Not quite sure what you mean,
20000MP are you saying megapixels? Thats not true, the resolution you need can be calculated from your render resolution and distance to the object.

For example if your rendering a 2048 x 1716 and your object will only every be half of the screen width / height in camera you only need at max a 2k texture for the entire thing.

Something that might be making your textures blury could be filtering. If you go to your texture node you can change to different and sharper filtering methods to get different results.

There is a strong chance that in your case that if you are using very high resolution textures with bad filtering and low samples that you will get a blurry result. If on a given pixel in a render there are hundreds of texture pixel values to filter you will get a blurry result if your filtering is bad.

Most renderers have a dedicated mip map texture format. For example with pixar renderman there is the TEX format which is esentially many tiffs, when generating a TEX texture from an exr or other image it will make various filtered versions of the texture into a fast acceleration structure so that at render time the appropriate filtered version is chosen for the given distance from camera and resolution. Also the renderer is able to load portions of the texture and then remove them from memory once it is used rather than loading the entire thing at once.

Here is a thread I found on filtering and mipmaping on CG society which covers things well:
https://forums.cgsociety.org/archive/...t-1046149.html