-Jade-'s profile

So AI has to be accessible to everyone globally and answer questions, give feedback or ideas across several topics. It cannot do this on its own, so it needs lots and lots of data from Non-AI sources to build up its understanding. Since there's a lot of stuff to remember across several topics, it's simply too much data to store on a simple hard drive, so companies have to build AI data centres, which you can imagine as buildings with loads and loads of servers just storing data.

When coming up with responses, AI relies on this data so to use it the servers have to do a bit of work, which uses up power. The power being supplied to these servers heat them up, but that heat has to actually go somewhere else it'll just melt the components.

There are 2 main ways they've considered to cool down these servers:

Power: The reason this doesn't work very well is because it's already very expensive, given how much power they're also using to keep the servers running

Water: Companies are using this as a cheaper alternative, because it can cool down the servers. They aren't pouring the water onto them (that would be dumb ) but the water takes in some of the heat and it causes the water to evaporate.

The water in your day to day life isn't pure H2O, because that would be pretty much impossible. There's a bunch of minerals in there too, but it's only the H2O that evaporates. This is an issue because as more of the water evaporates, the remaining water becomes more and more concentrated with these minerals that can actually damage the components inside. So this water has to be replaced with more clean drinking water, and that's why and where all our water is going