The video proves that the producer of the video has no idea what vinyl is actually doing, or the fact that the recordings were likely made on analog tape which doesn't have that bandwidth either. Look at the levels involved, and where the ultrasonic information is relative to the audible information, then consider what just a little distortion caused by a stylus mistracking, worn grooves, worn stylus, or all three, would look like. That's not original audio content at all, it's been added by the deficiencies of the playback system.
Vinyl has a practical physical limit to the maximum modulation at any frequency related to cutting stylus velocity, and cutting stylus shape, and that maximum falls rapidly as frequency goes up. If you increase the cutting stylus velocity with high level, high frequency information there is a point where the rear facets of the stylus slam into the wall of the groove just cut. In addition, the cutter head has very little ability to dissipate heat for very long, and with the inverse RIAA curve, there's a lot of HF energy there to generate heat. There is usually an 18kHz low pass filter in the chain somewhere to protect the cutter head from overheating, though it's not a brick wall, it could be 18dB/oct.
What all of this means is the extreme ultrasonic content seen in amateur spectrum analysis is very low in level, likely not cut into the groove, and is actually distortion products and noise. As to frequency response, there are many factors starting with the cartridge, including cable, preamp front-end, and RIAA network that impact the ability of the system to playback a flat composite response at levels well below maximum.
By contrast, any linear PCM digital system can reproduce its maximum capable output at any frequency up to the Nyquist filter. Distortion products at any frequency are extremely low, part of the noise floor. Composite record-play response is very flat throughout the entire audible spectrum.