A system for ad hoc endpoint device association for multimedia conferencing may include one or more processors and memory. The one or more processors may prepare to provide sets of video streams corresponding to a multimedia conference to first, second, and third participant devices, where the set of videos streams prepared for each participant device includes video streams associated with the other participant devices. The one or more processors may determine that the first and second participant devices are in a same physical area and may generate, in response to the determining, a location association between the first and second participant devices for the multimedia conference. The one or more processors may exclude the video stream associated with the third participant device from set of video streams prepared for the first participant device based at least in part on the location association between the first and second participant devices.