So it looks like we have the first independent confirmation about AMD Radeon's exact node. TSMC will be manufacturing the next generation AMD graphic cards (Arctic Islands and Greenland) on their 16nm FinFET node - something that was pretty much expected. It would appear however, that AMD will be reverting to Global Foundry's 14nm FinFET process for its Zen based processors (via KitGuru). While this aspect of the news (Zen) is also not really surprising, I would urge users to take it with a grain of salt, since it is very clearly not confirmed yet.
Not an official poster. @Wccftech
AMD Greenland graphics to be manufactured on the same node as the Pascal flagship - Zen possibly contracted to Global Fondries
One of the major problems (in my humble opinion) for graphic card manufacturers is that foundries like TSMC usually allot high priority to clients like Apple. This means that they get the first spot in a queue and the first try at the goods even after a specific node has matured to the point it can be used for high performance ASICs (which mobile SOCs do not really qualify as). TSMC lost quite a bit of share of Apple's A9 chips to Samsung Electronics and while it has retained a portion of the order, Qualcomm has decided to deal exclusively with Samsung this time. This reduction in clients and therefore the bump in the priority queue might (read: possible, not probable) translate to the next gen GPUs coming just a bit earlier.
TSMC’s 16FF+ (FinFET Plus) technology can provide above 65 percent higher speed, around 2 times the density, or 70 percent less power than its 28HPM technology. Comparing with 20SoC technology, 16FF+ provides extra 40% higher speed and 60% power saving. By leveraging the experience of 20SoC technology, TSMC 16FF+ shares the same metal backend process in order to quickly improve yield and demonstrate process maturity for time-to-market value. Now it is the 16FF+ variant that is of specific interest to us - because it is this variant that the AMD and Nvidia will most probably settle for.
TSMC and Nvidia have also confirmed on more than one occasion that the next generation (Pascal) GPUs will be produced on the 16nm FinFET+ node, with initial confirmation dating back approximately 9 months. AMD’s next generation Radeon graphics processor on the other hand, codenamed Arctic Islands, was not (initially) on the official list of products released by TSMC, so while their CEO have confirmed the use of a FinFET node (14/16) the exact specifics had remained to be seen. That particular mystery can now be considered solved since Digitimes reports the future Arctic Islands lineup from AMD will be fabricated on the same process (as always) as Nvidia's GP100 GPU. Interestingly, there is no word on Zen yet, which might indicate a reversion to Global Foundries as originally stated.
16nm FinFET tech entered into risk production and approached mature yields a while back, and now full fledged production has begun full steam ahead. More than 60 projects are underway, with known products in development including Avago, Freescale, LG, MediaTek, NVIDIA, AMD, Renesas and Xilinx. The list is obviously not exhaustive in nature but it finally includes the two companies our readers are most interested in: AMD and Nvidia. There have been rumors circulating about red and green flagships (namely Greenland and Pascal) with an insane amount of transistors (17-18 Billion) and I must say that unless TSMC has deliberately withheld details about its new process, I find the premise of a ~600mm^2 first generation 16nm FinFET+ product from any company very unlikely.
