Wiki 2016 Third quarter results discussion

[Fallenone]

The Q2 letter said 140 million shares. Is there a more recent source that you are using and/or are there assumptions in your number?

Straight 10-Q for Q2 earnings

"As of July 29, 2016, there were 148,692,605 shares of the registrant’s Common Stock outstanding."

The difference is the secondary offer in Q2

 

[doctoxics]

Straight 10-Q for Q2 earnings

"As of July 29, 2016, there were 148,692,605 shares of the registrant’s Common Stock outstanding."

The difference is the secondary offer in Q2

Thanks. Why would the letter use 140 million shares in the earnings calculations?

 

[Fallenone]

Thanks. Why would the letter use 140 million shares in the earnings calculations?

No idea. SEC reporting rules maybe.

 

[schonelucht]

Geat work! I think most of your assumptions are solid. I have a few nitpicks.

1. COGS for supplier parts (glider) has a 10% reduction compared to 1H 2016 and before due to doubled production rate of S and X and the reservation numbers of 3 giving Tesla leverage. This parameter will be part of the sensitivity analysis, using 5% and 7.5% as alternatives.

I continue to think 10% is a bit too optimistic and it shows up in your gross margin projections which are basically at where Tesla guided to be by the end of the year only.

6. Discount on base prices for 75 kWh Model S and Model X at 10%, for 90 kWh models at 30%. Based on my impression of the discounts I've seen

Are you modelling these discounts as having an impact on GM or on SG&A? It shouldn't matter for the earnings per share result of course.

8. Discount uptake rate for Model X 75 at 30%, X 90 at 10%. Lower for the 90 based on speculation on more 90 versions going to outside US customers who have been waiting for 3 years so no need to discount to push sales.

I think we can safely say there was no discount at all on the higher optioned X's.

 

[EinSV]

I updated my spreadsheet with the delivery numbers and the easiness of doing some sensitivity analysis on discount uptake rates, improvements on COGS, Model X's additional COGS compared to S (due to early production).

Assumptions:

A. On the automotive business
1. COGS for supplier parts (glider) has a 10% reduction compared to 1H 2016 and before due to doubled production rate of S and X and the reservation numbers of 3 giving Tesla leverage. This parameter will be part of the sensitivity analysis, using 5% and 7.5% as alternatives.

2. Fixed COGS cost reduction of 37% due to more production.

3. Model X costs 10k more in COGS to a comparable base Model S, down from $16.5k last quarter due to improved efficiency. This may sound huge, but I think this is reality, otherwise Model X's gross margin couldn't be that low in the past quarters. I also think this will continue to fall in the future. Alternative scenarios are 8k and 12k.

4. Distribution of Model S are 2700 for 60, 60D, 75, 75D. 2150 for 90D and P90D, 700 for P90DL

5. Distribution of Model X are 4300 for 75D, 2150 for 90D and P90D, 100 for P100DL

6. Discount on base prices for 75 kWh Model S and Model X at 10%, for 90 kWh models at 30%. Based on my impression of the discounts I've seen

7. Discount uptake rate for all Model S 30%. Also based on my impression of the aggressive sales starting in August outside of US and getting more and more apparent in the US in September.

8. Discount uptake rate for Model X 75 at 30%, X 90 at 10%. Lower for the 90 based on speculation on more 90 versions going to outside US customers who have been waiting for 3 years so no need to discount to push sales.

B. On the services and others business
1. Keeping it same as Q2, i.e. 88M. With an assumed 5% gross margin. This is because of

2. No significant ZEV, I expect this to be the same in Q3 based on reasons I've stated elsewhere and I have high confidence in this assumption.

3. Although there were some great news on TE, but I highly doubt any of those news made way into Q3 results.

4. Pure service (like serving owners and selling CPO) has razor thing margins. TE has better but still not taking the larger part of this top line.

C. On operating expenses
1. Assumed they went up with the same QoQ as Q2 over Q1. I am a bit worried about this one though. The longer they wait on ramping up OpEx, the higher the risk we would see a messy Model 3 early deliveries. Same goes with CapEx.

D. On interest and others
1. Interest income assumed the same as in Q2. Because I have no idea how to guess this.

2. Interest loss assumed going up with the same QoQ as Q2 over Q1.

3. Other expenses assumed to be the average over the past 6 quarters, as it seems fluctuating randomly to me.

E. On GAAP and non-GAAP differences
1. I don't know how the GAAP revenue will change this quarter. Not a bit. Therefore, I take two scenarios here. One is going along with the average of GAAP revenue from automotive being about 80% of non-GAAP, more of a business as usual scenario. The other one is going with 98% of the non-GAAP. Having GAAP and non-GAAP revenue the same is simply impossible because the wireless connection, future costs for the Superchargers, and other small things are not going to be affected by obsoleting Residual Value Guarantee.

2. Assumed service and others having no difference between GAAP and non-GAAP

3. Applied same change ratios for both GAAP and non-GAAP calculations on the OpEx.

F. Shares outstanding
1. 148.69 M


Results:
Base case
Glider 10% improved COGS, Model X additional COGS $10k

ASP for Model S $92.7k, gross margin 24.0%
ASP for Model X $112.2k, gross margin 23.3%

non-GAAP EPS $0.24
GAAP EPS with the 80% of non-GAAP automotive assumption $-0.78
GAAP EPS with the 98% of non-GAAP automotive assumption $-0.08

Best case
Glider 10% improved COGS, Model X additional COGS $8k

ASP for Model S $92.7k, gross margin 24.0%
ASP for Model X $112.2k, gross margin 26.0%

non-GAAP EPS $0.42
GAAP EPS with the 80% of non-GAAP automotive assumption $-0.63
GAAP EPS with the 98% of non-GAAP automotive assumption $0.10

Worst case
Glider 5% improved COGS, Model X additional COGS $12k

ASP for Model S $92.7k, gross margin 21.2%
ASP for Model X $112.2k, gross margin 18.4%

non-GAAP EPS $-0.35
GAAP EPS with the 80% of non-GAAP automotive assumption $-1.25
GAAP EPS with the 98% of non-GAAP automotive assumption $-0.66

Thank you for sharing this detailed and very helpful analysis.

I did have one question. How did you derive the average 30% discount for S90s, with an estimated 30% cars being sold with a discount? At the time I ordered a P90 in late August, the largest discount available on a P90 was about 23%, and these sorts of discounts were typically vehicles with a significant number of miles and in some cases damage that had been repaired. And there were almost no discounted non-P 90s available at that time (and I believe afterwards as well although I did not follow it that closely). I would expect the number of discounted non-P 90s and the average percentage discount to be quite low.

I am not sure that it matters at the end of the day since your overall projected margins and ASP seem very reasonable, but was just curious of your source/thinking on this.

 

[esk8mw]

Thank you for sharing this detailed and very helpful analysis.

I did have one question. How did you derive the average 30% discount for S90s, with an estimated 30% cars being sold with a discount? At the time I ordered a P90 in late August, the largest discount available on a P90 was about 23%, and these sorts of discounts were typically vehicles with a significant number of miles and in some cases damage that had been repaired. And there were almost no discounted non-P 90s available at that time (and I believe afterwards as well although I did not follow it that closely). I would expect the number of discounted non-P 90s and the average percentage discount to be quite low.

I am not sure that it matters at the end of the day since your overall projected margins and ASP seem very reasonable, but was just curious of your source/thinking on this.

I agree, I think the level of discount and number of discounted cars is off by quite a lot. 30% of S and 10% of X comes out to about 5500 inventory car cars and I think the number is far, far lower. I very seriously doubt that the entire inventory fleet is that large and of course not all cars were sold. I'd say more like 2500 total cars sold as inventory, with 75% at a discount (remember, the discounts only occurred for about a month) sounds right. I didn't follow X as close but 10% discount rate assumes over 800 Model X inventory sales. I don't think this many were sold as inventory and I know the deals were not very good.

As one of the inventory shoppers obsessively following the S deals, the level of discounts are also way high. The 75 discounts are close enough, I think $7500 to $9000 was about the average in the last couple weeks (less before that). I'd go with 8% but 10% is close enough. The 30% on 90s is way high, though. First, regular 90s were not discounted like the Ps, they were more like the 75s - 10-12% at the very most. The P90Ds pretty much topped out at 20% (which was quite rare) and were only for older pre refresh models with a lot of miles. When I was shopping, my OA told me $25k discount was needed on the loaded $140k cars to get to a lease number under $1000/mo (with $5k+ down) for the refresh models and that such discounts were very rare and disappeared pretty much instantly. That's just a 17% discount. There were waiting lists to grab these cars. Mine ended up being 27k off, or about 19%. The very best I heard from the pre refresh models was about $31k off, which is 22%. I bet maybe 50 fit this description total if I had to give a generous guess. A 30% discount is $42,000 off and I literally didn't hear of one person coming anywhere near this.

If I had to make guesses it would be as follows:

S 75/60: 8% discount, 1080 cars
S 90: 12% discount, 270 cars
S P90D: 20% discount, 360 cars
X models: 8% discount, 200 cars

500 regular discount inventory S/X: $110k ASP (remember, inventory cars used to skew almost all P models) and 3% discount.

 

[EinSV]

I agree, I think the level of discount and number of discounted cars is off by quite a lot. 30% of S and 10% of X comes out to about 5500 inventory car cars and I think the number is far, far lower. I very seriously doubt that the entire inventory fleet is that large and of course not all cars were sold. I'd say more like 2500 total cars sold as inventory, with 75% at a discount (remember, the discounts only occurred for about a month) sounds right. I didn't follow X as close but 10% discount rate assumes over 800 Model X inventory sales. I don't think this many were sold as inventory and I know the deals were not very good.

As one of the inventory shoppers obsessively following the S deals, the level of discounts are also way high. The 75 discounts are close enough, I think $7500 to $9000 was about the average in the last couple weeks (less before that). I'd go with 8% but 10% is close enough. The 30% on 90s is way high, though. First, regular 90s were not discounted like the Ps, they were more like the 75s - 10-12% at the very most. The P90Ds pretty much topped out at 20% (which was quite rare) and were only for older pre refresh models with a lot of miles. When I was shopping, my OA told me $25k discount was needed on the loaded $140k cars to get to a lease number under $1000/mo (with $5k+ down) for the refresh models and that such discounts were very rare and disappeared pretty much instantly. That's just a 17% discount. There were waiting lists to grab these cars. Mine ended up being 27k off, or about 19%. The very best I heard from the pre refresh models was about $31k off, which is 22%. I bet maybe 50 fit this description total if I had to give a generous guess. A 30% discount is $42,000 off and I literally didn't hear of one person coming anywhere near this.

If I had to make guesses it would be as follows:

S 75/60: 8% discount, 1080 cars
S 90: 12% discount, 270 cars
S P90D: 20% discount, 360 cars
X models: 8% discount, 200 cars

500 regular discount inventory S/X: $110k ASP (remember, inventory cars used to skew almost all P models) and 3% discount.

Nice analysis. I did not follow the S60/75 discounts very closely but your discussion on the S90/P90s seems spot on to me. IMO, the 20% discount guesstimate on P90s seems a little conservative (high), if anything.

 

[Fallenone]

@schonelucht @EinSV @esk8mw , thank you all for your input.
A few responses on your comments
1. I sure don't know how much discounts they would get from the suppliers of the glider. But I think this is the most sensitive parameter in my spreadsheet (dropping this to 7.5% results in $0.02 non-GAAP EPS in my spreadsheet with everything else being the same). Hope they can do as best as they can.

2. I am not modeling where the discount will go - GM or SG&A. Not very interested in this detail as I am trying to figure the EPS although I think it would go to SG&A.

3. The discount I give are for the base only, options not discounted. So it's not like a fully optioned SP90D selling for over 130k will be discounted to 91k, more like 103k (this comes much closer to the numbers provided by you guys).

4. I shared this with some other friends outside TMC and they all said my customer uptake rate of discounted cars are too high. The reason I went for these high uptake rate is Tesla's own website is showing customers cheaper, earlier delivery cars of similar choice of options even before the customer goes into the actual order page. I think what happened was, all showroom cars and some test drive cars were sold earlier. Tesla built replacements for them but did not hide these from the inventory list so many (I would think all) of these were sold very quickly, maybe even before hitting the stores. With 200+ stores and 2 cars each. I think there would be 800-1000 cars sold this way. And there are some more from other channels I think. This practice should have stopped too. As I was updating my spreadsheet I was browsing Tesla's website and cannot find any inventory cars suggested before ordering that have lower price.

5. One thing I didn't include previously was the referral discount. This would be 12M in net profit if 50% of the buyers had one. I'm not sure how prevelant this is, and actually I don't know how this would be accounted in the financial statements to be honest. But want to say here to let anyone wants to know.

6. In any case, I changed the numbers of discount % and uptake numbers based on @esk8mw 's numbers (translated them to discount on base price, so S75 still 10%, this translates to roughly 8% on all-in price with options). I also assumed no X90D was sold at a discount. And here are the results with different % of savings from glider

a. 10% savings from glider
non-GAAP EPS $0.69
GAAP (98% of non-GAAP revenue) EPS $0.36
GAAP (80% of non-GAAP revenue) EPS $-0.42

b. 7.5% savings from glider
non-GAAP EPS $0.47
GAAP (98% of non-GAAP revenue) EPS $0.14
GAAP (80% of non-GAAP revenue) EPS $-0.60

c. 5% savings from glider
non-GAAP EPS $0.25
GAAP (98% of non-GAAP revenue) EPS $-0.07
GAAP (80% of non-GAAP revenue) EPS $-0.77

 

[FluxCap]

Thanks for the heads-up on the thread.

 

[techmaven]

I went through and tried to take a gander with the deferred leasing and the deferred revenue. Is it true that:

automotive revenue = vehicles sales + deferred revenue - leasing

Ie. $1,181 million in automotive revenue - $151.6 million previously deferred revenue + $292.65 million deferred revenue current quarter = vehicle revenue?

Therefore vehicle revenue is $1,322.9 million in Q2?

As I run the numbers, I think the Model X had terrible gross margin in Q2. Like 10-11% on 4,638 vehicles. Bringing that up to anywhere close to the S would be a huge difference. I modeled an ASP drop on the S of $2,500 overall and a 1% margin decline, but moved the Model X margin to 15% lower than the Model S. As a result, I have them with a gross profit of $6.5 million from operations. No GAAP reconciliation factored in.

The next big areas of unknown are interest and interest expense. Interest expense in Q2 was $42 million, and I think the ABL had a lot to do with that. They paid down the ABL, and so if they bring down the interest expense to $35 million, and then the interest on the cash on hand might be double or more than what it was last quarter. I can easily see that they would have a slim positive result from operations, but negative with interest expense and taxes. But that's with $0 in ZEV credits.

Add in a modest amount of ZEV credits, like $100 million and my model goes from a loss of $0.18 to a profit of $0.48 per share. I think there is a very good possibility that we see a slim profit from ops and ZEV credits counteracts interest + taxes.

 

[schonelucht]

The next big areas of unknown are interest and interest expense. Interest expense in Q2 was $42 million, and I think the ABL had a lot to do with that. They paid down the ABL, and so if they bring down the interest expense to $35 million, and then the interest on the cash on hand might be double or more than what it was last quarter.

I think they will have used the asset backed line again near the end of the quarter. They have that line of credit specifically for end-of-quarter cars in transit coverage. A that number, we already know, was up Q2 on Q3 to 5500 cars. Otherwise cash position would be approximately $500M less still. Personally, I think the expansion of the ABL is more likely and part of the answer as to why their is no need to tap the capital markets while still keeping spending timelines.

 

[techmaven]

I think they will have used the asset backed line again near the end of the quarter. They have that line of credit specifically for end-of-quarter cars in transit coverage. A that number, we already know, was up Q2 on Q3 to 5500 cars. Otherwise cash position would be approximately $500M less still. Personally, I think the expansion of the ABL is more likely and part of the answer as to why their is no need to tap the capital markets while still keeping spending timelines.

Why would they us the ABL when they raised cash? In Q4, maybe. But they may not have touched it in Q3.

 

[schonelucht]

IIRC, they paid it down only after Q3 while they raised cash in Q2. By the same reasoning why didn't they pay it off even sooner?

 

[Mario Kadastik]

Why are people expecting the S gross margin to decline incrementally? Yes, there was the 60 and some discounts (how many discounted cars of the 24500 do you think Tesla sold? 1k?). The fixed costs were spread over far more cars so I would expect the increase from 14k to 24k to have increased the GM by at least a few percentage points. So even with the 60 and the discounts I would expect the GM to increase by 1-2%. And of course the X GM should climb far more. My guess would be around 20-22%.

So those guesses should bring the earnings closer to 0 even without necessarily accounting for the ZEV and we know there was a large amount of ZEV sold. So my guess is that we'll see GAAP positive EPS, the question is by how much...

 

[techmaven]

Why are people expecting the S gross margin to decline incrementally? Yes, there was the 60 and some discounts (how many discounted cars of the 24500 do you think Tesla sold? 1k?). The fixed costs were spread over far more cars so I would expect the increase from 14k to 24k to have increased the GM by at least a few percentage points. So even with the 60 and the discounts I would expect the GM to increase by 1-2%. And of course the X GM should climb far more. My guess would be around 20-22%.

I am not expecting the overall gross margin to decline materially. Tesla communicated in the last shareholder's letter that even with the 60, they expected gross margin increase by 2-3% through Q4. I do think that they refreshed the demo fleet extensively, which meant a lot of top line vehicles got big discounts. However, that's not likely more than 500 vehicles or so. And if they sold those vehicles at $115k, which a few were, that would mean they sold them at the average gross margin. Many were sold with much less than $30k off. There were delivery discounts and there where 75's sold at between 60 and 75 pricing. Those, probably 300-500 vehicles, were sold at higher gross margin than the 60's, since the 60's are 75's. There were additional discounts, but were relatively a few smattering ones, not material. The S's gross margin should be very good since they also got additional cost efficiencies. I played around with vehicle options pricing and I think the 60's get a gross margin of around 16-18%. The discounted 75's get higher gross margin.

The real story is that X wasn't really discounted beyond normal demo refresh and the crappy margin they probably had on 4,638 or so vehicles in Q2 isn't so bad now. I only modeled 18-20% gross margin on the X.

 

[Fallenone]

Why are people expecting the S gross margin to decline incrementally? Yes, there was the 60 and some discounts (how many discounted cars of the 24500 do you think Tesla sold? 1k?). The fixed costs were spread over far more cars so I would expect the increase from 14k to 24k to have increased the GM by at least a few percentage points. So even with the 60 and the discounts I would expect the GM to increase by 1-2%. And of course the X GM should climb far more. My guess would be around 20-22%.

So those guesses should bring the earnings closer to 0 even without necessarily accounting for the ZEV and we know there was a large amount of ZEV sold. So my guess is that we'll see GAAP positive EPS, the question is by how much...

While I don't think S gross margin will decline this quarter, there are two solid concerns that can support a possible declined gross margin of S: 1) S60, same COGS, lower selling price; 2) discounted sales. And regarding to the fixed costs, the tooling is not fixed cost but variable. So increase of production does not help here. Increase of production may help in Tesla's costs in sourcing the parts (I'm counting on this one) and the fixed cost of the factory itself. But the factory itself is not a substantial part of the COGS. Labor cost is hard to tell. The final assembly line is quite labor intensive and costs may not drop that much on a per car basis in response to increased production.

 

[surfside]

Labor cost is hard to tell. The final assembly line is quite labor intensive and costs may not drop that much on a per car basis in response to increased production.

To try to prove out this question, does anyone know if Tesla had a significant increase in hiring around the end of Q2 / early Q3? I seem to recall folks posting about a big Tesla job fair over the summer, but my recollection could be wrong. If there was no significant increase in job postings / hiring at Fremont, it would seem unlikely that there would be a big increase in labor costs. I personally believe one of the big reasons for the low X margins in Q2 was due to allocation of labor resources to building the Xs in such small quantities.

Curious for others' thoughts...

surfside

 

[Fallenone]

To try to prove out this question, does anyone know if Tesla had a significant increase in hiring around the end of Q2 / early Q3? I seem to recall folks posting about a big Tesla job fair over the summer, but my recollection could be wrong. If there was no significant increase in job postings / hiring at Fremont, it would seem unlikely that there would be a big increase in labor costs. I personally believe one of the big reasons for the low X margins in Q2 was due to allocation of labor resources to building the Xs in such small quantities.

Curious for others' thoughts...

surfside

There was a "massive" hiring event in H1 this year.

I think the low gross margin on X is inexperience caused inefficiencies. Too many not making pass QC initially and ended up being spare parts instead of sales. Would get better with time.

 

[AutoTechist]

Tesla Motors posted impressive third quarter results with highest number of vehicle deliveries!