Close Menu
Greentimes
    Qoui de neuf

    MOGA Festival 2025: A study estimates over 51.5 million MAD in economic benefits for Essaouira

    28 January 2026

    FM6SS and Nareva Services Strengthen Their Partnership with a 1.2 MW Solar Power Plant in Dakhla

    28 January 2026

    Rural Development: Morocco Strengthens Its Role as a Strategic Partner of IFAD

    28 January 2026
    Facebook X (Twitter) Instagram
    GreentimesGreentimes
    • Home
    • CSR
    • Energy Transition and Renewables
    • Sustainability
    • Climate Change
    • Analyses and Opinions
    • Datas
    • Dates & events
    • Last news
    • FR
    Wednesday 28 January 2026
    Facebook X (Twitter) Instagram
    Greentimes
    Home » ChatGPT: One training session equals three days of a city’s consumption
    Sustainability

    ChatGPT: One training session equals three days of a city’s consumption

    22 September 2025No Comments2 Mins Read
    Facebook Twitter Pinterest LinkedIn Tumblr Reddit WhatsApp Email
    Share
    Facebook Twitter LinkedIn Pinterest WhatsApp Email

    Artificial intelligence is not only powerful but also extremely energy-hungry. ChatGPT, in particular, requires massive amounts of electricity, both to process the millions of daily requests from its users and for its training.

    According to the study How Hungry is AI?, the daily requests of ChatGPT account for approximately 2.5 billion interactions, equating to nearly 7,832 GWh per year, the equivalent of the annual consumption of 1.6 million French households.

    However, the most significant environmental impact lies in the training of the models, which is essential for enabling them to comprehend data and generate relevant responses. The initial training of a model like GPT-4 would have consumed nearly 50 GWh, which is as much as San Francisco uses in three days or the annual consumption of about 4,000 French households. In more tangible terms, this would correspond to charging a smartphone 3.3 billion times.

    This consumption is not limited to training. Running ChatGPT in environments such as Google Search requires hundreds of thousands of high-performance servers, further exacerbating the ecological impact. The precise calculation remains difficult, as tech giants share little data on the actual consumption of their data centers.

    These figures highlight the major environmental challenges associated with artificial intelligence, urging a reconsideration of how these technologies are developed and deployed.

    artificial intelligence ChatGPT daily queries data training ecological challenges electricity usage energy consumption environmental impact GPT-4 Technology
    Previous ArticleWorld Cleanup Day: Khénifra Launches Morocco’s First Citizen Mobilization
    Next Article Climate: Africa bets on ambitious national plans to boost growth and employment

    Related Posts

    FM6SS and Nareva Services Strengthen Their Partnership with a 1.2 MW Solar Power Plant in Dakhla

    28 January 2026

    Rural Development: Morocco Strengthens Its Role as a Strategic Partner of IFAD

    28 January 2026

    OCP: Washington Reopens the Issue of Taxes on Moroccan Fertilizers

    27 January 2026
    Leave A Reply Cancel Reply

    Ne ratez pas
    CSR

    MOGA Festival 2025: A study estimates over 51.5 million MAD in economic benefits for Essaouira

    28 January 20260

    The MOGA Festival has published the results of a study focused on the socio-economic impact…

    FM6SS and Nareva Services Strengthen Their Partnership with a 1.2 MW Solar Power Plant in Dakhla

    28 January 2026

    Rural Development: Morocco Strengthens Its Role as a Strategic Partner of IFAD

    28 January 2026

    Natural Gas: How Morocco is Redrawing its Energy Map After the End of Algerian Flows

    28 January 2026
    Restez connecté
    • Facebook
    • Twitter
    • Pinterest
    • Instagram
    • YouTube
    • Vimeo
    Nos vidéos

    The circular economy at the heart of LabelVie group’s initiatives.

    10 July 2025

    Mounir El Bari: “Our great ecological challenge is access to the resource!”

    10 July 2025

    Driss Nahya: “Control must be strengthened to access the waste deposit.”

    10 July 2025

    Reda Boukallal: “Priority to the valorization of household waste”

    10 July 2025
    Facebook X (Twitter) Instagram Pinterest
     
    © 2026 Green Times.

    Type above and press Enter to search. Press Esc to cancel.