• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

TSMC May Lose 16 nm and 14 nm Market Share to Competitors in 2015: Chairman

It's about damn time for TSMC to man up to its problems. I've been saying this for years, NVIDIA should've gone with other fabs to produce their GPU's, seeing as TSMC just can't fix their problems on their own. Samsung is looking like a great candidate, as they've been leapfrogging almost everyone (except Intel of course) in process nodes and the apparent ease that they are accomplishing it every time.

TSMC always has lots of teething problems with their new processes unfortunately, which severely affect the new GPU's production and consequently pricing as well. I guess most companies have finally gotten tired of this and the perpetually-delayed 20nm process may have been the last drop in the bucket.

I know it's not easy to switch to new foundries due to the new libraries and such from the foundry that need to be adjusted to the GPU (or whatever processing unit is being made), but most likely in the long term this will result in huge progress on the GPU front. These are exciting times! :rockout:
 
NVidia needs to start making plans on running their own fabs.. look at Intel, they have full control and seem to be finding it easier to develop these new wafers. I always felt as if TSMC was one of the factors that kept some companies from doing better..

As for the drama above, I actually like this news not for the financial bit but for the advancement in die nm. Seeing TSMC struggle makes me sad since I want the best nm when I buy my technology. :)
 
I see this as a huge sign saying that mobile devices will continue to take jumps in performance per watt, while mainline GPU cores will either have to become more efficient at current process size, or become absurdly expensive to make up for the wafer competition, so bad either way. Also it tells me that AMD is going to have to start kissing bottom at other foundries if they want to continue down their current path of needing smaller nodes to make their dreams come true.

From all angles we are looking at a stall in system performance except by Intel whose fabs are looking awfully shiny and good right now. Perhaps AMD needs to start designing more customer chips for niche markets and lease out their IP on said projects to make more $$$ and get more clout with fabs.
 
Back
Top