Volta is already used, they wont re-use the name for a different architecture. Also, the architecture names almost certainly do indicate true different architectures. So they wont use different architecture names for the same chip used in different segments. They already have marketing names: Tesla, Titan, Geforce, Quadro.
It is completely expected that Nvidia will continue to diverge their gaming and compute architectures. Maxwell started this, by dropping the FP64 support for gaming parts. Pascal continued this, the GP100 has a fair amount of differences with the gaming cards. Volta has lots of HPC and DL specific aspects. It standards to reason that the next step is an architecture spinoff, so taking some of the compute advanced made in Volta but continuing with more gaming specific optimizations and changes, which could give us Ampere.
Also, 7nm is no where close for consumer level GPUS but for HPC/DL use 7nm is coming at end of the year. Nvidia will want to be pushing the availability boundary hard on tis to stay ahead of competition. Hence we get something like Turing