Due to the introduction of new services, the volume of data transferred in mobile networks is rapidly growing and operators periodically face the necessity to upgrade their network. Such upgrades allow them to increase the capacity and provide adequate Quality of Service (QoS). In this paper we propose a general framework for deriving the optimal date for a network upgrade. We show that this date is the result of a compromise between the decrease of upgrade investment cost with time and the loss of profit generated by insufficient capacity. The upgrade should hence be performed when the loss of profit, derived using analytical capacity expressions, exceeds the expected discount. The model presented herein accounts for the randomness of the demand and upgrading cost functions, and results are given for a HSDPA network.