ABSTRACT Predominant approaches in the rebel governance literature have looked at control over the means of violence as a prerogative of rebel-rulers, or armed/non-armed actors, somehow deterministically linked to territory. Here weapons have been understood as either autonomous technical-factors or as analytically invisible objects instrumental to human agencies and interactions aiming to territorial control. This paper challenges understandings of control over the means of violence as a central property radiating outwardly through hierarchically and geographically ordered spatial containers. It argues that the means of violence are relational networks among heterogeneous human-non-human entities – e.g. weapons, stockpiles, militarised architectures, forms, armed individuals/groups – that generate territory. These networks are controlled and stabilised via diffused techniques and rationalities of control. Drawing on the study of Ta’ang areas of Northern Shan State – among the few in Myanmar where well-established rebel movements have experienced official disarmament and later undertook a full-fledged re-armament – I find that controlling the means of violence occurs via turbulent combinations of technical objects, techniques and rationalities that relate to four main domains: narcotics eradication; institutionalisation; ethnonationality; and humanitarian security. Processes and practices through which attempts to control the means of violence are made entail alternative strategies to re-generate spatial organisational control and shape multiple shifting territories. Empirically exploring a highly under-researched case, the paper provides a view of the diffused character of controlling the means of violence and its mutually constitutive relations with territory, while illuminating also the role of weapons, other technical objects, and techniques.