> "The majority [of water] is used for big agro industries, such as soya, rice, and wood pulping. Now we have Google planning to use enormous quantities of water.”
This is like calling a rubber raft "enormous" after comparing it to an aircraft carrier. Agriculture consumes astronomically more water, often wastefully, for far less valuable purposes. I can't take these objections seriously.
So, for my part, I’m not so concerned about the water inputs that are consumed by datacenters. The most water efficient traditional power generation systems we have are combined cycle natural gas turbines, and they consume about 5-10x as much water to produce a megawatt of power than the data center would use to cool itself from using that megawatt of power, even if they used the worst “one and done” through cooling system.
Check the HARC report on this.
No, my major concern is about what happens to the waste? Does it just get dumped into the sewers? What else might that poison? Where do those sewers get their water treated, and where does that water go downstream? Because my “treated” water output back into the rivers and aquifers is your downstream water input for humans to drink, golf courses to water their grass, etc….
And what happens when there is a toxic spill? Does that go into creeks and rivers on the surface? Where do those creeks and rivers go? Where do they let out, in relation to the closest water intakes?
Does the toxic spill go into the karst geography below the land? Into the aquifers? What all is exactly down there, and how many decades or centuries might they be poisoned? I mean, you may be able to clean up a surface creek or river, but you can’t clean up the karst geography below the surface, or the aquifers.
You could easily poison an entire city with a single spill of toxic waste from a datacenter, and continue to poison that city for years and decades to come.
Anyone angry about DC water use should be much more angry about golf courses. They use far more water and often overuse pesticides that can run off and leach into ground water.
> Tapping into public drinking water is both easier and cheaper than paying the cost of treatment themselves. That means data centers end up competing with communities over their potable water supply.
My city has tiered pricing for water consumption based on gallons consumed per-household. Instead of the vague pearl clutching here, why can't municipalities provide preferred pricing to residential customers and have the data center consumers subsidize ongoing water treatment and infrastructure build out to support the extra capacity. Seems like a win-win.
This is as weak article. For example, it says, "A hyperscale data center can suck up between 1 and 5 million gallons of water a day..." but does not cite a source. It also says, "Data centers end up competing with communities over their potable water supply." Not really. Data centers have to buy potable water, so unless the water providers are selling water too cheaply, what's the problem?
I think that tech companies are often insanely greedy when it comes to building these data centers. They're asking for tax breaks they don't need and putting strain on local electrical grids that they expect other people to pay for, all while bringing little in the way of permanent employment increases to the areas where they're built. It's wrong.
That being said, water isn't like electricity, land, or a given year's tax revenues. It exists in a cycle and can be reused. It can be piped in from a potable water source, and you could probably develop a cooling loop that would return it back to the civic supply in potable form.
Other than somebody not wanting to because "b-b-b-b-but what about the shareholders!?", why not build that loop, put its components into mass production, and let economies of scale start cutting the cost while having less impact upon local water districts?
> "The majority [of water] is used for big agro industries, such as soya, rice, and wood pulping. Now we have Google planning to use enormous quantities of water.”
This is like calling a rubber raft "enormous" after comparing it to an aircraft carrier. Agriculture consumes astronomically more water, often wastefully, for far less valuable purposes. I can't take these objections seriously.
This is just the latest weird moral panic.
So, for my part, I’m not so concerned about the water inputs that are consumed by datacenters. The most water efficient traditional power generation systems we have are combined cycle natural gas turbines, and they consume about 5-10x as much water to produce a megawatt of power than the data center would use to cool itself from using that megawatt of power, even if they used the worst “one and done” through cooling system.
Check the HARC report on this.
No, my major concern is about what happens to the waste? Does it just get dumped into the sewers? What else might that poison? Where do those sewers get their water treated, and where does that water go downstream? Because my “treated” water output back into the rivers and aquifers is your downstream water input for humans to drink, golf courses to water their grass, etc….
And what happens when there is a toxic spill? Does that go into creeks and rivers on the surface? Where do those creeks and rivers go? Where do they let out, in relation to the closest water intakes?
Does the toxic spill go into the karst geography below the land? Into the aquifers? What all is exactly down there, and how many decades or centuries might they be poisoned? I mean, you may be able to clean up a surface creek or river, but you can’t clean up the karst geography below the surface, or the aquifers.
You could easily poison an entire city with a single spill of toxic waste from a datacenter, and continue to poison that city for years and decades to come.
Is this guy reviewing his own book?
Anyone angry about DC water use should be much more angry about golf courses. They use far more water and often overuse pesticides that can run off and leach into ground water.
> Tapping into public drinking water is both easier and cheaper than paying the cost of treatment themselves. That means data centers end up competing with communities over their potable water supply.
My city has tiered pricing for water consumption based on gallons consumed per-household. Instead of the vague pearl clutching here, why can't municipalities provide preferred pricing to residential customers and have the data center consumers subsidize ongoing water treatment and infrastructure build out to support the extra capacity. Seems like a win-win.
This is as weak article. For example, it says, "A hyperscale data center can suck up between 1 and 5 million gallons of water a day..." but does not cite a source. It also says, "Data centers end up competing with communities over their potable water supply." Not really. Data centers have to buy potable water, so unless the water providers are selling water too cheaply, what's the problem?
Sometimes I wonder how many resources things like AI actually save rather than consume (I don't see that reported on as much)
mmmhmm. Like cities in the desert and green lawns at the golf course, right?
Then use thermofluid instead of water, which is corrosive.
Then go to carbon-based computing to waste less heat.
I think that tech companies are often insanely greedy when it comes to building these data centers. They're asking for tax breaks they don't need and putting strain on local electrical grids that they expect other people to pay for, all while bringing little in the way of permanent employment increases to the areas where they're built. It's wrong.
That being said, water isn't like electricity, land, or a given year's tax revenues. It exists in a cycle and can be reused. It can be piped in from a potable water source, and you could probably develop a cooling loop that would return it back to the civic supply in potable form.
Other than somebody not wanting to because "b-b-b-b-but what about the shareholders!?", why not build that loop, put its components into mass production, and let economies of scale start cutting the cost while having less impact upon local water districts?
[dead]