That’s a great thing.
If the data center which requires the grid improvements is fully paying for the upgrades, I'm OK with it. OTOH, if they expect to fob off the cost of the otherwise-optional upgrade on the rest of the ratepayers, Fck-NO!
And for those who say "but it benefits the rest of the ratepayers, so they should pay too", NO! If you (person or corp) come in and demand we change something, you should be a net benefit for everyone, not just yourself.
Be a good citizen and help everyone, and you'll find support. Don't, and find people see you as an asshat and you'll find opposition.
More importantly, what says the grid improvements will go beyond upgrading the corridor to the new data centers? Sure, that might benefit customers right along that route, but not customers in general. IOW, even the "overall" improvements are likely to be limited, but the costs will be borne by everyone.
That said, the way grids are going, I consider the right solution to be primarily off-grid solar+battery on every house, maybe with local grids to share/level excess capacity, and maximum local solar/wind+battery for commercial. That would yield a robust power system far more resistant to human disasters relevant to civil defense and natural disasters; when every house can maintain basic communications, refrigeration, and sanitation on a near-indefinite basis (which is not possible with a grid-based collapse/destruction), disasters are far more survivable.
Your argument conveniently ignores that electricity is tiny fraction of the average American's budget, which is far more heavily weighted toward housing, gas, and groceries. This argument is extremely tenuous.
> for the sake of shareholder value
This is emotionally manipulative and without substance. Data centers are being built because there's demand for them from people who want to use AI, like me. These data centers drive down the cost of AI, which is valuable to me, and to society. The fact that it also happens to increase "shareholder value" is irrelevant.
No they aren't
They're being built so shareholders can profit off of people like you. The fact that you benefit from it does not actually factor in any way aside from the fact that you will pay them for that benefit
They aren't benevolently trying to help you out, they just want your money.
The fact that you want AI and think it will help you and society is actually what is irrelevant. If they weren't confident it would make them money it would not be built no matter how beneficial it would be.
> No they aren't
You literally said they are building data centers to cater to demand later on in your own comment:
> If they weren't confident it would make them money it would not be built
There you go. Your own point debunked by your own words.
> They aren't benevolently trying to help you out
I didn't say anything about the owners building stuff out of charity or benevolence. Are you responding to the wrong comment? Yours is more-or-less irrelevant to mine in addition to being self-contradictory.
> The fact that it (building data centers) also happens to increase "shareholder value" is irrelevant
It's not irrelevant. It is the primary reason they are being built
Your understanding about the cause and effect around why data centers are being built is completely backwards
> It's not irrelevant. It is the primary reason they are being built
...and without any implications, that fact is irrelevant. Do you understand what relevance means? It means that there's meaning, implication. You haven't given a single follow-on implication from that hypothesis that actually matters to the discussion.
> Your understanding about the cause and effect around why data centers are being built is completely backwards
No, this is just factually false.
Fact: for datacenters, without market demand, without a buyer, shareholder value cannot possibly be generated.
So, first there is a market, and then companies act on it to generate shareholder value.
You're living in a world detached from reality where "shareholder value" exists as its own separate entity that can generate market demand.
You need to take a high-school level economics course so that you don't make objectively incorrect statements like this.
If you think people's motivations are irrelevant then we just live in different worlds, and there's no point continuing
I never said that peoples' motivations are irrelevant in general. Just that they're irrelevant in this specific case. You're putting words in my mouth, which clearly indicates malice.
> then we just live in different worlds, and there's no point continuing
We do. I live in reality, where most people understand basic facts of economics, causality is meaningful, and you still haven't actually provided a reason why you think that the motivations matter, so as far as the rest of us in the real world know, they don't. You've had three comments to point that out, and you haven't, so the reasonable conclusion is that you don't have a reason, and you're running away to avoid that intellectual dissonance.
You assume people's psychological resistance to increased cost is entirely based on the proportional impact of those cost increases. Yes, housing/gas/groceries are much larger portions of the increase in cost, but datacenters are much more visible than poor foreign policy or starting new conflicts with overseas energy producers. Also, electricity traditionally has much less cost variation than gas/housing/groceries.
> This is emotionally manipulative and without substance. Data centers are being built because there's demand for them from people who want to use AI, like me.
At my company's internal conference we had a rep from one of our partners come speak, and he emphasized that a significant value-add of our company's relationship to our partners is that we're much more tech-forward. he went on to say "Execs in [partner industry] don't know what AI is or what it does, but they know that they want it."
We're seeing AI features being bolted into every application under the sun, whether it actually adds functionality or not. I think you bear a significant burden of proof to the claim that the demand for datacenters is in turn driven by consumer demand for AI, and company boards losing their minds over what they think AI can do and demanding their R&D divisions sink millions of dollars into AI spend to try and capitalize on their dreams.
I'm reminded of Bitcoin: for a hot minute every startup was talking about blockchain and GPUs were getting snapped up by miners left and right because of what people imagined the value could be. That's what I'm getting at when I say "shareholder value:" these developments seem to be less about what value a company is _actually_ producing (& thus the inherent cost-of-goods-sold demand), and more about the _possibility_ of what value a company _could_ produce.
This is a good point. I see that I misunderstood your original argument - I thought you were claiming that the increase was going to be financially meaningful, rather than what you actually said, which was the perception of and response to the increase. My apologies. Thank you for your thoughtful response!
tbf, I didn't state the psychological thesis up front either :)
Or two jobs and a "side-gig" or three, and still barely "squeaking by", and still unable to afford health care or any other such "luxuries" of life.
If the data centers hadn't been _quite_ so hostile to absolutely all their critics and instead had simply offered to pay for some local schools or libraries or something, at like 5% of the cost of the data center, this would have gone differently.
Instead Altman says it will destroy the world: https://www.techradar.com/pro/quote-of-the-day-by-sam-altman...
So it's bad for the world according to its proponents, and at the same time, a bad neighbour? Why would anyone not vote against it given those?
Datacenter is not bringing more economic activity to the area.
the questions are 1) who is footing the bill and 2) who is benefiting from that economic activity
if it's not the local community, then why should they agree to it much less pay for some of it
/s
I commute by rail to the tune of ~6000 miles a year. How long would it take for those gas turbines to make those greenhouse savings worth it?
Some napkin math, lets say a conservative 50MW/day? half what the XAI Colossus DC uses. So they'd be generating about 2MWh running 24/7.
Those gas turbines spit out ~700kg of CO2 per MWh.
My car would emit about 5 tonnes of CO2 over about 15k miles(2L hatchback).
The drive is far less direct than the train, so lets call it approximately 10k miles in commuting.
In a day, one of those gas turbines would produce about 16.8 metric tons of CO2. Over 3x what my car would produce. In a day. It takes me all year to save 4ish metric tons of CO2 from my personal carbon emissions.
Yeah, so screw your small town and screw the datacenter. We all share this world. Your small town isn't worth the damage.
I don't live there, but this attitude won't make you any friends.
>We all share this world.
Yes, tell me how me and you and Bezos and Musk are all equally sharing the world. We're all in this together and you care so much about the pollution of other peoples' small towns. That's why there's nice neighborhoods and not so nice neighbourhoods, nice areas to live in and not so nice areas to live in. Because some people have to deal with the pollution so that you can live in nice neighborhoods without any pollution while still enjoying all the luxuries of modern society that generate pollution except not in your back yard.
>My car would emit about 5 tonnes of CO2 over about 15k miles(2L hatchback).
And where was that car manufactured? In your back yard or in some industrial town with more pollution? Do you oppose car factories there too?
It's CO2, it's in my town too.
I'm saying, the minor economic boon for this small town is not remotely worth the pollution over the lifetime of that data center ran off turbines.
> while still enjoying all the luxuries of modern society
We could enjoy the modern luxuries a bit more without Altman Et. Al. working to centralize labor into expensive automated Capital that we'd all have to rent back. Hell, even with some pollution caused by relocating manufacturing stateside.
I firmly believe we need to reintroduce friction into our lives to provide them some meaning and experience beyond ordering food on apps and looking at ads.
> And where was that car manufactured?
Michigan.
That's why in my country the government is incentivizing via fiscal policies for employers to be cancelling all home office perks to employers. To get people back to spending money on commuting.
>Michigan
So what's your opinion on building car factories there? If we remove those factories, air there would be cleaner.
Are you arguing we did good so now we can cheat a little bit and have some ice cream? Like someone falling off the wagon on their diet.
Anything less is just insufficient.
(now you have two planning permission problems!)
2) what's the plan for handling the long-term waste, exactly? how many sites are available?
2) Dry cask storage. It works and it is extremely safe, you can leave the casks on site or bury them wherever you want. The most significant concern with nuclear waste is not the radiation, it is the possibility of radioactive materials leaking into the environment and that is exactly the problem dry casks solve.
2.1) Reprocessing, you can reduce the volume of nuclear waste by recycling fissile uranium and plutonium from spent fuel into new fuel pellets. The remaining waste is typically embedded in glass, which can be placed in dry cask storage (see above).
Nuclear waste is not the problem people think it is. We know how to deal with it and we have been dealing with it for a long time now. For comparison, solar panels do break down over time and solar panel waste includes several toxic heavy metals that can leak into the local environment. Nobody objects to solar power by pointing to the waste stream (let alone the waste created by all those batteries that you need to make solar power useful for industrial operations like DCs).
I'd imagine it's something like each American eating half as many cheeseburgers a year would offset all current and projected emissions from DC energy usage.
Yep. And that will decrease air quality locally, which is especially bad for people with asthma or other sensitivities. And increases greenhouse gases globally. And this natural gas generator will probably be less efficient than a normal power plant.
> This is economic growth and a good thing.
Not much. Sure, there will be some construction jobs while it is being built, but after that, there won't be many jobs for the DC. And there is a good chance the jobs that will exist won't go to local residents. And the owners of the DC aren't going to be buying much if anything from the local community.
Your other points are generally very well-thought-out, but this one is not.
Grid expansions take years. Building new nuclear power plants takes decades. And there's "slack" in the sense that utility prices can raise substantially before there's actually impetus to start planning a grid expansion, and then you still have to wait years for construction to finish.
And the party that takes the hit for those sustained elevated electricity prices is local residents, who in general have far less capital and are disproportionally hurt by it relative to the owns of the data centers.
You can make the argument that the electricity price increases are going to be lower than most people think and wouldn't meaningfully hurt the average local resident, but that's a separate argument. I don't buy that "if it pulls power from the grid that's a good thing" is valid.
If there isn't enough power to meet demand, then build out the grid first, or tax datacenter power consumption so that the inflection pain-point before grid capacity expands is less harmful to the residents.
Now, I'd argue that building has turbines have negative environmental effects, but that's a separate matter.
Got any hard numbers on the tax revenue? I am skeptical because typically these projects receive tax payer money.
> Not to mention the nuclear power plant being built in response to the DC
A nuclear power plant takes 6-10 years to build in the US, and sometimes even longer.
According to the county FAQ: https://www.boxeldercountyut.gov/661/Stratos-Page-FAQ-Page
> What tax revenue is expected?
> In Phase One, the county expects to receive $30 million in annual tax revenue. When fully built out, it could generate up to $108 million every year.
"Could" generate 108 million a year
Naw man. These numbers need to be more concrete upfront. This sounds like "under ideal conditions (that we won't ever achieve)" weasling
This is for a standard "warehouse sized" 100MW facility no one would know exists unless you told them it does which has been there for 20 or so years now. It generates far more taxes for the municipality than buildings of similar footprint next to it, as it's in a light industrial park like most of these things are. Scaled up to 1GW that's $60m/yr in simple "sales tax" on electricity consumption. Property tax is harder to dig through, but ignoring initial 10-20 year abatements typical for such projects it's roughly double that - but that's from memory.
I agree the tax abatements are bullshit. But those are bullshit for any sort of development and imo should be outright illegal at a federal level regardless of the project.
This doesn't account for any local income tax/etc. generated by jobs - where a datacenter of course will not be significant compared to most industrial land uses of similar acreage.
These datacenter facilities have typically been very tax revenue dense land uses for quite some time now without cities needing to invest very much in infrastructure to service them. It's why they generally have sailed through approvals until recently.
Wow, that's a lot of electricity. What market is that electricity purchased from?
The locality tax is in addition to everything else paid on the electric bill.
And yes, datacenters have always been about converting electrons to photons, density of photonic output and electron input steadily increasing over time.
The bucket of money is from the food truck, the only business that can be sustainable from a DC in the local area. The food truck is cash only, because they payment company now charges 50% because they are using AI to counter the spam and fraud from the other companies using AI.
(Farms don’t employ many people either.)
I can't eat the output of a data center, though.
I've been really enjoying eating soybeans. Absurdly good source of protein (yes, even for a bean), dirt cheap, decent source of fiber and a bunch of minerals, and pretty easy to just drop in a pressure cooker and eat however you like once it comes out.
Can't say I eat a lot of field corn or hay though.
Hay is the stalk of a grain plant. We take the grain off the stalk and make it into flour. A main ingredient in bread and pasta/noodle.
The point is that something tangible is produced.
And the data center will probably contribute to it dying faster. It'll produce only a handful of permanent local jobs, and the long term economic output will largely be for rich companies who won't be in your town. Even tax revenue will be tiny because of the exemptions they give it.
> They're using natural gas to generate electricity. Everyone involved has said that it won't pull power from the grid. But if it does that's a good thing. That means we'll need to expand our grid.
Everyone involved will lie in exchange for money. Put it in writing, make it legally binding, and then maybe I would believe them. They haven't done that, have they? Also, who do you think pays for expanding your grid?
What on earth is this "If I get in a car accident, it's actually a good thing because I get to buy a new, better car" logical reasoning?
It'd be like your widget maker harvesting wood for their widgets from public land and exhausting it so the public is left without this resource, then expecting the public to pay to arrange for more wood to be grown to replace the wood they used up.
I wonder how adding gas generators helps this graph go down.
Now you could argue that billionaires, their investors and loan sharks should be allowed to buy up gas —perfectly legal— but it's actively screwing up people's lives.
The externalities are not paid by the oligarchs who own the datacenter.
If the buildout was solar or wind or even hydro, then you might have a point, but new fossil fuel buildouts should NOT happen in the year 2026!