MS see xbox as the living-room win10 machine - of course it will be stuck with x86 for a while, perhaps through the end of the architecture. Sony are a wildcard - outside of BC reasons (not particularly strong with that company), next ps could go either way. A hypothetical Zen falter (I hope not, but one never knows with AMD) would make sony abandon x86 for the ps5, as intel chips are expensive and intel will be reluctant to play again that subsidize-into-oblivion game they tried with the mobile makers and that got them billions into the red for the respective division. The reality is x86 just doesn't have a healthy supply landscape, as simple as that.
Sure, ARM is closing in on x86 in the server space too, but if you want a high performance server today for Windows/Linux servers x86 is the best architechture.
Servers are intel's strongold - ARM has yet to prove themselves there. But I was referring to something else.
PowerPC and Itanium are dead and ARM does not have the single core performance that Intel Xeon has.
That's the thing - the majority of people don't use xeons in their daily routines - they use ARMs. Even those who spend hours in front of desktops spend about as much on their phones/tablets during the rest of their day. That is the base of the pyramid and thus the bulk of computing needs we as a civilization have. On the top of the pyramid are the supercomputing users with their massive throughput requirements, who get more and more GPGPU oriented (unsurpassed FLOPS/watt), and thus get less and less mindful toward the CPU architectures. Case in point - the newest POWER9/nvidia supercomputer design wins.
Supercomputers etc use ARM or other architechtures but Amazon and Azure are mostly x86.
Azure needs those windows VMs so they'll be stuck with x86. Google have been on the search for x86 alternatives for a while now, and their GPGPU and ASIC forays indicate they might have found an alternative, at least for some of their Machine Learning purposes, which might put google strictly in the supercomputer camp sooner than later.