Near-future quarterly financial projections

[ReflexFunds]

Thanks for this effort, ReflexFunds. True service to the community. Outstanding.

I notice one of your assumptions is an FSD take rate of 50%. I've seen other estimates of 25%. There's been an attempt to model FSD take by dividing deferred FSD revenue by the number of HW2+ units produced since Oct 2016.

Do you have a source estimating FSD take rate? This seems it would have an outsized effect on gross margin per unit.

Cheers!

My FSD take rates are from trying to reconcile prior revenue/gross profit/deliveries accounts and the $500m accumulated FSD deferred revenue balance.
Apparently a Deutsche Bank analyst reported that Tesla IR had told them FSD had a 27% take rate. But i didn't have enough information or confidence to give this report much weight - given that informal comments like this are often misinterpreted in translation. Was the 27% take rate in Q3 or cumulative since 2016? Who knows.

If my FSD take rates are too high, it means some of my other assumptions are too low (most likely AWD/P mix). Given I have kept these all flat QoQ it won't impact my Q4 forecasts if i had the mix wrong between FSD and AWD.

Hopefully FSD is indeed as low as 27%. This would imply much more potential for increased take rate as Tesla releases new FSD features this year.

 

[EVNow]

I think the ASP for all models will rise after the release of FSD

That's a good call. I'll increase the ASPs in my next update.

 

[EVNow]

Hopefully FSD is indeed as low as 27%. This would imply much more potential for increased take rate as Tesla releases new FSD features this year.

I don't know what the take rate is - but I doubt it has dramatically increased in the last few quarters (from 10% to 50%) - since the deferred amount itself is ~150M for a few quarters now, even with ever increasing deliveries.

There was apparently some leakage of some files that allowed people to calculate FSD take rate ~ 25% (IIRC). I'll update this post if I can find the info.

ps : Older cars upgradable to newer Tesla chip?

35% as of Q3 '17. A db with information about all cars was leaked.

 

[ReflexFunds]

I don't know what the take rate is - but I doubt it has dramatically increased in the last few quarters (from 10% to 50%) - since the deferred amount itself is ~150M for a few quarters now, even with ever increasing deliveries.

There was apparently some leakage of some files that allowed people to calculate FSD take rate ~ 25% (IIRC). I'll update this post if I can find the info.

In March the FSD they were selling had no available features - I expect take rate was very low. In Q2 they bundled NAV on Autopilot into FSD and in Q3 they added Enhanced Summon.
I'm sure there was a significant increase in FSD take rates when the product was restructured to include many of the EAP features.

There are many moving parts in deferred revenue - including additions and bookings, free supercharging (and significant variation QoQ on how many cars this has been provided to), one off bookings from discounted FSD/EAP upgrades etc.
On my numbers it looks like the deferred FSD gross additions per Q has been increasing despite likely increasing recognition percentages.

 

[Doggydogworld]

Its very possible that Tesla has calculated that US sales won't suffer at all by March, and the ROW will absorb production fully in Jan/Feb. With anecdotal evidence from S. Korea and rejuvenated tax policies in the UK and Germany, Tesla likely continues to remain Production limited throughout 2020Q1.

Very little production is going to Pier 80 so far. It's nothing like Q4. Either they've cut production dramatically or they're moving cars some other way. I've seen reports of truckloads heading south on I-5, but no idea how many. Model 3 wait is 5-8 weeks here in south TX, so at least some Jan/Feb production will head this way.

I read a claim that the tent with GA4 is now gone. That would really surprise me at this point, the flexibility it offers should be very valuable as they start the Model Y ramp. I could see them getting rid of it once the Model Y line is humming, though. Does anyone have any info on this?

As for FSD, a salesperson told me in mid-2018 take rate was very high for EAP but low for FSD. He also implied take rate was lower for Model 3 buyers, but they really hadn't delivered many 3s at that point.

 

[gerebgraus]

Very little production is going to Pier 80 so far. It's nothing like Q4. Either they've cut production dramatically or they're moving cars some other way. I've seen reports of truckloads heading south on I-5, but no idea how many. Model 3 wait is 5-8 weeks here in south TX, so at least some Jan/Feb production will head this way.

I read a claim that the tent with GA4 is now gone. That would really surprise me at this point, the flexibility it offers should be very valuable as they start the Model Y ramp. I could see them getting rid of it once the Model Y line is humming, though. Does anyone have any info on this?

As for FSD, a salesperson told me in mid-2018 take rate was very high for EAP but low for FSD. He also implied take rate was lower for Model 3 buyers, but they really hadn't delivered many 3s at that point.

Tesla has traditionally stopped production in the first week of January for maintainance, improvements, giving time for workers to recover. They tend to do crazy overtime in the last days of the year, etc. There is a possibility of course that some lines are still producing. On the other hand there were claims that Model Y has multiple technological advances. There should not be any reason not to apply these advances for Model 3, and in this case, they might need a much longer time to modify the production lines. Now that would make Q1 predictions more difficult. Production could be lower due to fewer production days, but costs could be less increasing profit.

 

[dc_h]

Very little production is going to Pier 80 so far. It's nothing like Q4. Either they've cut production dramatically or they're moving cars some other way. I've seen reports of truckloads heading south on I-5, but no idea how many. Model 3 wait is 5-8 weeks here in south TX, so at least some Jan/Feb production will head this way.

I read a claim that the tent with GA4 is now gone. That would really surprise me at this point, the flexibility it offers should be very valuable as they start the Model Y ramp. I could see them getting rid of it once the Model Y line is humming, though. Does anyone have any info on this?

As for FSD, a salesperson told me in mid-2018 take rate was very high for EAP but low for FSD. He also implied take rate was lower for Model 3 buyers, but they really hadn't delivered many 3s at that point.

The pier is 80% full-posted in gen thread. The tent is there shown by shorts on Twitter this week.

 

[Thekiwi]

Here is my corrected model.

The correction is to bring the auto-margin forecast inline with the previous margin calculation. There was an error earlier where for past quarters I was calculating the margin including the regulatory credits, but the forecast excluded the credits. This inflated the profit.

View attachment 499379

Just trying to get my head around leasing revenue, wondering how you are allocating model 3 lease revenue - how much per car leased per quarter? and how many quarters are you using before that leased revenue for that car finishes?

 

[FrankSG]

Just trying to get my head around leasing revenue, wondering how you are allocating model 3 lease revenue - how much per car leased per quarter? and how many quarters are you using before that leased revenue for that car finishes?

The way I've done it is to assume that leasing revenue is correlated to the value of lease vehicles on Tesla's balance sheet:

1. I take the "Operating lease vehicles" value from the balance sheet.
2. I assume this depreciates by a certain % each month.
3. I add the COGS of the cars leased that quarter to the "Operating lease vehicles" value.
4. I calculate revenue based on the "Operating lease vehicles" value.
5. Calculate profit with a margin assumption.

 

[EVNow]

Just trying to get my head around leasing revenue, wondering how you are allocating model 3 lease revenue - how much per car leased per quarter? and how many quarters are you using before that leased revenue for that car finishes?

You can see the formula I use in the sheet. Essentially I calculate the delta from last quarters lease revenue. Vehicles from 3 years back drop off and this quarters vehicles get added. Because Tesla sells most vehicles in the third quarter it’s a bit more complicated, but that’s the gist.

 

[Thekiwi]

What is everyone using for income tax rate on profitable quarters going forward? does Tesla have a large amount of offsetting tax losses from past quarters that will reduce the tax bill in 2020?

Judging by the current activity in the main investor thread this weekend, perhaps I was asking the wrong question about offsetting tax losses!

I do hesitate though before making any assumptions that Tesla will recognize a big one off gain due to accumulated losses in any particular quarter. Management, quite rightly IMO, likely has no desire to bring forward a one off profit item just to create a trailing 4 quarter profit to end up qualifying for index inclusion (inclusion of which has nothing to do with running Tesla). I think that is especially true when the same thing (trailing 4 quarter profit) will highly likely happen just 13 weeks later after Q1 earnings.

I think Tesla will recognize a one off profit from unrealized tax loss reclassification (or whatever the correct term is), when the business conditions reach the appropriate level. (Now that might actually be last quarter, in which case great!)

 

[EVNow]

I do hesitate though before making any assumptions that Tesla will recognize a big one off gain due to accumulated losses in any particular quarter. Management, quite rightly IMO, likely has no desire to bring forward a one off profit item just to create a trailing 4 quarter profit to end up qualifying for index inclusion (inclusion of which has nothing to do with running Tesla). I think that is especially true when the same thing (trailing 4 quarter profit) will highly likely happen just 13 weeks later after Q1 earnings.

I don't understand why some posters start hyping before ER and Delivery reports - bringing up some highly speculative, creative accounting options to get one time outsized profit. It just sets up some people to over leverage.

Last quarter a lot of missed the profit - but few bulls speculated on profit because of sudden change in margin & forex changes.

 

[Troy]

Here are my current production and delivery estimates for 2020.

There is a drone video from 4 days ago HERE that shows the Shanghai factory. It looks like they started Phase Two of the construction which includes the building for the '70 parts to 1' casting machine for Model Y. That's the building they are currently building the foundations for. I assumed it will be finished by the end of this year and they will do a test run in December. I think it makes sense for Tesla to use the same casting machine for Model 3 and Model Y at Fremont. I expect that to happen in 2021.

 

[FrankSG]

Here are my current production and delivery estimates for 2020.

There is a drone video from 4 days ago HERE that shows the Shanghai factory. It looks like they started Phase Two of the construction which includes the building for the '70 parts to 1' casting machine for Model Y. That's the building they are currently building the foundations for. I assumed it will be finished by the end of this year and they will do a test run in December. I think it makes sense for Tesla to use the same casting machine for Model 3 and Model Y at Fremont. I expect that to happen in 2021.

I think those Model Y numbers are possible, but I'm personally more conservative on those projecting a slower production ramp up, and about 65k produced in total in 2020.

However, those are some real pessimistic numbers in the rest of your model in my opinion. Only 245k M3 from Fremont, that's some extreme cannibalization.

I could see MY cannibalizing M3 to some extent in 2021, but there's bound to be some pent up MY demand in the form of reservations/orders that they will have to get through first. The Model S and X sales numbers didn't start dropping until M3 had sold >100k units, and it's still very much a question of how much of that was due to cannibalization from M3, and how much was due to the discontinuation of the SR versions.

I can't see them do anything less than 575k in 2020, with upside to 650k, perhaps even 700k. Get ready to adjust your model after 2020 guidance in six days.

 

[Troy]

However, those are some real pessimistic numbers in the rest of your model in my opinion. Only 245k M3 from Fremont, that's some extreme cannibalization.

I wasn't actually thinking of cannibalization. If you add up these 3 numbers, the total is 119K:

I wanted to keep this below 125K because I think the maximum capacity of the paint shop at Fremont is 500K/year which is 125K/Quarter.

Also, if these numbers were true, Model 3 sales in US+EU would drop by only 9%. In 2019, Tesla delivered 301K Model 3s worldwide. 34K of those were delivered in China. That means US+EU deliveries were 267K units. I assumed 242K deliveries in US+EU in 2020 which is not that big of a drop. I think it makes more sense for Tesla to drop the Model 3 SR+ at some point. 31% of Model 3s sold are Model 3 SR+. Discontinuing SR+ in Q4 2020 and making more Model Ys would be a good idea to increase gross margins and profitability.

 

[kengchang]

@Troy there's permit of South paint shop being retooled if you search for my post hisw

 

[mongo]

I wanted to keep this below 125K because I think the maximum capacity of the paint shop at Fremont is 500K/year which is 125K/Quarter

I'd rethink that assumption. Model 3 was planned to hit 10k/ week on its own while still making S/X. Elon proposed that Y will out sell 3,A and X combined. No way would they set up the Y line in Fremont with that type of constraint. They have also had building permits recently related to the paint shop.

 

[Troy]

I can't see them do anything less than 575k in 2020, with upside to 650k, perhaps even 700k. Get ready to adjust your model after 2020 guidance in six days.

OK, I started a poll on Twitter. You can find it here.

 

[FrankSG]

I wasn't actually thinking of cannibalization. If you add up these 3 numbers, the total is 119K:

I wanted to keep this below 125K because I think the maximum capacity of the paint shop at Fremont is 500K/year which is 125K/Quarter.

Also, if these numbers were true, Model 3 sales in US+EU would drop by only 9%. In 2019, Tesla delivered 301K Model 3s worldwide. 34K of those were delivered in China. That means US+EU deliveries were 267K units. I assumed 242K deliveries in US+EU in 2020 which is not that big of a drop. I think it makes more sense for Tesla to drop the Model 3 SR+ at some point. 31% of Model 3s sold are Model 3 SR+. Discontinuing SR+ in Q4 2020 and making more Model Ys would be a good idea to increase gross margins and profitability.

OK, I started a poll on Twitter. You can find it here.

That poll is useless. The choices you've given are already influencing people to think a certain way. People who have done no research at all might assume that 480k - 540k is a reasonable range (which imo it is not), and choose an option within that based on their level of optimism.

And some people who have done research but are more bullish than your most bullish option (like me), are going to disregard this poll completely because the choices are too low.

 

[The Accountant]

Here are my current production and delivery estimates for 2020.

There is a drone video from 4 days ago HERE that shows the Shanghai factory. It looks like they started Phase Two of the construction which includes the building for the '70 parts to 1' casting machine for Model Y. That's the building they are currently building the foundations for. I assumed it will be finished by the end of this year and they will do a test run in December. I think it makes sense for Tesla to use the same casting machine for Model 3 and Model Y at Fremont. I expect that to happen in 2021.

I know some think this number is too low. I believe you are coming from a Fremont supply constraint POV. I can understand your logic.
I think the numbers will be higher - but let's call this my bear case. Even with these numbers....Tesla is profitable in all quarters in 2020.
I think 99,250 deliveries in Q1 gets us a profit...and that would be awesome !!