Taiwan’s TSMC(Taiwan Semiconductor Manufacturing Company) is now the most sought after foundry in the world, with giants like Apple, Huawei, AMD, Nvidia etc, using their fabrication process for making chipsets for the respective device manufacturers. But now with Huawei reducing capacity due to a shortage in phone sales, a large chunk of the capacity was up for grabs, and both AMD and Nvidia have grabbed all they could.
According to the Digitimes report “A ramp-up in short lead-time orders placed by AMD and Nvidia will allow TSMC to post relatively strong results compared to other foundries in the first half of 2020”
AMD needs a lot of capacity for their RDNA2 based graphics cards, including Big Navi. AMD also needs capacity for ensuring that Microsoft and Sony have enough graphics chips for their Xbox Series X and Play Station 5 consoles. Nvidia on the other hand is gearing up to release it’s Ampere lineup that will be competing in the high end gaming market, by releasing a 4k 60fps gaming GPU, and who better than TSMC to fab them off?
Nvidia is reported to be using TSMC’s new 5nm process node, but not for its Ampere GPUs.Ampere will be using the 7nm node for now, but Nvidia may be preparing their 5nm process for another device, which could probably be their server GPU.While 5nm is on the horizon, AMD is increasing the pace of its 7nm process chip orders. Meanwhile, Nvidia is buying increased quantities of 16nm, 12nm and 7nm nodes, thanks to demand for products across its whole family. Products using 16nm and 12nm nodes include the Tegra X1 processor found in the Nintendo Switch, as well Nvidia’s GPUs based on Pascal and Volta architectures, while its 7nm orders are being used for Ampere.
Both Apple and AMD are looking to book ‘N5 Enhanced’ production capacity at TSMC for later in the year. This enhanced process was developed specifically for AMD, says the report. “N5” is the 5nm node that will be used by Apple and AMD starting this year, and going on through next year.