NVIDIA: 15-20% Intel chipset share by year end 'seems reasonable'.

Friday 23rd May 2008, 08:25:00 PM, written by Arun

NVIDIA executives talked at the JPMorgan Technology Conference on Tuesday, and several interesting tidbits came up, from their ambitious market share expectations in Intel chipsets to the technical reason for G92 yield problems and the revelation that G92/G92b will last another '6-8 to 12 months'.

NVIDIA's Michael Hara, VP of Investor Relations, and Daniel Vivoli, SVP of Marketing, were present at the conference. A variety of points were discussed, but most notably:

  • NVIDIA believes 25% share of their target addressable market for Intel chipsets to be a realistic goal. The part of the market they do not want to (nor can) attack is the ultra-low-end where quality and branding are both worthless and you compete on cost exclusively. When asked if that meant 15-20% overall share would be achievable, Hara answered that it 'seems reasonable'.
  • For the first time, a specific reason was given for the lower than expected G92 yields: Michael Hara claimed it was testing procedures, rather than manufacturing itself, which were the problem. If really true, that would imply they were likely too conservative. Changing the SKUs (and/or moving to the 55nm G92b) would also seem like it could help by making some targets less strict.
  • On G92 (and presumably G92b), Hara claimed 'that product has a lifecycle of another 6-8 to 12 months' and then continued on to say that 'the new stuff that is coming out the next 3 to 4 months is all on top of it, positioned above it'. Although this could just emphasize that OEM design wins take forever to phase out, it's more likely to hint at the lack of a GT2xx 65/55nm replacement of G92(b).
  • Optimism about Tesla, but clear emphasis on the fact just the current supercomputer market wouldn't be a big deal; revenue opportunities are very much about growing the market, personal supercomputers for scientists, reducing the importance of the CPU in many ways, etc.
  • On the subject of the handheld business, design wins for the APX 2500 will start rolling out in Q3/Q4 2008 in Personal Navigation Devices and Personal Media Players in Q4/Q1. Hara claimed there wasn't any other company that could effectively target PNDs, PMPs, and smartphones with the same chip. Pretty pleased with the variety of design wins so far. Pricing for this part over the long run should be between $8 and $12, margins starting out in the low 40s (Wait, what? Surely this must refer to the entire family, not just the APX 2500?)

They also didn't seem worried at all about Nehalem and GPU integration via Fusion and Intel's equivalent; certainly it's easier to understand that in the Intel market if they do manage to deliver on their 15-20% share goal for the end of this year; after all, if OEMs like their solutions that much, it would be very surprising if they adopted Intel's dual-chip MCM approach (that doesn't save any real cost) in every single one of their IGP-level products. They also seemed to think Hybrid SLI was a major advantage in that area.

Overall, nothing really unexpected except the Intel chipset share forecast, but a couple of noteworthy tidbits and a fair bit of optimism which only the future will tell us if it was warranted.

Discuss on the forums


nvidia ± g92, g92b, mcp7a, mcp7c

Related nvidia News

CUDA 4.0 and Parallel Nsight 2.0 released
NVIDIA Fermi GPU and Architecture Analysis
NVIDIA's Parallel Nsight finally released
NVIDIA GeForce GTX 460 - GF104 breaks cover
PhysX87, ancient tragedy in 5 acts by RWT
So long, Chris, and thanks for all the fish
NVIDIA GF100 graphics architecture details
NVIDIA Fermi: new GPU architecture, starting with GF100
NVIDIA release OpenCL GPU drivers for Linux and Windows
NVIDIA GeForce GTX 275 at $250 to fight HD 4890