I’ve recently found out that I’m a Xennial. Those that argue for the designation suggest that people born between 1977 and 1985 are distinct from Generation X and Millennials. The most notable feature of the Xennials is that we had an analogue childhood and a digital young adulthood. Some have called us the “Oregon Trail Generation” because we grew up playing the game before there was the Internet. Because trying to shoot pixelated squirrels and buffalo was the best part of my school day, the name fits.
Having lived without the internet – but still being able to adapt to it easily – is said to give Xennials a specific relationship to technology. We’re more comfortable using it than those who came before us, but we are still more wary of its total impact than later generations. I, and I expect most of my fellow Xennials, worry what might happen when there’s not enough people questioning if the benefits of a new technology is worth its cost.
In a short span of time, artificial intelligence (AI) has gone from the subject of science fiction movies to being incorporated into most people’s daily lives. AI has improved the efficiency of many tasks and services, allowed us to make better use of data and information, advanced innovation in nearly every field and given us another tool to use in making better decisions. In total, it has exploded the possibilities of the virtual realm. However, this digital technology still has very tangible physical requirements.
The computing power of AI comes from servers, which offer the processing, memory, local storage and network connectivity that underpin AI’s functionality. Typically, these servers are stored in data centers, “hyperscale” facilities ranging from 200 to 1,000 acres in size. It is estimated that $7 trillion dollars will be invested globally by 2030 into data centers. In the U.S., already home to many more facilities than any other country, the demand on the national energy grid will increase by 11%-12% in the next four years.
The need for land to build these data centers has put farming in the middle of the AI discourse. Farmers across the country have been offered “life-altering money” by investment firms looking to build data centers, with reported estimates ranging from $40,000 to $90,000 per acre. Although nondisclosure agreements make it difficult to know what various farmers have been offered, it tends to be significantly more than the land’s assessed value. Last year, an investor paid $615 million for 100 acres in northern Virginia. While other reports are more modest (a Kentucky farmer was offered $33 million for 650 acres, for example), they still include staggering sums.
It’s hard to begrudge any farmer that accepts that type of payday, especially after struggling to survive in a difficult industry. However, the erection of data centers hurts other farmers in the region. Such high land sales raise the assessment value and property tax in the area, making it harder for new farmers to enter the occupation, as well as for current farmers to pay their taxes. While data centers offer tax revenue to local communities, they also require a lot of energy to function, as well as water to cool the servers. Local aquifers sometimes come under threat, and the price of electricity has sometimes doubled or tripled for households in the vicinity. Not only is the land utilized in data center production lost from agricultural use for good, but it makes the surrounding area a less-than-ideal place to live. Nearby residents complain of the constant humming noise, which also drives away birds and wildlife, emptying local woods and hedgerows. Ultimately, the cost of AI is disproportionately borne by rural people.
Perhaps one of the most surprising stories of the land grab by data companies is how many farmers have turned down offers. While it is hard to gauge what percentage of landowners have refused to sell to investors, both because of nondisclosure agreements and because those who don’t sign receive more media attention, there are still many reported cases in which farmers have declared that they can’t be bought. A farmer in Mishicot, Wisconsin, turned down $80 million, a Pennsylvania farmer declined $15 million and the previously mentioned Kentucky farmer, age 82, did not accept the $33 million she was offered. Thanks to AI, it’s easy to find more examples on the internet.
The threat to farmland posed by AI’s computing needs is becoming part of the legislation discussion around the world. After failing to be enacted in 2023, The Farmland for Farmers Act (H.R. 8531) was reintroduced to the U.S. Congress in April by Senator Cory Booker (New Jersey) and Representative Jill Tokuda (Hawaii) and endorsed by over 80 organizations and farming advocacy groups. The National Family Farm Coalition (NFFC) points out that the number of corporate-owned agricultural properties has tripled since 2009, making land more difficult to secure by independent producers. Additionally, I sat in on a meeting with a group involved in Irish agriculture who seek to establish a land observatory in Ireland which would monitor how land changes ownership and work to keep it in possession of family farms.
The ability to push boundaries and advance our potential is, I would argue, something to celebrate about the human species. However, as an Oregon Trail-playing Xennial, I also worry that we won’t know when to stop. Most steps forward require leaving something behind. While it’s exciting what the future may hold, and I appreciate the accoutrements modern life has given me, I also believe that neither should come at the cost of farmers and other rural people.
At the risk of sounding my age, maybe it’s time to get off the computers and listen to the birds outside.
