The rapid expansion of artificial intelligence is creating an unexpected challenge for American universities: soaring demand for large properties that can support energy-intensive data centers. Institutions that once reserved sprawling campuses for classrooms and research labs now find themselves competing with tech giants for the same real estate.
George Washington University’s recent sale of its Virginia Science and Technology Campus to Amazon Data Services for $427 million exemplifies this trend. The 120-acre property in Ashburn, Virginia, lies in Loudoun County, the heart of Northern Virginia’s “Data Center Alley.” The university announced the sale in February without initially naming the buyer or disclosing the price, citing confidentiality. Amazon was later confirmed as the purchaser.
GW stated the transaction would strengthen its long-term finances. It can continue using the campus for up to five years while relocating programs and offices. A substantial portion of the proceeds will fund a new endowment, and eligible faculty will receive one-time payments: $2,500 for full-time benefits-eligible staff and $1,250 for part-time faculty.
Former GW professor Ellen Scully-Russ, who taught at the Virginia campus from 2009, described the property as struggling to become a central part of the university’s strategy. She noted a gradual reduction in resources, including unfilled faculty positions and eliminated teaching assistant roles due to budget constraints. Faculty and students were not informed about sale discussions in advance. “There was no transparency,” she told Fortune. She also questioned the long-term viability of large AI infrastructure projects, calling the trend potentially “a short-term bubble.”
The University of Michigan is taking a different approach with a planned $1.25 billion high-performance computing and research center in Ypsilanti Township. Developed in partnership with Los Alamos National Laboratory, the facility will support research in medicine, materials science, clean energy, engineering, and national security. It is expected to require about 50 megawatts of power initially, eventually reaching 100 megawatts. However, local residents and officials oppose the project, and the Ypsilanti Township Board of Trustees unanimously passed a resolution against it.
As AI continues to drive demand for data center infrastructure, universities must navigate the tension between preserving academic land and capitalizing on lucrative real estate opportunities. The outcome will shape the future of both higher education and the AI industry.