Focus · AI and climate
Scenarios and uncertainties
The calculation methods, which put the same question at anywhere from 0.012 to 1.14 gram of CO2e.
By Projet Celsius
AnalysisWhy do estimates of AI's impact differ?Read the analysis →
Focus areaAI emissions1.14gCO2efor a Le Chat answer of about one page, over the full life cycleMistral AI, July 2025Open the focus area →You are hereScenarios and uncertaintiesThe calculation methods, which put the same question at anywhere from 0.012 to 1.14 gram of CO2e.
Focus area2030 trajectories458gCO2e/kWhfor one kWh of electricity in the world in 2025, over the life cycleEmberOpen the focus area →The Cordemais thermal power station, in Loire-Atlantique, seen from the Loire. Photo Liberliger, Wikimedia Commons, CC BY-SA 4.0
Charts and maps of the focus area
Charts and maps →Focus area
ChartOne question · logOne question, 4 ways of counting CO2e0.3 Wh, France0.012Google, Gemini0.030.3 Wh, world0.14Mistral, Le Chat1.14 g92timesgap between the lowest and the highestEpoch AI; Ember; Google; Mistral AIChartCompany · logElectricity emissions, by contract or by gridMetaGoogleMicrosoftdepending on the contractsdepending on the grid1 to 4,400the gap at MetaMeta, 2025; Google and Microsoft, 2026Chart2019 to 2023 dataThe digital sector in global emissionsThe Shift Project3.5 %GreenIT3.4 %Freitag et al.2.1 to 3.9%Malmodin et al.1.4 %Shift Project; GreenIT; Freitag et al.; Malmodin et al.
Analyses and articles
Analyses
Where does this figure come from?
Climate20 October 2026"Meta's electricity emits only 1,358 tonnes of CO2e"In 2024, Meta reports 1,358 tonnes of CO2e for the electricity it buys, the annual footprint of 170 people in France. The same report also gives 5.97 million tonnes. Both of the 2 figures are accurate.AnalysisClimate12 November 2026Why do estimates of AI's impact differ?Calculated with French electricity, one question to an AI emits 0.012 gram of CO2e; Mistral AI publishes 1.14 gram, more than 90 times as much. Its water use ranges from 0.26 to 45 millilitres depending on the published measurements. These figures are accurate; they do not cover the same scope.On the Projet Celsius blog ↗
Sources for this focus area
- Meta, 2025 Environmental Data Index, October 2025 ↗
- Google, 2026 Environmental Report, June 2026 ↗
- Microsoft, 2026 Environmental Sustainability Report, Data Fact Sheet, July 2026 ↗
- Epoch AI, How much energy does ChatGPT use?, February 2025 ↗
- Ember, Global Electricity Review 2026, and annual data via Our World in Data ↗
- Elsworth et al. (Google), Measuring the environmental impact of delivering AI at Google scale, August 2025 ↗
- Mistral AI, with Carbone 4 and ADEME, Our contribution to a global environmental standard for AI, July 2025 ↗
- The Shift Project, Impact environnemental du numérique : tendances à 5 ans et gouvernance de la 5G, March 2021 ↗
- GreenIT, Impacts environnementaux du numérique dans le monde, February 2025 ↗
- Freitag et al., The real climate and transformative impact of ICT, Patterns, 2021 ↗
- Malmodin et al., ICT sector electricity consumption and greenhouse gas emissions - 2020 outcome, Telecommunications Policy, 2024 ↗
Across the hub
AI Impact DatabaseAll the figures, their methodology and their sources.
AI Ranking10 assistants rated out of 100: performance, ethics and environment.
AnalysesThe signed pieces: analyses, viewpoints, "Where does this figure come from?".
Charts and mapsThe reference visuals, with their data and sources.
ToolsOpen calculations, your assumptions on the left, the result on the right.
A question about a figure?
An error, a source to add, a methodology question, a press request: write to us. Guillaume Pakula, Clément Reynaud, Sébastien Pierfederici and Alexandre Veilly will reply.
Write to us →
Projet Celsius, which runs this hub, also helps companies measure and reduce the footprint of their AI use.
Accounting for AI in your Bilan Carbone® ↗
