Caporegime
Yeah that theory has been doing the rounds for a while and it certainly makes sense. What doesn't make sense is Turing. At all. There really is no reason for it to exist at this moment in time. Absolutely none. Even if it weren't stupidly priced, there's obviously no utilisation for it, and it will be a long time before we have a slew of games that do take advantage of RTX features. Even then, we probably aren't going to see full ray traced gaming this generation... nothing outside of 1080p anyway, at best. All we'll probably get is half baked pretty effects of some sort, and DLSS of course, but that's largely been derided by tech press as nothing you can't already achieve with upscaling 1440p to 1800p. I do think ray tracing holds great promise, but its realisation is years away.
I don't know, time will tell, but it's all a big mess right now and what with Intel's insanely priced 9-series, overall just an awful time to be buying new hardware.
It makes sense Turing was developed for the VFX market as Nvidia talked about it and so did investment websites like Nasdaq,so what this is Nvidia making a play for this market,like Fermi was a play for compute in commerical markets. With Maxwell onwards Nvidia had two sets of GPU lines - one for gaming and one for professional markets,so it looks like they have unified these lines to a degree,which saves on having so many lines,so this should help save a degree of costs. Hence all the fully enabled GPUs appear to be in the Quadro RTX lines,and the rest in the gaming cards,but OFC Nvidia needed to find a use for the RT cores and tensor cores,so hence RTX and DLSS. But since the focus was more for commercial usage,the software stack for gaming,etc was probably not quite there,hence why devs seem behind the curve on this.
Basically I see gamers as kind of funding the foray into the VFX market.
If RTX was developed for gaming first,it would have made more sense to produce another line of FP32 focussed cards with larger chips than Pascal,which would have been a bigger jump.Imagine the TU104 chip in the RTX2080 if had standard CUDA cores?? Then develope one high end RTX enabled chip,the TU102,and release this as a professional card for commercial RT work,and seed it to games devs,to get the software stack for games up and running,and then refine the tech and release 7NM cards in a year or so.