Originally posted by: Pantalaimon
So now the whole mobile graphics chip debacle has cost NVIDIA 300+ million dollars.
You can bet they won't be making that mistake again. These Q&A type lessons-learned opportunities have a way of really making companies more robust at a whole new level.
I've seen it firsthand, its like a wakeup call to management and the groups tasked with the job gain a whole new level of authority and budget to ensure protocols and methodologies are in place to prevent future occurrence.
It can go the other way though too if the internal paranoia over Q&A gets too far out of hand, with future products being so ridiculously over-engineered in the packaging dept so mush so that it results in unacceptably higher production costs (lower margins).
Its also a real opportunity for foundry business to migrate. Anyone else remember when Xilinx gave up on IBM's foundry process at 90nm because the
SiLK low-k dielectric they were using in the BEOL was giving Xilinx massive field failure issues?
Xilinx
shifted the bulk of their 90nm production to UMC thereafter even though IBM moved away from SiLK and adopted a more traditional carbon doped silicon oxide dielectric.
So who is going to benefit from Nvidia's lesson-learned here? Could be more line-items in the "pro" column for a shift to GF's down the road... In the meantime you can be pretty darn sure that the failure point in future gen NV chip is going to be anything but the packaging, and thats good news for us end-users (even if it means costs might be a skosh higher).