The long-running debate about the spatial implications of financialisation has given rise to a gap in the literature, between the standard (or convergence view) on the one hand, and the orthodox (or divergence) view on the other. This paper seeks to contribute to this debate by charting a middle course, differentiated from the current literature in both methodological and conceptual scope, with the main proposition being that: while financialisation has brought geographies into question, its spatial legacy lies not in any organised deterritorialisation of economy, but rather, in its structural, compositional and computational arrow of time (i.e. historical, irreversible time). The essay begins by reviewing the standard view, before presenting evidence to the contrary. Drawn from case studies in CIP deviations, banking networks, and the international division of labour, the alternative ‘complexity’ view is then explored, followed by a critical analysis of its strengths and shortcomings, as well as a discussion about the need for a reorientation of assumptions. The essay concludes with an empirical evaluation of the new theory.