This paper analyzes the ergodic capacity of secondary point‐to‐multipoint communications under the outage constraint of multiple primary user receivers (PU‐Rxs) and the secondary user (SU) maximum transmit power limit. We first derive analytical expressions of the ergodic capacity for three scheduling schemes, that is, round robin, Max‐signal‐to‐interference‐plus‐noise‐ratio (Max‐SINR) and Min‐SINR, and compare their performance. Numerical examples show that when the number of SU receivers (SU‐Rxs) increases in the presence of a single PU‐Rx and at high SINR, the Min‐SINR outperforms the Max‐SINR scheme. As the number of PU‐Rxs increases, the Max‐SINR performs better than the Min‐SINR scheduling. When the number of SU‐Rxs becomes large, the system capacity is enhanced but so is the feedback load. To exploit the Max‐SINR transmission while reducing the feedback load, we assume a threshold based on the channel quality where the SU‐Rxs with the instantaneous SINR above the threshold participate in the Max‐SINR scheduling; otherwise, an SU‐Rx is selected randomly. Then, an analytical expression of the average capacity is derived. Numerical results illustrate that the capacity with limited feedback is approximately the same as for the case of Max‐SINR with full feedback when the SU threshold for feedback condition is set to low and medium values. Copyright © 2016 John Wiley & Sons, Ltd.