Opposition to AI datacenters has emerged as a primary theme in US politics, one that – surprisingly – doesn't fall along party lines. We applaud people coming together for constructive debate on any issue, and agree that communities need to evaluate whether any economic benefits these datacenters bring is worth their costs. Still, we worry that a focus on datacenters obscures the larger impacts of AI on people's lives: the concentration of power of AI companies, and their widespread political and financial influence.
Local datacenter opposition is grounded in legitimate concerns about misallocation of land resources when housing is at a premium, pressures on already higher energy prices, and localized environmental impact. Unlike other resource-consuming and polluting industrial facilities, datacenters produce very few jobs . The fact that US opposition to datacenters seems to be most fierce among lower-income communities reflects righteous indignation with an inequitable bargain, where tech companies and developers profit from exploiting local resources but offer little in return. On a global scale, their carbon footprint could grow unsustainably if usage accelerates. And all this is in aid of a technology that many fear will propagate misinformation, take their jobs, or even cause existential risks for humanity.
For some, datacenter opposition may feel like the only tangible mechanism for registering their concern, disapproval, or even anger about AI. The problem is that this may be exactly what the AI companies are banking on. They can overcome the protest when it matters to them, and live with a significant fraction of proposals being defeated. More importantly, focusing political opponents on the datacenter issue obscures the bigger prize they're after.
While there is a staggering three-quarters of $1tn being spent on datacenter infrastructure by US companies this year alone, this investment should be taken in perspective . The market for enterprise software, for example, is about twice this size. And it's small compared with what these companies actually want.
AI companies have their eyes set on capturing all the value created by entire industries. The technology has arguably already conquered customer service and consumer sales. But on the horizon are bigger targets, such as enterprise software development, creative design, management and even legal services. In AI companies and their allies' vision of the future, AI replaces teachers and doctors . The companies would rather spend time fighting resistance to how fast they are building computing infrastructure than dealing with issues of how their products should be used in those fields, or how those fields should be protected from their products.
And while datacenter opposition campaigns have been successful in building widespread appeal , their effectiveness in the US is mixed. They seem to be most successful when organizing against speculative , early-stage datacenter proposals that have a relatively low likelihood to ever see fruition. Meanwhile, advanced-stage, well-capitalized datacenter projects have proven to have the resources to overcome local opposition. An OpenAI- and Oracle-backed facility in Saline township, Michigan, is breaking ground on construction even after local officials voted to reject it. The developers sued the town of 3,000 and forced a settlement that involved their project going forward. Meanwhile, the Trump administration, a vigorous ally of corporate AI, has signalled its willingness to advance AI infrastructure development by overriding state objections and even using federal lands .
[...] As for AI itself, the concentration of power and wealth in these tech companies is the greatest existential risk facing society today. This means we must limit corporate power, especially corporations' ability to exploit the public and manipulate our political system.
Opposing datacenters should be just a starting point. We can advocate for states to regulate AI, to reject irresponsible uses of the technology, and shape corporate behavior. We can fight for AI computation to be taxed , so that the public can capture some of the profit of AI use while also forcing AI companies to internalize more of the energy and environmental consequences associated with its use. And we all can join the global movement for Public AI , an alternative ecosystem for AI that is developed under public control with an incentive structure to create public benefit rather than private profit.
[...] Political organizers should call out and reject the AI companies' framing of the debate, and reorient campaign agendas around populist resistance to corporate concentration of wealth and power. When AI companies pump millions into legislative races, the result should not be hyperbolic discussion of AI superintelligence. And when a plot of land in a small town is pitched as a datacenter site, the debate should be about more than the local costs and benefits. It should include out-of-control money in politics, and Citizens United -proof solutions to limit corporate influence like public financing and state regulation .
We all have a vested interest in what's on the policy agenda, and what the outcomes are. Today, the greatest risk AI poses to society is the exacerbation of inequality and the concentration of wealth. The real problem is trillion-dollar AI companies and their trillionaire oligarchs cozying up to political power in Washington and governments worldwide, and using their money to enact their agenda over the popular will of the people. This is the issue we'd like to see put front and center, and it requires solutions much more extensive than slowing datacenter development.
« South Korea Making its Own Security-Centric AI Model | Google Breaks Alibaba’s Record For Europe’s Largest DMA Fine »
Related Stories
The Global Government Affairs team over at X dropped a bombshell Thursday night, revealing that the X Safety team had uncovered proof positive of a foreign influence campaign targeting the build out of AI and data centers in the United States:
It turns out that, after launching an investigation, the team uncovered a massive bot farm consisting of roughly 200,000 "suspected Chinese inauthentic accounts," including hundreds that appear to have been trying to manipulate debate in the U.S. over artificial intelligence, data centers, and the associated energy costs.
This development comes after President Trump warned last month that "other nations are trying to get us to slow down" and are spending money on "propaganda" to convince Americans that data centers aren't a positive thing.
According to X, about 200 accounts within the bot farm network were posting claims that data centers drive up household electricity prices and put excessive strain on the power grid. Some of the accounts, ironically, posted AI-generated cartoons depicting data center operators getting rich while ordinary American citizens were stuck paying the bill.
[...] The insidious thing about this kind of foreign influence operation is that it preys on the concerns that many Americans already have about the data centers popping up in their communities – debates about things like electric bills and the robustness of the power grid are already happening here. All a foreign adversary like China has to do is amplify those fears and sow distrust to make it harder for the U.S. to keep building out its AI infrastructure.
Related:
- Protesters Haul A Guillotine To City Council Meeting About A Potential AI Data Center
- Texas Halts Data Center Connections to Power Grid Amid Overwhelming Demand
- The Fight Against AI Data Centers is Important – but It's Just a Starting Point
- 'Cost Me the Election': Data Centers Trigger Voter Backlash
(Score: 1, Offtopic) by oregonjohn on Tuesday July 28, @06:01PM (2 children)
I'm curious how AI wars over corporate control will work out. Any sci-fi new or old works cover this type of scenario? It seems we are clearly in the thick of it now.
(Score: 4, Interesting) by JoeMerchant on Tuesday July 28, @09:49PM
Unfortunately, I think SciFi has mostly missed the nuance of "Lexical Prediction" which is an incredibly useful, yet dramatically limited, intelligence mimic - much the way Eliza was a "computer psychologist" in the 1960s. Eliza has a very small set of "typical next things" that a psychologist will say in a simple session and thus people interacting with Eliza "feel like" they're interacting with a psychologist on the other end of a text terminal connection.
LLMs use massive statistical models - sort of the same way AlphaZero uses predictive modeling to "look at" a game situation and predict what next move leads to victory most often from that point based on its prior experience of playing all possible outcomes, the LLMs train on Terrabytes of text inputs and from those infer what "token patterns" people most want to hear in response to the patterns they have input, and that's how they "win the game" by giving the expected/desired output patterns.
If the training set is good (SunTzu's Art of War, "Blue Ocean Strategy" by W. Chan Kim and Renée Mauborgne, "The Lean Startup" by Eric Ries, "Good to Great" by Jim Collins - along with millions of other relevant and non-misleading references) - then the token pattern recognizer has a chance of taking inputs from your chat prompt (the current game state) and synthesizing the next moves (tokens of text response) that should lead to the kinds of outcomes the training set contains - and hopefully it can distinguish winning from losing outcomes... not doing so is called "hallucination."
The interesting thing is: the worthwhile (accurate / factually based) training sets are from a world without LLM agents available, everything they might train on which includes them is pure speculation / mostly fantasy. So I would guess that well trained LLM agent CEOs might, eventually, perform well in traditional business roles in traditional business environments where training data is available. Expecting to extrapolate that accurately, or even beneficially, into the changing landscape of the future would be... optimistic in my view.
Meanwhile - spitting out graphing audio applications like I used to get paid to write 30+ years ago, in Python or Tauri or whatever... been there, done that, plenty of Dummies guides available that are good enough. LLMs are whippin' the Llamas' asses on those jobs.
🌻🌻🌻🌻✌️ [google.com]
(Score: 2) by Reziac on Wednesday July 29, @05:34AM
I always think of the Forbin Project.
"The truly liberal mind is by definition uncertain; it admits it may be wrong, but once set and the decision made the wavering stops, and no sort of hell can sway it."
-- D.F. Jones, The Fall of Colossus
And there is no Alkibiades to come back and save us from ourselves.
(Score: 0, Flamebait) by khallow on Tuesday July 28, @06:19PM (6 children)
(Score: 0) by Anonymous Coward on Tuesday July 28, @06:46PM
> ... Make more power plants.
You forgot to mention the increase CO2 concentration in the atmosphere. While nukes and solar/battery might be long term solutions for data center power, in the short term they are going to burn oil or natural gas, because that's what is available.
(Score: 5, Insightful) by VLM on Tuesday July 28, @08:17PM (4 children)
The financial industry does not really understand the AI bubble.
Your stereotypical muni/green bond (for solar and stuff) is 10 to 30 years.
Your stereotypical PPA on the demand side is negotiated for like 15 years.
For some crazy reason the OECD caps nuke plant project bonds at 15 years and that has supply/demand impacts on the entire industry such that for all practical purposes the nuke industry is built on the vague handwaving general idea of 15 year term bonds to pay for glow in the dark steam kettles. Worldwide, not just USA.
The point I'm making is the financial industry as a whole is "stuck" funding the energy sector as a whole for vaguely 15 year projects, and the AI bubble will pop long before then. The industry and its funders don't financially speaking understand funding a short term bubble, its going to crash them, and crash them hard.
After the bubble pops it'll be chaos. Kind of like the temporary deflation in the housing bubble in '08 exploded several banks, flipping the power switch off on hundreds of GW nationwide and not paying the bill anymore will result in some wild collapses.
Issue "totally safe and sorta but not really government backed security bonds that are financially engineered to be nearly risk free as long as nothing goes wrong, then the underlying security value goes to zero a couple years in and whoopsie..." its the same scam as '08 just operating in a different industry. Prepare for massive taxpayer funded bailouts by 2030. "Nobody ever could have forseen that an enormous financial bubble could pop..." uh huh...
It's possible in ten years we'll be a 100% solar country, not because of the growth of solar (beyond existing trends) but due to financial collapse. TVA is probably not going to be issuing bonds on their 100 year anniversary due to the AI crash, kinda crazy to think about they've been issuing since 1950s at least I think. Probably won't be quite that bad, but its going to be bad indeed.
It's going to be hard, post crash, to sell bonds for awhile, even for unrelated projects (repairing or mitigating hurricane damage or solar projects and stuff like that)
(Score: 1) by khallow on Wednesday July 29, @12:01PM (3 children)
As to the bubble and its eventual popping, I think the US has moved into an economic phase where bubbles are highly attractive - due to enormous government debt. The federal government is highly incentivized to juice up market bubbles (aside from the usual corruption). Look at it this way, the only time in recent decades that the US came close to neutral budget was in the 1999-2000 financial year when capital gains from the dotcom bubble surged high enough to almost eliminate the deficit for that year. When you have a huge debt, ongoing massive deficit spending, and a short term view, the lowest pain way to pay down that debt is via capital gains tax on a massive asset bubble. That it will soon transition into high deficit spending and jobless recovery is future me's problem and future me will at least be able to act heroic and stuff. That future borrowing is a large part of the reason the bond markets crater in a recessionary world.
I don't think the electricity providers will be hurt that much - at least the ones that aren't heavily investing in AI directly, of course.
(Score: 2) by VLM on Wednesday July 29, @02:42PM (2 children)
Can't have it both ways, either massive infrastructure investment is exploding energy bills, which seems geographically limited to corrupt states and not so in less corrupt states, or when the bubble pops they're going down just like the banks in '08.
Its possible we don't have a common frame of reference enough to even debate.
Ignore the data center/AI stuff and imagine a fidget spinner factory. Take out a loan for 15 years to expand the factory massively, which will be paid off monthly by selling fidget spinners. Fad bubble ends in five years. No longer paying the bank. Well the bank can repo the factory but then they have to mark to market that the $1B factory is worth at most $1M due to lack of demand. Now its a bank balance sheet problem and the bank collapses.
Its the same thing with electricity. The entire industry is built on steady slow growth and permanent demand. So keep on funding plants and transmission lines using 15 year bonds. Then flip off the revenue source when the bubble pops. Whoops no way to pay for those bonds anymore and they already raised rates to high as the public utility commission would let them (depending on local corruption levels, this is 'not much' all the way up to 'a lot')
(Score: 1) by khallow on Thursday July 30, @12:13AM (1 child)
I disagree. It depends on who pays for that infrastructure. In said corrupt states, it's a hot potato game that everyone involved will lose. In the states that aren't so corrupt, there's a simple solution - have the large data centers pay for the infrastructure they need. At that point, when drama hits the AI bubble, the electricity infrastructure is pretty much clear.
Another thing to keep in mind is that there will be future demands on electricity from data centers and computation. I think it's a bit like railroads in the 19th century. And electricity companies can discontinue to some degree infrastructure that didn't work out. And worst comes to worst? Someone else can buy the assets of the company and keep the lights on.
(Score: 2) by VLM on Thursday July 30, @01:18AM
Yeah, that's exactly the problem.
The powerco doesn't understand financial bubbles as a large significant customer.
Someone signs a long term purchase agreement for 15 years for xyz GW avg.
Turn around and wave those PPA in the face of bond buyers to sell them 15 year bonds to build another natgas plant or transmission lines or expand a coal burner or "whatever" to meet those purchasing agreements.
Make a profit off the spread for 15 years, as much as the state P.U.C. will allow.
This is how its done. How its pretty much always been done. It will not work if 1/4 of your revenue is a bubble that will disappear when the bankruptcies roll in.
The powerco doesn't "understand" that a large fraction of its demand will evaporate in 3 years when the bubble pops. How much revenue do you get on a 15 year purchase agreement two years after the counterparty declares bankruptcy? Nothing? Well you can't pay your construction bonds with "nothing". Thats going to be a BIG problem.
You can try options strategies and purchasing insuranceon the purchasing agreements; good luck; even if you get a counterparty, thats going to end up like the '08 crash where the system was not designed for every insurance policy to be cashed out at once.
You can't just say "have the customer pay for the expansion" because the system intentionally doesn't understand it. This is "new" territory for electrical companies. In some cases the state P.U.C. will make it illegal not to bankrupt themselves in the name of "universal service" even if they know the "15 year agreement" will hit the paper shredder within 5.
(Score: 5, Insightful) by Anonymous Coward on Tuesday July 28, @07:28PM (2 children)
The datacenters aren't being built so you can generate anime girls or have some LLM do your taxes. They are part of the surveillance state and will be used accordingly. No article about them ever brings that up.
Notice how fast they are being built, all the drama and pointless circle jerks. Trillions now invested and it sure as shit isn't for hallucinating chatbots.
(Score: 4, Funny) by SomeGuy on Wednesday July 29, @12:57AM (1 child)
Ok.
But can we still have the anime girls?
(Score: 2) by Bentonite on Wednesday July 29, @06:15AM
No - access to the slop copier will be taken away, unless you can cover the $10,000/month/used running costs.
There are plenty of actually decent drawn anime girls that already exist - why not look at those instead?
(Score: 3, Informative) by jb on Wednesday July 29, @05:15AM (5 children)
These are not "AI companies". They are LLM companies. Big difference. LLMs are not AI: there's no intelligence (artificial or otherwise) involved whatsoever.
For goodness sake stop pushing the enemy's propaganda for them by adopting their own false claims, even in your valid criticism of them.
(Score: 3, Touché) by Reziac on Wednesday July 29, @05:44AM (4 children)
I don't much give a flip what they do with their LLMs, by any name. That horse is out of the barn and running laps at the racetrack, and the barn is on fire.
What concerns me is when they come along and outbid my little local power co-op, so if we want electricity now we have to pay half again more, to no benefit. Or suck groundwater out of the desert and dry up private wells for miles around. Or fuck up a municipal sewage system because their first flush overwhelms it. That's the sort of thing that needs addressing. So far they have not demonstrated that they'd be entirely good neighbors.
Yonder is Detroit, with lots of vacant industrial land and a city that will GIVE it to you for the asking. And infrastructure already exists. Why not there instead of Broadview, Montana, with one power line and no water??
And there is no Alkibiades to come back and save us from ourselves.
(Score: 4, Insightful) by Bentonite on Wednesday July 29, @06:24AM (3 children)
It might be something to do with that there's a remote possibility of hiring enough people to ransack and burn the place down in Detroit, despite the barbed wire fence and multiple checkpoints, while in Broadview, Montana, there are like 150 people and it can be immediately realized if hundreds more show up and also, slowly and carefully, what's left of the wilderness/desert can be destroyed.
(Score: 2) by Reziac on Wednesday July 29, @06:55AM
Ouch, but yes, this has occurred to me.
And there is no Alkibiades to come back and save us from ourselves.
(Score: 3, Touché) by VLM on Wednesday July 29, @02:49PM (1 child)
People think Detroit is the size of Michigan, but most "on foot" rioters won't march more than five miles, which is no far in the grand scheme of things.
Bloomfield Hills is about 25 miles from the hood in Detroit. Put your data center about 30 miles away and they'll never finish looting and burning Bloomfield. Heck put it two miles away and it'll be safe enough.
People talk like you can't be safe from the hood unless you're 500 miles away and there's only a binary option of live across the street from Fent Park or live 500 miles away from Fent Park and there's no land in between.
However, break open the Google Maps and figure out how far apart the local version of Fent Park is from the closest, lets say, strip mall with a Hobby Lobby and a Panera and a chiropractor in it, thats about how far away you have to be minimally safe, build the data center a small multiple, figure a couple miles, it'll be fine.
Note that "around Detroit" does not exactly lack for available electrical power or water.
(Score: 3, Informative) by Reziac on Thursday July 30, @12:27AM
Yeah, even where scum and lowlife riot for weekend recreation, it generally does not penetrate into industrial areas. Come to it, fences topped with razor wire and an open buffer zone adequately discourage all but the most adventurous.
Don't needlessly build near old residential areas, and they won't even know you're there.
And there's nothing much to break or burn on a concrete slab building.
However, I can see the techdudes looking at the run-down parts of Detroit, and thinking anything they build will vanish into the sludge.
BTW Broadview was a real example:
https://www.kulr8.com/news/proposed-5-000-acre-ai-data-center-near-broadview-raises-questions-from-residents/article_18781d68-ebc4-4d32-845c-38fe62de9e9c.html [kulr8.com]
https://montanafreepress.org/2026/06/17/agreement-between-data-center-developer-and-union-federation-thin-on-details/ [montanafreepress.org]
Hey techdudes, since you already mostly own Colstrip, build over there, next to the coal-fired plant. Reclaimed strip mines would be the perfect place.
And there is no Alkibiades to come back and save us from ourselves.