see that’s crap because they need muni clean water; it’s gotta come from SOMEWHERE, and your muni water system wasn’t setup to feed a city AND a bunch of datacenters. so it’s gonna have to expand - and who’s gonna pay for that? NOT THE AI. The people will pay for it, every fucking time.
Agriculture uses so much more water. Just so we can grow corn so it can be turned into ethanol. “AI” uses so much more energy though. and it’ll only get worse, the more “advanced” the llms become, the more energy they’re going to require.
it typically uses non-potable water though, and that water is then typically evaporated as steam which then in precipitated back down, or it’s directly routed back into the stream it came from, which is the case in hydroelectric systems
we don’t have infinite potable water. it’s going to come from municipal supplies. they’re going to stick the cost to these communities, and probably stress cities that are already water-fucked even more:
I suggest you read the article I posted, or if you don’t feel like reading, maybe listening to what’s being said in the YouTube video.
There are some very real concerns with everything happening around AI. Water is quite simply not one of them. Focus your energy on criticisms based in reality.
I’d already seen Green’s video. I agree it’s a small thing compared to the power, but in a world obviously running out of easily accessible potable water, it’s only going to exacerbate problems for the communities that are already being hit by the relentless power costs due to AI as well.
This isn’t a one or the other thing. They’ll need water for cooling. It’s not recycled. it’s got to be clean, potable water.
And it depends on the area. Dude with the blog post isn’t considering the many areas with real scarcity. The water, it’s pollution, etc. are a very real threat to many ecosystems and people.
I mean, I’m certain that if you consider there to be many reputable scientific publications having studies to the contrary, you should at least be able to post a single one, right?
Or at least couldn’t make it to section V. They mentioned the common claim, for instance, that GPT 3 alone took three quarters of a million gallons of water to train (roughly the footprint of 3-400) cars.
But, please, keep congratulating yourself. I was selection based on the criteria of (1) easy to read (2) multiple environmental issues listed.
I’m 4 pages into a paper on the subject. Want something on air pollution, check out grok in Memphis. Something about water table distruction, check the rural New York one.
If you want me to cite something obscure like some strangeness between weak and strong NP-Completeness? Sure. That makes sense.
If someone wants me to cite a source regarding whether prime numbers are real I’m just going to tell them to stop being stupid and ask Google.
That is how simple this one is. I have hundreds of citations, both scientific and mass media. They took almost zero effort to find.
I would challenge you to avoid being lazy and stupid.
am intimately familiar with the shitshow in Memphis.
searching new york data center water table destruction gives some links to stories related to datacenter water use but nothing specific to NY or AI. duckduckgo and google (non ai) searched.
I would challenge your assumptions but you’re just an asshole. Have a great life.
The water-thing is mostly bullshit, see https://andymasley.substack.com/p/empire-of-ai-is-wildly-misleading / https://www.youtube.com/watch?v=H_c6MWk7PQc
see that’s crap because they need muni clean water; it’s gotta come from SOMEWHERE, and your muni water system wasn’t setup to feed a city AND a bunch of datacenters. so it’s gonna have to expand - and who’s gonna pay for that? NOT THE AI. The people will pay for it, every fucking time.
Agriculture uses so much more water. Just so we can grow corn so it can be turned into ethanol. “AI” uses so much more energy though. and it’ll only get worse, the more “advanced” the llms become, the more energy they’re going to require.
And creating energy is also using large amounts of water.
it typically uses non-potable water though, and that water is then typically evaporated as steam which then in precipitated back down, or it’s directly routed back into the stream it came from, which is the case in hydroelectric systems
The water that’s returned is significantly warmer though, which can have some pretty large effects on the local environment.
that is true
Energy use is a more real concern, yes.
there are entire cities facing water crisis already. water is going to be a huge concern going forward.
https://news.arizona.edu/news/phoenix-sees-increase-people-living-without-running-water-new-analysis-shows
https://www.forbes.com/sites/kavehmadani/2025/12/03/6-big-questions-about-irans-water-crisis-and-tehrans-day-zero/
https://wetravelwebond.com/13-u-s-cities-on-the-brink-of-a-water-crisis/
don’t believe the ‘it’s only a tiny bit of water’ bullshit
Most of the water used by “AI” is in energy generation though, so you’re both right.
to me it’s simply not an either:or it’s an yes:and
the utility customers of these municipalities aren’t being told about the construction of AI data centers, because so many come with NDA strings.
https://www.nbcnews.com/tech/tech-news/data-center-ai-google-amazon-nda-non-disclosure-agreement-colossus-rcna236423
we don’t have infinite potable water. it’s going to come from municipal supplies. they’re going to stick the cost to these communities, and probably stress cities that are already water-fucked even more:
https://wetravelwebond.com/13-u-s-cities-on-the-brink-of-a-water-crisis/
I suggest you read the article I posted, or if you don’t feel like reading, maybe listening to what’s being said in the YouTube video.
There are some very real concerns with everything happening around AI. Water is quite simply not one of them. Focus your energy on criticisms based in reality.
I’d already seen Green’s video. I agree it’s a small thing compared to the power, but in a world obviously running out of easily accessible potable water, it’s only going to exacerbate problems for the communities that are already being hit by the relentless power costs due to AI as well.
This isn’t a one or the other thing. They’ll need water for cooling. It’s not recycled. it’s got to be clean, potable water.
we do not live in an infinite muni-water supply.
And it depends on the area. Dude with the blog post isn’t considering the many areas with real scarcity. The water, it’s pollution, etc. are a very real threat to many ecosystems and people.
yuuuup.
https://news.arizona.edu/news/phoenix-sees-increase-people-living-without-running-water-new-analysis-shows
https://www.forbes.com/sites/kavehmadani/2025/12/03/6-big-questions-about-irans-water-crisis-and-tehrans-day-zero/
https://wetravelwebond.com/13-u-s-cities-on-the-brink-of-a-water-crisis/
Yeah, thanks, this random substack and YouTube video definitely are credible and trustworthy sources.
As opposed to your vibes?
As opposed to the many reputable news sources and scientific publications that run contrary?
I mean, I’m certain that if you consider there to be many reputable scientific publications having studies to the contrary, you should at least be able to post a single one, right?
Sure. Here is a decent one:
Environment and sustainability development: A ChatGPT perspective
Priyanka Bhaskar, Neha Seth
Applied Data Science and Smart Systems, 54-62, 2024
Point is finding it or the dozens like it take like 2 seconds on Google scholar.
Great!
Now, where in the paper did you find the primary claims about water consumption?
(There are none. You didn’t read the paper)
Well, I suppose one of us didn’t. XD
Or at least couldn’t make it to section V. They mentioned the common claim, for instance, that GPT 3 alone took three quarters of a million gallons of water to train (roughly the footprint of 3-400) cars.
But, please, keep congratulating yourself. I was selection based on the criteria of (1) easy to read (2) multiple environmental issues listed.
Once again, they make no primary claims. All they do is cite other papers.
if you can refute it please do.
don’t talk about it, do it.
Fuck man, read one of the myriad of studies. Not some blog blog post. XD
When it comes to ludicrous and bullshit “evidence” you don’t get to challenge the world to refute something.
normally that would be a sensible tact.
Yet, you came here to say “As opposed to the many reputable news sources and scientific publications that run contrary?”
with not even a cunthair of evidence for those contrary points. “nuh uh you should search I’m right but can’t be fucked to give a shred of evidence”
you fucking fuckheads
I’m 4 pages into a paper on the subject. Want something on air pollution, check out grok in Memphis. Something about water table distruction, check the rural New York one.
If you want me to cite something obscure like some strangeness between weak and strong NP-Completeness? Sure. That makes sense.
If someone wants me to cite a source regarding whether prime numbers are real I’m just going to tell them to stop being stupid and ask Google.
That is how simple this one is. I have hundreds of citations, both scientific and mass media. They took almost zero effort to find.
I would challenge you to avoid being lazy and stupid.
am intimately familiar with the shitshow in Memphis.
searching new york data center water table destruction gives some links to stories related to datacenter water use but nothing specific to NY or AI. duckduckgo and google (non ai) searched.
I would challenge your assumptions but you’re just an asshole. Have a great life.