This thesis investigates whether equities most exposed to the artificial intelligence build-out exhibit econometrically identifiable explosive price dynamics consistent with speculative bubble formation, and whether such dynamics propagate systematically along the AI value chain in a manner robust to the choice of detection method. A bubble is defined operationally, following the rational-bubble literature, as a non-stationary explosive deviation of prices from a random-walk benchmark. This definition relies on the statistical properties of the price series and is independent of any estimation of fundamental value. The empirical design, guided by a preliminary mapping of actors and commercial relationships along the AI value chain, relies on a constructed universe of thirty-one publicly listed equities, mapped ex ante to an eleven-segment taxonomy spanning upstream compute, midstream cloud, and downstream software.