Near-future quarterly financial projections

[Fact Checking]

BTW, thanks for doing the math on that. I hadn't done it recently. When I make a 2025 or 2030 valuation for TSLA, I have to assume 3.5% dilution per year. If you think about it, this makes for a significant hurdle to valuation; it impairs the per-share value quite substantially in the bull case. (In order to keep earnings per share steady, they have to grow total earnings by 3.5% every year.)

To be fair, I expected Tesla to stop equity financing after Q3-Q4 last year - boy was I wrong about that.

Nevertheless I do think this might have been the final equity financing round. Once Tesla is self-financing they might start the Amazon/Apple route that might within 5 years even store excess profits tax-free by buying back stock - i.e. I'm quite sure the 3.5% won't be sustained over 10 years - it might even be negative.

Growth companies dilute and then highly successful companies buy back stock. The net dilutive or anti-dilutive effect on a single TSLA share bought today is not possible to measure or estimate in any reliable fashion.

 

[tmoz]

If I'm reading this correctly, it sounds like if an RU or a stock option (ISO or non-qualified) is valued at the time of grant at the current stock price, then there the expense on the books for the company is zero.
Understanding the New Accounting Rules For Stock Options and Other Awards - FindLaw
I'm not a finance guy, so if I'm missing something, please point me to a clearer explanation.

 

[Doggydogworld]

If I'm reading this correctly, it sounds like if an RU or a stock option (ISO or non-qualified) is valued at the time of grant at the current stock price, then there the expense on the books for the company is zero.
Understanding the New Accounting Rules For Stock Options and Other Awards - FindLaw
I'm not a finance guy, so if I'm missing something, please point me to a clearer explanation.

That was the "old way". At-the-money options were counted as zero expense. It was absurd, but as @Fact Checking points out we did it that way forever until tech companies realized what a huge loophole it was and started really exploiting it. In defense of the FASB, employee stock options were rare in the 30s when the rules were put into place and there wsa no standard way to value them. Myron Scholes and Robert Merton won the 1997 Nobel Prize for Economics for the now-famous options pricing equation they developed with Fischer Black (who unfortunately died in 1995 and was thus ineligible). Their work dated from 1973 but took time to be accepted by the accounting profession. And then it took another decade to overcome lobbying from corporate execs who absolutely did not want to count the mega-millions in compensation they granted themselves as an expense on the income statement.

Anyway, near the bottom of your article is says that since 2005 you must use Black-Scholes or a similar method to estimate the fair market value of the options, then count that as an expense over the vesting period. Exactly the same as if you paid suppliers in options, or sold options in the market and used the cash to pay suppliers or employees. As it should be.

 

[tmoz]

That was the "old way". At-the-money options were counted as zero expense. It was absurd, but as @Fact Checking points out we did it that way forever until tech companies realized what a huge loophole it was and started really exploiting it. In defense of the FASB, employee stock options were rare in the 30s when the rules were put into place and there wsa no standard way to value them. Myron Scholes and Robert Merton won the 1997 Nobel Prize for Economics for the now-famous options pricing equation they developed with Fischer Black (who unfortunately died in 1995 and was thus ineligible). Their work dated from 1973 but took time to be accepted by the accounting profession. And then it took another decade to overcome lobbying from corporate execs who absolutely did not want to count the mega-millions in compensation they granted themselves as an expense on the income statement.

Anyway, near the bottom of your article is says that since 2005 you must use Black-Scholes or a similar method to estimate the fair market value of the options, then count that as an expense over the vesting period. Exactly the same as if you paid suppliers in options, or sold options in the market and used the cash to pay suppliers or employees. As it should be.

That is about the time when Intel switched from issuing ISOs to RUs.
Pre dot com bust, people who exercised ISOs (which normally expired after 10 yrs of issuance, and people had to exercise them or lose them), were hit with large paper gains because their grant price was way below the exaggerated pre-bust highs at the time, and had to pay hefty alt-min tax payments. When the bust came, and their actual gains deflated seriously, they were out large chunks of over-tax payments. It could take someone decades to get their overpayment back. Sometime in the early 2000s, a tax law was passed that allowed people with these overpayments to even up with the IRS over a 2 yr period. I think RUs were partly an answer to the alt-min tax problem, or perhaps easier managed expenses in light of the new tax law - I'm not sure which. Nobody has sympathy for people with large tax bills, but at Intel, ISOs were issued to just about every pay grade, not just the higher up VPs. Sorry, I but I ramble on for too long.

 

[EinSV]

Apologies if this has been covered but one possible additional source of revenue for Q2 that could be significant is recognition of a portion of EAP revenue based on the release of no-confirm Navigate on Autopilot on April 3.

It is possible this revenue was recognized when Nav on AP was first released but I doubt it because my impression is that Tesla has been conservative about recognizing AP revenue, and also recognition of deferred revenue has been lower in the past few quarters than I would expect if they recognized this revenue.

For the 9 months ending 9/30/18 Tesla recognized ~$72M in deferred revenue from auto sales and for the 12 months ending 12/31/18 they recognized ~$112M. $112M-$72M=$40M recognized in Q4. That includes a bunch of things like Supercharger access, OTA updates, internet connectivity, so is unlikely to also include Nav on AP. In Q1 2019 they only recognized another $37.4M in deferred auto sales revenue so I doubt deferred Nav on AP revenue was recognized in Q1.

According to the Q1 report, they expect to recognize $462.3M in deferred revenue from automotive sales over the next 12 months -- a significant uptick from the pace in 2018-Q1 2019.

They stopped offering the EAP package around 2/28 and sold about 400K+ AP2+ cars during the time they offered EAP. Only a portion of customers ordered EAP and I believe NOA is not yet enabled everywhere, but if Nav on AP is worth $1000 of the $5000 EAP package (WAG) this could easily be another $100M+ in revenue for Q2. I assume it would be high margin.

 

[Doggydogworld]

Apologies if this has been covered but one possible additional source of revenue for Q2 that could be significant is recognition of a portion of EAP revenue based on the release of no-confirm Navigate on Autopilot on April 3.

It is possible this revenue was recognized when Nav on AP was first released but I doubt it because my impression is that Tesla has been conservative about recognizing AP revenue, and also recognition of deferred revenue has been lower in the past few quarters than I would expect if they recognized this revenue.

For the 9 months ending 9/30/18 Tesla recognized ~$72M in deferred revenue from auto sales and for the 12 months ending 12/31/18 they recognized ~$112M. $112M-$72M=$40M recognized in Q4. That includes a bunch of things like Supercharger access, OTA updates, internet connectivity, so is unlikely to also include Nav on AP. In Q1 2019 they only recognized another $37.4M in deferred auto sales revenue so I doubt deferred Nav on AP revenue was recognized in Q1.

According to the Q1 report, they expect to recognize $462.3M in deferred revenue from automotive sales over the next 12 months -- a significant uptick from the pace in 2018-Q1 2019.

They stopped offering the EAP package around 2/28 and sold about 400K+ AP2+ cars during the time they offered EAP. Only a portion of customers ordered EAP and I believe NOA is not yet enabled everywhere, but if Nav on AP is worth $1000 of the $5000 EAP package (WAG) this could easily be another $100M+ in revenue for Q2. I assume it would be high margin.

People have mentioned it, but not with in depth numbers like yours. Non-EAP/FSD piece should roughly scale with fleet size, which doubled in 2019 vs. 2018. So 112*2=224m of non-EAP/FSD in 2019? Maybe Supercharging won't double, since the older fleet has a lot more free Supercharging than recent additions. So let's say an even $200m, leaving an expected 262m for EAP/FSD in 12 months or about 65m per quarter.

There's also an effect on current sales. They used to recognize maybe 2-3k of EAP and 0k of FSD on each sale. Now they recognize an embedded 2k or so of AP and maybe 3k of FSD. FSD take rate is pretty low, though, so the change from the old EAP/FSD approach to the new AP/FSD is probably close to revenue neutral.

 

[EinSV]

People have mentioned it, but not with in depth numbers like yours. Non-EAP/FSD piece should roughly scale with fleet size, which doubled in 2019 vs. 2018. So 112*2=224m of non-EAP/FSD in 2019? Maybe Supercharging won't double, since the older fleet has a lot more free Supercharging than recent additions. So let's say an even $200m, leaving an expected 262m for EAP/FSD in 12 months or about 65m per quarter.

There's also an effect on current sales. They used to recognize maybe 2-3k of EAP and 0k of FSD on each sale. Now they recognize an embedded 2k or so of AP and maybe 3k of FSD. FSD take rate is pretty low, though, so the change from the old EAP/FSD approach to the new AP/FSD is probably close to revenue neutral.

Just to make sure we are on the same page, I'm thinking of a one-time recognition in Q2 2019 of EAP revenue for pre- 2/28/19 sales due to enabling Navigate on Autopilot. Nav on AP was an important part of EAP that should be worth a significant amount of $ per car.

Per the data from Lex Fridman in the link below, as of July 7 there were 528,000 AP2 cars. Deducting 95,200 for Q2 and 35,000 (WAG) for March 2019 gets us to ~400,000 AP2 cars sold under the old "EAP" system. Lots of variables including the take rate for EAP, how much revenue can be recognized for Nav on AP per car ($1000+?), and percentage of EAP cars in jurisdictions with Nav on AP enabled (well over 50%), but seems like this could easily be $100M-$200M in Q2 revenue.

Tesla Vehicle Deliveries and Autopilot Mileage Statistics | MIT Human-Centered AI

 

[moneyburner1]

Deferred Revenue is an interesting Topic !

Can somebody explain why there are 2 lines for Deferred Revenue - Balance sheet from March 2019?

Deferred revenue 762,810

Deferred revenue, net of current portion 1,157,343

And are there any corresponding Cogs activated? Or will - if all Features are delivered - this straight go to Revenue and drop into the Profit line?

Thanks for your help

cheers

 

[EVNow]

Just to make sure we are on the same page, I'm thinking of a one-time recognition in Q2 2019 of EAP revenue for pre- 2/28/19 sales due to enabling Navigate on Autopilot. Nav on AP was an important part of EAP that should be worth a significant amount of $ per car.

I was earlier thinking they will send out enhanced summon to the fleet - so they can recognize that revenue. But that didn't happen. IIRC, once summon is done full EAP can be recognized. Also, with FC in the coming months, wonder how much of the FSD revenue can be recognized (probably also tied to when customers get their cars upgraded to HW3).

 

[Doggydogworld]

Deferred Revenue is an interesting Topic !

Can somebody explain why there are 2 lines for Deferred Revenue - Balance sheet from March 2019?

Deferred revenue 762,810

Deferred revenue, net of current portion 1,157,343

And are there any corresponding Cogs activated? Or will - if all Features are delivered - this straight go to Revenue and drop into the Profit line?

Thanks for your help

cheers

The 763m current liability relates to revenue they expect to recognize in the next 12 months. The other 1157m is 12+ months. The total 1920m is a mix of RVG-related mess and the EAP/FSD/Supercharger/etc. stuff mentioned above. The RVG stuff has significant COGS, EAP/FSD has almost none.

@EinSV - we're talking about the same thing, I'm just triangulating from the other side. You're looking at a feature they deployed and estimating it's value. I treat the ~262m of EAP/FSD they expect to recognize in the next 12 months as a cookie jar. It's ~65m per quarter on average, but they have a lot of discretion over which quarter to recognize it thus will tend to put it where it will do the most good. 100m is almost 200 bpp of margin uplift, which Q2 could use. But Q3 may need it even more assuming Europe goes in heavy for SR+.

Perhaps they'll take 100m in each Q2 and Q3 and let Q4 fend for itself since EOY margins don't need much help. Q1 will suck, as usual, but by then Model Y will be the story.

 

[moneyburner1]

This is a useful alternate way to look at the numbers. Thanks for this. If you see luvb2b's model, he has a lot of different margins as well.

Gross margin is what they report - and as such they will be managing that number. Tesla probably manage their pricing in such a way as to get ~20% margin after accounting for non-ZEV credits (i.e. regulatory). I split it for leasing/non-leasing to make it easier to calculate the sales cost of revenue.

In the spreadsheet you have increased the ASP - are you thinking the ASP will be higher than my estimate or was that just a test ?

Dear EVNow,

thanks for your Input

in my model the ASP is a function of the "real" automotive deliveries excluding ieasing and the "real" automotive Sales less shown adjustments. Only the split of the ASP between S&X and Model 3 is done by myself - and for sure open for discussion.


Of course I agree that the gross margin is managed by Tesla and normally a meaningful number.

But I think the "real" ASP and the "real" cogs need to be looked at and controlled by each category. Key Driver for the ASP is the pricing incl. rebates and the number of options sold.

For Cogs the driver are the number of Options sold (but you need to differentiate that correlation to the ASP - very low for FSD - higher for battery size), the volume to absorb the fixed cost, the efficency of production and the quality / scrab cost.

I hope / think that my "real" ASP and "real" Cogs give a more stable and "real" picture. That helps to predict the next quarter(s).

Looking at the result I think the numbers make mostly sense and I fully agree with your estimate for Q2 (give or take a few millions ;-) )

Cheers

 

[EVNow]

But I think the "real" ASP and the "real" cogs need to be looked at and controlled by each category. Key Driver for the ASP is the pricing incl. rebates and the number of options sold.

For Cogs the driver are the number of Options sold (but you need to differentiate that correlation to the ASP - very low for FSD - higher for battery size), the volume to absorb the fixed cost, the efficency of production and the quality / scrab cost.

I hope / think that my "real" ASP and "real" Cogs give a more stable and "real" picture. That helps to predict the next quarter(s).

Yes, if we can predict these individually better - then we can model better.

BTW, why is COGS for S&X in Q1 suddenly so much higher ? I know they changed the mix and discontinued lower level models for a while ...

 

[moneyburner1]

Yes, if we can predict these individually better - then we can model better.

BTW, why is COGS for S&X in Q1 suddenly so much higher ? I know they changed the mix and discontinued lower level models for a while ...

As I said, the split between Model 3 and S&X was done by myself and can be discussed….
I left the Cogs for Model 3 constant, as the volume was more or less constant.
To reach the reported cost the S&X Cogs had to be increased.
Key Driver is for sure the lower productions volume ~25.000 each quarter last year and now only 14.163 S&X.

this for sure has driven up the depreciation per car but also the other fix Overhead cost had to be absorbed by the lower volume (more than 40 % less).
I assume that Tesla was not able to react that fast also on production worker.
Of Course there could be also other cost in the Cogs - like the switch to the raven production.

But as I said, the total cost are reported - so the cost Need to be split between Model 3 and S&X.

 

[Doggydogworld]

Yes, if we can predict these individually better - then we can model better.

BTW, why is COGS for S&X in Q1 suddenly so much higher ? I know they changed the mix and discontinued lower level models for a while ...

Dropping to one shift held direct labor cost per car roughly the same (might have even improved slightly). Unless they add severance costs for 2nd shift to current quarter to COGS, that is. Fixed cost per car almost doubled, of course. Supplier costs per car may have increased on lower volume, e.g. their 18650 deal with Panasonic may require them to absorb battery cell fixed costs.

Also, they took 75D off the web site in early Q1 and presumably only built 100Ds. Those obviously have higher COGS, but they sold 75Ds out of inventory which should have had lower Q3/Q4 COGS. We don't know how they calculate COGS, though. In my mind they assign a cost to each VIN based on configuration and quarterly averages for component, labor and fixed costs. They add that VIN's cost to finished goods inventory when produced, then subtract it from FGI and add it to COGS when sold. There are lots of other ways to do it, though, e.g. by SKU using FIFO or whatever.

 

[EVNow]

As I said, the split between Model 3 and S&X was done by myself and can be discussed….
I left the Cogs for Model 3 constant, as the volume was more or less constant.
To reach the reported cost the S&X Cogs had to be increased.
Key Driver is for sure the lower productions volume ~25.000 each quarter last year and now only 14.163 S&X.

this for sure has driven up the depreciation per car but also the other fix Overhead cost had to be absorbed by the lower volume (more than 40 % less).
I assume that Tesla was not able to react that fast also on production worker.
Of Course there could be also other cost in the Cogs - like the switch to the raven production.

But as I said, the total cost are reported - so the cost Need to be split between Model 3 and S&X.

You can look at luvb2b's model to see what cogs for s&x he has calculated. On 102k ASP, he calculates 14.5% margin (ex reg credits) or cogs of < 90k.

 

[Stars aligned]

The 763m current liability relates to revenue they expect to recognize in the next 12 months. The other 1157m is 12+ months. The total 1920m is a mix of RVG-related mess and the EAP/FSD/Supercharger/etc. stuff mentioned above. The RVG stuff has significant COGS, EAP/FSD has almost none.

@EinSV - we're talking about the same thing, I'm just triangulating from the other side. You're looking at a feature they deployed and estimating it's value. I treat the ~262m of EAP/FSD they expect to recognize in the next 12 months as a cookie jar. It's ~65m per quarter on average, but they have a lot of discretion over which quarter to recognize it thus will tend to put it where it will do the most good. 100m is almost 200 bpp of margin uplift, which Q2 could use. But Q3 may need it even more assuming Europe goes in heavy for SR+.

Perhaps they'll take 100m in each Q2 and Q3 and let Q4 fend for itself since EOY margins don't need much help. Q1 will suck, as usual, but by then Model Y will be the story.

That's actually very interesting. Do you have a view how much of the 763m could be recognised in Q2? 25% maybe? And could this be a game changer for GAAP earnings figure?

 

[Stars aligned]

Interestingly enough, deferred revenues were reduced from 1015m to 630m 2017 to 2018 y/y, but increased from 630 to 762 from Q4 to Q1. Similarly, Deferred revenue net of current portion was reduced from 1177m to 990m y/y, but increased from 990 to 1157 q/q. It has also to be noted that, although deferred revenues have been increased q/q substantial, sales were quite low.

I see that there was a change in accounting standards, leading to a reduction of deferred revenue of 436m and net of current of 429m as per end of 2018.

Nevertheless, Gerome stated in his email to employees that "quality is also reaching record highs" and in Q1 Tesla was not able to deliver. Net/net, might deferred revenues increase in Q2 at a much lower rate, compared to revenues, and therefore contribute to better earnings?

Edit: I always found that wild swing from profits in Q3 and Q4 to such a big loss in Q1 suspicious. If we look at this from the above angle, than 2018 earnings have been helped by 436+429m=865m, average 217m per quarter. Accounting games.....

 

[EVNow]

Edit: I always found that wild swing from profits in Q3 and Q4 to such a big loss in Q1 suspicious. If we look at this from the above angle, than 2018 earnings have been helped by 436+429m=865m, average 217m per quarter. Accounting games.....

Hmmm … when you have such a big drop in deliveries, you will see such "swings". Your revenue and operating profit goes down while your fixed costs remain.

 

[moneyburner1]

You can look at luvb2b's model to see what cogs for s&x he has calculated. On 102k ASP, he calculates 14.5% margin (ex reg credits) or cogs of < 90k.

Thanks, good idea - I used his S&X data as starting point and then calculated the Model 3 cogs.

This makes more sense !

Except the Cogs M3 for Q1 2019. I don`t understand the Driver for an increase vs Q4 2018 - especially as the produced volume increased slightly.

I think you are aware that your and luv2b`s number differ in Q2 in Sales and Cogs - not so much in Gross Profit...

 

[ProfTournesol]

Why the low margin? The Q1 margin is not realistic. There were serious issues, the margin will be more in the range of 18%-19%. Especially with the recent price adjustment, I cannot imagine the will accept a margin of 14%.

Also note: Elon changed his behaviour and is now in a state of under promise and over deliver. I expect a positive Q2.