Date of Submission
Spring 5-11-2026
Degree Type
Dissertation/Thesis
Degree Name
Bachelor of Architecture
Department
Architecture
Committee Chair/First Advisor
Ameen Farooq
Abstract
The materialization of AI infrastructure has outpaced all regulatory frameworks, allowing data centers to proliferate with limited oversight despite their growing environmental and social impacts. Through a review of artificial intelligence’s technological development, the rise of big tech, and contemporary data center expansion, this thesis situates AI infrastructure within a broader pattern of deregulation, delayed accountability, and market-driven governance. Using a qualitative mixed-method approach combining historical analysis, policy critique, regional case studies, and research-through-design, the study examines how existing land use, utility, and environmental frameworks have failed to keep pace with the scale of data center sprawl; The study also analyzes what harms to water systems, energy grids, and residential communities result from this governance gap. The central question asks whether the physical design of data centers can itself function as a regulatory mechanism, embedding transparency resource reciprocity, and public benefit directly into the building typology. The proposed response is a vertical mixed-use data center situated in Seattle’s waterfront district, which integrates closed-loop cooling, heat waste recovery, civic programming, and modular hardware components into a single architectural framework. The project argues that architecture is not a substitute for policy, but that when infrastructure is embedded in the city and made visible, poor governance becomes harder to sustain.
Included in
Architectural Technology Commons, Environmental Design Commons, Urban, Community and Regional Planning Commons
Comments
Portions of this thesis are adapted from a co-authored conference paper: Sterchi, A. & Farooq, A. (2026). 'Virtual Matter: Governing AI's Physical Infrastructure.' Proceedings of the 27th Annual International Conference on Digital Government Research (dg.o 2026), forthcoming June 2026, published by TU Delft OPEN Publishing under CC-BY 4.0. The co-author has approved this thesis for upload.