Green Acres
People in rural communities have something in common with shareholders of tech stocks. They are both mad as hell about all these datacenters.
I live in a farm community, population 1,100. I talk to my neighbors about the kinds of things you’d expect … horses, feed, trucks, fertilizer, etc., but a new topic has entered the chat: AI datacenters. Technically, the “AI” qualifier is a misnomer, as there is more to a modern datacenter than simply running ML models, but this is how the general public perceives them. If there’s a datacenter coming to a former corn field near you, it’s just one more reason to fear and loathe AI.
The public’s backlash against AI is an ongoing story that really began as the “techlash”, as it was first called sometime around 2018. The AI sector of the tech industry has, up to now, primarily drawn the scorn of critical researchers, civil society, privacy advocates, artists & creators whose IP has been scraped, and public interest technologists (like me), to name a few. But now the circle has widened to include, well, the general public. And that’s a lot of people.
TIME Magazine ran a cover story a couple weeks ago to dig into this growing angst, and the root of it seems to be the lowest level of the so-called AI stack: the physical infrastructure, aka datacenters. TIME frames the issue:
As physical manifestations of the [tech] industry’s heedlessness, the data centers powering AI systems have become the focus of protests. From Virginia to Indiana to Arizona, 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.
One of the people interviewed for the piece was Wisconsin state representative Francesca Hong, who is running for governor and, like several other candidates from both parties, is out talking to voters about datacenter restrictions. As she describes it,
What used to be a niche issue is now a rallying cry, because so many voters are “concerned and pissed off” about the projects, she says. “They want Wisconsin to be a hostile environment to the construction of AI data centers.”
People really hate these things. They’re big, they’re loud, they’re ugly, they produce near-zero jobs, they consume huge amounts of water & electricity, and building them on farmland makes them a detestable accelerant to what was already a decline in the rural way of life. The thing datacenters require is cheap, available land, and (permitting & zoning aside), farm owners are often the only ones standing in the way of groundbreaking.
Farmers are underappreciated for their knowledge of technology and commodity markets, among other things. It’s not unusual for a farmer to have the moral equivalent of a Bloomberg terminal in his/her tractor cab, checking spot prices and executing trades from the middle of a field, while the tractor uses GPS to drive itself in perfectly straight rows, maximizing crop yield and minimizing wasted seed, herbicide, and fertilizer. The farm-tech community on the internet is endlessly fascinating, but physical infrastructure is one element of the tech universe for which rural communities have no love.
Regardless, some farmers are selling or leasing their land to tech companies anyway, often because there is no other financially comparable option. Farming is not a great business, and family farming in particular is a tough place to be. Unlike corporate agriculture, in which scale economies, a balance sheet, and huge cash flows can help weather the cyclical nature of commodity markets, family farms are in a struggle for survival. The financial pressures come from everywhere: rising input prices (costs of equipment, seed, fertilizer, etc.) combined with stable or declining commodity prices, squeezing margins or worse yet, producing losses. This past November, 60 Minutes did a piece titled, “The Family Farm”, in which several multi-generational farmers talked about the financial struggles they’re facing, and the personal pressures that don’t appear on a P&L:
For many of these families, farming is more than a livelihood. It is a heritage passed from parent to child, often with the hope that each generation will build upon the last. The sudden shock to the soybean market, followed by long stretches of uncertainty, has transformed that hope into a fear shared across rural America: the fear of being the generation that loses the farm.
Supporting data are everywhere. Purdue University’s monthly Ag Economy Barometer for February shows farmer sentiment declining, while farm bankruptcies are sharply on the rise. Farmers are unhappy with the government (both parties) for allowing decades of mergers & acquisitions to create Big Tech’s heartland cousin, “Big Ag”, which is essentially a small number of huge companies engaging in what antitrust lawyers call “dominant firm conduct”. Farmers are unhappy with the Trump administration for starting trade wars that produce retaliatory effects like boycotts of US products or denied access to foreign markets. Farmers are unhappy with what they perceive as a system tuned to the interests of corporations and not farmers or rural areas. As an example, my internet out here blows, but I digress. The lure of selling a farm to a tech company is understandable.
For their part, technology companies are playing both sides. They make big promises, like carbon-neutral pledges that eventually run into challenges, while working behind the scenes to conceal information about water and electricity consumption from people in the communities where they operate. This is not new. In 2021, The Dalles, OR, a city about 90 minutes east of Portland on the Columbia River, sued Oregon’s largest newspaper on behalf of Google to halt publication of water usage figures for the company’s local datacenter. The public backlash was swift, the 13-month legal fight was dropped, and the published figures revealed that the datacenter consumed 274.5 million gallons of water that year. Citizens were not pleased.
The battle has now come to the Great Lakes. This past September, an environmental group sued the city of Racine, WI to force disclosure of Microsoft’s planned water usage for a datacenter in nearby Mt. Pleasant, about an hour’s drive from my house. The result produced publication of some eye-popping numbers:
The first phase of Microsoft’s data center campus would use a peak of 234,000 gallons per day or 2.8 million gallons per year. Under subsequent phases, the campus would use a peak of 702,000 gallons per day or 8.4 million gallons annually.
Seattle is one of the tech industry’s most active epicenters, and while living there I admit I was in a bit of a bubble until about a year ago. My return to the Midwest has been illuminating to say the least, but I did not expect the datacenter issue to become so front & center, so fast, in this part of the country, and that’s simply because I wasn’t paying attention, nor was I listening to what real people are saying in communities like this one as the industry races to build now and ask questions later.
This confluence of factors has been years in the making. Small farms have been in trouble for decades, while real estate developers have been circling these parcels like vultures in anticipation of financial distress forcing farmers to finally take their calls. The AI datacenter buildout simply introduces a new participant into a process that will continue the demise of these farms, and that is a big loss, not just for rural communities but for what so many of us consider emblematic representations of American culture and history. And yes, I’ve read Zinn and know all about how this land was stolen in the first place, but that is a longer conversation.
Keep a sharp eye on what’s happening in your community, as there may be a real estate deal in the works where the public officials who’ve sworn to protect your health, safety, and welfare are nonetheless tied up by a tech company’s NDA that bars them from telling you what’s really going on. As we head into election season this coming fall, AI datacenters will be a tier-1, bipartisan policy issue for countless races across America. In a country desperately in need of unifying policy issues, this one seems primed to put warring parties on the same side.
/end