With the proliferation of head-mounted displays, cloud computing platforms, and machine learning algorithms, the next-generation of AR/VR applications require research in several directions - more capable hardware, more proficient software and algorithms, and novel network protocols. While the first two problems have received considerable attention, the networking component is the least explored of these three. This paper discusses the networking challenges encountered by the AR/VR community that experiments with novel hardware, software, and computing platforms in a real-world environment. In this collaborative work, we discuss the current networking challenges both quantitatively (by analyzing AR/VR network interactions of head-mounted displays) and quantitatively (by distributing a targeted community survey among AR/VR researchers). We show that the cloud-provided network services are not ideal for the next-generation AR/VR applications. We then present a Named Data Networking (NDN) based framework that can address these challenges by offering a hybrid edge-cloud model for the execution of AR/VR computational tasks.