Keep in mind that there will be considerable expense that goes along with that revenue. These computers are costing billions and they become obsolete within a few years. Still, looking forward to the ramp.I think wall st analysts who are 100% focused on valuing the company like any other car company from the 1980s... are going to be caught completely unaware by Teslas long term revenue from FSD subscriptions.
They absolutely 100% guaranteed will not see what is happening until there is a big jump in FSD revenue reflected in a quarterly financial statement.
Services and other revenue in Q4 2023 was approximately 10% of total revenue, almost exactly the same as the full year figure.
From the end year deck:
In other words, FSD subscription revenue has been tiny up until now. As Tesla sell more cars at a lower price, the percentage of buyers who have the up-front $12,000 to buy out FSD completely is bound to fall, so I would expect that the percentage of FSD revenue that is subscription based will rise.
If Tesla continue to have no desire to impress wall st, they can continue to not draw attention to subscription revenue until it becomes awkward not to.
As someone with TSLA stock, I'd love to see official figures on revenue from superchargers, and from FSD subscriptions. However I get the impression Elon hates having people go through the company data with too fine granularity, so I reckon they will lump all this stuff together and hide it for a while, just as I expect they will not give us exact cybertruck production figures until its tens of thousands a month.
I didn't honestly expect 2024 to be the year of FSD revenue growth, but here we are!
Only if you can also give a reason why you're doing it this way. Example. FSD wants to turn on the first block past the alley instead of continuing to the main intersection and then turning. The street off the alley is wide with two big lanes and has few to no parked cars. The street FSD wants to turn on is technically a few metres shorter, but has cars parked on both sides and room for one car between them. So the larger street will be safer and quicker. (FSD also wants to go down that narrow street at 30 mph which is about double the safe speed).
it probably needs the new 800V / 48V architecture which I think will be in all future Gen3 vehicles.
In NZ there is another plan
Ah, finally the power of exponential growth and just its beginning. No more hardcoded limits only AI self reinforcement.
Actually it's probably the H100 coming online. It has been a year since H100 started to ship and it takes 9-12 months to get a datacenter online from beginning to end. Of course Tesla didn't get the allocation of 10k H100 since release date as Musk was complaining how they are unobtainium at the time. So lets say they got 10k worth 6 months ago and it took them 6 months to set up a datacenter using them(and this is 2x the speed of industry standard).
Well Tesla actually already worked out 90%+ of all the problems from self driving before V12. They went from fixing thousands of problems to now finding specific problems. Remember how brutal V1 was. People said Tesla couldn't even drive straight. Now just imagine if they fix 1 or 2 problems on your drive and you'll be disengagement free for weeks. And another problem fixed will lead to disengagement free for months. We are at the tail end of FSD training...now if only they get map data to be just as good as self driving....
Waiting to buy back in at $160 or $150 after P&D and earnings later in April.
