This paper models the banknote printing costs of a central bank in a society that uses progressively less cash. In such a setting, the central bank runs the risk of overinvesting when it introduces a new technology, for example, when it switches from paper to polymer notes. This paper shows that simple durability/cost rules of thumb are unhelpful in determining the viability of a switch to polymer, and that the size of the decrease in currency demand matters, albeit not in a monotonous way. A key factor is how the fall in demand compares with the note replacement rate. Simulations for three Nordic countries illustrate our findings.