New rumours circulating with the Ampere-based RTX 30 series graphics cards being made by Samsung 8nm LPP process, contrary to the earlier assuming TSMC 7nm. The original speculation had weight since the Ampere A100 was already known to use TSMC 7nm node.
According to the source, Samsung has been trying to get an upper hand over the Taiwan Semiconductor Manufacturing Company by providing chip fabrication for cheap. TSMC was getting a lot of orders from multiple manufacturers- including AMD and Nvidia. Nvidia usually used Samsung for its low-end offerings while TSMC was traditionally for mid-to-high-end graphics cards.
For AMD, all of its chips are made by the TSMC, including the Ryzen 3000 series XT Matisse Refresh CPUs and its graphics chips.
As of now, nothing is set on stone. Most likely, Samsung may have a few more chips than usual to make, depending on Nvidia. Samsung’s 8nm process is based on its 10nm process, according to the prolific leaker’s tweet. He even went on to say all will be made on Samsung 8nm. But AMD relies on TSMC for 7nm manufacturing and therefore took whatever slots they can. This might have been a tipping point for Nvidia? Who knows. From a layman’s point-of-view, it makes sense to diversify, starting with the coming RTX 30 Series. It will be interesting to see how it plays out in the end.
What is known so far is that Nvidia will release RTX 3090, RTX 3080/ Ti/Super and maybe RTX 3070. The Founders Edition has a very different cooling design- both internally and externally, while its AIC partners are using a newer variant of its cooling system most likely with custom PCB layout.