Neural circuits utilize a host of homeostatic plasticity mechanisms, including synaptic scaling, to maintain stability in circuits undergoing experience-dependent remodeling necessary for information processing. During synaptic scaling, compensatory adaptations in synaptic strength are induced after chronic manipulations in neuronal firing, but our understanding of this process is largely limited to its initial induction. How these homeostatic synaptic adaptations evolve when activity renormalizes and their impact on subsequent homeostatic compensation are both poorly understood. To examine these issues, we investigated whether a previous history of homeostatic scaling in networks of cultured hippocampal neurons altered their subsequent homeostatic responses to chronic activity manipulations. Unexpectedly, we found that a history of synaptic scaling strongly suppressed future scaling to the same, and even opposite, activity challenges. This history-dependent suppression was specific for future homeostatic compensation, as networks with a prior scaling history showed no deficits in the chemical induction of long-term potentiation (cLTP), a Hebbian form of synaptic plasticity. Hippocampal neurons with a prior scaling history exhibited normal engagement of activity-dependent signaling during subsequent activity challenges (as assessed by examination of the ERK/MAPK pathway) but demonstrated widespread alterations in activity-dependent transcriptional regulation to further activity changes. We also investigated what features of synaptic scaling are most tightly associated with history-dependent suppression and found that the resetting of synaptic weights upon activity renormalization plays a key role: extending this resetting period or eliminating it altogether both abolished the suppressing effects of prior scaling history on future homeostatic scaling. Taken together, our data show that the history of homeostatic signaling in neural circuits plays a key role in shaping future compensatory adaptations to chronic changes in network activity.