Hacker News Reader

Comment by pjmlp | Hacker News Reader

What a surprise with all the wars going on, and AI depleting Earth resources, what a change from about the pandemic era when everyone was into paper straws and cups and promising to be a better person, because that is what was going to change anything.

View on HN

All data centers in aggregate (AI and all other uses) use about 1.5% of electricity production, which itself is about 20% of total energy use.

So when people are focusing on AI above all other energy uses, it doesn't really paint an accurate picture of what's going on.

I've heard many different groups tell me their small fraction is not the small fraction that matters.

It's not really about which one matters. They all matter. But here is a rough breakdown of global fossil fuel energy usage:

* Electricity: 27%

* Industry: 24%

* Transportation: 15%

* Agriculture & land use: 11%

* Buildings: 7%

Then within electricity, data centers use about 1.5% of global electricity. Within data centers, AI accounts for somewhere between 15-20% of energy use.

So if you take 27% × 1.5% × ~17%, you find that AI is currently responsible for something like 0.07% of global fossil fuel emissions.

It definitely matters in the "every bit matters" sense, but also the numbers paint a really different picture than you'd get from statement like the one we started with.

You can split up every single industries/topics/&c. into "yeah but it only use 1.5% of energy", "yeah but it only produces 1.5% of the co2"

Guess what happens when you add them up...

This kind of logic only works if the percentages for each industry are all equally that small, so you can treat them as all equally bad, but they are absolutely not.

They're all that small if you split them as OP did. Just look at "transportation", it's like 25% of co2 emitted globally, but once you break it down:

Aviation is 2.5%: https://ourworldindata.org/global-aviation-emissions

Shipping industry is 3%: https://www.transportenvironment.org/topics/ships

Large truck freight is 3%: https://www.statista.com/statistics/1414750/carbon-dioxide-e...

Medium truck freight is 1%

The single biggest non divisible sector you can realistically come up with is "personal transportation", but even that is only 10% of global co2. You can look at other sectors like "industry" and "energy" and I can guarantee you will be able to easily split things down into sub categories which have <5% impact on global co2 emissions.

But they didn't split it up like that. They said all data center emissions irrespective of what those data centers are used for — which can be an extremely wide variety of things since data centers basically run our entire network information internet economy. That's much more like saying all transportation emissions instead of splitting it up by type of transportation. Yes, that doesn't include the full life cycle emissions of creating the data centers. But I'm pretty sure that transportation, as a proportion of emissions, doesn't also include the full life cycle emissions of producing the cars, trucks, boats, and airplanes in the first place.

Also, I think it's worth pointing out that the sectors you list are like 1.5 to 2x larger than the one he gave and the largest nondivisble sector, you listed is literally 10x, which I think does more to prove his point than yours.

Also, by your logic, literally any new sector of the economy that uses any amount of energy, basically at all, should be banned, because it "all contributes." that's a consistent position to take and there are certainly people that hold that position, but that at that point seems like a fundamental axiological difference that I and probably OP are simply not going to agree with you on.

>Guess what happens when you add them up...

I'll guess, they add up to 100%?

I don't see what's the insight here.

Thats OP's point - you need to reduce usage everywhere and pointing out that AI is only 1.5% doesn't take away from the fact that usage needs to be reduced there as well.

Wasn't crypto a significant percentage as well? And that was before the AI buildout started.

Not even close. Crypto has always been able to cut their own emissions before needing lots of compute.

AI on the other hand cannot, and still needs thousands of wasteful data centers.

It will normalize though once everyone is out of a job

What otheer industries are hyping the need for tens of gigawatts, maybe hundreds? On top of that they are hyping the idea of building utterly unrealistic space stations that would cost 10 times what the ISS cost. So maybe people are focusing on the dishonesty instead of the energy use. One or the other I suppose.

AI is nothing compared to automobiles and heating, construction and shipping.

It uses a bunch of energy, but not so much compared to moving yourself around in a car of plane.

A bunch of energy, water and Earth rare materials, nothing really to care about.

It is also much more likely to use renewable energy. Data centers look at the local energy mix when planning where to put one. (though they are perhaps taking energy that would otherwise be shipped to a different city/state)

This is definitely not always the case[0], let's not pretend these companies give a fuck about the communities in which their data centers operate[1].

0: https://www.theguardian.com/technology/2026/jan/15/elon-musk...

1: https://www.texastribune.org/2025/10/09/texas-hood-county-cr...

Community has nothing to do with it. They are asking about energy mix because that is important to enough of their customers that it is worth looking at.

Yes. The obsession with demonizing AI/data centre loads seems to be a deliberate distraction from the much, much larger carbon loads of the economy at large relative to which IT power consumption is a tiny proportion.

I think it's much less cynical than that. People both fear and dislike AI, recognize that the "it may destroy my livelihood and commodify human creativity" complaint falls on deaf ears, and are latching onto anything resembling a credible ethical complaint that people may actually listen to.

Animal agriculture is around 15% of global emissions, and AI is probably .1% to .5%, but sure, stop using LLMs. That will solve the problem.

Indeed lets tear down forests to build soya fields for all that fake meat tofu.

You must have been misinformed, they tear down forests to grow soya to use as feedstock (mostly for beef). Nothing to do with fake meat tofu, quite the opposite actually.

The point stands.

"I'm not going to stop openly blaming vegan products after learning it's actually beef manufacturing that consumes all this soy" or how should one read this comment?

What was your point? My point is that if more people were vegan there would be less deforestation. Instead of eating the beef that eats the soy, you eat the soy directly, which is much more efficient. It's unclear if you think it's false or if you had a totally different point.

Why bring the vegans into that when they are the ones consuming the fewest resources?

But the real meat requires much more soy for animal feed so it's still a better choice.

This is a common fallacy. https://news.ycombinator.com/item?id=47275088#47280922

> Animal agriculture is around 15% of global emissions

The majority of which is methane, which only has a 7-12 year life. Which means — unless for some reason you started eating way more animals than you did yesterday — that your emissions today simply replace your emissions from 12 years ago. In other words, it is a stable system, unlike carbon, which basically sticks around forever.

But aren’t we doing just that? Many more people eating more meat than ever before?

Methane has a shorter shelf life, but is far more potent, meaning that any increase is worrying, and decreases could have a dramatic impact.

I'm not. If you are, well, I guess you can't control other people.

Yes, what a bright future where we feed our bodies with a soy slop, and our brain with AI slop.

Most people pushing back against data centers simply don't want invite something into their city that will use up resources (likely raising prices), while the big selling point is that it will put them out of work. You can say that won't actually happen and everyone will keep their jobs, but that has not been the messaging. CEOs want to know how many people they can get rid of once they start using AI. Why would anyone sign up to have that in their backyard?

> AI is nothing compared to automobiles and heating, construction and shipping.

When the oil in your frying pan is smoking, adding a tiny bit more heat may be unwise.

Yeah that assumes that AI is an absolute negative. What if it impacts positively? I mean you gotta spend money to make money.

I would be fine if LLMs disappeared tomorrow, but if I couldn't heat my house, I'd freeze to death. But I guess some would argue that everyone needs to live in a city with district heating.

The heating of your house is worth the energy of many thousands of llm queries. So only turning down the thermostat by the smallest of steps would compensate your llm use.

So, no you don't have to freeze to death to easily compensate your llm use.

Which in turn are also relatively small compared to the damages of cattle and fishing.

Seriously adapting our diets around being more sustainable. I'm not advocating for veganism or such, but at least to understand that eating a burger pollutes as much as driving a large vehicle for 50 miles and that maybe we can substitute that with poultry or eggs or cheese many times.

What about putting trees down for soya fields?

Speaking of deforestation:

> Globally, 77% of soya is produced for animal feed, 19.2% for direct human consumption and 3.8% for industry (biodiesel, lubricants, etc.).

https://www.deforestationimportee.ecologie.gouv.fr/en/affect...

Poultry protein efficiency is 21% and beef 3%

> We find that reallocating the agricultural land used for beef feed to poultry feed production can meet the caloric and protein demands of ≈120 and ≈140 million additional people consuming the mean American diet, respectively, roughly 40% of current US population.

https://iopscience.iop.org/article/10.1088/1748-9326/11/10/1...

Edit:

Note the soy usage vary around regions. The first link is from a French gov page, after the previous citation it gives the world repartition:

> In animal feed, the largest consumer of soybeans is chicken (37% of world production), followed by pork (20.2%), aquaculture products (5.6%), dairy products (1.4%) and beef (0.5%).

Which is quite different than the French reparation :

> flesh and egg poultry account for 44% of total [imported] soy, then diary/mixed cows (36%) then flesh cow (8%) and pigs (6%)

(fr) https://chaire-bea.vetagro-sup.fr/en-france-les-animaux-dele...

Soy protein are 97.9% as digestible as beef

> protein qualify can be scored in terms of its Protein Digestibility-Corrected Amino Acid Score (PDCAAS). Soy achieves a PDCAAS of 0.92, comparable to beef at 0.94.

https://ourworldindata.org/land-use-diets

The DIAAS is sensibly worse (soy 0.898 - beef 1.116) but still far from the beef protein efficiency cited above (3%)

https://en.wikipedia.org/wiki/Digestible_Indispensable_Amino...

> AI is nothing compared to automobiles and heating, construction and shipping.

Let's stop building homes to live in and use motor vehicles for travel. Smartphone with LLM access will fulfill all our needs.

As we all know, global warming wasn't happening until AI data centers started being built. They are honestly a drop in the bucket

Reply on news.ycombinator.com