This article may give a clue. While it is not comparing directly to ARMv9 it explains why vector processing has advantages over SIMD processing. SVE2 is still to my knowledge a form of SIMD. Taxonomy can be a bit hard when different approaches start copying ideas from each other.

You can organize vector registers into much larger units on RISC-V.

https://itnext.io/advantages-of-risc-v-vector-processing-over-x86-simd-c1b72f3a3e82