A fragment of a low orbit collision can't end up in a higher circular orbit, because the old orbit and new orbit inevitably intersect at the point/height of the

A fragment of a low orbit collision can't end up in a higher circular orbit, because the old orbit and new orbit inevitably intersect at the point/height of the collision, so in the best/worst case the new (more energetic) orbit's perigee/low point (which is the one that matters for the drag and deorbiting - and small fragments deorbit faster than large satellites) happens to be exactly is at the height of the collision and only it's apogee is raised; and in every other case the new orbit has a much higher apogee but the perigee is even lower, which means that the fragment deorbits on its own even faster. So no, if satellite swarms are in orbits low enough to deorbit on their own within some years (e.g. 550km for Starlink), then a mass collision doesn't result in a Kessler syndrome as the resulting debris is effectively self-clearing. Doing the same at twice the height (e.g. 1100km original Starlink plan) would have that risk though.