Graphics processing unit (GPU) giant NVIDIA might be on to something big this year, as rumors point to its next flagship GPU being announced sometime in April.

The said flagship graphics card, which is used primarily for gaming and professional software, is none other than the long-awaited GeForce GTX 2080 that is set to replace the GTX 1080 released back in 2016. One thing to note is that the rumors are about NVIDIA's announcement of the GTX 2080 and not the actual release, though that may come shortly after the announcement.

Speculations began from the readers of German site 3DCenter, in which the obsoletion of a previous GPU architecture has just been discovered. This was the GP102 that powers the GTX 1080 Ti, an upgraded version of the GTX 1080. The said architecture is no longer being produced at the moment, meaning that current-gen GTX 1080 Ti and TITAN X graphics cards would stop being made as well since there could be a replacement on the way, hence the rumors.

The new architecture is said to be code-named Ampere after the previous generation being dubbed Pascal, and the GP10X architectures will be replaced with GA102 (Enthusiast), GA104 (HighEnd), GA106 (Midrange), GA107 (Mainstream) and GA108 (LowCost) architectures.

There are also some expectations that NVIDIA would be utilizing a new and much faster video random-access memory (VRAM) components for the GTX 20XX series of graphics cards starting with the 2080 at the top. This was due to NVIDIA inking a deal with Samsung that would see the company making its new 14 nm GPUs and as such, might have VRAM modules from the Korean company.

This might ultimately mean the replacement of graphics double data rate 5 (GDDR5) currently used today for the faster and better GDDR6 VRAM. Estimates of performance increase range from 10–30 percent compared with the previous GTX 10XX series counterparts.

PC owners in the market for an upgrade or a new GPU will do well to wait until April before making their decisions on which GPU to purchase.