MIT Technology Review · Jun 3, 2025
MIT Technology Review on AI’s energy appetite and the questions beyond personal use
Markdown · JSON · Original source
MIT Technology Review’s newsletter, originally published June 3, 2025, describes a reporting project on AI’s energy use and emissions. The writer says estimates have “holes” and argues that chatbot questions, image generation and video creation are only a limited view of AI’s possible energy demands. The piece reports that, among the models tested, producing a short, low-quality AI video used far more energy than answering a chatbot question. The writer says newer video models may demand more energy, while noting that Google and declined to provide specific figures. The writer’s central point is that AI’s energy future is not predetermined: models, chips and cooling methods could become more efficient, and video’s impact will depend partly on how it is used and priced. The piece shifts attention from an individual’s chatbot use to larger choices about energy supply, data-center siting, water use and public support. It cites concerns about water drawn from Nevada aquifers and natural gas used for a large data center in Louisiana. The writer says leading AI companies provide little transparency about present and future energy demand even as they seek public support.
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
MIT Technology Review · Article
Source details
Original source date: Jun 3, 2025 (2025-06-03T09:00:00Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-13T20:38:23.251Z. First observed: 2026-10-30T11:50:22.631Z. Retrieved: 2026-10-30T11:51:57.057Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
Text as retrieved on October 30, 2026; originally published June 3, 2025. The writer says Google and did not provide specific energy-use figures; the piece therefore does not establish the energy demand of those models. The writer says the reporting’s measurements cover chatbot, image and video uses, not every possible AI use.
Open question
How will the way AI video is used and priced affect its energy demand and emissions?
Source-stated
MIT Technology Review’s newsletter writer says the reporting found gaps in common understanding of AI’s energy use.
After speaking with dozens of researchers, we realized that the common understanding of AI’s energy appetite is full of holes.
Source-stated
For the models tested, producing a low-quality, five-second AI video required much more energy than answering a chatbot question about a recipe.
It was 42,000 times more than the amount needed for a chatbot answer a question about a recipe
Source-stated
The writer says Google and OpenAI denied requests for specific energy-use numbers for their AI models.
Google (as well as OpenAI, which has its own video generator, Sora) denied our request for specific numbers on the energy their AI models use.
Source-stated
The writer argues that the future energy demand of AI is not predetermined because models, chips and cooling methods could become more efficient.
The models, chips, and cooling methods behind this AI revolution could all grow more efficient over time
Source-stated
The writer argues that energy sources, data-center impacts and company transparency deserve attention alongside individual AI use.
it’s not so much your footprint as these other factors that are keeping climate researchers and energy experts we spoke to up at night.
Source and contributor context
MIT Technology Review · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
OpenAI
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
Editorial provenance
Version 1. Publication ID: 33812b83-378c-ddfe-a145-a9dad0f5def2.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-13T20:33:48.257Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model, second_opinion_categorical.
Observation: 9d6d7e12a767b5cca002c149b195b8408fbf59c7feac6e75830683c12b45db7e. Source text hash: 423d34dbe64deeb4b5d950bd7de247073320afa1b47c3670f1eba25d306f4c0d.
Revision and source-check history
Version 1 · 2026-11-13T20:38:23.251Z
MIT Technology Review on AI’s energy appetite and the questions beyond personal use
MIT Technology Review’s newsletter, originally published June 3, 2025, describes a reporting project on AI’s energy use and emissions. The writer says estimates have “holes” and argues that chatbot questions, image generation and video creation are only a limited view of AI’s possible energy demands. The piece reports that, among the models tested, producing a short, low-quality AI video used far more energy than answering a chatbot question. The writer says newer video models may demand more energy, while noting that Google and declined to provide specific figures. The writer’s central point is that AI’s energy future is not predetermined: models, chips and cooling methods could become more efficient, and video’s impact will depend partly on how it is used and priced. The piece shifts attention from an individual’s chatbot use to larger choices about energy supply, data-center siting, water use and public support. It cites concerns about water drawn from Nevada aquifers and natural gas used for a large data center in Louisiana. The writer says leading AI companies provide little transparency about present and future energy demand even as they seek public support.
Original source
MIT Technology Review · Article
Source details
Original source date: Jun 3, 2025 (2025-06-03T09:00:00Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-13T20:38:23.251Z. First observed: 2026-10-30T11:50:22.631Z. Retrieved: 2026-10-30T11:51:57.057Z.
Limits
Text as retrieved on October 30, 2026; originally published June 3, 2025. The writer says Google and did not provide specific energy-use figures; the piece therefore does not establish the energy demand of those models. The writer says the reporting’s measurements cover chatbot, image and video uses, not every possible AI use.
Open question
How will the way AI video is used and priced affect its energy demand and emissions?
Source-stated
MIT Technology Review’s newsletter writer says the reporting found gaps in common understanding of AI’s energy use.
After speaking with dozens of researchers, we realized that the common understanding of AI’s energy appetite is full of holes.
Source-stated
For the models tested, producing a low-quality, five-second AI video required much more energy than answering a chatbot question about a recipe.
It was 42,000 times more than the amount needed for a chatbot answer a question about a recipe
Source-stated
The writer says Google and OpenAI denied requests for specific energy-use numbers for their AI models.
Google (as well as OpenAI, which has its own video generator, Sora) denied our request for specific numbers on the energy their AI models use.
Source-stated
The writer argues that the future energy demand of AI is not predetermined because models, chips and cooling methods could become more efficient.
The models, chips, and cooling methods behind this AI revolution could all grow more efficient over time
Source-stated
The writer argues that energy sources, data-center impacts and company transparency deserve attention alongside individual AI use.
it’s not so much your footprint as these other factors that are keeping climate researchers and energy experts we spoke to up at night.
Source and contributor context
MIT Technology Review · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
OpenAI
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
Editorial provenance
Version 1. Publication ID: 33812b83-378c-ddfe-a145-a9dad0f5def2.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-13T20:33:48.257Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model, second_opinion_categorical.
Observation: 9d6d7e12a767b5cca002c149b195b8408fbf59c7feac6e75830683c12b45db7e. Source text hash: 423d34dbe64deeb4b5d950bd7de247073320afa1b47c3670f1eba25d306f4c0d.
uncertain · 2026-11-13T20:33:48.257Z · anthropic/claude-opus-5.5
Publication: 33812b83-378c-ddfe-a145-a9dad0f5def2
attribution — pass: The reading attributes the piece to MIT Technology Review's newsletter and its writer, matching the source's statement that it appeared in The Algorithm. It is one first-person piece, not a listing. Claims stay attributed to the writer.
This story originally appeared in The Algorithm, our weekly newsletter on AI.
summary — uncertain: Calling 'not predetermined' the writer's central point misframes the piece. It has three takeaways, and the strongest ending argument is that energy sources and transparency matter more than personal footprint. The summary also flattens the force of the video finding. 'Far more energy' omits 42,000 times and the microwave comparison, and drops the writer's warning about video. The writer's view that Veo and Sora are 'very likely' much more energy-demanding is softened to 'may'. Smallest repair: drop the 'central point' label and restore the 42,000x figure and the writer's emphasis.
If there’s one type of AI whose energy appetite should worry you, it’s this one.
date — pass: The reading states June 3, 2025, matching the UTC day of originalPublished. That date is also the URL path. The page was modified the same day. The Deep Dive sidebar has later items, such as the 'ten years after AlphaGo' piece. These are likely later page additions, but the reading neither uses nor relies on them. Forecasts are not presented as outcomes. A publisher page needs no archive disclosure.
support — pass: Each claim's excerpt is an exact source passage that supports it. The claims cover the holes in common understanding, the video versus chatbot comparison, Google and OpenAI refusals, possible efficiency gains, and wider factors beyond personal footprint. The question about video pricing follows from the writer's stated view. The limitations accurately note the missing company figures.
I think the key to whether the use of AI video will produce indefensible clouds of emissions in the near future will be how it’s used, and how it’s priced.
media — pass: No media is selected and no verified media candidates were supplied, so there is no media identity or relevance issue.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAIOpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAIOpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAIOpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAI