NPO limits

Physics (and money, as usual :-) )

The capacitance is determined by the area of the electrodes, the distance between them, and the dielectric constant of the insulating material. To increase the capacitance, you must either increase the area of the electrodes (makes the part bigger), stack up multiple layers of electrodes (makes the part thicker), reduce the distance between the electrodes (makes the part thinner but the thinner insulation means a lower voltage limit), or use an insulator with a higher dielectric constant.

NP0/C0G caps use an insulator whose dielectric constant is relatively low, but which doesn't change much with temperature or applied voltage. Hence, you get a relatively low amount of capacitance per volume, but it's quite stable.

High-capacitance ceramic caps use materials which have a higher dielectric constant, giving more capacitance per layer (all else being equal). The tradeoff is that these materials are less stable in the face of temperature change, and the dielectric constant drops a lot (and so does the capacitance) if they're standing off more than a few volts. Good for bypass caps, bad for predictable low-distortion signal-path filters.

In principle it should be possible to fabricate NP0/C0G caps of much larger capacitance than 0.1 uF - but they'd be so big and expensive that there isn't much market for them. Or, you could make high capacitance in a small volume by using really thin dielectric layers... but the maximum voltage would be so low (and the reliability might be quite poor due to the thin layer being prone to develop pinhole defects) that once again there isn't enough market demand to have them made as standard components.

If somebody develops an insulator with a stunningly high dielectric constant which is temperature- and voltage- insensitive, physically stable, costs $0.02 per square meter, and doesn't cause fatal beak warts in migratory song-birds, then things may be very different in the future :-)