L5 is not going to be an overnight event with 1 release of new OTA software. When Tesla gets to the point of "feature complete FSD" that they release, that will still be officially L2 for a while -- i.e. legally the driver remains responsible and may need to take over any moment.
That will be followed by a long iterative fleet-learning process, where driver interruption data is streamed back to the mothership for the next iteration learning of the NN. Release that OTA, then rinse and repeat, then repeat, then repeat, many times*... At some point in that process it will be re-classified as L3 (at which point the driver is no longer at full responsibility all the time and the car needs to ask the driver to take over sometimes) and more iterations will repeat. And so on and so forth to reach L4 and eventually L5.
My point is, its a long process which merely starts by the release of "feature complete FSD" as Elon calls it.
* anybody who paid attention to the presentations on "autopilot day" should understand this
It's getting fairly OT at this point- so might be better to direct any further discussion to one of the autonomy threads elsewhere on the board.... but your estimation of how fleet data and driver feedback is used is.... very exaggerated from how it's really used... (see among other things Greentheonlys debunking of how much shadow mode actually does compared to what a lot of folks think it does).
And of course, the autonomy day presentation you mention folks should've paid attention to included Elon saying he was confident they'd have feature complete end of 2019 (they didn't) and they'd have legally approved Robotaxis this year (they won't)- so as he himself pointed out in the 60 minutes interview- if he's giving you a date on something he has never done before, he really has no idea what he's really talking about.... (which begs the question why he keeps doing it if he knows that...)
Also actual L5 is basically not gonna happen with this sensor suite- there's nowhere near enough redundancy or ability to handle bad weather... (it already often drops from NoA on highways down to just regular AP in moderate to heavy rain, let alone snow... because there's only a single camera (and in most cars a single non-heated front radar) pointing in most directions- so there's roughly 0 chance it'll ever handle L5 where by definition it operates on all weather conditions....
Best you'll get is L4, where part of the operational domain definition includes "within certain weather conditions" or similar language.... (and to be fair- L4 is all Tesla ever promised in the actual feature description when purchasing FSD for 2016 thorough early 2019 buyers- so they're in the clear there if they at least get that far... and buyers since March 2019 have been promised even less- just a specific list of things rather than general autonomy, and no promise of it ever being better than L2 on the sales screen).
So do not count on that impacting financials or the SP in the short term.
Absolutely- I expect the company, and SP, to continue growing even without them ever reaching L5- even a legit highway only L3, something I think is absolutely possible by next year- would put them considerably ahead of any other car maker though.
The longer term question I think would be what impact would it have if they eventually conclude they can't manage the robotaxi level of autonomy they envisioned with the current sensor suite?
If they've otherwise executed brilliantly, have 4 or 5 gigafactories churning out cars that still sell as fast as they can be built, have terafactories making a ton of next-gen batteries, and they're still generations ahead of everyone else on all other aspects of EVs and the energy business continues growing significantly.... does that really hurt them much? (other than maybe owing some refunds to FSD buyers)?
I don't think it really does other than temporarily.