[NextGen_Gamer]

It seems that NVIDIA has officially confirmed that "Fermi" will only be released in 2010. This is a quote from NVIDIA's CEO, Jen-Hsun Huang, in their financial conference call yesterday:

“Next year it is going to be an interesting first quarter because, in fact, we will need more wafers than ever in Q1. The reason for that is because – and I mean more 40nm wafers than ever in Q1 –we are […] fully ramping Fermi for three different product lines: GeForce, Quadro and Tesla."

NVIDIA's first fiscal quarter for 2011 begins on January 26th, 2010 and ends on April 26th, 2010. So the earliest we will see "Fermi" will be late January, while it may not come until April 2010. More than likely we will have a paper launch at CES in January, which would correspond to the rumors saying NVIDIA is trying hard to get working samples ready for CES.

[power666]

I still suspect that nVidia will do a preview or paper launch in December, though volume availability in January has been long suspected. nVidia will likely follow the same plan they did last year with the GTX 295: demo/preview in December with availability in January.

The December show off will be only of the Geforce based part with the Quadro and Tesla card demos likely appearing in January for a February or March availability. Those markets focus more on stability than performance so more testing will be done on both hardware and driver levels. With Fermi supporting ECC, it'll likely need a unique board layout instead of one based on an existing Geforce PCB.

The news of massive 40 nm orders could also mean something else: nVidia is readying a midrange Fermi based card that isn't the 512 shader monstrosity they've alluded to. It has been awhile since nVidia introduced a true midrange chip as they've been using more high end designs and just accepting lower margins (or even losses). A 256 shader design with a 256 bit wide GDDR5 memory bus would be cheaper to make than the current GTX 260 and GTX 275's and threaten the Radeon 5770 and 5850 in performance. (That is assuming Fermi's shaders are just as efficient as the GT200's.)

[watzup_ken]

Its a no brainer that they will need more wafer in the future since they are late and they will need to ramp up production. I think based on what we heard Fermi should be a bigger chip than a GTX285 which should also translate to a need for more wafer especially at the start.

[Neo_XF]

Translation ==>> Even if Fermi will end up as a great product (read: not only faster), ATI's head start will mean (aside from the huge profits they're racking up) that they will be ready to one-up nVidia once more soon after. I mean, those news/rumours about them skipping 32nm for GPUs really aren't all that surprising...

Which, in the end, actually means quite bad news for us all...

[Mario26JT]

Originally Posted by Neo_XF

Translation ==>> Even if Fermi will end up as a great product (read: not only faster), ATI's head start will mean (aside from the huge profits they're racking up) that they will be ready to one-up nVidia once more soon after. I mean, those news/rumours about them skipping 32nm for GPUs really aren't all that surprising...

Which, in the end, actually means quite bad news for us all...

Head start and huge profits?Nope,no way - ati 5xxx series selling in very limited amounts,and that suits Nv perfectly in that case.Until ati cant get anough fawers for proper profits,Nv probably release they new vga's..At the moment,all situation better for Nv.An im highly doubt,that Nv dont release anything until Christmas...
Ati be ready to one-up Nv once more?Nope,coz IF true,Fermi is very highly scalable,whta that means you know..Nv was capable fight (and actualy win) with 3 year old technology,so i have no doubt,they come ok this time again