In this article, the authors study the effect of peak interference power constraint given by the primary receiver on the performance of multi-hop cognitive amplify-and-forward (AF) relay networks. The athours assume that all involved channels are subject to independent, not necessarily identically distributed Nakagami-m fading and the secondary multi-hop relay network operates in channel state information-assisted AF mode. An analysis of the system performance in terms of outage probability and symbol error rate (SER) is presented. Accordingly, closed-form expressions for the tightly bounded outage probability and SER are formulated which are used for quantifying the impact of the fading channels, the interference power constraint and the number of hops on system performance. More importantly, an asymptotic performance analysis, which intuitively reveals benefits of cooperative diversity of the secondary relay network, is established. The analysis shows that the diversity gain of the considered cognitive relay networks is equal to the minimum of the fading severity parameters of all relaying hops. Also, the interference power constraint imposed by the primary receiver only affects the coding gain of the secondary relay network.