What Computing Takes

What Computing Takes follows the material demands of big data and AI from data centres, electricity, cooling, and water to semiconductors, critical minerals, manufacturing, and global supply chains. BDSL’s earlier work on data energetics established one point of departure by examining how rapid gains in computational efficiency coexist with expanding infrastructure and demand, raising questions associated with the Jevons paradox. The project extends that inquiry across the material dependencies of hyperscale computing, asking how growing demand for energy, hardware, resources, and capital connects facilities that appear local to systems operating at global scale.

Field research locates these dependencies in particular places. BDSL’s work has followed data and preservation infrastructure across Korea, China, Estonia, and North and South America, including visits and interviews involving the Internet Archive, national web archives, Naver’s hydroelectric-powered data centre, and the e-Estonia program. New research extends this inquiry into semiconductor production, critical minerals, and the supply chains connecting extraction and fabrication to hyperscale infrastructure, while examining how data centres alter local demands for electricity, water, land, and the acoustic environment. The project asks what the expansion of computation requires of the places and communities that supply its resources, host its facilities, and participate in its production, and how these local transformations become entangled with industrial policy, resource security, and geopolitics.