We used to have a good tradition going on with quarterly results discussion thread. Hopefully this post can be a start of a new episode. I will first try to estimate the number of cars, from there the NON-GAAP revenue and financial results.
Number of Model S delivered :
Everyone knows I believe in VIN counting based on two assertions (1) every car delivered has a VIN (2) Tesla assigns VINS mostly in a regular pattern. For the S we only really have one known challenge to the latter rule the last 2 years, but we have thousands of reports that confirm it. Therefore I think any differentiation from rule (2) is likely small enough not to make a significant impact on the results. The tricky thing is to decide which cars are delivered in which quarter though and also how much of the VIN gaps are going to inventory and how much of older VIN gaps got sold out of inventory.
Cars in the range 144k-146k were delivered both in Q2 and Q3, but more in the latter quarter so I count 2000 cars in this range. 147k-157k were delivered exclusively in Q3 so 10 000 cars. Finally the range 158k-159k are delivered both in Q3 and Q4, I count 1000 cars. Total 13k Model S. However Tesla made a real push to lease inventory in the US this quarter. I think this could add up to a net another 1000 cars to the total. Be aware that I've seen less being shifted through ev-cpo.com, but not all cars make it to the website. So 14k Model S. At the same time, about 5% of the VINs in sequence are not backed by a produced car and about an additional 5% are generally in loaner/demo/inventory service. The lower bound of deliveries could be as low as 11k but realistically I think 12k is the real minimum.
The above number seems consistent with what we know. Europe is on track to do nearly 4k Model S this quarter which puts it right around its historical average of 1/3th of global deliveries. The US scorecard has 5275 units so far for the quarter which is unseasonably high but I suspect that's really part of the abnormally large number of cars in the pipeline at the end of last quarter. Add in a regular 'last quarter of the month' and we get somewhere to 9000 cars in the US. This seems high, because it leaves very little cars to the rest of the world, but if we do the same exercise for first quarter (12420 delivered - 3814 in EU - 6490 US = 2116 ROW) and last quarter (9764 delivered - 3046 EU - 5700 US = 1018 ROW) we can see that either ROW is indeed not a very important market or inside-ev overestimates US a bit. Either way our conclusion remains that the numbers remain plausible.
Number of Model X delivered :
This is a tougher cookie to crack, because production has been chaotic with some models/geographies not yet being served. The highest X vin assigned with a plausible September delivery is 19600. 7242 of those were already delivered in earlier quarters. Another 1000 serve as demo/loaner/inventory (4 per store). Some VINs are for 5 seaters or from 90D orders who upgraded to P100D with restricted availability. Estimating either is nearly impossible. I am taking a stab and saying that another 1000 VINs are like this. This leaves at most 10k Model X to be delivered this quarter.
Evinside tells us we may be looking at 5K deliveries in the US, again that's consistent. EU is at a much lower base which would imply an inconsistency, however the very first sign we have for September shows exceptionally strong deliveries. The X may also simply be a more popular car in China than Europe which also partly would complain the lackluster number we had for S deliveries in the above paragraph for ROW (Rest Of World) Again, 10k Model X are not inconsistent with other numbers but this too seems to be about the upper bound. Deliveries as low as 8k are still possible.
Revenue
Let's go with what I consider an optimistic scenario : 14k Model S deliveries and 10k Model S. For the S I am taking a baseline ASP of $95 000 for the last quarter in 2015. For the X we take $5 000 more for US models and $15 000 for ROW models (half of them will be signature). Average baseline X ASP therefore is $105 000. However, the 60kWh was introduced recently and takes up over half of the orders this quarter. Their ASP is $9500 less than what people would have normally ordered which they may make up with some extra options. I discount therefore the ASP for half of the cars (the 60kWh) for $5 000 (minus $9500 plus $4500 in options) or $2 500 over the full fleet.
Non GAAP automotive revenue therefore is $1.295B for the S and $1.025B for the X. Together $2,32B. Services and other may be $100M but it will be a wash. ZEV credits another $50M. All total $2.47B.
Cost of goods
Automotive margin was 21,9% last quarter on a non-GAAP basis, assuming guidance holds we should see 23% this quarter. Tesla runs its services department as a non-profit so 0% there, while ZEV credits are pure profit. Cost of goods therefore is approx $1.8B on automative, $100M on services and 0% on ZEV, total of $1.9B leaving a $570M gross profit for the quarter.
Expenses
R&D is difficult to scale down. Most is in headcount and with Tesla in a deadly hiring race for talent, I don't think they'll reduce their engineering force. On the contrary, Elon is known for wanting to hire talent! Therefore I am assuming $200M in R&D. In the third quarter, Tesla increased the number of service centers from 128 to 144. While not a perfect proxy, it shows that the structure of the company is growing and so must the expenses associated with it. A linear increase would set Selling, General and Administrative costs at $361M, but assuming some cost saving come with scale, $350M seems more reasonable. Finally, interest expense is rapidly increasing for Tesla. $46M last quarter so very conservatively $60M. The differences in non-GAAP versus GAAP on various of these numbers is a but muddled with me, but most seems to come from stock based compensation programs. If anyone could shed cleared light here, it would be most appreciated. Currently I am going with excluding stock based compensations for a total of $75M , a modest decline from $81M assuming most milestones related to Model X production had been met before the quarter)
Earnings per share (non-GAAP)
$570M gross profit - $200M R&D - $350M SG&A - $60M interest + $75M GAAP reconciliation = $35M earnings on a non-GAAP basis or $0,25/share. This would be in line with Elon's leaked mail.
Possible surprises
* Tesla sold a lot more inventory. Every 1k in inventory adds about $20M to the earnings
* Tesla sold a lot less cars. Every miss by 1k removes about $23M from the earnings
* Tesla misses/improves on the gross margin. Every 1% makes a difference in $23M to the bottom line
* Tesla does not sell ZEV credits. This would take $50M of the earnings
* Tesla's credit became significantly more expensive (it neared the borrowing limit with one partner already). Could remove $10M from the earnings
* SG&A costs are not yet under control as witnessed from the letter, could remove up to $15M from the earnings
* Cost savings are possible in the R&D from Model 3 design phase finished, could add up to $15M to the earnings
* Tesla needs to start making negative adjustments for the cost of cars turned in under the Resale Value Guarantee. Up to $85.6M outstanding liabilities for cash payments are current this quarter. If they need to buy half of the cars and take a bath of 10% on each, that's at most $5M.
* Tesla sold Tesla Energy products at higher margin than 0. Ramp up on those were only planned for next quarter and we know Tesla is not likely to be early. Still, just $25M of sales at 20% GM is $5M to the bottom line.
Looking forward to corrections, things I overlooked and discussion.
Number of Model S delivered :
Everyone knows I believe in VIN counting based on two assertions (1) every car delivered has a VIN (2) Tesla assigns VINS mostly in a regular pattern. For the S we only really have one known challenge to the latter rule the last 2 years, but we have thousands of reports that confirm it. Therefore I think any differentiation from rule (2) is likely small enough not to make a significant impact on the results. The tricky thing is to decide which cars are delivered in which quarter though and also how much of the VIN gaps are going to inventory and how much of older VIN gaps got sold out of inventory.
Cars in the range 144k-146k were delivered both in Q2 and Q3, but more in the latter quarter so I count 2000 cars in this range. 147k-157k were delivered exclusively in Q3 so 10 000 cars. Finally the range 158k-159k are delivered both in Q3 and Q4, I count 1000 cars. Total 13k Model S. However Tesla made a real push to lease inventory in the US this quarter. I think this could add up to a net another 1000 cars to the total. Be aware that I've seen less being shifted through ev-cpo.com, but not all cars make it to the website. So 14k Model S. At the same time, about 5% of the VINs in sequence are not backed by a produced car and about an additional 5% are generally in loaner/demo/inventory service. The lower bound of deliveries could be as low as 11k but realistically I think 12k is the real minimum.
The above number seems consistent with what we know. Europe is on track to do nearly 4k Model S this quarter which puts it right around its historical average of 1/3th of global deliveries. The US scorecard has 5275 units so far for the quarter which is unseasonably high but I suspect that's really part of the abnormally large number of cars in the pipeline at the end of last quarter. Add in a regular 'last quarter of the month' and we get somewhere to 9000 cars in the US. This seems high, because it leaves very little cars to the rest of the world, but if we do the same exercise for first quarter (12420 delivered - 3814 in EU - 6490 US = 2116 ROW) and last quarter (9764 delivered - 3046 EU - 5700 US = 1018 ROW) we can see that either ROW is indeed not a very important market or inside-ev overestimates US a bit. Either way our conclusion remains that the numbers remain plausible.
Number of Model X delivered :
This is a tougher cookie to crack, because production has been chaotic with some models/geographies not yet being served. The highest X vin assigned with a plausible September delivery is 19600. 7242 of those were already delivered in earlier quarters. Another 1000 serve as demo/loaner/inventory (4 per store). Some VINs are for 5 seaters or from 90D orders who upgraded to P100D with restricted availability. Estimating either is nearly impossible. I am taking a stab and saying that another 1000 VINs are like this. This leaves at most 10k Model X to be delivered this quarter.
Evinside tells us we may be looking at 5K deliveries in the US, again that's consistent. EU is at a much lower base which would imply an inconsistency, however the very first sign we have for September shows exceptionally strong deliveries. The X may also simply be a more popular car in China than Europe which also partly would complain the lackluster number we had for S deliveries in the above paragraph for ROW (Rest Of World) Again, 10k Model X are not inconsistent with other numbers but this too seems to be about the upper bound. Deliveries as low as 8k are still possible.
Revenue
Let's go with what I consider an optimistic scenario : 14k Model S deliveries and 10k Model S. For the S I am taking a baseline ASP of $95 000 for the last quarter in 2015. For the X we take $5 000 more for US models and $15 000 for ROW models (half of them will be signature). Average baseline X ASP therefore is $105 000. However, the 60kWh was introduced recently and takes up over half of the orders this quarter. Their ASP is $9500 less than what people would have normally ordered which they may make up with some extra options. I discount therefore the ASP for half of the cars (the 60kWh) for $5 000 (minus $9500 plus $4500 in options) or $2 500 over the full fleet.
Non GAAP automotive revenue therefore is $1.295B for the S and $1.025B for the X. Together $2,32B. Services and other may be $100M but it will be a wash. ZEV credits another $50M. All total $2.47B.
Cost of goods
Automotive margin was 21,9% last quarter on a non-GAAP basis, assuming guidance holds we should see 23% this quarter. Tesla runs its services department as a non-profit so 0% there, while ZEV credits are pure profit. Cost of goods therefore is approx $1.8B on automative, $100M on services and 0% on ZEV, total of $1.9B leaving a $570M gross profit for the quarter.
Expenses
R&D is difficult to scale down. Most is in headcount and with Tesla in a deadly hiring race for talent, I don't think they'll reduce their engineering force. On the contrary, Elon is known for wanting to hire talent! Therefore I am assuming $200M in R&D. In the third quarter, Tesla increased the number of service centers from 128 to 144. While not a perfect proxy, it shows that the structure of the company is growing and so must the expenses associated with it. A linear increase would set Selling, General and Administrative costs at $361M, but assuming some cost saving come with scale, $350M seems more reasonable. Finally, interest expense is rapidly increasing for Tesla. $46M last quarter so very conservatively $60M. The differences in non-GAAP versus GAAP on various of these numbers is a but muddled with me, but most seems to come from stock based compensation programs. If anyone could shed cleared light here, it would be most appreciated. Currently I am going with excluding stock based compensations for a total of $75M , a modest decline from $81M assuming most milestones related to Model X production had been met before the quarter)
Earnings per share (non-GAAP)
$570M gross profit - $200M R&D - $350M SG&A - $60M interest + $75M GAAP reconciliation = $35M earnings on a non-GAAP basis or $0,25/share. This would be in line with Elon's leaked mail.
Possible surprises
* Tesla sold a lot more inventory. Every 1k in inventory adds about $20M to the earnings
* Tesla sold a lot less cars. Every miss by 1k removes about $23M from the earnings
* Tesla misses/improves on the gross margin. Every 1% makes a difference in $23M to the bottom line
* Tesla does not sell ZEV credits. This would take $50M of the earnings
* Tesla's credit became significantly more expensive (it neared the borrowing limit with one partner already). Could remove $10M from the earnings
* SG&A costs are not yet under control as witnessed from the letter, could remove up to $15M from the earnings
* Cost savings are possible in the R&D from Model 3 design phase finished, could add up to $15M to the earnings
* Tesla needs to start making negative adjustments for the cost of cars turned in under the Resale Value Guarantee. Up to $85.6M outstanding liabilities for cash payments are current this quarter. If they need to buy half of the cars and take a bath of 10% on each, that's at most $5M.
* Tesla sold Tesla Energy products at higher margin than 0. Ramp up on those were only planned for next quarter and we know Tesla is not likely to be early. Still, just $25M of sales at 20% GM is $5M to the bottom line.
Looking forward to corrections, things I overlooked and discussion.
