Data Centers Are Not “Campuses”
In recent months, a particular word has begun to circulate with striking emphasis in press releases across the United States. Tech companies and government officials are announcing the construction of “data center campuses.” Not factories, factories, warehouses or complexes, but “campuses”.
In October, OpenAI and Oracle unveiled a $7 billion data center in Michigan, celebrated as the state’s first “hyperscale campus.” Last month, officials in Wythe County, Virginia, announced the creation of a “data center campus” as a marker of economic progress. This month, AVAIO Digital announced plans to create a large-scale, AI-ready “electrical data center and campus” in Little Rock, Arkansas.
None of these look like something most people would recognize as a campus. These are industrial installations designed to host large-scale calculations. They consume immense amounts of electricity and water. They employ very few people on a permanent basis. They are generally isolated from the surrounding community, both physically and economically. Calling them campuses is ridiculous – unless, of course, the aim is to give the appearance of civic legitimacy, to obscure how few humans are actually involved, and to gloss over questions of the true public value of these operations.
The word “campus” comes from the Latin “field”, but in English, campus means much more than a plot of land with buildings. It is a place ordered for people engaged in shared work. Universities and hospitals are called campuses because they bring people together in long-term relationships of teaching, healing, work, and deliberation. More than a field between buildings, a campus is a place for targeted human interactions.
Corporate adoption of the word “campus” is not new. Tech companies such as Apple and Google have long described their headquarters as “campuses,” as have pharmaceutical companies and government agencies, borrowing the language of universities to describe large, centralized workspaces. These earlier uses still referred to places where thousands of people gathered daily to work, where cafeterias served meals, where collaboration took place in person, and where the site functioned as a center of human activity. The word has expanded, but it has retained a connection to its original meaning as a place of congregation.
Data centers break this connection entirely. A data center does not bring together people. According to a 2024 report from the Virginia Joint Commission on Audit and Legislative Review, these buildings typically occupy hundreds of thousands of square feet on hundreds of acres of land while housing only a few dozen permanent workers, many of whom are contractors, with most operations managed remotely. In spaces measured in football fields, the human presence is counted in the dozens. Calling such facilities “campus” does more than extend a metaphor; this normalizes an economic activity that requires almost no people.
A data center is a highly sophisticated industrial facility designed to house servers and processing equipment in which digital information is stored and transmitted. Its value, especially in the era of generative AI, derives from automation, redundancy, and uninterrupted operation rather than human participation. Yet tech companies persist in borrowing the language of community life to describe facilities that, by design, are indifferent to it. This appropriation is not accidental. He does political work.
Data centers are upending many of our basic assumptions about how new industrial facilities operate. A single hyperscale facility can consume as much electricity as a mid-sized city while employing relatively few people. The Michigan facility, for example, will use enough energy to serve about a million homes, while the most generous employment projections suggest a maximum of 450 jobs. Local communities provide land, electricity, water, roads and public subsidies. The resulting profits are slim, narrowly distributed and often speculative.
When leaders and governors call these institutions “campuses,” they borrow the moral authority of institutions that belong to communities. Universities and hospitals justify their presence with enduring obligations to students, patients, workers and neighbors. To describe an industrial server warehouse as a campus is to appropriate this legitimacy while circumventing moral obligations.
Language functions here as infrastructure. This shapes how power is interpreted and consent is obtained. In 36 states, legislatures have provided grants to facilitate the creation of data center sites. Here in North Carolina, for example, the legislature has codified into law tax exemptions on energy used by qualifying data centers. When a legislature is asked to subsidize a campus, the word itself implies permanence, shared benefit, and civic contribution, even though, in the case of a data center, the facility may be largely empty of people.
The technology sector has become adept at this: absorbing relational language and redeploying it to describe operations that systematically reduce the need for human presence. Ride-hailing platforms describe themselves as part of a “sharing economy” while dissolving labor protections. Social media companies post “community guidelines” for platforms that often de-emphasize the actual community. Virtual environments promise bodiless presence and commitment-free connection.
The data center “campus” belongs to this same lexicon. This allows political leaders to reassure the public that something clearly civic is being built, even if material reality goes in the opposite direction. It offers the image of a community while only delivering machines. This suggests development while truth is extraction. Calling these vast sites “campuses” rather than “data factories” or “computer mines” makes relevant questions less likely to be asked. The word provides a sort of camouflage for tech companies and their government collaborators.
If an industrial facility can be described as a campus, the opposition begins to seem unreasonable. Who would object to a campus? The transformation of language precedes the transformation of land and place.
True flagship institutions are difficult to counterfeit. We write this article as professors in a university village and as advocates for our local rural hospitals. In communities like ours across the country, universities and hospitals are rooted in the places they call home. They shape economic opportunities for our neighbors. They require sustained investment in human resources. Sometimes they generate conflicts as well as benefits because they are places of real human encounters.
Data centers, on the other hand, are intentionally locationless, although they take over land in rural counties with little participation from surrounding urban dwellers. Their labor requirements are minimal. Instead, they seek cheap energy, permissive regulation, and public subsidies – using the land more like strip mines or CAFOs than universities or hospitals. Data centers aren’t so much community-owned institutions as infrastructure, like a pipeline, that just runs through them.
The pipeline analogy is telling: Proponents of data center expansion often point to secondary benefits such as temporary construction jobs or symbolic inclusion in the future of artificial intelligence. But as with pipelines, the benefits are real but fleeting, while the costs persist. Data centers put long-term strain on power grids and water systems, competing with communities for these resources. They lock up energy-intensive infrastructure at a time when our places and all their creatures need restraint. They exclude other land uses that could have supported more sustainable jobs or social value. Algorithms and AI increasingly shape our experiences and relationships on an individual level, while the data centers that give rise to them erode the natural resources and spaces that make up our communities on a macro level.
Some digital infrastructure is necessary for modern life. But the words are not ornamental. They are part of the moral ecology of a place. And with artificial intelligence, we have begun to accept a vocabulary that disguises invasion as belonging and automation as community. We have begun to allow words that once denoted shared work and mutual accountability to be repurposed for projects that actively diminish both.
A campus is a place where people come together. If the space is for machines with only a few people to operate them, we should make that clear.
