In the early 2000s it was x86 being a threat in the datacenter to Sun, PA-RISC, POWER etc.
Most companies switched to x86 15-20 years ago and it's pretty entrenched now.
They are just now switching -- or freshly switched -- to ARM, and the technical differences between ARM and RISC-V being very minimal, ARM won't be entrenched before RISC-V is competitive around 2025 (designs on the drawing boards now).
True, I probably went too far back, still, though it seems like ARM was closer to where x86 was back then than RISC-V is to ARM now, at least if we compare actual chips that are available.
> entrenched before RISC-V is competitive around 2025 (designs on the drawing boards now).
I'm sure people working on POWER, Itanium etc had similar plans at some point in the past. Are there significantly more resources being invested into RISC-V compared to ARM now and if not is there something inherently superior about RISC-V which would outweigh that?
What do you mean by "closer?" A modern arm chip would of course blow an x86 server from 200X out of the water, performance-wise. But by 200X, at least for large values of X, x86 was dominant in the server market (apparently they were around 90% of server sales in 2007 https://www.cnet.com/tech/tech-industry/despite-its-aging-de... ).
ARM is starting to make some headway, but it is significantly less than half. And AMD does seem to have revitalized the x86 market a bit. It isn't a given that we'll ever see ARM dominance.
And the world has changed, right? If someone is interested in running their workloads on ARM, then they apparently have a pretty portable stack. ARM adopters at the moment will specifically select for those with the least lock-in, who'll try out RISC-V when it comes around.
