Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla looking to recognize more FSD revenue for 2020 Q1

This site may earn commission on affiliate links.

Mardak

Active Member
Oct 13, 2018
2,786
6,647
USA
elon 2020 q1 fsd.png
Elon Musk on Twitter

Q1 typically has lower revenue from vehicle deliveries, so if Tesla wants to boost their Q1 numbers, recognizing deferred revenue seems like a reasonable alternative especially with over $1 billion (not necessarily all FSD related) according to their FY2019 update.

Unclear if FSD visualization preview counts as a feature for recognizing some revenue. For 2019Q3 -> Q4 the total deferred revenue still increased although this was limited to just US market. If some revenue could be recognized, making the visualization available to the rest of world seems like an "easy" change (even if the visualizations are not quite right as seen in the existing US behavior).

"Reverse summon" (i.e., "park seek mode" ??) was recently mentioned, so that definitely would be a feature allowing recognition of revenue. Zachary Kirkhorn said Tesla was able to recognize $30 million of deferred revenue when Smart Summon was released in the US, so one would guess a similar proportion could be recognized with the "reverse."

And as always, there's stopping for traffic lights and stop signs and navigating through intersections for "feature complete."

Any other hints at what "FSD features later this month" might be referring to?
 
  • Like
Reactions: destructure00
Any other hints at what "FSD features later this month" might be referring to?

Probably a reference to reverse summon. And based on the FSD website, I would guess that Tesla is probably trying to finish and release "traffic light and stop sign response" at some point too. We know they have the visualization for it and they have been working on it for over a year.
 
I presume Tesla had to make an accounting provision for the costs of upgrading HW2 and HW2.5 to HW3 for those relevant vehicles with FSD since Elon promised to do it for free. As those vehicle get upgraded, they can unwind that accounting provision, which is probably more of a P&L impact than a revenue impact.

I tried to do my part to upgrade, and the appointment was today, except they cancelled it because of parts non-availability. :mad:
 
We should remember that any "features" that Elon is talking about have to go to Early Access first, several iterations and only then, go out to the fleet. So "releasing more features later this month" probably does not mean that we will get the features at the end of March, it just means that EA will get them... if there is no delay.

We all remember what happened with Smart Summon. Elon promised the feature "next week" and that became "next month" and that became "EA will get it next week" and that became "we need to work on it some more" and that became "EA will get version 2 soon" .... It was several months before we actually got it.

I would be shocked if Tesla releases Reverse Summon to the entire fleet this month.
 
We should remember that any "features" that Elon is talking about have to go to Early Access first, several iterations and only then, go out to the fleet. So "releasing more features later this month" probably does not mean that we will get the features at the end of March, it just means that EA will get them... if there is no delay.

We all remember what happened with Smart Summon. Elon promised the feature "next week" and that became "next month" and that became "EA will get it next week" and that became "we need to work on it some more" and that became "EA will get version 2 soon" .... It was several months before we actually got it.

I would be shocked if Tesla releases Reverse Summon to the entire fleet this month.

IMHO "Reverse summon" won't be released to vehicles without HW3 (aka FSD computer), so never "the whole fleet".

I also though one of the reasons for the Smart Summon delay was making it work on HW2 and HW2.5 vehicles because they'd promised it to EAP customers. They probably really regretted that promise, and it probably was such a distraction as to put the native HW3 support project back several months.
 
IMHO "Reverse summon" won't be released to vehicles without HW3 (aka FSD computer), so never "the whole fleet".

I also though one of the reasons for the Smart Summon delay was making it work on HW2 and HW2.5 vehicles because they'd promised it to EAP customers. They probably really regretted that promise, and it probably was such a distraction as to put the native HW3 support project back several months.

True, Reverse Summon will only be for HW3.

I've heard anecdotal evidence that AP is a lot better on HW3. I am still waiting on my HW3 upgrade. So it is possible that we will see a big leap forward on HW3 after the FSD rewrite.
 
I've heard anecdotal evidence that AP is a lot better on HW3. I am still waiting on my HW3 upgrade.
At least from our FSD computer retrofit, there's some cases of slightly improved Autopilot and some cases of slight regressions, but nothing that's "a lot better" other than fixing the "move out of passing lane" bug with HW2.5.

At least from Full Self-Driving Computer Installations it seems like your Model 3 in North America MCU2 + HW2.5 is "Scheduling Now," so request it with "Tesla app by selecting 'Schedule Service’ > ‘Accessories' > ‘Upgrades & Installations’ and add ‘FSD Computer Upgrade' in the description."
 
  • Like
Reactions: diplomat33
As those vehicle get upgraded, they can unwind that accounting provision, which is probably more of a P&L impact than a revenue impact.
I wonder if FSD visualizations being restricted to FSD computer is an accounting trick where if Tesla can recognize some FSD Capability deferred revenue, having Service complete a FSD retrofit and updating software to allow FSD preview visualizations is effectively a revenue generating procedure.
 
  • Informative
Reactions: pilotSteve
I think the FSD rewrite should be Elon's last chance at convincing us that FSD is soon possible. I am assuming that these new features Elon is referring to will be derived from the FSD rewrite. If these new features are not significantly better than what we have now with smart summon and navigate on autopilot, it doesn't look like FSD will come within the next year or two. If that's the case, then once again Elon will be way off on the FSD timeline. It is sad that we're still nowhere close to FSD, yet FSD is 50% of teslas overall mission. Elon has said that 2 of Tesla's major goals is electrification and autonomy.
 
Last edited:
I think the FSD rewrite should be Elon's last chance at convincing us that FSD is soon possible. I am assuming that these new features Elon is referring to will be derived from the FSD rewrite. If these new features are not significantly better than what we have now with smart summon and navigate on autopilot, it doesn't look like FSD will come within the next year or two. If that's the case, then once again Elon will be way off on the FSD timeline. It is sad that we're still nowhere close to FSD, yet FSD is 50% of teslas overall mission. Elon has said that 2 of Tesla's major goals is electrification and autonomy.

The rewrite will likely bring some regressions, and take time to refine. I am certainly not expecting a major breakthrough this year, but think it will be interesting to participate in the incremental progress.
 
The rewrite will likely bring some regressions, and take time to refine.
It'll be interesting to see if @verygreen notices B node doing something different than a plain copy as suggested in a tweet from when FSD visualization preview came out: green on Twitter

Tesla could reduce regressions by having the B node run the rewrite while still actively driving with the original code as before. But if new features are dependent on the rewrite, Tesla is then incentivized to deploy potentially buggier code to recognize deferred revenue.

Although a potential hybrid approach could be to keep existing Autopilot driving behavior as before and have B node run the rewrite to power "dramatically improved" visualizations and "more FSD features" that didn't exist before without conflicting with the well-tested existing driving decisions. And behind the scenes, this would allow Autopilot to detect B node driving behavior regressions without actually exposing the regression.
 
  • Like
Reactions: Kmartyn
Yes, I realize the intended functionality of B node is for redundancy during autonomous driving, but while human drivers are required, Tesla could safely make use of it for additional functionality. Making use of those resources could help Tesla experiment with ideas before deciding to keep and optimize to run on a single node with redundancy.
 
  • Like
Reactions: ramonneke
Looks like green found some new assets in 2020.12 that might be used for indicating when Autopilot will stop at red lights.
2020.12 stop.png


Seems likely it'll be based on the existing red light and stop sign warning behavior, which seems to currently rely on OpenStreetMap data having the highway=traffic_signals and highway=stop features (in addition to the neural network detecting the red light / stop sign and stop line).
 
  • Informative
Reactions: diplomat33