We are pretty far off topic here, if OP wants this discussed elsewhere just speak up.
Its not a "good enough" before assembly, actually the opposite, its better in a sense. When done as a bare shaft it is concentric to the bearing journals, period. Once you add bearings, spacers, nuts, rotors, housings etc the tolerance stack up from all that -- if done not quite right -- can create taper runout. Now you go and grind the taper, sure it runs out good now, but you still have the misalignment between the front and rear bearings that was the root cause, (or whatever the cause was) all you did was hide it. Now are the bearings going to have a long prosperous life, no, but my taper runout looks great!!
When done right, taper is ground bare shaft to the right AT class--confirmed with air gauge, runs out about a micron +/-. If dual contact, that relationship is correct relative to the face. Test bar in the taper should be good as well and really confirms the results. Once the entire spindle is built, I expect to see that taper runout not move much, maybe a micron, This proves that all the other parts are right and the assembly is healthy. At this point, no grind is necessary because all the bearings are set right, no spacer issues etc etc. Test run and install.
Grinding in assembly is usually covering an issue elsewhere which still exits afterwards.
As with everything, there are sometimes exceptions. Ive had situations where sub 1/2 micron max TIR was required at the tooling taper after assembly, now you are getting to where the bearings are limiting factors so you must "cheat" and hit it after assembly. This is really air bearing territory, but a topic for another day.
I have never seen an oem grind in assembly. Think about it this way, next time it needs to be rebuilt, the taper would not run true to the journals and therefore must be ground again..every time.