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

Near-future quarterly financial projections

This site may earn commission on affiliate links.
IMHO you are underestimating sales of S and X in Q3, and therefore in Q4-19.
Perhaps add 25% for both the main automotive sales season and the benefit of the Raven refresh, so an extra 4K units each quarter.
The letter makes it clear, all they want to do with S/X is to reduce inventory levels. In the call Musk specifically said they are not sure why the orders for S/X has fallen. They are not going to increase production until the orders materialize. I don't think they are seeing increased orders after refresh - otherwise Musk would have said so. 18k in Q3 I think is already on the optimistic side.

Your production estimate for Model 3 in Q4-19 doesn't appear to take account of the China factory coming online,
Yes, I'm assuming volume production only in Q1 '20 from China. Any sales from there in Q4 would be an upside. Good thing is we'll know when the GF3 cars actually start getting delivered, so we can adjust the forecast.

Of course Q1-20 is the big challenge. Since China factory just came online with a lot of immediate non-seasonal demand, you can assume a full contribution of (say) 30K vehicles from it, both production and sales.
Elsewhere in the Northern Hemisphere, Jan-Feb is a really lousy time to be selling cars (and March isn't great), so Tesla should have a mild end-quarter inventory build up, and some drop-off in sales (from Q4-19). So perhaps take Q1-19 sales and add 30K China Model 3 to the mix, and call that a forecast?
I've not thought about Q1 much - but, hopefully GF3 sales can offset reduced sales due to seasonality in US and we can get 90k+ sales. But, because GF3 is online, you start seeing that depreciation etc start hitting the p&l. The margin might be bad in the beginning. We'll probably see some loss in Q1 '20.
 
A few small notes from the 10-Q:

Deferred auto revenue (non lease)
It looks like $76m of the Dec-18 deferred auto revenue balance was booked in Q2 vs $37m in Q1. I think the Q1 number was mostly related to non FSD deferred revenue (superchargers, internet, OTA updates etc), so i'd guess the $40m QoQ growth in Q2 is booked from FSD deferred revenue due to the wider NoA rollout. This would have been a boost to auto gross margin.

The overall deferred auto revenue balance increased $150m QoQ despite the $76m booked from the Dec-18 balance - so likely around $230m gross new deferred auto revenue generated in the quarter (or $2.6k per new car sales after excluding leased cars).

The 12 month expected revenue recognition balance increased $105m QoQ to $567m.
If we assume $160m annually for non FSD deferred revenue recognition, I guess around $400m deferred FSD revenue is expected to be booked in the next 12 months. This $400m is likely mostly going to be booked in Q3 and Q4 upon Enhanced Summon, Stop light, Traffic Light and City driving autopilot release.​

Credit revenue
Tesla still have $140m of deferred unbooked credit revenue. I think this is the Q1 FCA payment. My guess is they haven’t booked any credit revenue in Europe yet. I think the Q1 credit revenue boost may have been due to a one off sale of US credits to GM. Hard to know for sure though.
Maxwell
This is booked in the Auto segment and included c.6 weeks of Maxwell numbers. Maxwell 1Q revenue was $15m and COGs $17m so this will be a drag to Auto gross margin.
Another adjustment for auto buyback option after the price cuts.
Q2 Auto segment reduction was: revenue $64m, COGs by $50m and Gross profit by $15m. This wasn't part of the $117m exceptionals disclosed by Tesla in the shareholder letter.
Restructuring expense breakdown
Silevo writedown - $47m intangible asset + $15m equipment
$49m facility closure costs
$6m employee termination
Warranty costs
Tesla incurred $67m warranty costs in Q2 (inc leased cars). This is just $100 per car per quarter so looks under good control. Warranty provision was an average $1.7k per new car (after excluding cars sold on lease) - Model S/X will be higher and Model 3 will be lower.
Customer deposits
Reduced $137m QoQ. I expect this was likely a one off impact as Tesla cleared out the Model 3 reservation book. There's a better chance Tesla will start growing the balance again from here.
Liquidity
In addition to the $5bn cash balance, Tesla has $1bn of credit lines available and $0.8bn of warehouse lines (to fund future auto leases)
Capex guidance
"Capital expenditures in 2019 are projected to be approximately $1.5 to 2.0 billion, to continue to develop our main projects as planned including further capacity expansion and automation for our current Model 3 production, Gigafactory Shanghai, Model Y and Tesla Semi, as well as to further expand our Supercharger and vehicle service and repair networks. Given the breadth of our various planned projects and our focus on cost efficiency, as we make progress on such projects we may find that our actual spend may be less than previously expected."​
 
Last edited:
I believe Tesla has under provisioned for warranty. One full set of screen recalls exceeds the whole reserve. That said, I also think doubling the warranty reserve does not really hurt too much in the long run.

How would screen recalls cost $941m? Often recall costs will be covered by a supplier if the part was faulty.
 
Last edited:
  • Helpful
Reactions: immunogold
i need some help. probably @brian45011 , @neroden , @ReflexFunds, or @Fact Checking could help.

first refer to stock-based compensation table in recent 10q, top of page 31. Inline XBRL Viewer

now refer to stock-based compensation table in prior year q2 10q, top of page 28. tsla-10q_20180630.htm

there is a shift in the allocation of sbc from cost of revenues into sg&a. for 18q2, the recent 10q shows 21,661. in the original 10q we had 16,081. the total amount is the same at 197,344 because the reduction is taken from sbc in sg&a.

my questions:
a) so assuming that some expense was shifted from sg&a up to cost of revenues, why don't the historical financials reflect that change in the 19q2 filing?

b) how can the categorization change historically from cogs to sg&a?
 
warning: extremely boring post about deferred revenue. drink coffee first.
tl;dr: much of tesla's claimed increase in automotive gross margin seems to be from changes in one-time accounting adjustments and accounting (deferral & recognition) of deferred revenue.

A few small notes from the 10-Q:
Deferred auto revenue (non lease)

It looks like $76m of the Dec-18 deferred auto revenue balance was booked in Q2 vs $37m in Q1. I think the Q1 number was mostly related to non FSD deferred revenue (superchargers, internet, OTA updates etc), so i'd guess the $40m QoQ growth in Q2 is booked from FSD deferred revenue due to the wider NoA rollout. This would have been a boost to auto gross margin.

The overall deferred auto revenue balance increased $150m QoQ despite the $76m booked from the Dec-18 balance - so likely around $230m gross new deferred auto revenue generated in the quarter (or $2.6k per new car sales after excluding leased cars).


almost, but not quite. they changed the way the disclosures are made, the 76.1m deferred revenue you estimate them booking is of the balance that was deferred as of dec 31 2018. from the 10q:

Revenue recognized from the deferred revenue balance as of December 31, 2018 was $113.5 million for the six months ended June 30, 2019. From the deferred revenue balance as of January 1, 2018, revenue recognized during the six months ended June 30, 2018 was $44.5 million.


if you go back and review the 2018 disclosures which are much better, you see the total deferred revenue recognized in the first 6 months was $49.4m vs the $44.5m which is just the portion of the deferred revenue from the jan 1 2018 balance that was recognized.

similarly we have to make an adjustment for 19q2 to estimate the amount of this deferred revenue recognized. the method i chose is to take:

(amount recognized 19q2/balance end of 18q4)*(deferred revenue eop 19q1 - deferred revenue eop 18q4) + ($76.1m amount disclosed as booked out of 18q4 balance in 19q2).

note the 76.1m comes from taking 113.5m (6mos ended 19q2 recognition) - 37.4m (19q1 recognition).

the actual amount of deferred revenue recognized was thus ~90m. but to some extent this is also supercharger, connectivity, and software updates which are normal. so what we're really interested in is the change vs baselines. i'm trying to understand all of that but my suspicion is that another huge chunk of the claimed 200bps gross margin increase stems from the deferral & recognition of deferred revenue. if some one can put together cogent thoughts on that please go ahead. here's a table of data you can work with.

here are a couple thoughts as i edit the post (refer to my table below).
1. 2019q1 deferred revenue per vehicle sold is unusually high due to promotions for autopilot and fsd during the quarter, which would result in additional deferral of revenue not related to unit sales of autos.
2. the 2019q2 deferred revenue per vehicle sold picture is confounded by changing the definition and pricing of autopilot, and also a free supercharging promotion to clear inventory cars.
3. it sure likes they were using deferred revenue as a slush fund to me. for example, 18q3 had a lower mix of vehicles with free supercharging. free internet connectivity was reduced for 18q3 as well. i expect as more model 3 were in the mix that the eap/fsd take rate went down from 18q3 to 18q2. and yet the table shows deferred revenue per vehicle increased in q3 to the highest level seen in 2018. more deferred revenue means less current recognition and more to pull in future periods. as 18q3 was their best quarter, this is where one would expect them to shove the most into the cookie jar.
4. if we assume the 18q4/19q1 recognition are more representative of the baseline amounts that should be recognized per quarter, then we could ballpark ~50m excess deferred revenue was recognized in 19q2 (~90m - ~40m). that would be a 1% boost to auto gross margin.

Screenshot 2019-07-29 13.48.31.png
 
Last edited:
ok after some further work, i think netting out all the one time effects i can think of, the auto gross margin did improve about 2%. whew that was hard to see.

basically i worked it out like this:
19q1 auto revenue - credits - one time accounting adjustments - one time autopilot/fsd recognition on eap/fsd promotion =
3,723,861 - 215,981 - (-500,500) - 50,000 = 3,958,380
19q1 auto cogs - one time accounting adjustment = 3,382,101
adjusted gross margin ~14.5%

19q2 auto revenue - credits - one time accounting adjustments - excess pull forward of deferred revenue =
5,376,389 - 111,219 - (-64,100) - 50,000 = 5,279,270
19q2 auto cogs - one time accounting adjustment = 4,409,685
adjusted gross margin ~16.5%
 
Roughly 500,000 cars on the road. 941M / 500,000 cars == <$2000/screen, whch isn't enough. Do the math.
Roughly 500,000 cars on the road. 941M / 500,000 cars == <$2000/screen, whch isn't enough. Do the math.
And what makes you think all 500k of those cars need a screen replacement? I'm guessing that most of the screen problems have occurred in the last year or two. Is this only a Model 3 problem, or has anyone with an S or an X had the problem? Furthermore you didn't indicate why you think it's all Tesla's responsibility.
 
  • Disagree
Reactions: neroden
Roughly 500,000 cars on the road. 941M / 500,000 cars == <$2000/screen, whch isn't enough. Do the math.
More like <450k S/X total (sept 2018, 106k X, 250k S), some of which are out of warranty, some of which have a newer version of the screen.
No requirement for new units, they can pull, replace the front glass and display and reinstall. With a contract house on the back end, it would be a mobile ranger swap. In a more ideal world, simultaneously with an AP3 retrofit.
 
And what makes you think all 500k of those cars need a screen replacement? I'm guessing that most of the screen problems have occurred in the last year or two. Is this only a Model 3 problem, or has anyone with an S or an X had the problem? Furthermore you didn't indicate why you think it's all Tesla's responsibility.

READ WHAT I SAID. I said "if there's a comprehensive screen replacement needed under warranty, the warranty reserve won't cover it." This has already been required for Model X and Model S (and they're trying to deny valid warranty claims, which is very uncool -- they're losing arbitrations), not yet for Model 3, but it's a definite possiblity.
Same thing if there's a drivetrain replacement needed (which happened on Model S).
Same thing if there's a four-door-handle replacement needed (which happened on Model S).

I don't expect all of these things to happen, but the odds are that SOME thing is going to need a comprehensive replacement. Warranty reserve is just too low at 3% of auto revenues. I'd be comfortable with 5%.
 
  • Informative
Reactions: hobbes
the 2019q2 deferred revenue per vehicle sold picture is confounded by changing the definition and pricing of autopilot,
the 2019q2 deferred revenue per vehicle sold picture is confounded by changing the definition and pricing of autopilot,
Since they moved around items in AP/FSD (and made the FSD definition simpler) - what happens to all the old deferred revenue ? Does that follow new definition or does it have to follow the definition under which it was received ?

I'm trying to figure out whether - the increase in revenue recognition in Q2 can be attributed to clear and simpler definition of AP & FSD in Q2. AP can be completely recognized and FSD can be now partially recognized. Also, whether they can recognize the trend will continue in Q3.
 
Here is the comparison between my cash flow estimate and the actual. Many individual items are quite different and I've to dig in to figure out why.

- Depreciation is more by $100M ( FYI : @Doggydogworld )
- Accounts receivables actually increased i.e. they must have delivered even more cars in the final days of Q2 compared to Q1.
- Accounts payables went up only by $77M, not $165M. (FYI : @Fact Checking)
- Inventory went down $100M less than estimated. This is because raw materials, WIP and service parts went up.
- There is a sudden increase by ~$190M of "Operating Lease vehicles" (Model 3 leases ?)
- Customer deposits down significantly by ~135M
- Ofcourse, they spent only $250M on capex, not $500M.

TSLAQ2CashFlowComp.png
 
More like <450k S/X total (sept 2018, 106k X, 250k S), some of which are out of warranty, some of which have a newer version of the screen.
No requirement for new units, they can pull, replace the front glass and display and reinstall. With a contract house on the back end, it would be a mobile ranger swap. In a more ideal world, simultaneously with an AP3 retrofit.
I thought the UV fix works and a screen swap is not needed?
 
  • Like
Reactions: Esme Es Mejor
Since they moved around items in AP/FSD (and made the FSD definition simpler) - what happens to all the old deferred revenue ? Does that follow new definition or does it have to follow the definition under which it was received ?
Revenue deferred under old FSD can only be recognized as they deliver old FSD features. Same with EAP. Otherwise it'd be a huge loophole - just redefine FSD to be "can follow lanes on the highway" and recognize all deferred revs.

As for cash flows, PP&E depreciation went from 299 to 334m. The rest of the jump is probably impairment (Silevo, inventory, etc.). I was surprised AR went up that much, I expected flattish. The impact of operating lease vehicles on the cash flow statement is a nightmare, but Model 3 leasing did and will continue to hurt OCF/FCF. It's mostly offset by financing cash inflow, though.
 
  • Informative
Reactions: Fact Checking
warning: extremely boring post about deferred revenue. drink coffee first.
tl;dr: much of tesla's claimed increase in automotive gross margin seems to be from changes in one-time accounting adjustments and accounting (deferral & recognition) of deferred revenue.




almost, but not quite. they changed the way the disclosures are made, the 76.1m deferred revenue you estimate them booking is of the balance that was deferred as of dec 31 2018. from the 10q:

Revenue recognized from the deferred revenue balance as of December 31, 2018 was $113.5 million for the six months ended June 30, 2019. From the deferred revenue balance as of January 1, 2018, revenue recognized during the six months ended June 30, 2018 was $44.5 million.


if you go back and review the 2018 disclosures which are much better, you see the total deferred revenue recognized in the first 6 months was $49.4m vs the $44.5m which is just the portion of the deferred revenue from the jan 1 2018 balance that was recognized.

similarly we have to make an adjustment for 19q2 to estimate the amount of this deferred revenue recognized. the method i chose is to take:

(amount recognized 19q2/balance end of 18q4)*(deferred revenue eop 19q1 - deferred revenue eop 18q4) + ($76.1m amount disclosed as booked out of 18q4 balance in 19q2).

note the 76.1m comes from taking 113.5m (6mos ended 19q2 recognition) - 37.4m (19q1 recognition).

the actual amount of deferred revenue recognized was thus ~90m. but to some extent this is also supercharger, connectivity, and software updates which are normal. so what we're really interested in is the change vs baselines. i'm trying to understand all of that but my suspicion is that another huge chunk of the claimed 200bps gross margin increase stems from the deferral & recognition of deferred revenue. if some one can put together cogent thoughts on that please go ahead. here's a table of data you can work with.

here are a couple thoughts as i edit the post (refer to my table below).
1. 2019q1 deferred revenue per vehicle sold is unusually high due to promotions for autopilot and fsd during the quarter, which would result in additional deferral of revenue not related to unit sales of autos.
2. the 2019q2 deferred revenue per vehicle sold picture is confounded by changing the definition and pricing of autopilot, and also a free supercharging promotion to clear inventory cars.
3. it sure likes they were using deferred revenue as a slush fund to me. for example, 18q3 had a lower mix of vehicles with free supercharging. free internet connectivity was reduced for 18q3 as well. i expect as more model 3 were in the mix that the eap/fsd take rate went down from 18q3 to 18q2. and yet the table shows deferred revenue per vehicle increased in q3 to the highest level seen in 2018. more deferred revenue means less current recognition and more to pull in future periods. as 18q3 was their best quarter, this is where one would expect them to shove the most into the cookie jar.
4. if we assume the 18q4/19q1 recognition are more representative of the baseline amounts that should be recognized per quarter, then we could ballpark ~50m excess deferred revenue was recognized in 19q2 (~90m - ~40m). that would be a 1% boost to auto gross margin.

View attachment 435361

I largely agree with this, but a few points:
  • In 3Q18 Tesla still had unlimited supercharging referral bonus and also had FSD on sale. Both of these were removed in early Q4 so I think this is the main reason for the step down in Q4. I don't think Tesla has flexibility to just randomly defer revenue, it has to be tied to specific features, costs, milestones etc.
  • Your deferred revenue per vehicle should be per cash sale vehicle - leased vehicles do not create upfront deferred revenue
  • The recognised per vehicle should be divided by the total customer fleet size (excluding leased fleet), the number of new car sales in a quarter isn't really relevant to this metric. The greater fleet size likely explained c.$5m of the step up in Q2.
 
Last edited:
Your deferred revenue per vehicle should be per cash sale vehicle - leased vehicles do not create upfront deferred revenue
Since anyone can buy FSD at any time, even this is suspicious. I'm not even sure most of the people who buy FSD, buy when they buy the car. It is one of those things that can be easily put off - so as to not increase the initial cost - and bought later.
 
I largely agree with this, but a few points:
  • In 3Q18 Tesla still had unlimited supercharging referral bonus and also had FSD on sale. Both of these were removed in early Q4 so I think this is the main reason for the step down in Q4. I don't think Tesla has flexibility to just randomly defer revenue, it has to be tied to specific features, costs, milestones etc.
in the best of all possible worlds i agree. however there is some flex around exactly how milestones are defined, or how tight the performance of the features must be to meet the definitions, or even how to time feature releases. i've seen other companies use these 'opportunities' and wouldn't put it out of the realm of possibility.

Your deferred revenue per vehicle should be per cash sale vehicle - leased vehicles do not create upfront deferred revenue

thanks did not know that, but it makes perfect sense.

The recognised per vehicle should be divided by the total customer fleet size (excluding leased fleet), the number of new car sales in a quarter isn't really relevant to this metric. The greater fleet size likely explained c.$5m of the step up in Q2.

actually the recognized per vehicle i know is not meaningful of how revenue is recognized vs the fleet. i was instead trying to get an impact to how it would make this period's asp look - effectively how much extra revenue per vehicle being sold this period we would see from recognition of deferred revenue.

here's an updated table. the adjustments for leased vehicles advised by @ReflexFunds smooth out the deferred revenue calculations nicely. speaking of which, @ReflexFunds do you know how margins might be estimated on the deferred revenue that's recognized? thanks!
Screenshot 2019-07-30 09.34.20.png