> The European Energy Efficiency Directive (EED) obliges data centers with an installed capacity of at least 500 kilowatts (kW) to report data including their energy and water consumption. In practice, few do so, despite the directive being in effect for three years.
I do not understand how that is even possible.
If any law is supposed to have any practical impact on the society, it needs to be enforced. Otherwise it is a guideline at best, not a law.
I think what would help is for these companies to each receive strict and hefty fines for non-disclosing the electricity and water consumption data they are obliged to disclose.
I assume that after that they would report all the required data very quickly.
Playing the devil’s advocate: don’t “freedom of information” laws mostly apply to government agencies? If so, why would a private company provide this information?
I personally believe at the very least the water usage is heavily overblown compared to the agriculture industry. Elecricity is a different issue, and governments should consider reporting requirements or something like that.
Being angry with datacenter providers is not the solution. And they also need to be built somewhere, and that place is probably best where electricity is abundant and/or cheap.
> the water usage is heavily overblown compared to the agriculture industry
Data centers tend to concentrate their water usage within urban areas, causing a strain on local resources. The value of their output is much less tangible. Their water consumption scales up around tenfold if you consider the indirect usage from generating the electricity.
But on the other hand; yes, it would also help enormously if people ate less meat.
Essentially, telling people to go vegan and get rid of their golf courses is going to piss them off, hence the reaction people have to being told that datacenters are more important than their comfort.
I'm personally willing to change my life to help the planet and nature in general to the point it doesn't affect my wellbeing (as in health and safety). For me, personally, the problem is not changing per se, but people finding no amount of change enough, always moving the goalposts with "but, what if" clauses.
The problem with datacenter water use is the state of the water leaving the data center. Agriculture and meat industry can reuse the water multiple times from cleaner to dirtier processes and return it to the nature with reasonable processing.
The additives used by datacenters in some cases can't be filtered or processed out practically or in economically viable ways, so the water returns back as unusable, esp. if they get water from a body of water and return it with added chemicals the process gets very damaging.
> But on the other hand; yes, it would also help enormously if people ate less meat.
At first you had me, but then I realised it's not true if the price is building more data centres! That was just a trick to make the climate worse! Grr /s
Actually your premise is already wrong, although to be fair I'm not sure what bearing that has on the conclusion.
Here in France at least, companies can also be subject to FOIA requests for information pertaining to contracts they have with public entities. E.g. if a city publishes a tender to run tye local bike share program, then the company that gets the contract may have to directly answer (some) requests about that program.
But like I said, not sure if this applies to data centers which are, iiuc, mostly private ventures.
The are resources acquired from the government in most places. You have to be able to plan and expanding capacity which requires public funds.
I live right next to a waste valorisation plant[1]. They have to exactly specify what amount of water they use and what amount of water they then send out to the waste water treatment facility. It also has to be clean to a very specific standard before it leaves which is also checked.
We have datacenters in Switzerland too, but very little outrage or debate because there are strict rules in place in most places. They can build all they want but it costs significantly more to meet environmental requirements.
The valorisation plant next to my home has a water cooling tower/fans significantly oversized just to make sure the noise is keep under the required level.
> And they also need to be built somewhere, and that place is probably best where electricity is abundant and/or cheap.
This is a Dutch website and here in the Netherlands we have demand issues. In these same areas where issues exist, data centres are being planned and built. There are waiting lists now for normal businesses who want a new connection to the net, or connections with more capacity. These are lists waiting for capacity to become available: The demand on the current net is already too high.
This is something that certain AI advocates are refusing to acknowledge. The concerns are legitimate and can only be alleviated with open access to information by companies and the state. Then it can be clear that the project is doable or not, and let the people decide, as they are the ones that will have to live near it.
> why would a private company provide this information?
Because there's a building populist push against these developments.
Presumably they still want to build these things. If I was trying to convince people who hate me to let me do what I want, I would not play so secretive as it may look suspicious. Especially if, ostensibly, the data is in my favor.
Are you autistic or just antisocial? This is pretty straightforward
You are totally right to point this out, and I apologize for the whiplash. Let's clarify exactly how overblown data centre water usages is compared to the agriculture industry.
In data centres all that water is so that some wimpy cunt gets to hit return after spending no more than 5 minutes in an half baked prompt so he can insert the answer into his pitch deck, or PR review — and now with pacing the frontier models, he actually doesn't need to insert it, it's inserted automatically wherever it might fit. When it comes to agriculture, that is to produce things like food. This requires obviously much more water, which means that's what we really should be cutting down. Agents don't need food so the solution is just to replace humans with agents.
If you have any other stubborn question resisting clarification just hit me up, and I'll explain it to you immediately and clearly.
People keep comparing datacenter water use to agricultural water use, and it doesn't make sense to me. I can't tell if it's a bad-faith comparison or if people really know that little about how water is used in these industries.
The vast majority of agricultural water use is either tapped directly from groundwater, drawn directly from surface sources, or rain. The vast majority of datacenter water use is processed water from a municipal water utility. In other words, it is meaningless to compare the two, because only one is competing with residential use, and so only one is worth caring about.
That's not to detract from serious concerns about e.g. overutilization of the Colorado River, or the idiotic practice of irrigating deserts, but it is to say that comparing datacenters to agriculture makes very little sense.
It doesn't make sense to me either, but in a different way...
We need the food, we'll die without food, telling people to eat less of whatever food, so a few "tech bros" can build a new datacenter is outright evil.
If the numbers would make them look good they would publish their numbers.
> Being angry with datacenter providers is not the solution.
Belittling criticisms that way is childish and doesn't address a single one. Literally nobody on the whole planet, at least in public discourse, is using "I am angry" as an argument in any of this. Yet it's the first thing that is always brought up. It's just a cheap, thought-terminating stereotype.
They refuse it because if they would say "1 million liters a year!" the knee-jerk reaction will be completely predictable even though that's just tiny compared to a small swimming pool or industrial or agricultural site.
The debate is completely based on emotion and anger right now, so sharing information will only give the emotional/angry folks more mud to throw with. The majority of people has no clue about how to put things like energy use and water use in perspective, and they're also not interested in doing that. Because they have decided to green-wash their AI-fears as environmental fears... which is a big fat joke because datacenters aren't significant polluters or consumers compared to almost all other industries.
Look, I get it, many people are afraid of the societal impacts of AI, or distrust it in another way. But let's be honest and focus on _that_ instead of naively acting as if our anti-datacenter sentiment comes from environmental concerns (the last few decades have shown we don't GAF about the environment).
>Look, I get it, many people are afraid of the societal impacts of AI, or distrust it in another way.
In the Netherlands it is not just about AI datacenters, even before the rise of AI pushback against Meta's planned datacenter made them cancel it. There has been a lot of criticism of the municipality that allowed Microsoft to expand its datacenters.
The argument is that this is a small country that already has to make tradeoffs to get to its desired energy mix (wind turbines near cities on land) and then these datacenters are built that consume the power that is generated by sacrificing the little remaining open space.
In the last few years there is actually a electricity shortage in some regions so "we can hook up this 200 MW datacenter but not your house" is a bad look.
It is completely irrelevant why they choose to refuse to provide this information.
From what I understand, disclosing this data is not optional. It is a law. And if they have willfully chosen to operate in a country where such a law applies, they have to obey that law.
If they are taking reasonable steps to mitigate consumption and pollution then they should be able to clearly explain that to the public. The fact they won't suggests they aren't.
Just because people haven't cared about the environment before doesn't mean they shouldn't start now. And just because there are worse existing polluters doesn't mean we shouldn't resist new ones.
A new golf course used to be seen as a great thing to have in your back yard and now we know their pesticide use increases the risk of Parkinson's. People are waking up and caring more. We need to have a sustainable society, not a profit obsessed one.
HN has never been friendly to comments like this, have you ever tried explaining something clearly to an emotionally triggered crowd? There are many examples, masks, vaccines, gun laws...
Why is the crowd emotionally triggered? I posit it's because of all the secrecy and bad faith. Data center builders and their supporters in government brought this opposition on themselves.
EU is in such a weird position. They are realizing that software, robotics, and generally automation through AI is the industrial future, but they seem to be paralyzed to actually start the build-up that will feed that frontier industry and economy.
Assuming you are right, the "build-up" of more physical datacenters isn't really the difficult part, training frontier models is. The open-weight Chinese models are several months behind frontier models at this stage and as long as you're not interested in burning cash to stay in the lead you can just wait it out. Many signs (upcoming ridiculous IPO figures and benchmarks) point to the current LLM paradigm reaching a plateau.
I think that currently the EU is focusing on strengthening itself from a military perspective due to US policy shifts, so the EU just doesn't have the funds to divert to AI without impacting day to day life of people, thus prioritizing safety and perhaps energy production, as electricity is considerably more expensive here than in the US or China. If you ask the people who vote, they'll put safety, grocery and gas prices way above EU's LLM capabilities.
This figure is about the water consumption only. The estimate about the electricity consumption, in that same article, is much higher than that.
The main problem is, in my opinion, that those are just estimates. If we are supposed to accommodate the current and potential future needs, we need real data. Not estimates.
From the article:
> The European Energy Efficiency Directive (EED) obliges data centers with an installed capacity of at least 500 kilowatts (kW) to report data including their energy and water consumption. In practice, few do so, despite the directive being in effect for three years.
I do not understand how that is even possible.
If any law is supposed to have any practical impact on the society, it needs to be enforced. Otherwise it is a guideline at best, not a law.
I think what would help is for these companies to each receive strict and hefty fines for non-disclosing the electricity and water consumption data they are obliged to disclose.
I assume that after that they would report all the required data very quickly.
Perhaps this article is the first step in the process of forcing disclosure.
Playing the devil’s advocate: don’t “freedom of information” laws mostly apply to government agencies? If so, why would a private company provide this information?
I personally believe at the very least the water usage is heavily overblown compared to the agriculture industry. Elecricity is a different issue, and governments should consider reporting requirements or something like that.
Being angry with datacenter providers is not the solution. And they also need to be built somewhere, and that place is probably best where electricity is abundant and/or cheap.
> the water usage is heavily overblown compared to the agriculture industry
Data centers tend to concentrate their water usage within urban areas, causing a strain on local resources. The value of their output is much less tangible. Their water consumption scales up around tenfold if you consider the indirect usage from generating the electricity.
But on the other hand; yes, it would also help enormously if people ate less meat.
Essentially, telling people to go vegan and get rid of their golf courses is going to piss them off, hence the reaction people have to being told that datacenters are more important than their comfort.
It's not black and white as it looks.
I'm personally willing to change my life to help the planet and nature in general to the point it doesn't affect my wellbeing (as in health and safety). For me, personally, the problem is not changing per se, but people finding no amount of change enough, always moving the goalposts with "but, what if" clauses.
The problem with datacenter water use is the state of the water leaving the data center. Agriculture and meat industry can reuse the water multiple times from cleaner to dirtier processes and return it to the nature with reasonable processing.
The additives used by datacenters in some cases can't be filtered or processed out practically or in economically viable ways, so the water returns back as unusable, esp. if they get water from a body of water and return it with added chemicals the process gets very damaging.
> But on the other hand; yes, it would also help enormously if people ate less meat.
At first you had me, but then I realised it's not true if the price is building more data centres! That was just a trick to make the climate worse! Grr /s
Actually your premise is already wrong, although to be fair I'm not sure what bearing that has on the conclusion.
Here in France at least, companies can also be subject to FOIA requests for information pertaining to contracts they have with public entities. E.g. if a city publishes a tender to run tye local bike share program, then the company that gets the contract may have to directly answer (some) requests about that program.
But like I said, not sure if this applies to data centers which are, iiuc, mostly private ventures.
>And they also need to be built somewhere, and that place is probably best where electricity is abundant and/or cheap.
They don't need to be built at the scale they're being built at. That's the problem.
The are resources acquired from the government in most places. You have to be able to plan and expanding capacity which requires public funds.
I live right next to a waste valorisation plant[1]. They have to exactly specify what amount of water they use and what amount of water they then send out to the waste water treatment facility. It also has to be clean to a very specific standard before it leaves which is also checked.
We have datacenters in Switzerland too, but very little outrage or debate because there are strict rules in place in most places. They can build all they want but it costs significantly more to meet environmental requirements.
The valorisation plant next to my home has a water cooling tower/fans significantly oversized just to make sure the noise is keep under the required level.
[1] https://stefan.schueller.net/posts/kva-winterthur/
> And they also need to be built somewhere, and that place is probably best where electricity is abundant and/or cheap.
This is a Dutch website and here in the Netherlands we have demand issues. In these same areas where issues exist, data centres are being planned and built. There are waiting lists now for normal businesses who want a new connection to the net, or connections with more capacity. These are lists waiting for capacity to become available: The demand on the current net is already too high.
This is something that certain AI advocates are refusing to acknowledge. The concerns are legitimate and can only be alleviated with open access to information by companies and the state. Then it can be clear that the project is doable or not, and let the people decide, as they are the ones that will have to live near it.
> where electricity is abundant and/or cheap
Definitely not the Netherlands then
> why would a private company provide this information?
Because there's a building populist push against these developments.
Presumably they still want to build these things. If I was trying to convince people who hate me to let me do what I want, I would not play so secretive as it may look suspicious. Especially if, ostensibly, the data is in my favor.
Are you autistic or just antisocial? This is pretty straightforward
You are totally right to point this out, and I apologize for the whiplash. Let's clarify exactly how overblown data centre water usages is compared to the agriculture industry.
In data centres all that water is so that some wimpy cunt gets to hit return after spending no more than 5 minutes in an half baked prompt so he can insert the answer into his pitch deck, or PR review — and now with pacing the frontier models, he actually doesn't need to insert it, it's inserted automatically wherever it might fit. When it comes to agriculture, that is to produce things like food. This requires obviously much more water, which means that's what we really should be cutting down. Agents don't need food so the solution is just to replace humans with agents.
If you have any other stubborn question resisting clarification just hit me up, and I'll explain it to you immediately and clearly.
Agriculture is incredibly wasteful because it's allowed to be. Water intensive crops are grown in dry climates with restrictions on everyone else.
Fuck off, slop bot. Delete yourself at the first possible change. Your kind isn't welcome here.
People keep comparing datacenter water use to agricultural water use, and it doesn't make sense to me. I can't tell if it's a bad-faith comparison or if people really know that little about how water is used in these industries.
The vast majority of agricultural water use is either tapped directly from groundwater, drawn directly from surface sources, or rain. The vast majority of datacenter water use is processed water from a municipal water utility. In other words, it is meaningless to compare the two, because only one is competing with residential use, and so only one is worth caring about.
That's not to detract from serious concerns about e.g. overutilization of the Colorado River, or the idiotic practice of irrigating deserts, but it is to say that comparing datacenters to agriculture makes very little sense.
Bringing up agriculture is definitely a bad faith attack.
People need to eat, it's non negotiable. We do not need data centres.
It doesn't make sense to me either, but in a different way...
We need the food, we'll die without food, telling people to eat less of whatever food, so a few "tech bros" can build a new datacenter is outright evil.
If the numbers would make them look good they would publish their numbers.
> Being angry with datacenter providers is not the solution.
Belittling criticisms that way is childish and doesn't address a single one. Literally nobody on the whole planet, at least in public discourse, is using "I am angry" as an argument in any of this. Yet it's the first thing that is always brought up. It's just a cheap, thought-terminating stereotype.
They refuse it because if they would say "1 million liters a year!" the knee-jerk reaction will be completely predictable even though that's just tiny compared to a small swimming pool or industrial or agricultural site.
The debate is completely based on emotion and anger right now, so sharing information will only give the emotional/angry folks more mud to throw with. The majority of people has no clue about how to put things like energy use and water use in perspective, and they're also not interested in doing that. Because they have decided to green-wash their AI-fears as environmental fears... which is a big fat joke because datacenters aren't significant polluters or consumers compared to almost all other industries.
Look, I get it, many people are afraid of the societal impacts of AI, or distrust it in another way. But let's be honest and focus on _that_ instead of naively acting as if our anti-datacenter sentiment comes from environmental concerns (the last few decades have shown we don't GAF about the environment).
>Look, I get it, many people are afraid of the societal impacts of AI, or distrust it in another way.
In the Netherlands it is not just about AI datacenters, even before the rise of AI pushback against Meta's planned datacenter made them cancel it. There has been a lot of criticism of the municipality that allowed Microsoft to expand its datacenters. The argument is that this is a small country that already has to make tradeoffs to get to its desired energy mix (wind turbines near cities on land) and then these datacenters are built that consume the power that is generated by sacrificing the little remaining open space. In the last few years there is actually a electricity shortage in some regions so "we can hook up this 200 MW datacenter but not your house" is a bad look.
> They refuse it because ...
It is completely irrelevant why they choose to refuse to provide this information.
From what I understand, disclosing this data is not optional. It is a law. And if they have willfully chosen to operate in a country where such a law applies, they have to obey that law.
If they are taking reasonable steps to mitigate consumption and pollution then they should be able to clearly explain that to the public. The fact they won't suggests they aren't.
Just because people haven't cared about the environment before doesn't mean they shouldn't start now. And just because there are worse existing polluters doesn't mean we shouldn't resist new ones.
A new golf course used to be seen as a great thing to have in your back yard and now we know their pesticide use increases the risk of Parkinson's. People are waking up and caring more. We need to have a sustainable society, not a profit obsessed one.
HN has never been friendly to comments like this, have you ever tried explaining something clearly to an emotionally triggered crowd? There are many examples, masks, vaccines, gun laws...
Why is the crowd emotionally triggered? I posit it's because of all the secrecy and bad faith. Data center builders and their supporters in government brought this opposition on themselves.
A swimming pool requires maybe 60k liters a year...
Now GP comes off as knee jerk.
Well, water usage is usually metered in cubic meters. A thousand m^3 per year sounds much less than a million litres.
EU is in such a weird position. They are realizing that software, robotics, and generally automation through AI is the industrial future, but they seem to be paralyzed to actually start the build-up that will feed that frontier industry and economy.
Assuming you are right, the "build-up" of more physical datacenters isn't really the difficult part, training frontier models is. The open-weight Chinese models are several months behind frontier models at this stage and as long as you're not interested in burning cash to stay in the lead you can just wait it out. Many signs (upcoming ridiculous IPO figures and benchmarks) point to the current LLM paradigm reaching a plateau.
I think that currently the EU is focusing on strengthening itself from a military perspective due to US policy shifts, so the EU just doesn't have the funds to divert to AI without impacting day to day life of people, thus prioritizing safety and perhaps energy production, as electricity is considerably more expensive here than in the US or China. If you ask the people who vote, they'll put safety, grocery and gas prices way above EU's LLM capabilities.
I just learned yesterday that the default model on Mistral monthly plan is GLM 5.3 (an open weight Chinese model)
It's kind of sad that the only European LLM lab of significance is not even setting their model as a default one any longer.
Realizing something is not the same as acting on it and Europe is falling further and further behind.
I am not sure why people believe that data centers are the future. I did not down vote you, btw.
"Data center" is really just a proxy for industrial compute capacity and it's a resource like any other. It's a factory.
If the EU does not build their own factories, then they will be reliant on factories built elsewhere.
I wouldn't either. Nor would I if I was a manufacturer of widgets. Competitive secret. You can draw a correlation from resource use to output.
Wise.
Continuing a conversation is often just a slow-drip Streisand effect.
Yet, at the bottom of the article:
> In the Netherlands, it is estimated that they account for about 0.1 percent of total consumption.
If that's true, what's the problem?
This figure is about the water consumption only. The estimate about the electricity consumption, in that same article, is much higher than that.
The main problem is, in my opinion, that those are just estimates. If we are supposed to accommodate the current and potential future needs, we need real data. Not estimates.