As reported by the DigiTimes portal, the TSMC foundry and other Taiwanese technology companies have secured the business from Nvidia for the future GPUs of Hopper architectures and the Ada Lovelace GPUs, known as the NVIDIA RTX 4000 series. The latter is the one expected. It hits the market next year and delivers twice the performance, at the cost of higher power consumption than a larger power supply could require than many enthusiasts typically have.
With the current generation of Ampere-based GPUs, Nvidia split the chip production contracts. Enterprise GPUs and High Performance Computing (HPC) are sourced from TSMC and consumer GPUs from Samsung.
TSMC will manufacture the RTX 4000 known as Ada Lovelace
According to the article published by DigiTimes, the H100 will be the first GPU based on the Hopper architecture. It will be manufactured in TSMC's 5nm process with advanced CoWoS packaging. The Hopper series of GPUs will be ready to be used in various applications such as supercomputing clusters, data centers and other similar high performance environments, not so much as consumer.
Insights revealed by the Taiwanese industry shed some light on Ada Lovelace GPUs. GPUs based on the Ada Lovelace architecture will be used to power the GeForce RTX 40 series graphics cards, which will be aimed at enthusiasts and gamers. These GPUs, also manufactured by TSMC, are expected to be available from Q2022 XNUMX at the earliest.
While it arrives, new Turing and Ampere architecture GPU models are expected to arrive in between targeting the PC gamers sector. They will be reissues and renovations of the old GPUs since there is demand for them and they are not reaching the players. The 2060GB GeForce RTX 12 is scheduled to go on sale in the coming weeks, while the new Ti or Super variants of the RTX 30 series will be available for notebooks and desktops in early 2022.
Source: Guru3D
