Don't we think subscriptions, with beta 9 in the wild or out of beta, would mean they can raise the recognition?
If they can raise the recognized value of already sold FSD from the last five years from 60% that should make up for most/all short term shortfall? Either all in one quarter for a gigantic profit and then temporary lower numbers for a couple of years or just raise it with something like 5% per quarter. That would take two years to get to 100% recognized at which time subscriptions should be almost caught up.
They probably can't go to 100% the moment they start subscriptions. But it should be worth something so maybe to 70 or 80%?
This one is murky because we don't have a lot of visibility as to the details of the variables are actually influencing the recognition rate on up front FSD purchases. Since Tesla is continuously improving the software and releasing OTA updates there will always be a material portion of the up-front sale that will be deferred over the useful life of each vehicle. We still don't really know what that useful life is on these vehicles as we haven't seen full generational trade-ins of model years (i.e., is it four years or ten years). That may be on the horizon though, which may give the accounting team at Tesla more data to be able to hone in on what the appropriate recognition rate is on the deferral.
Also, the spec sheet for FSD has evolved over time. The argument for recognition of higher % of revenues has generally been that certain features had not yet been released, but once they are that more of the revenue could be recognized (i.e., save for the component that relates to on-going maintenance and OTA updates, which again is anyone's guess as to what the mix of deferred revenue is between those two components). That theory doesn't appear to have played out though with some of the recent City Streets releases not having triggered significant recognition of deferred revenues.
As for the exact % recognition moving on release of V9, I would wager that we will instead see price increases with the same % recognition. If we consider that take rate on up-front purchase may drop with a newly launched subscription model, maybe they sufficiently counterbalance each other in the US. Also to be seen if FSD subscription requires drivers to first have paid for EAP. Something like $4k up-front plus monthly model would be another way to manage the near-term impact. For example, you could see a halving of take rate on up-front FSD, but with $4k required on EAP to subscribe, that could be entirely absorbed.
Based on new car sales, yes. However, the subscription will also be available to 75% of cars that have already been sold.
I am of the view that take rate on the existing fleet won't be material. Similar view to Gary Black's points on the fact that subscription pricing will likely be 199 and not 299+ like some believe, because you could always add FSD up-front to your financing cost and it has a residual value on trade-in or re-sale. In other words, it fundamentally isn't JUST a cost decision (unless you're buying vehicle cash outright).
Ultimately, I don't see offering a subscription being a *
near term* needle mover on existing fleet given that the decision making wasn't purely financial, but rather functional (i.e., the utility of FSD is still be proven). Offering it does allow people to try out FSD without a larger commitment (i.e., save for the fact there may be an up-front EAP requirement).
*
Long term* I do expect it to be boon to margins as more people become comfortable with self-driving and on the assumption that functionally it achieves its march of 9s. When we start to see global take rates on subscription model above 30%, and we start to see up-take on the legacy fleet, then we can get very interesting.