The price includes any Tesla discounts already offered including the larger discounts on new inventory cars which have been demo cars. EV-CPO data for this purpose is manipulated by Tesla as it only covers the cars Tesla want you to see. It may be that the figure is skewed by models Tesla is having trouble selling and so appearing for sale, but you’d hope they weren’t making large volumes for stock with an undesirable and expensive config. Tesla may sell for lower still, may offer higher trade in or subsidise lease deals ending, etc too.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Near-future quarterly financial projections
- Thread starter luvb2b
- Start date
The price includes any Tesla discounts already offered including the larger discounts on new inventory cars which have been demo cars. EV-CPO data for this purpose is manipulated by Tesla as it only covers the cars Tesla want you to see. It may be that the figure is skewed by models Tesla is having trouble selling and so appearing for sale, but you’d hope they weren’t making large volumes for stock with an undesirable and expensive config. Tesla may sell for lower still, may offer higher trade in or subsidise lease deals ending, etc too.
Fact Checking
Well-Known Member
BTW., this is the current wording of EAP:
"Enhanced Autopilot
Option Selected $5,000
Additional features, including:
Option Selected $5,000
Additional features, including:
- Auto Lane Change
- Autosteer
- Traffic-Aware Cruise Control
- Autopark
- Summon
Enhanced Autopilot includes additional driver assistance features. Every driver is responsible for remaining alert and active when using Autopilot, and must be prepared to take action at any time."
I believe 100% of this is present in all vehicle platforms. Was Model 3 Summon enabled before July 1st? If yes then this might have been fully recognized already, in Q2.
FSD is characterized currently as:
"Full Self-Driving Capability
Option Selected $3,000
In the future, Model 3 will be capable of conducting trips with no action required by the person in the driver’s seat.
Requires Enhanced Autopilot
$5,000 upgrade if added after delivery
This functionality is dependent upon extensive software validation and regulatory approval. It is not possible to know exactly when it will be available, as this is highly dependent on local regulatory approval, which may vary widely by jurisdiction."
Option Selected $3,000
In the future, Model 3 will be capable of conducting trips with no action required by the person in the driver’s seat.
Requires Enhanced Autopilot
$5,000 upgrade if added after delivery
This functionality is dependent upon extensive software validation and regulatory approval. It is not possible to know exactly when it will be available, as this is highly dependent on local regulatory approval, which may vary widely by jurisdiction."
Arguably the on-ram-off-ramp functionality rumored to be in V9 can conduct many types of highway-only trips with no action required by the person in the driver's seat. So if Tesla decides so they could recognize some of the FSD revenue - say 25% of it?
The FSD deferred revenue could be pretty significant. According to the Troy Tracker the take-rate is a surprisingly high 23.6%. This might be due to the expectation that only owners of the FSD option are going to receive the Tesla AI Chip hardware upgrade for free.
Tesla probably also moved a fair amount of FSD demand forward by pricing the delivery-time option at $3,000, while the post-delivery price is $5,000, which is a significant price gradient.
So I'd not be surprised if the revenue recognition would be in the $100m range...
That list might be the currently available features.
From Autopilot
Auto lane change with on-ramp to off-ramp may be the last missing feature provided in v9 (depending if summon is good enough (limited movement range/ no path following)). Current lane change is auto after being driver commanded. The $5,500 if you order now definitely looks like a last chance for people to get in and lets Tesla pull forward more EAP adoptions/ revenue.
I don't see mention of FSD features and I don't expect them before HW 3 (which EAP does not noticeably benefit from/ require).
The deferred line had a major restructuring between Q4 17 and Q1 due to the accounting changes.
Wading through the 10-Q for Q1. Due to the change in accounting practices, the deferred revenue line lost a lot of least related items. Fortunately, they broke out the adjustment.
In thousands:
End of 2017: 1,177,799
Adjustment: (429,771)
Start of 2018: 748,028
End of Q1: 536,465 a drop of 211,563 (due to recognizing the
)
End of Q2: 576,321 a gain of 39,856 ($980 per car, not sure if that is a valid metric)
In thousands:
End of 2017: 1,177,799
Adjustment: (429,771)
Start of 2018: 748,028
End of Q1: 536,465 a drop of 211,563 (due to recognizing the
On March 15, we released a significant Autopilot update, which has been well received by our customers. Also, our mapping architecture has been upgraded and establishes a key platform to enable safer driving and the transition towards full autonomy. The latest mapping software in our cars is dramatically simpler and faster, providing a better user experience and superior performance.
End of Q2: 576,321 a gain of 39,856 ($980 per car, not sure if that is a valid metric)
Hi, everybody. Here are the links to the Q3 2018 Tesla Production and Delivery Estimates Competition.
Google form to enter and edit your estimates. There are 5 categories. You can enter as many as you want. Open the same form again until the end of 30th, Pacific Time, to edit your data.
Google sheet to view the estimates.
The 5 categories are as follows:
In Q2, @yak-55 had the most accurate overall estimates with 95.9% average in 5 categories followed by @Bokonon at 92.6% average. The most accurate single estimate was @Bokonon's Model 3 production estimate at 99.9%. Actual production was 28,578 units, the estimate was 28,600.
Google form to enter and edit your estimates. There are 5 categories. You can enter as many as you want. Open the same form again until the end of 30th, Pacific Time, to edit your data.
Google sheet to view the estimates.
The 5 categories are as follows:
- Model S + Model X production in Q3 2018
- Model 3 production in Q3 2018
- Global Model S deliveries in Q3 2018
- Global Model X deliveries in Q3 2018
- Global Model 3 deliveries in Q3 2018
In Q2, @yak-55 had the most accurate overall estimates with 95.9% average in 5 categories followed by @Bokonon at 92.6% average. The most accurate single estimate was @Bokonon's Model 3 production estimate at 99.9%. Actual production was 28,578 units, the estimate was 28,600.
I think that assumptions of Q3 deliveries in models discussed so far are too low based on the latest available production information from Electrek.
Based on very reasonable assumptions below the deliveries can hit 58k of M3 and 29k of MS/MX.
- assume that Tesla can deliver all production reported on through 10/23/2018 (Electrek reported on 10/21/2018 + 2 days)
- assume that all Q2 in transit to customers cars are delivered in Q3
- assume that inventory at the end of Q3 is equal to same at the end of Q2. This is very conservative, so could provide improved fidelity for projected deliveries, or an upside for more enthusiastic observers
Calculations:
Based on very reasonable assumptions below the deliveries can hit 58k of M3 and 29k of MS/MX.
- assume that Tesla can deliver all production reported on through 10/23/2018 (Electrek reported on 10/21/2018 + 2 days)
- assume that all Q2 in transit to customers cars are delivered in Q3
- assume that inventory at the end of Q3 is equal to same at the end of Q2. This is very conservative, so could provide improved fidelity for projected deliveries, or an upside for more enthusiastic observers
Calculations:
Fact Checking
Well-Known Member
Ok, that looks useful.
Here's their Q1 and Q2 progress report about 'autonomy' features completed:
- Q1: "We are also making progress with our self-driving technology. In March 2018, we released a significant Autopilot update and we upgraded our mapping architecture, which establishes a key platform to enable safer driving and the transition towards full autonomy. This latest mapping software is dramatically simpler and faster, providing a better user experience and superior performance."
- Q2: "Finally, we continue to make progress with our self-driving technology. Our neural net and functionality continue to improve, and we frequently release minor software updates and from time to time release significant version updates, including most recently in March 2018. While we are subject to regulatory constraints over which we have no control, our ultimate goal is to achieve full autonomy."
This also makes your $980 per car estimate relevant, as we can use it to back-calculate and estimate the true revenue recognition in Q1. The $211.5m drop you calculated is the net of revenue recognition versus new residual revenue deferred on new cars delivered.
If we assume that EAP and FSD take-rate is similar between S/X and 3 (which might not be valid but is the best we can do right now), then we can scale the $980 value with number of deliveries in Q1, i.e. the increase in deferred revenue was possibly around $980 * 29,980 = $29.4m.
So the original self-driving revenue recognition in Q1 was probably close to $211.5m+ $29.4m = $240.9m.
The original deferred revenue would have been $777.3m, and they recognized $240.9m from this, which is 30.99% of the full sum - a suspiciously round number. So they could have recognized either 30% or 33.33% of the residual revenue, applied over the whole fleet which received the software updates and still had deferred revenue outstanding.
Also note that while the March update was significant (it's when AP2 leaped ahead of AP1 in AutoPilot quality), and the map updates were nice, the upcoming V9 update is even more significant:
- On-ramp-off-ramp self-driving feature that does lane changes, highway transitions and exits automatically. This is very close to a functional self-driving feature, limited to certain geographical regions (U.S. and Canada, highways-only for the time being).
- Complete rework of the UI on every platform.
- Significant additions of functionality for the Model 3.
I.e. I think the level of revenue recognition per car should be at least as significant as the one in Q1 - but they could recognize a larger chunk as well.
So I think Q3 estimates need to take this into account.
Now let's try to estimate how much the Q3 revenue recognition could be:
- The lowest end of their guidance is 25k+50k = 75k deliveries.
- In Q2 the residual per unit revenue was $980, applied to 75k units this gives an increase in deferred revenue of $73.5m.
- The total deferred revenue would thus be $576.3+$73.5m (assuming it contains no changes from other sources of deferred revenue), which is $649.8m.
- 30% of that is $194.9m, 50% of it is $324.9m.
- This would all be recognized with very little or no cost of revenue, as these are near-100% margin software features. I.e. net income and profits would be increased by this amount directly. This alone could account for an approximate EPS contribution of +$1.10-$1.90, a significant BFD I'd say ...
- Every +1k beat in Q3 deliveries would improve deferred revenue by +$1m. So if they do say 80k deliveries or 85k deliveries in Q3 then that's +$5m-$10m in deferred revenue.
- I think a good argument could be made for higher than 50% level of deferred revenue recognition if they really wanted to use up this resource. It's not like customers are going to demand money back for EAP which is widely praised and fully functional, and the FSD feature's contribution to deferred revenue is probably still a minority part and a lot of AutoPilot work they implemented was already working towards full self-driving - so Tesla probably has wide discretion to recognize a big chunk of this deferred revenue. 66.67% would probably be stretching it but it's possibly still defensible.
- Some show-stopper bug in V9 that delays the fleet-wide rollout of the V9 software by the end of September. They are certainly cutting it close ...
- Profit might already be healthy even without deferred revenue and Tesla might want to carry over more deferred revenue to Q4 or to Q1/2019.
- Since deferred revenue recognition is highly discretionary in this case @luvb2b might be uneasy to incorporate any of it in his estimates, as this is an unknowable depending on the state of mind of the board.
@Fact Checking
Nice mathing!
FSD is a separate option and there are no FSD specific features at this point, so I'd say it is still 100% deferred. For EAP, they could probably adjust based on roll out percentage, if they wanted. Beyond that EAP is pretty much complete (pending user feedback).
If they go 100% complete on EAP, deferred will drop by $980 per car previously sold with EAP, and gross profit in each Q3 car with EAP will go up by $980. (I think the the $980 (20%) is actually low based on feature difficulty). Q3 EAP sales would not improve the deferred line, nor add to it.
Nice mathing!
FSD is a separate option and there are no FSD specific features at this point, so I'd say it is still 100% deferred. For EAP, they could probably adjust based on roll out percentage, if they wanted. Beyond that EAP is pretty much complete (pending user feedback).
If they go 100% complete on EAP, deferred will drop by $980 per car previously sold with EAP, and gross profit in each Q3 car with EAP will go up by $980. (I think the the $980 (20%) is actually low based on feature difficulty). Q3 EAP sales would not improve the deferred line, nor add to it.
Zhelko Dimic
Careful bull
Delivere everything produced by September 23rd?I think that assumptions of Q3 deliveries in models discussed so far are too low based on the latest available production information from Electrek.
Based on very reasonable assumptions below the deliveries can hit 58k of M3 and 29k of MS/MX.
- assume that Tesla can deliver all production reported on through 10/23/2018 (Electrek reported on 10/21/2018 + 2 days)
- assume that all Q2 in transit to customers cars are delivered in Q3
- assume that inventory at the end of Q3 is equal to same at the end of Q2. This is very conservative, so could provide improved fidelity for projected deliveries, or an upside for more enthusiastic observers
Calculations:
View attachment 338484
That is NOT conservative, it's wildly optimistic.
"wildly optimistic" - based on what?
My assumption is based on the past record and is consistent with the anecdotal reports of deliveries happening within 1,2, 3 days after completion of the production. For example, in-transit MS/MX at the end of Q4 2017 were 2,520. My calculation implies 2,287 for Q3.
As far as revenue recognition is concerned, I interpret wording from the 10-K (p. 43) shown below as EAP revenue being recognized over 8-year period following the sale of a car.
http://ir.tesla.com/static-files/0fbefe56-326c-412e-a33c-aa1b342e9469
Automotive revenue includes revenues related to deliveries of new vehicles, sales of regulatory credits to other automotive manufacturers and specific other elements that meet the definition of a deliverable under multiple-element accounting guidance, including free internet connectivity, free access to our Supercharger network and future free over-the-air software updates. These other elements are valued on a stand-alone basis, and we recognize their revenue over our performance period, which is generally the eight-year life of the vehicle, except for internet connectivity, which is over the free four-year period. If we sell a deliverable separately, we use that pricing to determine its fair value; otherwise, we use our best estimated selling price by considering costs used to develop and deliver the service, third-party pricing of similar options and other information that may be available. At the time of revenue recognition, we record a reserve against revenue for estimated future product returns. Such estimates are based on historical experience and were immaterial in all periods presented. In addition, any fees that are paid or payable by us to a customer’s lender, when we arrange the financing, would be recognized as an offset against automotive sales revenue, in accordance with ASC 605-50, Customer Payments and Incentives.
http://ir.tesla.com/static-files/0fbefe56-326c-412e-a33c-aa1b342e9469
Automotive revenue includes revenues related to deliveries of new vehicles, sales of regulatory credits to other automotive manufacturers and specific other elements that meet the definition of a deliverable under multiple-element accounting guidance, including free internet connectivity, free access to our Supercharger network and future free over-the-air software updates. These other elements are valued on a stand-alone basis, and we recognize their revenue over our performance period, which is generally the eight-year life of the vehicle, except for internet connectivity, which is over the free four-year period. If we sell a deliverable separately, we use that pricing to determine its fair value; otherwise, we use our best estimated selling price by considering costs used to develop and deliver the service, third-party pricing of similar options and other information that may be available. At the time of revenue recognition, we record a reserve against revenue for estimated future product returns. Such estimates are based on historical experience and were immaterial in all periods presented. In addition, any fees that are paid or payable by us to a customer’s lender, when we arrange the financing, would be recognized as an offset against automotive sales revenue, in accordance with ASC 605-50, Customer Payments and Incentives.
Fact Checking
Well-Known Member
I don't think this part is about EAP and FSD revenue recognition. These are fixed type 'software+hardware+updates services' that apply to every car.
EAP and FSD are separate options, with specific pricing, and separate 'deliverables'. Deferred revenue is generated if a feature is sold that Tesla doesn't declare "fully finished" yet. Say initial sales of FSD were probably deferred to a level of 100%, and HW 2.0 EAP sales were possibly only recognized to a level of 50% until this March, when the recognition rate increased to 75% and 25% was deferred.
It doesn't make sense to defer EAP revenue for 8 years, as the R&D cost was already paid, and you want to recognize it when the feature is done. Updating it is a service.
(Also a reply to @mongo: )
Also, I think parts of FSD revenue were also recognized already, if you read the language I cited - it's talking about meeting a milestone towards 'full autonomy':
- Q1: "We are also making progress with our self-driving technology. In March 2018, we released a significant Autopilot update and we upgraded our mapping architecture, which establishes a key platform to enable safer driving and the transition towards full autonomy. This latest mapping software is dramatically simpler and faster, providing a better user experience and superior performance."
- Q2: "Finally, we continue to make progress with our self-driving technology. Our neural net and functionality continue to improve, and we frequently release minor software updates and from time to time release significant version updates, including most recently in March 2018. While we are subject to regulatory constraints over which we have no control, our ultimate goal is to achieve full autonomy."
This is going to be even more the case with V9 in Q3, which is very close to a fully autonomous self-driving vehicle, on highways: on a good day it can drive hundreds of miles without the driver having to intervene - which is exactly what the FSD option promises.
So I fully expect Tesla to recognize not just EAP but also years of FSD revenue with a certain percentage (say 50%), and if they don't it's not because they couldn't, but because they chose to defer that to Q4 or Q1/2019.
Last edited:
You took the words out of my keyboard regarding EAP/FSD being separately priced features and thus not part of the 8 year wording.
However, I'm going to disagree with you on the FSD recognition. Yes, development is being made toward FSD, but there are zero features a customer gets at this point by purchasing FSD that they don't get with EAP. Regardless of how much EAP progress is getting us closer to FSD, there is no value added to the vehicle for having FSD, thus I say they cannot claim any FSD revenue.
Fact Checking
Well-Known Member
Maybe it doesn't matter, if for example the historic FSD take rate was 1% then for a fleet of 200,000 this would add up to 6 million dollars of deferred revenue - not worth worrying about.
But I think there's value: at minimum FSD option owners are going to get the Tesla AI Chip hardware upgrade for free, and they also got $2,000 off the post-delivery FSD price. You are right that there's no immediate utility though - but is that a requirement to recognize revenue? Are there any mandatory guidelines regarding revenue recognition for software features?
Sure, there is benefit to pre-order of FSD/EAP, not arguing that. FSD will get HW3, but that is a prerequisite for FSD software and not needed for EAP, so not really a bonus. Once HW3 exists, all cars will be built with it by default.
If there is no utility requirement, then this discussion is fruitless since Tesla could arbitrarily claim any percentage of EAP/FSD sales. I think it has to be separate in the event they never succeed, since that would require a refund of the monies.
Fact Checking
Well-Known Member
I'm just guessing here - but the GAAP requirements don't seem to limit deliverables to features with 'utility':
So it appears to me that there's considerable flexibility in recognizing such revenue, assuming customers are unlikely to request a refund - which appears to be the case here.
In the unlikely case of the feature not being fully done in the end it could still be categorized as an extraordinary event, which it would be.
But these are just guesses ...
(out of my depth here, so feel free to ignore)
Is that for a feature that, when implemented, has no real utility? Whereas FSD and EAP have specified feature sets to judge against?
Or is this Tesla sandbagging?
ValueAnalyst
Closed
Tesla management has been conservative in recognizing Autopilot revenue; I'm not sure if this will change.
Also, they are cutting really, really close to quarter-end with the wide release...
The financial potential on FSD is not in deferred revenue, since most owners have not yet paid for it, so it's not on balance sheet. Since Tesla has been so mum about the FSD feature rollout, I wouldn't expect much impact from FSD cash inflows on the income statement.
Major Autopilot revenue recognition seems to be a 4Q18/2019 story, in my opinion.
The excerpt I quoted addresses, among other things, features "deliverable separately", having specific pricing and delivered via future free OTA software updates. As it was included in 2017 10-K, it covers specific items revenue for which was deferred in the covered period. What OTHER than EAP features you think this description might cover?
It totally makes sense to defer the associated revenue as R&D cost that was incurred DOES NOT cover all of the EAP features, and additional R&D will be required in future to develop additional features of EAP, as explicitly stated by Tesla.
Last edited:
@ValueAnalyst How can I put "Disagree" reaction to the "Funny" you marked my post with and "Funny" to your "Disagree" above? TMC needs to improve their platform.
Last edited:
