If I drink one glass of wine per day, it’s probably fine. If I drink three bottles per day, I’m gonna start having problems.
[1] https://ourworldindata.org/grapher/monthly-spending-data-cen...
You have people who eat meat at every meal losing their mind over water usage and people who set their AC to 68 panicking over energy usage.
Can you spot the difference?
Not for nothing that YouTube alone is responsible for a double-digit amount of all internet traffic.
Just because the bubble here uses AI doesn't mean it's even remotely valuable for the 80%. But they've done a great job convincing a specific group of people this.
It's actually significantly more environmentally friendly than heating to above 68F in the winter in most cases, since most AC is done with heat pumps but most heating is oil/gas (100% fossil fuel consumption) or resistive electric heaters (dramatically less efficient than heat pumps). On top of that AC use aligns strongly with solar generation output while heating load does the opposite, which means that even when heating is done with electricity it comes disproportionately from gas turbines at natural gas peaker plants.
And even better: all of that data is aggregated together and stored in its entirety several times over across multiple data centers because everyone shares your data with everyone else, such that if any single of more than a dozen data centers is ever compromised, it all gets leaked to whomever compromised it.
So do many other things which use less water.
The estimates I've seen place the water consumption of water at such a low level that a whole month subscription consumes roughly as much water as 1-120 grams* of beef.
> Setting my AC to 68 keeps me cool.
What, is 70 not good enough?
Air con, and basically all things connected to the power grid, also "costs [your] community money, damages the environment directly and indirectly, and further enriches a few already rich people."
* yes, one to one hundred and twenty, there's that much uncertainty in the estimates for the AI.
In what country do you not have an electric bill?
Build the data centre, sure, but if it takes you five years to build out the grid and sufficient generating capacity in a non stupid way in order to do that, sorry, but that's by far my preferred route to get there.
[1] https://www.nytimes.com/2026/08/08/climate/amazon-data-cente...
I can't read your link, paywall, but my question regarding Musk's infamous polluting DC, I wonder why it had to be built near so many people. It's not like there's going to be a need for a huge pool of workers to commute to it.
On the other hand, how does that combine with his apparent narcissistic need to be seen? Is his need so simple it can be met with "line go up"?
Think about it, the list is just about endless by now it's hard to tell the lies from the intentions and the character deficits.
So, for this point, I don't care that the words and actions he emits have no apparent rational long-term coherency, but I do care about the coherency of the system behind them which generates them.
Think of it like me wanting to learn how an LLM works so I don't get surprised by it doing more than a Markov chain.
I can't read the link either, but at a glance, the first one wasn't. It's in clearly industrial land a long way from any houses, with its closest neighbors as far as I can tell being a steel mill, a few existing power plants, and a wastewater plant.
The second one is practically right in the middle of a bunch of neighborhoods, and for that, yeah, I want to know as well. I'm guessing the land was cheap, industrial-zoned, and near a substation?
(Not to excuse what Musk is doing in either case - it's fucked and has wide ranging effects. But I think it's important to clarify the first because I've seen a repeated trend of people acting like a new DC is going to be built in the middle of a peaceful picket-fence neighborhood, and then you look it up, and it's actually being built on top of a former steel mill a mile away from the nearest house. It gets annoying.)
AC and meat are bad for the environment but they deliver personal benefit. Having a DC in your community has no appreciable personal benefit, only negatives. That's it.
I'm replying to someone who themselves argued the analogies.
> Having a DC in your community has no appreciable personal benefit, only negatives.
To the extent that statement is true, it applies equally to most primary and secondary economic activity.
Having a farm in your community has no appreciable personal benefit: for scale, the UK famously gets a lot of their sheep meat from New Zealand, as far away as it is physically possible to be.
An abattoir similarly has no substantial local benefit, but worse has much detriment owing to the psychological damage it does to the workers, leading to higher rates of arrest for violent criminal offences: https://journals.sagepub.com/doi/abs/10.1177/108602660933816...
Unless it's at the scale where it becomes a company town, having a factory, a warehouse, etc. these things are all heavily automated or mechanised and employ only a few of the locals as a percentage of the population. Putting anything that isn't service-sector in your community (i.e. schools, hospitals, shops, restaurants, hairdressers, etc.) is going to be of "no appreciable personal benefit" when considered in isolation.
Almost all primary (farming, fishing, mining) and secondary (manufacturing) comes with negatives (pollution, higher wear on the roads and utilities).
I put in a GPU to run some local LLM stuff. Now the whole thing generates enough heat and noise that I notice it if i sit near the cabinet it’s in.
Scale matters.
AI use has driven this to 20, 50, 100kW per rack or more. Supposedly 1000kW for the hyperscalers.
Extrapolate to the whole building and you have a lot of waste heat that has to go somewhere.
Instead of being similar to light industry, the hyperscale AI datacentre is using as much energy as a large steelworks.
One can critique systems they are still interacting with, to suggest otherwise is just so incredibly lazy.
So, my take is that maybe the pro-DC side can do better at seeing multiple perspectives instead of acting like anything negatively said about DCs is a psychosis.
Those things are both a result of a failure of government, not the existence of data centres. If there are externalised costs the company wants to put on people then your government should extract them from the company as a condition of permission.
Data centres are factories that turn electricity into heat and money. They're basically the nicest factories possible. There is no fundamental reason they need to harm people, any issues data centres cause should be easy to address as a fraction of the already massive sums involved.
In a metal foundry some of the physical externality is intrinsic to doing the process. It's a continuous industrial chemical operation emitting smoke, metal fumes, VOCs, dust, sound etc. You can capture 99% of pollution, but you're still melting, casting, shaking out and finishing tonnes of hot metal and sand. Complete physical isolation is expensive and imperfect.
With data centres, most local externalities are engineering choices. With enough money and sensible siting, you could plausibly build one whose effect on neighbouring residents is extremely close to nil. The biggest remaining externalities occur upstream in the electricity/water/grid system - and those are unusually amenable to being charged directly to the operator.
If a local data centre is impacting you negatively it's a failure of your government because it didn't need to be that way. It's not an intrinsic part of data centre operation. They turn electricity into heat and money, a comparatively simple factory to have completely deal with its own externalities.
I'm not saying tylenol is bad. I'm not saying it's good. I'm just pointing out that your position is equivalent to opposing a person saying "lets no shove 20 tylenol down someone's throat" with the argument "but one has reduced the fever".
Existing datacenters at the scale we were used to were just fine for what was expected of them.
We have supply chain routing software already. Humans wrote it, humans can improve it. Getting ai in the loop won’t change the realities of the supply chain, except maybe they’ll hire less programmers.
> maybe i'm making a commodity trading bot that more accurately prices wheat futures
How does that benefit me? Further financializing food production seems like a bad thing.
Not to mention all of what you’ve described could be run on a $1000 consumer GPU in a PC somewhere.
Making supply chain routing software cheaper, improving it for less money → all of these things directly make the products in grocery stores cheaper.
"Further financializing food production seems like a bad thing." →
Why assume rather than approaching with curiosity? Commodity futures are, again, critical to global supply chains and food markets. Farmers in eastern Europe actually look at futures markets to determine when and if to plant their harvest, cargo ships that transport your food buy futures in oil & gas so that they can have stable fuel costs to provide you with your food, etc. etc. Speculating on the future gives rise to the corrections necessary for providing things that people want in the future.
No one was objecting when pornhub was building out data centers. Anything they like to consume, no matter how frivolous, is accepted as morally better on a metaphysical level.
> further enriches a few already rich people.
Meat eaters damaging the Earth is okay because it's "equal"? 300 million people deciding they're entitled to consume anything they want in ever-increasing quantities is a problem because... there's 300 million of them!
And it’s not like non-meat is fine - the corn industry uses a blinding amount of water, as do almonds, etc.
We have a finite budget for water usage, and i’d allocate that to non-meat food > virtually anything else > meat > data centers > golf courses.
> No one was objecting when pornhub was building out data centers
I’m not entirely sure of the scale of pornhub’s operations, but i feel like one large ai data center uses more resources than all of pornhub’s data center usage around the globe. But maybe i’m wrong there, and the hub was just flying under the radar.
Switching from meat to non-meat unequivocally lowers your GHG and water footprint and there's really no room for quibbling on this unless you are switching to a diet of exclusively almonds
Data centers also generate local tax revenue. In fact, across industries, they have the best tax revenue to tax resource burden ratio, using very little government services while paying significant property tax.
And the environmental damage, whatever that is, is negligible relative to other industries. Their only output is some noise pollution nearby and whatever emissions are produced in the power plants that give them electricity. All industries require electricity in proportion to how much economic value they produce. So unless you want degrowth and no industry, you're going to have energy consumption.
In fact, rising energy consumption is a sign of a growing economy. That's why China has seen its energy consumption grow far faster than any other major economy while its economy has also grown far faster.
They do use water but their water usage is mininscule compared to agribusiness, or even fast food restaurants and golf courses. And the potable water usage can be brought down. They just use the water for cooling. So they could use treated wastewater or recycle the water they use. Newer data centers use these more water efficient cooling systems. So water usage is not a major long-term concern. Not one that justifies crippling this key emerging industry.
How does it damage the environment? By using energy? Like literally any other human enterprise?
What's wrong with it creating wealth? You don't have to spend money on it, it's none of your concern what other people do.
* Take a bunch of their tax money with no benefit to them
* lower their property value
* Make electricity and water more expensive (both of which make other things they use more expensive also)
* can cause other environmental impacts that raise inconvenience and expenses both directly and indirectly
so some AI douchebags can have a shiny datacenter to enable their bosses' stated desire to replace them with the very same AI might be opposed to the current AI drive datacenter boom.
Whatever actual benefits AI has, or tech in general, the way it is being approached right now is a disgusting and blatant attempt to steal as much as possible from everyone. People don't like that, and the like even less how the people doing it pretend they don't understand the dislike while building bunkers to protect themselves from the anger.
it sounds like you just want to ban American industry altogether, we've quietly let extreme conservatism sneak up on us in the US it seems.
They aren't tax positive. The deal might "theoretically" be tax positive in 30 years if things go exactly according to some plan. In the short term they are usually associated with large tax breaks for some number of years, accompanied a guarantee to occupy the data-center for less than the payback period (so the tax breaks can be extended later with threats to move on so those tax paybacks never happen). They also come with the locality footing the bill for a lot of infrastructure build-out.
That's just a bad use of my tax dollars. Let the datacenter people pay for themselves at the very least - my taxes don't need to go subsidize some pile of shit AI fanatic. If you like AI, pay for your AI, don't take it from me. It has nothing to do with me being pro- or anti-industry, however if an already profitable industry can't exist without my tax subsidizing it just so some rich guy can be a little bit richer... well maybe it doesn't need to exist.
I have no idea how any of this is conservatism... I'm opposed to a specific set of changes, not all changes. Perhaps you are just trying to throw shade because I think you should pay for your own toys?
Account for the runtime of the extra mining, production of materials... the scale blows up exponentially when externalities of all those extra screws and zip ties are appropriately accounted. An issue with Americans in particular as capitalism propaganda completely ignores externalities like kids in textile sweatshops not given a choice of college or the mines
3c global increase by 2050 is locked in. Meaning another 10-15 degrees local warming.
If Exxon could predict the min-max of the exact trend in the 1970s.
Physical reality doesn't give a shit about our rhetorical ability to blame a corporation or rich person; everyone consuming this shit is part of the math of the problem.
5x in a decade. The way it’s talked about you would think it’s orders of magnitude.
The question isn't whether it's possible for something to become problematic past a certain scale. The question is whether data centers are at that point, or will be anytime soon.
The actual facts and figures that are presented in opposition tend to be stripped of the larger context, which makes me skeptical. The graph you linked looks wild, but in a country that spends two trillion dollars per year on all construction, is the jump from $1 billion/month to $4 billion/month very important?
If you gave me one night, a pot of coffee, and few scruples, I could probably dump hundreds if not thousands of fake posts on the internet. Hundreds would easily break through and who knows, a handful might even go viral on Twitter. Yeah I could do that 5 years ago, but not nearly at the same scale. Maybe a few dozen?
Back to DCs: it’s like saying “well we had coal fire plants years ago.” Ok…well there are more now generating way more power and spewing way more pollution. It’s not the same situation anymore. This DC build out is massive.
This DC buildout is massive? Yep. So massive that it merits substantial public opposition? Citation needed on that part. And any citation that compares current or projected DC activity with past DC activity is missing the point.
Two concrete examples: the above graph showing a multi-fold increase in construction expenditures is useless, because it gives no indication of the absolute significance of the current level. It tells me very little. For a better example, here's a graph of RAM prices over the past year and a half: https://whereismyram.com/us/price-chart. Here, we can see that data center buildout is big enough to massively affect the global market for an important commodity, and we can translate that into some sort of understanding of how this affects ordinary people.
Completely agreed that the "this existed before" argument is tiring. But so is the "it's a lot bigger now" argument. Neither one actually says anything interesting. They're both statements of opinion (this is fine, versus this is bad) hiding inside obvious statements of fact.
I just want something concrete to judge. The data center debate rages but almost never in terms that help me figure out what to think about it. RAM prices going through the roof is something I can easily contextualize and shows the impact on the broader world. Data center construction spending going up a few hundred percent isn't and doesn't, and when I look up the numbers to contextualize it then it doesn't really look all that significant.
You are assigning additional value to my comments - I get it but I’m trying to say clearly that I’m not saying those things. All I am talking about is acting as if scale is not a part of the conversation. Anything beyond that I have yet to weigh in on. i’m happy to elaborate if you want me to on any particular point, I am not being deliberately opaque here. I’m just not entirely sure what you are looking for.
Honestly we’re getting off track IMO. I think we’re just running in circles here. Appreciate the conversation nonetheless, have a good rest of your week
It is true that wine doesn't do anyone any good (unless you count it making people more social or something). But the effects of one glass a night is indistinguishable from zero. There are definitely statistically significant morbidity effects at 3.
I've been shocked reading public health proclamations that "anything above zero" is bad, and then reading the actual underlying paper they refer to, which clearly states exactly the opposite finding. I think it is an extension of the public policy belief that the public cannot be trusted with moderation, and so therefore must be communicated to in absolutes, and nobody will read the underlying paper anyway.
All that said, no one I've ever met said a glass of wine on Friday nights isn't "drinking in moderation". I think the medical establishment is quite well aligned on this. One source is the CDC, which defines moderate drinking to be roughly an order of magnitude more alcohol consumption than "one drink every Friday night."
https://www.cdc.gov/alcohol/about-alcohol-use/moderate-alcoh...
> Many people metabolize the alcohol more slowly, and get less of a "hit" from drinking, which makes them less susceptible to addiction. Others, like me, metabolize it very fast and that makes it quite a bit more addictive.
One thing to keep in mind when discussing "drinking in moderation": I think people will surprised to learn that each country has their own (often, very different) idea of what it means. For example, healthy alcohol consumption levels are persistently higher from UK public health sources, compared to the United States (mostly CDC). My sense (that I cannot prove): If your median alcohol consumption is higher than global average, your national health agency will have higher values for "drinking in moderation". I would love to see a comparison of wealthy (primarily) Judeo-Christian countries and their recommended limits.
As for what moderate is: I can absolutely believe it'cultural. I was intending to post from a US/EU perspective, as that's what I'm most familiar with, but your explanation makes sense.
This is the opposite of how it works, ethanol is the active ingredient for all of the fun parts of drinking, the metabolite, acetaldehyde, is the cause of all of the unfun parts (nausea, headache, flushing).
A significant percentage of East Asians have one or both of a superactive alcohol dehydrogenase (which converts ethanol to acetaldehyde) and a defective aldehyde dehydrogenase (which converts acetaldehyde to acetic acid), and they have significantly lower rates of alcoholism.
People want to conflate their personal desires and perceived social benefits with health benefits because mental health and physical health have overlaps
However they need to be separate to actually make a claim about alcohol as there are millions of people in functional communities that reject alcohol and have non of the negative effects
(1) https://www.who.int/europe/news/item/04-01-2023-no-level-of-...
This would be a misquote, but you didn't use quotes. The WHO said that they "can't identify" a safe level of alcohol consumption and that there are no studies that would indicate a safe level of alcohol consumption.
> The new WHO statement clarifies: currently available evidence cannot indicate the existence of a threshold at which the carcinogenic effects of alcohol “switch on” and start to manifest in the human body.
I could see why you would say that from this:
> “We cannot talk about a so-called safe level of alcohol use. It doesn’t matter how much you drink – the risk to the drinker’s health starts from the first drop of any alcoholic beverage. The only thing that we can say for sure is that the more you drink, the more harmful it is – or, in other words, the less you drink, the safer it is,” explains Dr Carina Ferreira-Borges,
But that seems to be a mixture of some language difficulty and some intentional hyperbole.
How do you not then clonclude that there IS a safe amount such that you would then consume it.
What about this study would make anyone who is not an alcohol drinker become one?
The only answer here that is actually logical is that
you accept that there are risks to drinking alcohol, and that those risks are less important to you than whatever social personal benefit that you personally get out of it
Moreover, the WHO never changed their stance when later big studies showed that the "lowest risk" level is not actually at zero drinks.
The timeline was this:
- In 2018, a study was published in the Lancet that was the Global Burden of Disease analysis. They claimed the modeled minimum risk exposure was zero drinks. But the magnitude effects were extremely small: moving from zero to one drink a day pushed the incidence rate of alcohol-associated health effects from 914 per 100,000 people to 918 per 100,000 people. This was the basis of the "no safe level."
- Four years later, an updated Global Burden of Disease analysis, also published in the Lancet, actually reversed course on this. After controlling for more factors, they found risk was actually slightly J-shaped, and for 40+ year olds the theoretical minimum risk was about 0.5 drinks per day. The WHO ignored this updated study when they wrote their opinion piece in 2023.
- Another big study with ~600k participants published in 2018 in the Lancet shows lowest total risk consumption levels at 1 drink per day, with mortality rising above that level. The WHO also ignored this.
- The JAMA study in 2023, with >4 million participants, showed no benefit to low-volume drinking, and also no increase in harms from low volume (1 drink daily) drinking. Again, the WHO ignored it.
Several other studies have shown either no discernible risk change between zero and 1 drink per day, or lowest possible risk at 1 drink per day.
But the WHO and much of the public health policy community ignores almost all of the actual credible science on this because it does not fit with their ideology. It's a huge problem with public health policy: the actual scientists and statisticians are doing great work, but if it's not what the policy people want to hear, they discard it and keep repeating their mantra.
So in the end results are in all practical implementation of this science is exactly the same: Even one drink of alcohol is the threshold
But if you are overweight or a preclinical obese, any amount of alcohol put you closer to clinical obesity.
For alcohol, almost every big study has shown either no discernible effects going from zero to 1 drink per day, or actual slight benefits going from 0 to 1 that then reverse when increasing beyond that. There is not strong consensus which of these is true, but there's definitely no real evidence that 1 drink per day is harmful as measured by all-cause mortality.
See GBD 2020 in the Lancet (which showed a slight J-curve), Wood in the Lancet 2018 (optimally low risk at 100g ethanol per week, or ~1 drink per day), Zhao in Jama (no discernible negative effects at 1 drink per day, no benefits either, no J-curve), and the latest National Acadamies review showing lowest all-cause mortality at about 1 drink per day (so, a J-curve).
Alcoholism has tons of negative effects. But these seem to start at around 2 drinks per day and rise rapidly from there.
nation states trying to slow progress. I was pretty perplexed as well, now we're concerned about data centers? They're everywhere already.
Oh wait, it's a FUD campaign sewn by competing nations using social media so they can, hopefully, catch up. That makes a lot more sense.
https://www.nytimes.com/2026/07/09/business/china-russia-ai-...
Just as complicit in obfuscation of reality through quantity of time spent in a topic: https://ourworldindata.org/does-the-news-reflect-what-we-die...
Thermal efficiency is in the 30s, so roughly 2x the electrical output as waste heat. Looks like waste heat from nuclear is still quite a bit more than from data centers.
Data centers are growing quickly and that's not including the waste heat from generating the energy the consume, just taking a quick look at the relative direct emission.
Although to be fair, when was the last time we actually managed to build a new nuclear plant?
https://en.wikipedia.org/wiki/Vogtle_Electric_Generating_Pla...