In short
One ordinary text question to a chatbot uses a fraction of a watt-hour on published figures, roughly what an oven uses in about a second. Images, video, long documents and reasoning modes can use many times more, and the global total is growing fast, so both "it's nothing" and "it's a disaster" miss the point.
Steps
- Use published per-prompt figures as a rough guide, not an exact count.
- Compare like with like, such as energy per question against energy per minute of another activity.
- Count the heavy uses separately, especially images, video and long documents.
- Remember that totals matter because of scale, even when each use is small.
Per question, small; per image or video, much larger; in total, significant. Holding all three ideas at once is the honest way to think about AI's energy.
How does one question compare?
| What | Energy | Source |
|---|---|---|
| Median Gemini text prompt | 0.24 Wh | Google (2025) |
| Average ChatGPT query | about 0.34 Wh | OpenAI (2025) |
| One short AI video from an open model | about 90 Wh | Delavande, Pierrard and Luccioni (2025) |
OpenAI compares its figure to an oven running for a little over one second, or a high-efficiency light bulb for a couple of minutes. The video figure comes from tests of one open text-to-video model on research hardware; the same study found a spread of over a thousandfold between models.
Why does the total still matter?
Because of scale. The International Energy Agency estimates data centres used about 415 TWh of electricity in 2024, around 1.5% of the world's electricity, and projects about 945 TWh by 2030, with AI the most important driver. Billions of small requests, plus the growth of images, video and AI agents, add up.
Common mistakes
- Comparing different boundaries. Some figures count only chips; others include cooling and idle machines.
- Using old numbers. Google reports its median prompt's energy fell 33-fold in one year, so a figure from a few years ago may be far off.
- Ignoring the type of task. "AI" covers everything from a short reply to a minute of video.
To estimate your own use, see work out your own AI carbon footprint. For the big picture, read why green AI matters now.
Sources
- The gentle singularity, Sam Altman (OpenAI) (2025)
- Measuring the environmental impact of delivering AI at Google scale, Elsworth et al., Google (2025)
- Video killed the energy budget: characterizing the latency and power regimes of open text-to-video models, Delavande, Pierrard and Luccioni (arXiv 2509.19222) (2025)
- Energy and AI, International Energy Agency (2025)
Related
How do I work out my own AI carbon footprint?
A rough personal AI footprint: count your text prompts, images and long tasks, multiply by published per-request figures, and treat it as a range.
How do I check a claim about AI's energy use?
Six questions to test any claim about AI energy, water or carbon: per what, who says so, what's included, how old, compared with what, total or per use.
Why green AI matters now
AI is driving rapid growth in data-centre electricity demand, straining grids and raising costs. The evidence on scale, energy, cost and policy in 2026.
Green AI: a plain-English guide for everyone
What using AI really costs in energy and water, which everyday habits matter and which don't, how to judge claims, and how to have your say.