I've heard that the tiny photosites also contribute to the sharpness issue, but I won't pretend to understand why.
A 4K crop of a 12K BRAW file looks 1080p-ish.
- When not stopped down, most lenses are not that sharp, especially older ones.
- Most sensors employ an anti-aliasing (lowpass) filter that blurs the image.
- The smaller the photosites, the worse the signal-to-noise ratio.
- Bayer interpolation.
- Less then optimal post-processing, like aggressive denoising and sharpening with too large of a radius
- Compression artifacts and color sub-sampling.
- Analog outputs are usually bandwidth-limited (lowpass filtered) for EMI compliance reasons. Not an issue with an all-digital workflow of course.