Then there would be a "production prior to X date" drive unit in the catalog for these "reject" motors they actually put into cars.
Just like there is for the suspension parts they replaced with newer parts in December (and again in June) on RWD cars for example.
But there's only 1 production rear drive unit PN, ever, for all LR models. Ditto front DU.
That means it got the software flash to unlock more power and add the red underline. Same drive units though.
I see. So for those suspension parts, the old part is in the catalog but it says "superseded by" part x?
For the motors: I'm not sure that it requires a separate part number, as the screening was only made better as time went by. All units, even the first production units possibly, are 100% identical; they are just binned based on "output" or whatever metric that is relevant for reliability (ones that aren't a pass are discarded - that's binning), and the production limits have been tightened over time. The part number doesn't have to change - they've just made it better through screening (adjusting screening limits to be tighter over time) to avoid field failures/RMAs. It's different than a design change. Happens in production environments all the time. Since the less well screened motors ONLY ever went in certain vehicles (possibly RWD and AWD), it's ok.
In any case I guess Elon wasn't lying in his tweet! Just playing up the Performance model! "Lot sorted for highest sigma output" just means they reject ALL parts which fall outside the production limit line - they don't go in any vehicle. There is I guess the question of "double the burn-in"...double the burn-in relative to what? Maybe double the burn-in relative to early production motors, just to make sure they can take future uncorking. Probably they found they could do this without limiting yields as manufacturing improved. But they can't go back and screen motors that have already shipped to early RWD & early AWD customers!
You might say: But the RWD max output is the same as the performance, so it must be just as good as a Performance motor! Not necessarily - it just has less margin to failure with the as-of-today software limited max current. But Tesla knows such motors were screened sufficiently (perhaps not receiving quite the same screening or burn-in though with their early test programs) to sustain the existing software current limit. But to allow future uncorking of the Performance (this speculation would imply the Performance is not currently uncorked, otherwise Performance rear motor output would be higher than RWD right now), they started to set more aggressive screening limits for ALL motors (with minimal yield loss presumably). Now they can proceed with further uncorking on Performance line only, knowing that ALL motors installed (or installed in the future) in such vehicles will be able to take it. They cannot do the same for AWD or RWD vehicles because some have early motors (with the same part number), which were screened with a looser limit.
* Edit - I suppose since even the original limit was good enough for RWD output, they could boost the AWD rear motor to the same level as the current Performance power limit....hmm. But I don't think they will - I think any increase in current & power probably increases probability of drive unit failure, so why would they take that on...
No assumptions needed, the above is a completely standard production flow as a product matures and process control improves, but it's all speculation.