My initial reaction is that they are looking for an ARM competitor, and they have finally realized x86 will never be that. Not necessarily to get into mobile, but you have an awful lot of low compute general CPU products out there, and ARM has pretty much been thrown in all of them - IOT devices, embedded controllers, etc.
There's no good reason RISC-V couldn't be used there, except for the fact that ARM already has the foothold, and a 20? year head start. I can't really see this making a push against Intel's x86 side - I think this is all about expanding market share into different areas, because x86 has reached it's market potential, and the only growth there is just in following typical ebb and flow of economies and population growth.
I would bet, had nVidia not snatched up ARM, Intel would be looking at cozy up ARM hard instead. I'm actually surprised they hadn't... AMD did years ago, and had some x86 CPUs with ARM cores. They didn't do much with it, apart from make some hybrid x86/ARM server packages, but they have the licensing and such in place pre-nVidia buyout.
Intel has always just been so blinded by x86 - everything had to be x86 for them... even their first real GPU effort (larrabee). I guess when all you have is a hammer... I can't help but think that was a marketing/C-level decision to leverage their existing IP portfolio, and all the engineers just had to suffer through trying to pound that round peg into the square hole.
All that said - in five years, this could look like prescient genius. It all depends on how nVidia treats ARM and it's existing customer base. nVidia doesn't exactly have a great reputation when it comes to open technology standards, and they could very well stagnate the ARM market - existing customers locked into grandfathered licenses of old revisions, but new revisions come with so many strings that it isn't appealing or cost effective. The ball really is in nVidia's court on this one.
