What if it's not inherently identical?
If I say that three pencils all have "length" I obviously can't mean that they all have some sort of disembodied labels to that effect, with identical names but unique values. Rather, wouldn't it make more sense to say that all three are objects whose shapes are easily comparable to each other, but which won't impress any such concept onto me unless I do the comparing?
Similarly, one could also describe a basketball as having "length" in a more geometrical and abstract sense - and that's exactly what would make it "more abstract"; that it's much less obvious to compare a sphere to a rod in that way. Which is why mathematicians make the big bucks.
The labels of "length" in essence, then, would themselves be in my own head but only after I'd derived them from some metaphysical fact; from something I'd observed about a certain arrangement of matter. How does it go - "entities are the only primary existents"?
So in one sense, you seem to be right, but not in another sense.
You guys seem to have been playing ring-around-the-rosy between intrinsicism and subjectivism ever since the OP framed the question as whether universals are "in here" or "out there". I believe Rand would've rejected that whole dichotomy (and if she wouldn't have then I think she should've).
Existence is Identity.
Consciousness is Identification.
I agree with @MisterSwig's use of "concrete", as opposed to @Eiuol's (I think - he did get a bit metaphysical towards the end, there). I don't remember Rand using "concrete" to mean specifically physical and extraspective; just individual and directly perceivable (even introspectively perceivable).
I would agree wholeheartedly with @intrinsicist's assertions about metaphysical universals if he'd phrased it as "similar" or "comparable" instead of insisting on "identical".
Ultimately, no two things in the universe are perfectly identical. Even if someone were to hypothetically clone some object with something like a Replicator from Star Trek - if the cloned object was physical then it could not share its parent's location, which would necessarily leave them at least one difference. But we don't need things to be perfectly identical in order to form concepts about them; just close enough.
The valid extent of anything we learn about things with such approximate similarities is not necessarily cut and dry, no matter how well we do our part of the process.
Just look up the main problem with experimenting on mice instead of people. At the end of the day, as similar as we are, a man is not a mouse.
Well. Some of us aren't, at least.