[wmf]

x86 is hampered by backwards-compatibility. Could anyone comment on some of the more technical CPU architecture reasons for this?

The reasons are mostly (1) x86 has strong memory ordering which is slower than weak ordering and (2) x86 needs a "complex" decoder to parse all the prefixes and crack instructions into uops (but ARMv8 cores require two decoders and may crack shifts) and (3) x86 is so hairy that it requires a massive test suite and dozens of errata still slip through, but the overhead of these is really small compared to Intel's advantages in money, manpower, and accumulated expertise.

Do A64 CPUs have potential to outpace x86_64 processors in power consumption and performance simply because of the architecture?

IMO no. It's cheaper to reach parity (Apple's processor R&D is much lower than Intel's), but I don't see why they'd outpace.