Near-future quarterly financial projections

[ReflexFunds]

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.



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



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!
View attachment 435692

Thanks for doing the work, interesting table.

It's very hard to know how they allocate costs to the auto deferred revenue, I would guess:

FSD: Possibly they accrue a reserve within COGs for computer replacement costs later this year, maybe c.$1k per FSD sale. Some of these costs would be deferred together with the deferred revenue, some would be booked upfront in proportion to their revenue recognition. I would guess the actual software costs for FSD development are all in R&D and SG&A, but its possible something gets allocated to COGs.

Supercharging: They have 3 different types of revenue here and may account for them all differently: 1) Customer supercharging one off purchases (Not clear if this is included in Auto revenue or Service revenue - likely low gross margin revenue). 2) Customer supercharging from unlimited free supercharging offer (Tesla's deferred revenue should be at third-party equivalent pricing/gross margins, so likely low gross margin revenue, unless customers are using significantly less/more charging than the deferred revenue reserve - I think there's a good chance they used to under reserve for supercharging costs and this is now hitting at a negative gross margin) and 3) Customer supercharging from unlimited free supercharging as customer referral rewards (it is possible Tesla record these costs at SG&A sales costs in which case gross margin could be 100%, but equally the accounting could be the same as for 2).

Internet Connectivity: Again should be booked at third party equivalent pricing. Likely low gross margin.

OTA software update commitments: I'd guess the corresponding costs are within R&D and SG&A, so this could be 100% gross margin revenue when recognised.

​

This is Tesla's comment from the 10-Q

"Other features and services such as access to our Supercharger network, internet connectivity and over-the-air software updates are provisioned upon control transfer of a vehicle and recognized over time on a straight-line basis as we have a stand-ready obligation to deliver such services to the customer. We recognize revenue related to these other features and services over the performance period, which is generally the expected ownership life of the vehicle or the eight-year life of the vehicle. Revenue related to Autopilot and full self-driving features is recognized when functionality is delivered to the customer. For our obligations related to automotive sales, we estimate standalone 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."​

 

[EVNow]

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.

I don't see any impairments explained in 10-Q. So we have a $100M dollar mystery still.

 

[schonelucht]

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.
​

I think customer deposit movements are dominated by full payments for cars in transit that just didn't make the delivery by the end of the quarter. Hence this reduction is more a function of the reduction of cars in transit than a draw-dawn on the reservation book.

 

[Doggydogworld]

I don't see any impairments explained in 10-Q. So we have a $100M dollar mystery still.

During the second quarter of 2019, we recognized $47.0 million in impairment related to IPR&D intangible asset as we abandoned further development efforts (refer to Note 4, Goodwill and Intangible Assets for details) and $15.0 million for the related equipment within the energy generation and storage segment. We also incurred a loss of $48.8 million for closing operations in certain facilities. On the statement of cash flows, these amounts were presented in the captions in which such amounts would have been recorded absent the impairment charges.​

I don't really understand the last line. There is no R&D line on the cash flow statement to put the 47m impairment, for example. I'd expect all 111m of these to be in the Depreciation, Amortization and impairments line. The 35m inventory writedown, not mentioned here, has its own line item.

 

[EVNow]

I don't really understand the last line. There is no R&D line on the cash flow statement to put the 47m impairment, for example. I'd expect all 111m of these to be in the Depreciation, Amortization and impairments line. The 35m inventory writedown, not mentioned here, has its own line item.

Thanks for that. The pdf copy doesn't let me search properly, downloaded the rtf version (I usually just look at excel version).

Seems to me these items under Note 19 – Restructuring and Other that cost Tesla $117 on P&L are mostly non-cash and thus come under impairment in the cash flow.

Employee termination : 6.2 M (cash, not paid)
Intangible : $47 M (non-cash)
equipment : $15 M (non-cash)
closing ops : $48.8M (?)

So, atleast 62M is in impairment. Depending on whether the closing ops was cash or non-cash, that would also be in impairment (or may be a part that is non-cash).

 

[ReflexFunds]

Thanks for that. The pdf copy doesn't let me search properly, downloaded the rtf version (I usually just look at excel version).

Seems to me these items under Note 19 – Restructuring and Other that cost Tesla $117 on P&L are mostly non-cash and thus come under impairment in the cash flow.

Employee termination : 6.2 M (cash, not paid)
Intangible : $47 M (non-cash)
equipment : $15 M (non-cash)
closing ops : $48.8M (?)

So, atleast 62M is in impairment. Depending on whether the closing ops was cash or non-cash, that would also be in impairment (or may be a part that is non-cash).

I would guess the closing opps cost is a mix of capex impairments and operating lease penalties (could also be dark leases where Tesla provisions for continued monthly lease payments but does not pay penalty upfront)

 

[EVNow]

I think customer deposit movements are dominated by full payments for cars in transit that just didn't make the delivery by the end of the quarter. Hence this reduction is more a function of the reduction of cars in transit than a draw-dawn on the reservation book.

If this was the case, we should have seen a jump in Q1 because of high # of transit vehicles. But deposits actually went down $25M.

BTW, I remember Tesla taking full price deposits for orders from China. Do they ask for full big deposits from EU too ?

 

[schonelucht]

If this was the case, we should have seen a jump in Q1 because of high # of transit vehicles. But deposits actually went down $25M.

That's because Q1 was the very first quarter they started to deliver CN+EU Model 3's almost exclusively to reservation holders. There is no such big effect anymore in Q2 and follow up quarters.

 

[FrankSG]

I believe it was @EVNow who asked a week or two ago in the daily thread if I was planning on updating my model. I finally got around to it today:

I've now included leases and credits. Especially because I didn't account for leases nor heavy S+X inventory discounting in Q2 is why my estimates turned out to be quite wrong.

It seems like my expectations are close to luvb2b's now, but slightly pessimistic compared to EVNow's. I think EVNow's M3 ASP drop of just 1k per car is too optimistic. It might be sort of accurate if you just look at only the price decrease, but you have to keep in mind that during Q1 and Q2 overseas deliveries barely had any SR+ cars in them. SR+ mix dropped slightly from June to July on eu-evs.com, but I'd expect there to be a much higher mix of SR+ M3s going overseas in August and especially September.

I also think EVNow's prediction for 650M in SG&A in Q4 is too low. I think there's going to be a sizeable increase because of Giga 3 ramp up. Even if they're going to be more capital efficient, and even if workers in China are cheaper than the US, I can't imagine them not increasing SG&A at least somewhat from 650M / quarter.

 

[EVNow]

I believe it was @EVNow who asked a week or two ago in the daily thread if I was planning on updating my model. I finally got around to it today:

Yes, it was I - but the idea was to see what you come up with for Q2 before the ER

It seems like my expectations are close to luvb2b's now, but slightly pessimistic compared to EVNow's.

Actually my GAAP Loss is higher than your (by ~15M). You are probably comparing your GAAP EPS to my non-GAAP. You should add non-GAAP in your sheet, since that is what the analysts forecast and the market looks for.

I think EVNow's M3 ASP drop of just 1k per car is too optimistic. It might be sort of accurate if you just look at only the price decrease, but you have to keep in mind that during Q1 and Q2 overseas deliveries barely had any SR+ cars in them. SR+ mix dropped slightly from June to July on eu-evs.com, but I'd expect there to be a much higher mix of SR+ M3s going overseas in August and especially September.

Basic difference is in regulatory credits. My ASP includes credits - yours & b2b's don't. So you are assuming $50k ASP for M3 they talked about in Q2 is pre-credit. I assume it to be post credit. That is why you see the difference in s&x ASP we have calculated. Your ASP for M3 in Q3 is same as mine once you add the credit.

The reason I think the ASP will not be much lower is
- higher reg credit in Q3
- higher recognition of deferred FSD/AP revenue in Q3. I expect V10 to come out in Q3 and with that more of FSD revenue recognized. Also possibly higher take of FSD.

I also think EVNow's prediction for 650M in SG&A in Q4 is too low. I think there's going to be a sizeable increase because of Giga 3 ramp up. Even if they're going to be more capital efficient, and even if workers in China are cheaper than the US, I can't imagine them not increasing SG&A at least somewhat from 650M / quarter.

The trend of SG&A is downward/flat. They have been reducing their sales people (and adding service). I don't expect GF3 to disrupt that. There is a theory that this is because of reduction in SG&A from SolarCity masking increase in auto. Possible - but I'm not so sure. They may have added delivery folks to handle higher volume - but that might be under service/COGS ? I had estimated flat $705M of SG&A for Q2 (Q1 was $703M). Turns out they slashed it by $50M. I'm going to optimistically assume flat SG&A until we see evidence of increase in SG&A. In general corporations have a budget for SG&A and they manage within that. I find it difficult to believe that Tesla would increase their SG&A budget when they are talking about closing sales locations and moving sales online.

Most of GF3 labor should be accounted for in COGS once production starts. Not sure how they account for labor before production starts - do they account for it in SG&A and move to COGS once production starts ? @Doggydogworld

 

[generalenthu]

Most of GF3 labor should be accounted for in COGS once production starts. Not sure how they account for labor before production starts - do they account for it in SG&A and move to COGS once production starts ?

All construction and setup costs until production starts are capitalized and then depreciated after production starts.

 

[FrankSG]

Actually my GAAP Loss is higher than your (by ~15M). You are probably comparing your GAAP EPS to my non-GAAP. You should add non-GAAP in your sheet, since that is what the analysts forecast and the market looks for.

Yeah, we're actually extremely close for Q3, especially Revenues and Cost of revenues. I was mostly referring to Q4 where I'm a little more pessimistic, largely due to me expecting an increase in OPEX.

Basic difference is in regulatory credits. My ASP includes credits - yours & b2b's don't. So you are assuming $50k ASP for M3 they talked about in Q2 is pre-credit. I assume it to be post credit. That is why you see the difference in s&x ASP we have calculated. Your ASP for M3 in Q3 is same as mine once you add the credit.

The reason I think the ASP will not be much lower is
- higher reg credit in Q3
- higher recognition of deferred FSD/AP revenue in Q3. I expect V10 to come out in Q3 and with that more of FSD revenue recognized. Also possibly higher take of FSD.

Ah, my bad. I didn't realise this. Q2 credits were very low, so this looks pretty reasonable then yeah.

The trend of SG&A is downward/flat. They have been reducing their sales people (and adding service). I don't expect GF3 to disrupt that. There is a theory that this is because of reduction in SG&A from SolarCity masking increase in auto. Possible - but I'm not so sure. They may have added delivery folks to handle higher volume - but that might be under service/COGS ? I had estimated flat $705M of SG&A for Q2 (Q1 was $703M). Turns out they slashed it by $50M. I'm going to optimistically assume flat SG&A until we see evidence of increase in SG&A. In general corporations have a budget for SG&A and they manage within that. I find it difficult to believe that Tesla would increase their SG&A budget when they are talking about closing sales locations and moving sales online.

Most of GF3 labor should be accounted for in COGS once production starts. Not sure how they account for labor before production starts - do they account for it in SG&A and move to COGS once production starts ? @Doggydogworld

It'd be awesome if you're right, and I was surprised by reductions in Q2, but I just don't see this being possible.

People actually assembling cars, sure. That's going into COGS. But what about people installing equipment, tuning equipment, and improving equipment and general factory efficiency after production starts? That's either R&D or SG&A. Same for HR, finance people, managers, etc. and whatever other people they need in and around Giga 3.

Looking at M3 ramp, SG&A increased from $537,757 to $750,759 in 1 year. Like I said, I'm sure they'll be more capital efficient, won't have to add as much to service / delivery capacity, and labor in China is cheaper. But nonetheless, I can't imagine it staying flat in the coming 3 quarters.

 

[schonelucht]

BTW, I remember Tesla taking full price deposits for orders from China. Do they ask for full big deposits from EU too ?

No. But they ask for full pre-payment up front. Typical scenario could be a customer gets a delivery date near the end of the quarter, pays the full amount for the car a few days early by bank transfer as instructed and then something comes up that pushes the delivery date a few days into the new quarter. At that point the full payment of the car will show up in customer deposits for that quarter. That includes taxes and VAT so it's going to be a fully 120% + of the car value recorded. A few hunderd of these cases across Europe will quickly dominate the few remaining reservation holders going out.

 

[EVNow]

It'd be awesome if you're right, and I was surprised by reductions in Q2, but I just don't see this being possible.

People actually assembling cars, sure. That's going into COGS. But what about people installing equipment, tuning equipment, and improving equipment and general factory efficiency after production starts? That's either R&D or SG&A.

As @generalenthu noted above, all of this gets capitalized and added to cost of plant & machinery (i.e. goes to Capex). Then, it gets depreciated like other machinery & plant items over the useful life of the assets.

Same for HR, finance people, managers, etc. and whatever other people they need in and around Giga 3.

Some of this could be in SG&A - though if those people are purely connected with the factory they will be charged to COGS. Either way don't expect a lot of people like this. Even if there are a 100 people, and get paid $5k/quarter (equivalent to about $80k US salary) - it is just 0.5M per quarter.

Looking at M3 ramp, SG&A increased from $537,757 to $750,759 in 1 year. Like I said, I'm sure they'll be more capital efficient, won't have to add as much to service / delivery capacity, and labor in China is cheaper. But nonetheless, I can't imagine it staying flat in the coming 3 quarters.

I don't know why SG&A went up by $200M in 2017/18 - but its not because of GF1. May be they hired a lot of people in sales assuming a faster 3 rampup ? There was a big jump in Q2 '18 (from 686 to 750) - may be the 10Q for Q2 has something on it ? Tesla doesn't provide any breakdowns of SG&A, so difficult to tell.

ps : Looks like the increase in SG&A was primarily due to more office space expense, more headcount in '18. In '19, the big decrease was because of headcount reduction. So, it should stay that way. I don't expect them to be hiring a lot of salespeople anytime soon.

SG&A expenses decreased $103.5 million, or 14%, in the three months ended June 30, 2019 as compared to the three months ended June 30, 2018. The decrease was primarily due to an $91.0 million decrease in employee and labor related expenses from decreased headcount.​

 

[FrankSG]

As @generalenthu noted above, all of this gets capitalized and added to cost of plant & machinery (i.e. goes to Capex). Then, it gets depreciated like other machinery & plant items over the useful life of the assets.


Some of this could be in SG&A - though if those people are purely connected with the factory they will be charged to COGS. Either way don't expect a lot of people like this. Even if there are a 100 people, and get paid $5k/quarter (equivalent to about $80k US salary) - it is just 0.5M per quarter.



I don't know why SG&A went up by $200M in 2017/18 - but its not because of GF1. May be they hired a lot of people in sales assuming a faster 3 rampup ? There was a big jump in Q2 '18 (from 686 to 750) - may be the 10Q for Q2 has something on it ? Tesla doesn't provide any breakdowns of SG&A, so difficult to tell.

ps : Looks like the increase in SG&A was primarily due to more office space expense, more headcount in '18. In '19, the big decrease was because of headcount reduction. So, it should stay that way. I don't expect them to be hiring a lot of salespeople anytime soon.

SG&A expenses decreased $103.5 million, or 14%, in the three months ended June 30, 2019 as compared to the three months ended June 30, 2018. The decrease was primarily due to an $91.0 million decrease in employee and labor related expenses from decreased headcount.​

How about Model Y ramp. Where do you think SG&A will be in two years in mid 2021 after they're producing ~5k MY / week?

 

[EVNow]

How about Model Y ramp. Where do you think SG&A will be in two years in mid 2021 after they're producing ~5k MY / week?

I think it will go up by $100M to $200M - mainly from stock options. Model Y rampup should also produce capitalized expenses that are amortized/depreciated over the years.

BTW, when solar starts producing sizable amount of solar roofs, we have to see if the SG&A for solar goes up. Same with storage. But then, they will start contributing meaningful margin too, which should offset any increase in SG&A.

 

[FrankSG]

BTW, when solar starts producing sizable amount of solar roofs, we have to see if the SG&A for solar goes up. Same with storage. But then, they will start contributing meaningful margin too, which should offset any increase in SG&A.

Yeah that's fair.

I think it will go up by $100M to $200M - mainly from stock options. Model Y rampup should also produce capitalized expenses that are amortized/depreciated over the years.

Mainly from stock options? Does that mean you think actual salaries of workers will go into COGS, but expenses from stock options granted into SG&A? :S

And then what do you think is the difference between Giga 3 ramp and Model Y ramp that makes Giga 3 ramp not increase SG&A, but MY ramp increase SG&A by $100-200M?

 

[EVNow]

Yeah that's fair.



Mainly from stock options? Does that mean you think actual salaries of workers will go into COGS, but expenses from stock options granted into SG&A? :S

And then what do you think is the difference between Giga 3 ramp and Model Y ramp that makes Giga 3 ramp not increase SG&A, but MY ramp increase SG&A by $100-200M?

Salaries & options of production workers goes into COGS. But corporate stock options go into SG&A. SG&A has the largest stock based compensation increase in Q2 (compared to Q2 '18) of $45M. Cogs increased by $16M and R&D by $26M.

Also, if Musk's performance award actually happens, there will be a one time big hit. But for that SP should go over $400

SG&A expenses decreased $86.0 million, or 6%, in the six months ended June 30, 2019 as compared to the six months ended June 30, 2018. The decrease was primarily due to a $145.4 million decrease in employee and labor related expenses from decreased headcount, partially offset by a $47.0 million increase in stock-based compensation expense related to the 2018 CEO Performance Award and stock awards granted for new hires and refresher employee stock grants.​

 

[trentbridge]

I think the "services" loss estimated by FrankSG for Q3 is too high. There was a significant improvement in margins in Q2 and I don't see that reversing. My theory is that Tesla sold a boatload of Model 3s in Q4 by taking in used Model S at a generous trade-in value..and then Tesla dropped the price of the new Model S in January and again in February - reducing the value of these used vehicles...I think Q1 saw the brunt of lousy profits from selling this devalued inventory:

• Q1 Shareholder Update
  This increase (in loses) is primarily attributed to used car inventory revaluation and reduction of ASP relating to new car pricing actions.
  

"and the smoke cleared in Q2 so I'm more optimistic about "services" in Q3.

 

[FrankSG]

I think the "services" loss estimated by FrankSG for Q3 is too high. There was a significant improvement in margins in Q2 and I don't see that reversing. My theory is that Tesla sold a boatload of Model 3s in Q4 by taking in used Model S at a generous trade-in value..and then Tesla dropped the price of the new Model S in January and again in February - reducing the value of these used vehicles...I think Q1 saw the brunt of lousy profits from selling this devalued inventory:

• Q1 Shareholder Update
  This increase (in loses) is primarily attributed to used car inventory revaluation and reduction of ASP relating to new car pricing actions.

"and the smoke cleared in Q2 so I'm more optimistic about "services" in Q3.

You could be right. Improved service margins from -40% to -25%, was definitely one of the highlights for me in Q2 earnings, my models had Tesla losing almost a billion $ this year on service before I adjusted them to what they are now.

I need to see another quarter of evidence before I further improve service margins in my models in the near term though. I guess with this model specifically I like to err slightly on the side of being conservative.