We've gone rounds on this before, but I disagree. Bayer spatial sampling interpolated to RGB color with high quality algorithms does not produce significantly lower resolution for red and blue channels. It is slightly less than what full discrete color channel sampling would be with a 3-chip camera, but you could not see the difference. Your eyes sample color in the same ratios as a Bayer sensor. My Bolex can easily produce a high resolution 4:4:4 RGB color gamut well beyond current YUV sub sampled video standards. It far exceeds P3, BT.709, and Adobe RGB gamut. Resolution is determined by the full luminance resolution of the sensor, not the color filter pattern. There is no direct comparison to the 50% reduction in actual horizontal pixel resolution for 4:2:2 sub-sampled encoded video formats. The loss of resolution on encoded Bayer video cameras is due to low quality debayer algorithms necessary for real tiime high speed frame processing with limited available processor power.
I think there might be something to your argument. Arri has provided UHD 4:4:4 on the Alexa Mini for some time and I have never heard anyone complain about it being a waste of space even though the sensor is only 2.8K.
If it’s technically possible and advantageous, I would always prefer a 4:4:4 debayered file for the same reason I would prefer the option to record a 12 bit RAW image. That extra headroom is likely to save a shot one day so I’m willing to pay for the extra drive space.