Although the Switch already has three models, the core hardware is basically the same, and the main computing chips are all fromNVIDIA’s Tegra X1 (T210+T214).
Although Nintendo game consoles do not rely on stacking hardware to attract consumers, the Tegra X1’s backward process (20nm, currently upgraded to 16nm.), weak performance hinders the game’s picture quality and resolution to a higher level.
According to the latest news, NVIDIA employees confirmed the existence of the Tegra T239 chip. In fact, as early as a year ago, Nintendo was revealed to be using the Tegra T239 to develop a new game console.
The prototype of the Tegra T239 is the Orin chip, which has been adopted by some advanced new energy vehicles as the autonomous driving computing center. However, the most entry-level version of the Orin family that supports L2-level assisted driving consumes 15 watts, and the CPU part is a 12-core Cortex-A78E. Obviously, the T239 needs to be significantly modified.
This revelation pointed out that the T239 is actually an 8-core Cortex A78 chip, and the GPU part adopts the Ampere architecture, which may introduce some advanced features of the Ada Lovelace architecture (DLSS3?).
As a comparison, the Tegra X1 CPU part uses 4x Cortex-A57 + 4x Cortex-A53, and the GPU is still the old Maxwell (256 CUDA). Even if the size of the T239 GPU remains the same, the performance will definitely be improved several times.
After the hardware is strengthened, the 4K resolution output (Dock mode) and 1080P resolution that players expect will be realized without pressure.
Further reading:
