Re: Nvidia may be in trouble according to this article
He's got some of Charlie's magical NDA-less sources I'm sure.
I just have to wonder how Charlie's sources are so accurate. If they can't be trusted to keep an NDA they may not be able to be trusted at all.
Charlie got his 2% yield number at first from a poorly translated Chinese website. Then much later claimed he got this number from insider sources... Well which is it? Yields were probably lousy but it remains to be proven.
It didn't take a genius to figure out that there was no way nvidia would launch Fermi in November 09. But that is really the only thing he has definitely gotten right.
We know the card is going to draw more than 225 W but other than that there has been no proof yet that it is knocking on 300 W's door.
Everyone says oh he's been right on everything up until now. But really the only thing he has gotten right so far is the delay timeframe. We're still waiting to see heat, power draw, performance, and finally initial availability.
Also his latest update seems to be an admission that he was wrong on the 448 cuda core issue. His so-called sources say the GTX 480 had 512 cuda cores.