The standard critique against the rapid expansion of generative AI has long settled on a single vector: the prohibitive, scaling cost of centralized infrastructure. The cycle of training and deploying high-precision models has locked the industry into a structural over-dependence on specialized, capital-intensive data centres dominated by a single hardware gatekeeper.
We’d like to use cookies to understand what readers enjoy and, as we add advertising, to measure and fund our journalism. Nothing is stored until you choose, and you can change your mind at any time under Cookie settings.
Cookie policy