HARMS AND DANGERS —DATA CENTERS AND AI
Water Issues
Threats to drinking water sources such as aquifers, rivers and lakes
In West Texas, three massive projects are all drawing from the Ogallala aquifer, the world’s largest. It is already depleting faster than it recharges. Recent research suggests that training and using AI could consume more water than the global bottled water industry.Depletion of water could leave communities without enough
Rural areas that rely on wells could run dry. Agriculture could suffer.Energy Issues
U.S. grid was not designed for this level of energy demand
Data centers run 24/7 and require massive amounts of energy to run AI. The U.S. grid infrastructure is old and should be upgraded, not stressed further.Consumer’s energy costs will necessarily rise, and opposition grows
The Trump administration is throttling wind and solar – the least expensive forms of energy. Big tech is pushing the costs of added energy onto consumers, who are demanding relief.Building Their Own Harmful Energy-Generating Sources
Power generation on-site because they can’t rely on the electric grid
- natural gas plants that emit chemicals harmful to health
- small nuclear reactors, relatively untested, requiring all spent nuclear fuel be stored on-site. The federal government is supporting this.
Pollution Issues
- Emissions from backup generators that emit methane.
- Water in closed loop systems requires a lot of chemical additives and then it needs to be dumped.
- Noise pollution
- Pollution issues threaten the value of people’s land and homes
Increasing Potential for AI Bubble to Burst
The costs of AI are rising rapidly and the math doesn’t make sense
- Graphics chips (GPUs) that power AI have doubled in price since 2023. The memory chips that go inside them are up 200%. Chip runs have been pre-bought to be installed in GPUs that don’t exist yet, to be installed in massive datacenters to feed a demand for AI that doesn’t exist, for companies that have no hope of ever turning a profit.
- Those expensive chips don’t last. They make up about 40% of a data center’s initial cost. AI training burns out chips in 18 to 36 months. If $264 billion of this year’s spending goes to chips (40% of $660 billion), and those chips need replacement in two years, these companies will face another $264 billion bill in 2028. And again in 2030. And again in 2032.
Most of these companies will spend more than they make
The rule of the AI game is just to get as many GPUs as you possibly can, as much real estate and power as you possibly can, and get it online as fast as you can. Because there’s no prize for second place. The big tech companies will borrow roughly half of what they’re spending on AI. If the AI boom slows in two years, you’re left with:- Buildings you can’t easily repurpose
- Massive depreciation expenses that eat profits
- Debt payments that still have to be made
- No quick and easy way out
They are investing in each other’s projects, creating more risk
- The big tech companies like OpenAI, Nvidia, Microsoft, Oracle, and AMD are investing in each other. For example, Oracle is committed to providing $300 billion worth of computing infrastructure to OpenAI, the company behind ChatGPT. OpenAI doesn’t really make money. It loses money, burning cash to train better AI models. They’re constructing massive data centers and filling them with equipment on the promise that OpenAI’s revenue will eventually materialize.
- These massive deals are circulating the same dollars, but it looks like it’s revenue from sales. This practice inflates stock value, giving the illusion of explosive growth, which encourages even more investment in these companies, including investments from working people’s pension funds. If AI has a big crash, it could trigger a devastating chain reaction, causing a widespread collapse similar to the 2008 Great Financial Crisis. Local communities could be stuck with data centers with diminished value that don’t generate tax revenue. Pension losses could harm many workers.
Federal Government’s Use of AI Threatens War
The Trump Administration is incentivizing companies to further develop AI capabilities that can be applied to weapons directed against China. The US is relying on advanced weaponry, using AI capabilities, to prepare to ‘win’ a war against China in order to maintain global hegemony.How are People Responding?
People all across the country are fighting the building of data centers
Activists stalled $98 billion in data-center projects in the second quarter of 2025 alone, researchers at Data Center Watch found. “Every day I hear from someone with a different reason for fighting a data center,” says Saul Levin, a D.C.-based organizer.
“I’m not sure we can trust these guys as an industry overall to make changes without further pressure,” says Maria Raine, who, along with her husband Matthew, sued Open AI after their son Adam died by suicide after seeking guidance from ChatGPT. (Open AI argued in legal filings that Raine misused the chatbot.) TIME spoke to a broad array of AI skeptics, from disparate regions, ideologies, and professions, but bound by a shared mission: to stop, or at least slow, a technology infiltrating almost every aspect of our lives.A movement is growing to call for a moratorium on data centers
A growing list of states have introduced bills, and many local governments have also pressed pause on the building of data centers. Food and Water Watch has a campaign to call for a national moratorium. Bernie Sanders will be introducing a bill calling for a nationwide moratorium on data centers.
Karel Riley works with the People’s Tribune, and its bilingual sister publication, Tribuno del Pueblo, as a writer and contributor on human rights and women’s issues. “I’ve been a feminist since early adulthood. As a clerical worker, I joined a union drive with AFSCME seeking comparable wages to men for female-dominated jobs, and we were partially successful. In the mid-80’s our union participated in the historic Hormel strike in Minnesota. Later, I joined others in support of a local welfare rights organization,” she says.
