"To spit out a five-second clip, the researchers found that it takes the equivalent of running a microwave for over an hour."
...
" the energy demands of text-to-video generators quadruple when the length of a generated video doubles — indicating that the power required for increasingly sophisticated generations doesn’t scale linearly."
Makes sense as they have to track the state of all frames I guess to maintain coherence.
https://futurism.com/future-society/ai-power-usage-text-to-video-generator
Koen Hufkens, PhD
npub1nl…uzenl
2025-09-26 18:23:42 UTC
Author Public Key
npub1nl4qthywk2gy8ez0mlthmjwknk9cpnaq67haj3f4nxeeydmq0pfsduzenlSeen on
wss://relay.momostr.pinkPublished at
2025-09-26 18:23:42 UTCEvent JSON
{
"id": "d9d06b4a8beddf57397dd9a5011c5433be3c10db8c505925dd74dd69729bc169",
"pubkey": "9fea05dc8eb29043e44fdfd77dc9d69d8b80cfa0d7afd9453599b39237607853",
"created_at": 1758911022,
"kind": 1,
"tags": [
[
"proxy",
"https://mastodon.social/@koen_hufkens/115271992748766726",
"web"
],
[
"proxy",
"https://mastodon.social/users/koen_hufkens/statuses/115271992748766726",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://mastodon.social/users/koen_hufkens/statuses/115271992748766726",
"pink.momostr"
],
[
"-"
]
],
"content": "\"To spit out a five-second clip, the researchers found that it takes the equivalent of running a microwave for over an hour.\" \n... \n\" the energy demands of text-to-video generators quadruple when the length of a generated video doubles — indicating that the power required for increasingly sophisticated generations doesn’t scale linearly.\"\n\nMakes sense as they have to track the state of all frames I guess to maintain coherence.\n\nhttps://futurism.com/future-society/ai-power-usage-text-to-video-generator",
"sig": "6cb27776edc9d21edea955b47b143fad3b1102d3cdd3c8c1e4199800ec22305463e0bf0760b0f0ba02206bf67b2fdc9fe09863784b7a0fa8fc78515d4d4f8ed7"
}