Nvidia: Fermi in Q1 2010
There have been two camps of rumours on Fermi’s launch date. The first have been saying Fermi is all set to launch in November/December, 2009. The second consider it entirely unlikely, and suggest a late Q1 2010 release, at best.
Now, we have the word straight from the boss. Nvidia CEO Jen Hsun-Huang suggests Fermi production will only be ramped in 2010.
More details next page.
There have been two camps of rumours on Fermi’s launch date. The first
have been saying Fermi is all set to launch in November/December, 2009.
The second consider it entirely unlikely, and suggest a late Q1 2010
release, at best.
Now, we have the word straight from the boss.
Nvidia CEO Jen Hsun-Huang suggests Fermi production will only be ramped
At a conference with financial analysts, JHH said, “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”.
Now, Q1 on Nvidia’s 2011 fiscal year refers to January 26th 2010 to April 26th 2010. This leaves us with a late Q1 2010 for availability, at best case scenario. If indeed something goes a little bit off plans – as they often do with chip development – we are looking at Q2 2010.
It is still possible Nvidia will demonstrate or maybe release a limited quantity of Fermi chips earlier in Q1 2010. Say, perhaps, only a handful of Tesla boards, as that is where the highest margins lie. But widespread availability across all three product lines – we are talking March 2010, at earliest. This date is also supported by a SemiAccurate report which claims Fermi A2 only taped out in WW42, at best, which is late October. We are looking at well into Q2 2010 if Nvidia does in fact require an A3 revision (which is likely).
That is, of course, a massive delay, and well beyond the 2009 release Nvidia and a few reporters have been insisting for a long, long time. This also gives ATI an incredible 6 months or more lead time for its HD 5800 series – that is more than half a generation! They will have to make the most of it – first, by fixing all supply issues. By the time Fermi does eventually come around, ATI will be preparing for either a Evergreen shrink or a next generation architecture.
With numerous controversies, woeful delays, and virtually no competitive Geforce products till Q1 2010, things are looking gloomy. But there’s a silver lining. Nvidia reported a strong Q3 FY 2010 showing – with a revenue of $903 million. Nvidia recovered from previous quarter’s loss to turn a $61.7 million profit. Incredibly, Nvidia expected revenue to be up in the final quarter (October 26th to January 26th, 2010). We are quite certain these don’t have much to do with Geforce products – GTX 200 cards are EOL, GT21x cards are woefully uncompetitive (though we have to give Nvidia’s PR credit). So, from where does Nvidia expect this “increase in revenue”? Tesla? Quadro? Tegra? More evidence that Nvidia is not as reliant on Geforce products anymore?