In Ontario, a data centre playbook has been introduced, sparking debates on AI data centre construction. A computer science professor at the University of Waterloo, Martin Karsten, distinguishes between traditional and AI data centres based on the types of computers they house.
Karsten explains that traditional data centres focus on storing, processing, and retrieving information with defined control flows, while AI data centres handle massive computations with simpler control flows. Energy consumption serves as a key differentiator between the two types, with AI data centres potentially consuming gigawatts of electricity due to their massive installations.
Efforts to cool these facilities often involve water, leading to concerns about environmental impact. However, modern data centres are adopting closed-loop cooling systems to minimize water usage. Despite only a few hyperscale data centres currently in Canada, numerous facilities are proposed or under construction nationwide, triggering protests over resource consumption.
Various municipalities in Ontario are considering moratoriums on data centres, with some already taking action. Green Party Leader Mike Schreiner and environmental groups advocate for pausing data centre projects due to their significant land, energy, and water requirements. Ontario recently unveiled a new data centre framework aimed at promoting economic growth while ensuring data sovereignty.
Karsten emphasizes the ubiquitous nature of data centres in modern society, highlighting their role in telecommunications, healthcare, emergency services, and everyday technology use. Despite controversies, the reliance on data centres across various sectors underscores their integral role in daily life.