I think they should take full advantage of their Ax cadence and beat intel like a drum by releasing Mx every spring and Mx pro/max every fall.
...but Apple aren't really in competition with Intel. Intel sell processors to computer makers. Apple sell computers and have shown no interest in selling their processors in competition with Intel. Intel have huge resources, and x86 has a vast user base in commercial PCs and servers who wouldn't switch to Mac if Tim Cook turned up on their doorstep with a bag of cash and a crate of beer, plus other user bases in cheap home PCs and gaming, which Apple has never shown any interest in trying to conquer. It would be a stupid, Quixotic move on Apple's part if they made it their mission to "defeat" Intel, which would just generate more of the sort of marketing FUD from Intel that we've seen already.
The main way Apple threaten Intel is that they give
credibility to the idea that ARM and other non-x86 ISAs are good for more than just mobile phones. The next move in
that war is for Microsoft, Dell, Lenovo, HP et. al. to think "hey, we sell more PCs than Apple, perhaps
we could make our own processors, or work with Qualcomm or someone..." - which Intel can counter with a well-funded FUD campaign since the target customers are unlikely to actually
try ARM for themselves and will be happy with an excuse to stay in their comfort zone.
However, Apple Silicon/ARM does have a built-in advantage over x86 that should keep it ahead of the game even as x86 moves to newer processes: An ARM core is a RISC processor core. An x86 core is (roughly speaking) a RISC-like processor core
plus a hardware x86-to-RISC instruction translator - so it is always going to be fundamentally bigger and more complex, which will allow ARM to use less power, cram in more cores, more specialist accelerators etc. on the same die. This works for Apple because they have a history of breaking backward compatibility every 10 years or so and their remaining user base is used to that. Rosetta is great, but it only has to work as a transitional solution for a few years - in 5 years time any Mac software still reliant on x86 will be dead. A big part of the PC world's huge, profitable customer base won't accept that - and wouldn't accept the level of "breakage" that happens with every major MacOS release, let alone a processor switch.
If Apple were like Wintel, they'd have kept
Apple II compatibility until around Snow Leopard, people on MR would still be moaning about the removal of Classic from Mavericks and Carbon would be available as an optional download for Monterrey.
Intel
could go for a Rosetta-type solution - drop the x86 instruction decode and rely on everything being pre-translated in software - but then (a) that would have to be the solution for the foreseeable future, because some of that x86 code isn't going away and (b) they'd kiss good by to their monopoly and be in direct competition with every ARM chip maker, given that both Apple and Microsoft already have pretty good x86-to-ARM translation tech.
Long-term, x86 is dying anyway - its only real selling point is the popularity and legacy of its ISA (which is reflected in the RISC-core + x86 decoder architecture) and
that is becoming less relevant as more and more code is written entirely in 64-bit clean high-level languages and/or runs as bytecode on a virtual machine (Java, Android, Microsoft CLR, etc) or via a JIT-compiled scripting language. x86 failed dismally in the mobile market - which grew to be a significant part of the personal computing market - and some people kinda failed to notice when their phones started doing image processing and AI tricks that put full-size PCs to shame. The only reason it is still around is because there are enough legacy applications and highly conservative customers to keep it in slow decay - and therefore a lot of profit still to be extracted - for another 10 years or so.
. 
