The M855/SS109 projectile has a lot of things going against it for making a precise bullet. 1) It's an FMJ. It's harder to hold consistent wall thickness when the ogive is the bottom of the cup vs. when it's a flat base (even at the back of a boat tail it's still a flat base for a punch to hit in the draw process). 2) It has a boat tail. Flat-base bullets are easier to get to shoot accurately because you have 1 less feature who's centerline and CoM have to be in line with the rest of the bullet. And 3) it has a 2-piece core. 2a) One piece of the core is not malleable at all. If the steel penetrator lands in the jacket in a fuzzed up orientation, it's going to remain as such.
Harder to get a concentric jacket, harder to get all inside and outside features of the bullet on the same centerline, and usually nobody's looking for the utmost in precision from that bullet, so I'd imagine a lot of manufacturers just slap them together and roll with it. Whatever doesn't make the mil spec gets dumped as cheap blasting ammo on the civilian market. If they shoot worse than 2 MOA someone was asleep at the wheel (or didn't care/need to care) on the bullet press.
ETA: shooting so close (25m) you really gotta watch out for parallax error, too. Very small changes in distance to target have much greater impacts on parallax error the closer you are.
Whether or not the dispersion is always vertical will probably be played out with a larger sample size. Shoot 10-20 in the same group and see if it's still a vertical line. If it is, I'd honestly look at optics or the rifle before the ammo/bullets. I haven't seen a bullet yet that will sling consistently in a line (within 300yd), vertical, horizontal, or otherwise. If they're bad it's a big round group once you get enough shots into it.