There's plenty of water for data centers - Matt Yglesias \ stacker news ~AI

pull down to refresh

There's plenty of water for data centers - Matt Yglesias www.slowboring.com/p/theres-plenty-of-water-for-data-centers

362 sats \ 4 comments \ @Scoresby 13h AI

Matt Yglesias doesn't get much play on SN, but this is a well-researched article.

tl;dr - people who say data centers are using too much water in the US are stupid.

He's got lots of fun facts to back it up (and, of course, a number of nods toward YIMBYism).

I was triggered by the "need more power instead" thing so I decided to check (first with (local, no worries, it just took my macbook battery) AI, but ultimately with some AI helpers and manual compilation instead because them AI OCR thingies mix up punctuation between dots and commas.)

I compared caiso (so this is just Cali, but the article was specifically mentioning Cali too) recordings of "current system capacity" in their plannings, between 2023-2024¹ and 2024-2025²:

	2025	2024	Growth
Nuclear	2,300	2,300	0
Natural Gas	30,216	29,938	278
Hydro	12,601	12,597	4
Solar	21,627	19,766	1,861
Wind	8,334	8,506	-172
Biogas	289	289	0
Biomass	439	434	5
Geothermal	1,682	1,682	0
Battery Storage	13,012	9,267	3,745
Hybrid	2,796	2,101	695
Other	4,218	4,250	-32
Total	97,514	91,130	6,384

So while we shouldn't expect year-on-year growth (or maybe not even decade-on-decade growth) for nuclear, so that makes sense, it's funny (to me) that more than half of the capacity growth is caused by batteries. I doubt that simply adding batteries will help with powering desalination plants (or data centers) if you require an actual increase in capacity, but it does help with stabilization of course. ³

Also a decrease in Wind is interesting, but I feel that that's a global trend at the moment due to equipment breakdown and servicing issues.

Either way, per the graph provided in the linked ⁴ of the linked article ⁵

, and the mention in the latter that reverse osmosis filtration is thought to be very close to the thermodynamic limit of ~1 kWh/m3, let's do maffs:

Population CA: 39,538,223 ⁶, let's call it 40M.
Size of a bathtub: 40 gallons = m3: 0.15
Days in a year: 365
Millions of cubic meters of water needed for Cali people to not smell: 40 * 365 * 0.15 = 2190
Minimum electricity needed for desalination at the thermodynamic limit: 2190 GWh, or roughly 22x the current total capacity.

At a rate of 6.3 GW capacity growth per year, in 347 years from now, the current population of Cali can take a bath every day using desalinated water

Maybe, there's a point that there's a need for more power, faster; there's no use in having desalination tech if there's no power to make it real.

0 sats \ 0 replies \ @optimism 50m

I actually effed up the math because I read the capacity to be in MWh but it's in MW. Don't need AI to make mistakes.

trying again from the last bullet point:

Minimum electricity needed for desalination at the thermodynamic limit: 2190 GWh, at 24/7 operation this would mean 0.25 GW, or roughly 0.25% of the current capacity

At a rate of 6.3 GW capacity growth per year, it takes 14.5 days to provide the additional electricity for this.

So what that means is that the electricity is actually a non issue.

But please do double check my calc.

202 sats \ 0 replies \ @denlillaapan 7h

People who yap about water, in general, are pretty stupid

0 sats \ 0 replies \ @Macoy31 3h outlawed

stackers have outlawed this. turn on wild west mode in your /settings to see outlawed content.

Footnotes