AI data centers reshape university land use and local governance

The AI data center land rush is hitting a surprising source: university campuses. George Washington University sold its Virginia campus to Amazon Data Services for $427 million, while the University of Michigan is planning a $1.25 billion research computing center with Los Alamos National Laboratory. For AI builders, these deals signal where future compute capacity will be built and how public-private partnerships are reshaping the landscape.

The $427 Million Sale That Signals a Shift

In February, George Washington University announced the sale of its roughly 120-acre Virginia Science and Technology Campus in Ashburn to Amazon Data Services for $427 million. The campus sits in Loudoun County, at the center of Northern Virginia's "Data Center Alley." The university initially declined to identify the buyer, citing confidentiality, but a student newspaper uncovered the details, leading to a lack of transparency that former faculty criticized.

GWU said the proceeds would strengthen its financial position: a significant portion goes into a new endowment, and eligible faculty and staff receive one-time bonus payments. But former professor Ellen Scully-Russ told Fortune the sale capped years of resource erosion at the Ashburn campus, describing it as a "drip, drip, drip" of lost faculty and eliminated teaching assistants. She warned that the broader trend of monetizing land for AI infrastructure lacks accountability.

UMich's $1.25 Billion Research Computing Hub

Meanwhile, the University of Michigan and Los Alamos National Laboratory are planning a $1.25 billion high-performance computing center in Ypsilanti Township. The university is careful to label it a "specialized research hub" rather than a commercial data center, saying the compute power will go toward scientific research in medicine, materials science, clean energy, engineering, and national security rather than cloud or consumer applications. The facility will start at about 50 megawatts and eventually reach 100 megawatts.

UMich has acquired roughly 150.5 acres across three land purchases approved in 2024 and 2025. However, the Ypsilanti Township Board of Trustees unanimously passed a resolution opposing the project, citing community governance concerns. This echoes a broader backlash against data centers as the U.S. has over 4,000 data centers operating and 3,000 more planned.

What This Means for AI Builders

These land deals directly affect where AI data centers get built and who controls the compute. Universities sitting on large parcels near power and fiber infrastructure are becoming key players. For founders and product teams, this creates both opportunities and risks:

Public-private partnerships with universities could offer cheaper or subsidized compute access in exchange for research collaboration.
Local opposition in places like Ypsilanti could delay or block new capacity, tightening supply and raising costs for builders reliant on specific regions.
University endowments funded by land sales may shift research priorities toward AI, potentially channeling more academic talent into commercial AI work but also risking a brain drain from core academic programs.

Uncertainty and Caveats

The evidence here relies heavily on a single Fortune article and its linked sources. Project scopes, exact land areas, timelines, and financial details may vary. The article notes limited transparency in university decision-making around these deals, so the full picture remains unclear. Local opposition and regulatory changes could alter outcomes significantly. Builders should monitor these deals as indicators of compute supply dynamics, but not treat them as settled facts.