Universities selling campuses for AI data centers
AI infrastructure race threatens physical presence of higher education
1New casualties in the AI infrastructure race
American universities have begun selling campus properties to meet the infrastructure demands created by the AI boom. Large parcels with access to electricity, water, and fiber infrastructure have become valuable for data centers. George Washington University's $427 million sale of its Virginia campus to Amazon exemplifies this trend. Universities' status as major landowners makes them attractive targets for data center companies seeking prime locations.
2GWU's surprise sale
George Washington University sold its 120-acre Virginia Science and Technology Campus to Amazon Data Services. Initially, the sale price and buyer were not disclosed, but student journalism uncovered the details. The campus is located in 'Data Center Alley,' home to the world's highest concentration of data centers. The university claimed the sale would strengthen its long-term financial position. Faculty and staff were kept in the dark about the transaction.
3Decision conflicting with academic mission
Former professor Ellen Scully-Russ stated the sale resulted from GWU's years of neglect toward its Virginia campus. Since 2009, the campus failed to find a sustainable role. Resources were gradually reduced, and teaching assistants were eliminated due to budget cuts. The professor emphasized the lack of transparency in the decision-making process, with no consultation of the academic community. Post-sale informational meetings were deemed 'too late' by affected parties.
4Financial pressures on higher education
Universities face challenges from declining federal research funding, changing international student policies, and broader financial difficulties. GWU announced it would allocate a significant portion of sale proceeds to a new endowment fund, with one-time payments of $2,500 for full-time faculty and $1,250 for part-timers. However, these sales raise fundamental questions about institutional priorities - AI infrastructure versus education.
5Michigan's alternative approach
The University of Michigan is taking a different approach to the AI infrastructure craze. It's building a $1.25 billion high-performance computing and research center in Ypsilanti Township. The facility will support research ranging from medicine to materials science, clean energy, and national security. The university avoids calling it a 'data center,' but the project faces local opposition from community members.
6Erosion of academic resources
Scully-Russ described the resource drain at the Virginia campus as happening 'drip by drip.' Over five to six years, faculty positions were lost through attrition without replacement. Although newer and better equipped in some aspects than GWU's Washington DC facilities, the campus never achieved strategic importance. The professor remarked, 'It never really got any legs.' The sale appears as the logical conclusion of this neglect.
7AI's impact on higher education
The sale of university land for AI infrastructure raises fundamental questions about higher education's future. Scully-Russ warns, 'We're chipping away at our higher ed resources to make room for this technology to expand.' She expresses concern about the lack of accountability and thoughtful progression, believing data centers represent a 'short-term bubble' that technological advances could soon render obsolete.
8Education or investment?
Universities' land-use decisions appear to conflict with their core missions. Properties valuable for AI infrastructure simultaneously serve as educational and research spaces. The GWU case demonstrates how financial pressures reshape academic priorities. Michigan's joint project offers an alternative model of university participation in this trend. In both scenarios, higher education's physical and intellectual presence undergoes significant transformation.
Original article: Fortune | FORTUNE — Üniversiteler AI veri merkezleri için kampüslerini satıyor