AI data center project sucked 29M gallons of water
tomshardware.com> the 29 million gallons were consumed during temporary construction activities, including concrete work, dust control, and site preparation
Misleading title
Is there a AI data center construction process that doesn't involve these costs?
Is there a large scale construction project that doesn’t incur these costs? Is the AI data center substantially worse on these metrics that other comparable projects? Or are we talking about it only because it’s AI-related?
> Or are we talking about it only because it’s AI-related?
We're absolutely only talking about it because it's AI.
From about 5 minutes of digging, I found the below which perhaps helps to put the 29m gallons in context.
> The Fayette County Water System has a total production capacity of 22.8 million gallons per day (MGD).
Source: https://fayettega.org/doing-business/global-access-infrastru...
So...keeping things simple and using 30d months:
Using 29M gallons over 15 months = 29,000,000 / (15 * 30) = 64,444 gallons per day avg
Based on 22.8m daily production capacity that's less than 0.3% of the total production capacity per day.
(Happy to be corrected if my napkin maths is wrong / i'm missing something here!)
I think it's important to know the total costs environmentally for these outfits
But what criteria do we use to judge whether these numbers are too high or too low? How do they compare to other construction projects of this size?
By whether or not it is disrupting other users access to the resource like what happened in the article? By how much it effects our very finite water table. Things like that sound reasonable to me
The title doesn't mention a time horizon it states what was used which it was. This is not misleading.
During the construction, not operating. In scale it's 44 Olympic pools.
[dupe] Discussion on source: https://news.ycombinator.com/item?id=48079333
So not much at all. Lets say enough for growing 4 tonnes of almonds.