NVIDIA seems to be planning to move its current generation GPU, Ampere to a smaller fabrication process. A new report from DigiTimes is saying that NVIDIA is planning on moving the consumer-grade Ampere to TSMC’s 7nm process by 2021.
NVIDIA currently utilizes Samsung’s 8nm process to make the Ampere GPUs, which seem to currently be in short supply. According to the report, the shift to TSMC’s 7nm is coming because TSMC has become “friendly” to NVIDIA. TSMC is moving to 5nm and has already begun the work towards its 2nm process, leaving room in the 7nm facilities. In the meantime, NVIDIA is looking for better yields.
The whole shift should be relatively easy for NVIDIA, as it is already using the process for the non-consumer Ampere GPUs. It is being said that NVIDIA could use the process for the SUPER/Ti variants. This would mean it would keep Samsung’s 8nm for the normal consumer Ampere cards.
Thoughts? Hit us up in the comments, down below!