Sorry for my bad english !
I have the feeling that most of the members who responded on this topic did so too quickly. This is why I would like to relaunch the subject, so that we think about it more seriously. Because it could be a fantastic and extremely affordable project.
It is clear that if we make only one link connection, that the speed could not meet the demand. However, what about when using all 4 channels (on a small module such as the FS1000A), for example. I would even like to take the idea even further.
Imagine that using an fpga or a microcontroller, we operate the RF module with a dynamic approach. For example, instead of defining a module with a static configuration of its addressing, we instead define channels, say 16 channels. Each channel represents a unique address on the module to send / receive a data packet (4 bits x Clock of the RF module) X our 16 channels defined in the micros. Finally, you kinda get the idea. And with registry chipset to created sufficient buffer memory. We would, hypothetically, be able to generate a stream that would probably transmit 4K video quality. Using 2 RF modules, to separate the sound and the image, with 2 micros to reconstruct the packets and a 3rd micro to recombine and synchronize the sound and the image. With few components, the project becomes, in my opinion, something possible.
Of course, it is only hypotheses that I launch like that, because for my part, it is a project that I would like to happen. If the project comes to fruition, imagine the possibilities that this approach would offer. To open up new horizons on how to optimize a data flow by harnessing the potential of modules with a completely different, but innovative approach!