520
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 15 Feb 2026
520 points (98.9% liked)
Technology
81209 readers
4111 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
Haven't seen anyone else ask this question, so I will. What on earth does an AI data center need all of this storage for?
The only significant use of AI is for text generation to my knowledge. Video gen is the only thing that takes up any significant space, and current models can only produce short video clips before they go off the rails. Also, very few people are interested in video gen. It's an expensive toy without much real world utility. Is there something I'm missing? Are these AI companies planning to scrape every video off the net and store them independently for training?
Booking out this much HDD capacity would only make sense to me if 5 TikTok or YouTube competitors all came onto the scene at once. Not AI. AI needs fast, parallelized compute and high performance memory to hold the models it's running. Text slop requires negligible storage.
Caching. I'm guessing they download the internet to large storage arrays and then run neural network learning on them. Its probable they download all videos too and transcribe them later. So its not so much used for hosting content, but for teaching models.
They want to log ALL user data in the world. Every convo you ever had with the ai, in text or audio, images video, etc.
It may be a little tinfoil hat like, but you cannot convince me that these companies are shoving AI in literally everything, buying all the hardware in existence, and building data centets on land that no one wants them at, just to "make a better ai for the consumer." I believe this is an attempt at hardcore tracking and surveillence.
I think it's a combination of that and the worry that there will be one winning ubercorp that practically merges with the US Government.
I mean, they are all pushing all their chips in at the same time. It's like they know it's now or never.
And govs are pushing it because it's yet another arms race.
The logical course of action is to collectively just not do AI. But you can't prevent your adversaries from doing such, so it ends up being like some sort of prisoner's dilemma.
Even if they didn't, they probably don't want to seem like they're falling behind, so once one person goes all in, so do the others.
Unfortunately the most refined usage, and money dumped sector of Ai by far is image recognition
Its where all the money from Amazon and others has been dumped (Flock Surveillance, Amazon Ring, FedEx Trucks, Wal~Mart and your choice of store)
Surveillance, ALPR, Facial recognition, gait recognition, etc. It takes a massive amount of data
Go see how much space you need if you want to secure data from 10,000 cameras at 480p/720p/1080p for a few weeks, let alone a year or two
Gotcha, wasn't thinking about surveillance. We are the social credit system – China ain't got shit on our authoritarian future.