He got those numbers from Nvidia. It was in their financial report.
We all know that. But his argument is that this proves bad Turing sales, which is pretty far-fetched.
It would be nice if we had some empirical observations that make this hypothesis more probable. For example: there aren't many RTX users among gaming forum peers, AIBs don't invest in making Turing cards, OEMs don't use them etc. But none of that is happening.
Turing launch looks exactly like every previous successful Nvidia launch did - with 2 major exceptions/circumstances:
1) It happens after the crypto mining collapse, so there are some question marks over Nvidia's financial situation (which can also be said about many other companies - including AMD and few graphics card AIBs)
2) Unlike previous generations, this one puts features before performance, so it's not as easy to evaluate. Obviously, people understand GPU fps graphs better than GPU technology*. And while performance is objective, image realism is not.
(*) which in case RTRT is still surprising. Ray Tracing is a very intuitive technology compared to what GPUs normally do. It's a much better model of what actually happens around us.
And, more importantly, it's been the dominating rendering method for decades.
For the last 2 years quite a few people on this forum have been raving about Zen rendering potential. But the RTX discussions have shown that some of them have no idea how rendering works.
I find this
deeply disturbing.