year. It is something to monitor over time, without urgency, which our [AI carbon footprint calculator](https://projetcelsius.com/en/tools/ai-carbon-footprint-calculator/) quantifies from your volumes and which other providers also measure, compared in our article on [who can measure the impact of AI in a company](https://projetcelsius.com/en/blog/measure-ai-footprint-in-your-company/). On the digital side, the most effective lever remains extending the lifespan of equipment: going from 4 to 6 years cuts the annualised manufacturing share of the [digital carbon footprint](https://projetcelsius.com/en/blog/website-carbon-footprint/) by about 33%.

## Will AI become a larger emission source in future?

In many companies, AI use **doubles every year**. For now, efficiency gains in models do not offset that growth, because they are absorbed by the increasing size of models. If these 2 trends continue, digital could become a significant emission source within **5 to 10 years**.

Including AI in your [Bilan Carbone®](https://projetcelsius.com/en/blog/bilan-carbone-method-explained/) (the French carbon accounting method) now gives you a **baseline** against which to measure this change. Methodologies exist, although they are still young: the [ADEME](https://www.ademe.fr/) digital reference framework and the database of the French organisation Boavizta.

## 5 levers to reduce your company's digital carbon footprint

In descending order of impact:

- Extend the lifespan of equipment from 4 to 6 years: the annualised manufacturing footprint falls by **about 33%**. This is the number 1 lever, and it saves money.
- Include refurbished equipment in the IT purchasing policy: a refurbished laptop avoids about 77% of the annual climate impact of a new one, and 73% for a desktop (ADEME, the French Agency for Ecological Transition, 2022)
- Choose a cloud hosting provider in France or Scandinavia: emissions are **5 to 10 times lower** than with a US data centre powered by natural gas
- Use AI selectively: a chatbot that answers questions the FAQ already covers wastes energy. Reserve AI for high-value tasks
- Choose smaller models when the task allows it: a small model uses 5 to 10 times less energy than a frontier model on simple tasks

## Key takeaways

For most organisations, AI has a real but still marginal carbon cost. Within digital, manufacturing equipment weighs far more than the use of cloud services; what needs watching in the coming years is how AI use develops.

- An AI query uses about as much energy as a Google search (~0.3 Wh each): the old factor of 10 has not survived recent measurements, and usage remains marginal at the scale of a company's carbon footprint.
- AI accounts for **less than 1% of the total footprint** of most SMEs, even when it is used intensively
- The most effective digital lever is extending the lifespan of equipment: going from 4 to 6 years cuts the annualised manufacturing footprint by about 33%.
- The choice of data centre matters: hosting in France or Scandinavia emits 5 to 10 times less than hosting in the United States, for the electricity consumed.
- Google's emissions have risen by 81% since 2019. The issue is real at the macro level, even if it is not yet a priority for an SME.

> **Need to include digital in your carbon footprint assessment?** We include digital emissions (equipment + cloud + AI) in our [Bilan Carbone® assessments](https://projetcelsius.com/en/blog/bilan-carbone-cost-france/) using ADEME and Boavizta methodologies. We give you a baseline and reduction levers in proportion to the real stakes. [Contact us](https://projetcelsius.com/en/contact/) for a first conversation.

## Frequently asked questions

## Frequently asked questions

### How much energy does a ChatGPT query use?

About **0.3 Wh** for a short text query. The long-standing estimate of 2-3 Wh (10 times a Google search) has been **divided by about 10** by recent analyses: Google measured a median Gemini query at ~0.24 Wh, and Sam Altman puts ChatGPT at ~0.34 Wh. An AI query and a Google search are now of the **same order of magnitude**.

### Is AI a priority emission source in a company's carbon footprint?

No, not for most companies. The priorities remain purchases ([Scope 3](https://projetcelsius.com/en/blog/scope-3-emissions-guide/)), energy and travel. AI accounts for **less than 1%** of the footprint in most cases. It is something to include and monitor, without urgency.

### How do you calculate the carbon footprint of AI tools?

ADEME's digital reference framework and the Boavizta database provide emission factors for each type of cloud service. In practice, your cloud provider (AWS, Azure, GCP) also publishes emissions data by service. These figures go into your [Bilan Carbone®](https://projetcelsius.com/en/blog/bilan-carbone-method-explained/) under digital services in Scope 3.

### Should we stop using AI for climate reasons?

No. The aim is to use AI **selectively** rather than systematically, and to choose hosting providers with a low-carbon electricity mix. Giving up AI to save less than a tonne a year would be a poor trade-off as long as purchases, energy and business travel have not been addressed.

### Is a data centre in France lower-carbon than one in the US?

Yes, **5 to 10 times lower** for the electricity consumed, thanks to the French electricity mix (95% low-carbon in 2025 according to RTE, the French grid operator, mainly nuclear). It is a simple but rarely used lever: choose a hosting provider with data centres in France or Scandinavia.

### Can AI help reduce a company's emissions?

Yes, if it is used well: it can optimise logistics, analyse consumption data and model reduction scenarios. Its net effect depends on what it saves compared with what it consumes.

### What is the carbon footprint of generative AI?

The footprint arises in 2 phases: training, which happens once, and use (inference, meaning running a trained model to answer a query), which recurs with every query. The LCA published by Mistral puts Mistral Large 2 at 20,400 tCO2e, including training and the first 18 months of use. A short text query uses about 0.3 Wh, the same order of magnitude as a Google search: Google measured 0.24 Wh for a median Gemini query. Worldwide, data centres accounted for about 1.5% of electricity consumption in 2024, and the IEA projects a share of around 3% in 2030.

### How can a company reduce the carbon footprint of its AI use?

The first lever is targeted use: reserving AI for high-value tasks rather than using it by default (for example, to answer questions that an FAQ already covers). Next come a smaller model, where the task allows it, which uses 5 to 10 times less energy than a frontier model on simple tasks, and hosting in France or Scandinavia. The latter cuts emissions by up to a factor of 5 to 10 compared with a US data centre powered by gas. Within a company's digital carbon footprint, extending the lifespan of equipment from 4 to 6 years matters more: the annualised manufacturing footprint falls by about 33%.

## Sources

**Sources**: [Epoch AI - How much energy does ChatGPT use?](https://epoch.ai/gradient-updates/how-much-energy-does-chatgpt-use) (the old figure of ~3 Wh was about 10 times too high); [Google - Measuring the environmental impact of AI inference, arXiv 2508.15734](https://arxiv.org/abs/2508.15734) (median Gemini query ~0.24 Wh); [IEA - Energy and AI](https://www.iea.org/reports/energy-and-ai) (data centres ~1.5% of global electricity in 2024, ~3% in 2030); [Mistral - Life cycle assessment of Mistral Large 2](https://mistral.ai/news/our-contribution-to-a-global-environmental-standard-for-ai) (20,400 tCO2e, training and 18 months of use included); [ESG Dive, based on Google's 2026 environmental report](https://www.esgdive.com/news/googles-emissions-continue-to-climb-due-to-ai-buildout/824373/) (2025 emissions: +18% year on year, +81% compared with 2019).
