There wasn't really a legacy software advantage for x86 in the Mac arena either.
Indirectly there was and that's all that matters: The x86 legacy advantage was unassailably strong in the wintel world. And wintel had the lions's share of teh sales. As a result, Intel had more money than the competitors to invest in both processor design and process technology, the result of which is that intel eventually overran their competitors.
The processors for macs just couldn't keep up because Motorola and then IBM didn't have the volume and margins in their chip business to be able to compete with Intel.
The world is a bit different now, but is it different enough to matter?
The whole expensive x86 front end decoder cost used to matter on the desktop, but eventually the large number of parallel functional units and the OoO logic to keep them filled started to dominate massively. Then it used to matter on phones, but now it's pretty much reaching the stage where phone processors are so large and powerful that similar things are happening.
But the low end still exists (below phones), so ARM will never be squeezed out by Intel. There will always be a market for some noddy core with 2K of RAM and on that scale the decoder matters.
So, ARM is there. While nothing like as rich as Intel, they put most of their developement into the CPU tech, not process. The world has also hit diminishing returns in CPU design. Back in the past, there were "easy" developments like the caches, MHz wars, the transition to superscalar, out of order execution and vector instruction units. Once those topped off, the next bit was tweaking the cores for more IPC, e.g. the Core 2 to i7 transition, but returns have really diminished in that regard. Now it's got to the stage where it takes massive effort to get a few percent IPC better.
In recent times Intel have dominated IPC. However, while competitors may not ever catch up completely, it's easier for them to close the gap than it is for Intel to keep it open because Intel have already taken their low-hanging fruit.
Still, Intel have one of the best CPU design teams out there, which is always going to be an advantage.
Then there's the process tech. This is another area where intel lead, but the world has been losing fabs at a shocking rate. Previously, Intel was the 800lb gorilla up against a lot of other smaller chipmakers with smaller market sizes. Everyone else has been consolidating so intel now has fewer, but much larger competitors. This will make life harder for Intel relative to the past.
Intel is not a gun for hire. This has positives and negatives. On the plus side, they bend all their resources to fabbing the top end PC chips. On the minus side, the major phone manufacturers can't get custom chips like they can with ARM, which means that unless they are very lucky, they're paying for things they don't want or have lower integration if they go with intel.
It also means that the other people can test out new processes on smaller chips. Large area makes the probability of damaging defects go way up. Other fabs can do smaller chips on new processes which keeps the proportion of defective units lower.