A.I. tools fueled a 34% spike in Microsoft’s water consumption, and one city with its data centers is concerned about the effect on residential supply
A.I. tools fueled a 34% spike in Microsoft’s water consumption, and one city with its data centers is concerned about the effect on residential supply
“They were pretty secretive on what they’re doing out there,” says a former mayor about Microsoft’s data center projects.
I feel like we’re missing the forest for the trees here, this is all in service of a feature in bing that was supposed to revolutionize search but actually caused a 0% increase in Bing’s market share from q1 to q3 https://www.zdnet.com/google-amp/article/bings-search-market-share-fails-to-budge-despite-ai-push/
AI Tools - plural. ChatGPT (and OpenAI as a whole) predominantly runs on Azure infrastructure. Microsoft also owns GitHub with its associated copilot. And now all the Microsoft product specific copilots.
Not trying to defend their usage, but there are several forests here that are quite visible.
Oh wow didn't know github was owned by Ms
I'd use it all the time if it didn't require an account.
I've been using the Codium plugin in VSCode for a little bit now and find it useful. It's free, but does require an account.
If it's just for cooling, wouldn't they just be able to pull water directly from a lake and then return the same water into the lake? Why is any consumption happening?
In theory, yes. Of course, the same holds true for a lot of things which we currently use clean water for! The water needs of agriculture, toilets, carwashes, and many more could be addressed through so-called graywater (e.g.: pumped lakewater, rooftop rainwater) if we really sat down and wanted to make it happen.
The reason that we don't do these things is rather mundane: it's cheaper and easier to tap into the shared drinking water infrastructure than it is to collect your own water and roll your own silos/filtration tech. That might change as the world changes -- something has to give eventually if we use more groundwater than we replenish, but much like clean drinking water, I don't think it's a problem we should ask individual entities to solve. Governments would generally be much more suited to efficiently collecting drainwater, scrubbing it, distributing it, and mandating usage in wasteful commercial applications.
A lot of problems we don’t solve boil down to “it’s boring and expensive” lol it’s sad when you think about it. Everyone says they want infrastructure investment because they think it sounds mature or whatever, but when the day comes, they shake their heads.
The water isn't being consumed. It's going through the same process all the water in the city is going through.
Pulled from the river, cleaned, used for cooling at the data centers, and returned to the river via the waste water system.
The only loss is the energy/resources to treat the water.
I gotta wonder though, water used for server cooling is basically just run through metal fixtures and returned right? Couldn't it be possible to force some kind of maintenance and cleanliness standards onto the equipment and just have the water return to the supply? Is there any reason that water wouldn't be just as drinkable after?
They use evaporative cooling on days where it is over 85f
Microsoft’s data centers currently use adiabatic cooling, which relies on outside air to cool down temperatures inside. It’s a system that uses less electricity than air conditioning and less water than cooling towers. But when temperatures rise above 85 degrees Fahrenheit, outside air isn’t very helpful. At that point, an evaporative cooling system kicks in, which uses water. It works like a “swamp cooler” — cooling the air by pushing it over or through water-soaked screens.
https://local.microsoft.com/wp-content/uploads/2021/11/Datacenter-water-consumption-fact-sheet.pdf
They probably treat the water to prevent mineral and bacterial build up. No matter how sanitary it is it will probably require some amount of treatment before it can be put back into public drinking supplies.
Well, all sewer water requires treatment before it's used again but this water doesn't go into the sewer, it's evaporative cooling so it goes into the air.
No, but I assume you'd have to build extra infrastructure for that, which is expensive. They might now consider it worth it if they continue to need that much water, though.
if you think that's wild, animal agriculture uses 13 TRILLION litres or gallons of water a year. Compared with fracking, which uses 220 billiion.
i don't think that's right. can you substantiate that?
oh, you know what? it's actually worse. significantly worse. It's 34-76 trillion gallons annually, https://www.cowspiracy.com/facts
Animal agriculture water consumption ranges from 34-76 trillion gallons annually. [ii] [xv]
"Summary of Estimated Water Use in the United States in 2005". United States Geological Service
Pimentel, David, et al. "Water Resources: Agricultural and Environmental Issues". BioScience. (2004) 54 (10): 909-918
Rain World moment
This is the best summary I could come up with:
But one thing Microsoft-backed OpenAI needed for its technology was plenty of water, pulled from the watershed of the Raccoon and Des Moines rivers in central Iowa to cool a powerful supercomputer as it helped teach its AI systems how to mimic human writing.
Few people in Iowa knew about its status as a birthplace of OpenAI’s most advanced large language model, GPT-4, before a top Microsoft executive said in a speech it “was literally made next to cornfields west of Des Moines.”
In response to questions from The Associated Press, Microsoft said in a statement this week that it is investing in research to measure AI’s energy and carbon footprint “while working on ways to make large systems more efficient, in both training and application.”
Microsoft first said it was developing one of the world’s most powerful supercomputers for OpenAI in 2020, declining to reveal its location to AP at the time but describing it as a “single system” with more than 285,000 cores of conventional semiconductors, and 10,000 graphics processors — a kind of chip that’s become crucial to AI workloads.
It wasn’t until late May that Microsoft’s president, Brad Smith, disclosed that it had built its “advanced AI supercomputing data center” in Iowa, exclusively to enable OpenAI to train what has become its fourth-generation model, GPT-4.
In some ways, West Des Moines is a relatively efficient place to train a powerful AI system, especially compared to Microsoft’s data centers in Arizona that consume far more water for the same computing demand.
The original article contains 1,229 words, the summary contains 252 words. Saved 79%. I'm a bot and I'm open source!
Rain World time
For a second, I thought that Keegan-Michael Key was gonna have to lecture someone