Near-future quarterly financial projections

[EVNow]

When I was volunteering at the local delivery center at the end of last month, it became clear to me that about half the people there were temp staff (not counting the volunteers). I wonder if they are counted in that 45k, when the snapshot was taken?

Most companies don't count temps when they talk about employee counts.

 

[RobStark]

474 mail room workers? Uhhhh

More like a snap shot of General Motors circa 1962.

 

[ThereAre4Cars]

Musk suggested at shareholder meeting in May their battery costs were connected to market prices of raw materials:

"We think at the cell level probably we can do better than $100/kWh maybe later this year … depending upon [stable] commodity prices"

Cobalt spot price has dumped this year (to my surprise): from 40$\lb in June to 28$\lb today: Cobalt Prices and Cobalt Price Charts - InvestmentMine

Nickel has dropped as well: From 15.5k in June to 12.5k today: Nickel PRICE Today | Nickel Spot Price Chart | Live Price of Nickel per Ounce | Markets Insider

Lithium has dropped as well: From 130 to 117. Lithium | 2016-2018 | Data | Chart | Calendar | Forecast | News

Not sure what graphite did.

Those 4 are definitely the most significant inputs. That's pretty monumental importance. Back when I was less lazy I figured out the raw materials were ~50$-60$\kWh. I lost that work, but if you take 10% off that and get 5$\kWh, you are looking at 400$ a car or up to 30M$ for the quarter. That's worth even more than an ill-considered tweet.

 

[RobStark]

This is for large flake battery grade graphite.

 

[Heimdallr]

Musk suggested at shareholder meeting in May their battery costs were connected to market prices of raw materials:

"We think at the cell level probably we can do better than $100/kWh maybe later this year … depending upon [stable] commodity prices"

Cobalt spot price has dumped this year (to my surprise): from 40$\lb in June to 28$\lb today: Cobalt Prices and Cobalt Price Charts - InvestmentMine

Nickel has dropped as well: From 15.5k in June to 12.5k today: Nickel PRICE Today | Nickel Spot Price Chart | Live Price of Nickel per Ounce | Markets Insider

Lithium has dropped as well: From 130 to 117. Lithium | 2016-2018 | Data | Chart | Calendar | Forecast | News

Not sure what graphite did.

Those 4 are definitely the most significant inputs. That's pretty monumental importance. Back when I was less lazy I figured out the raw materials were ~50$-60$\kWh. I lost that work, but if you take 10% off that and get 5$\kWh, you are looking at 400$ a car or up to 30M$ for the quarter. That's worth even more than an ill-considered tweet.

Thanks, great insight! I wasn't really considering raw material costs because my thinking was that if there was a huge rise or drop in costs, we would probably hear about it. The $5 reduction per kWh from lower commodity pricing makes it even more likely that Tesla is going to go below $100 per kWh on the cell level by the end of this year. It's also interesting to know that Tesla specifically mentions commodity prices. I was under the impression that due to their massive long-term supply contracts, they aren't susceptible to sudden price changes in lithium for example.

 

[Fact Checking]

When I was volunteering at the local delivery center at the end of last month, it became clear to me that about half the people there were temp staff (not counting the volunteers).

I'm wondering whether they were true temps, or Tesla employees brought in temporarily for the '7 day delivery push'?

Since the Fremont factory didn't do a production rush at the end of the quarter, it could possibly have freed up factory workers to volunteer for a week in key delivery centers.

 

[luvb2b]

i did find that model... all-in opex grows by about 28-38k per employee. i guess the rest of the costs must go in cogs.

just using that crude number vs. 17q4 reporting, i'd expect ~200-300m more in opex vs. 17q4 levels. add 50m for elon's stock-based comp as well as that was a new thing in 18q1. so the expected steady state run rate of opex would be i think 1.3-1.4b per quarter.

you won't necessarily see the full amount this quarter depending on when hires were made, but there is the possibility of a 100m upside surprise vs my opex estimate.

Here's a pretty good analysis of the structure of Tesla's labor costs:

Peeking Behind Tesla’s Labor Curtain | CleanTechnica​

It has an extensive list of sources and clearly laid out methodology. Here's the estimated composition of labor, by type:

That's with a headcount of 37.5K.

So if those numbers are accurate, and if most new hires were related to 'making cars', then that's an opex increase of about ~$200m per quarter. If these new hires started in early Q3 and the 45k headcount was for end of Q3, and if the headcount increase was linear, this is an opex upside of ~$100m for Q3 and ~$200m for Q4.

 

[brian45011]

i did find that model... all-in opex grows by about 28-38k per employee. i guess the rest of the costs must go in cogs.

just using that crude number vs. 17q4 reporting, i'd expect ~200-300m more in opex vs. 17q4 levels. add 50m for elon's stock-based comp as well as that was a new thing in 18q1. so the expected steady state run rate of opex would be i think 1.3-1.4b per quarter.

you won't necessarily see the full amount this quarter depending on when hires were made, but there is the possibility of a 100m upside surprise vs my opex estimate.

Might want to add another $13 million to one time SG&A expense for 4Q18

Oregon claws back $13 million from Tesla over inflated solar credits, report says

Not sure how the claw back washes through the VIEs.

 

[brian45011]

So this is what they wrote in the Q2 update letter:

"Cash outflow from operating activities in Q2 2018 was $130 million, which was significantly better than outflows of $398 million in Q1. This improvement occurred despite a substantial increase in finished goods vehicle inventory of $579 million as a result of the timing of deliveries. Model 3 gross profit excluding non-cash items shifted from negative in Q1 to positive in Q2, driving significant improvement in cash profitability. Additionally, significant improvement in our other working capital needs helped to mitigate the impact of inventory growth."​

Good point! Here's that line:

                                                                  Q2         Q1
Automotive gross profit excluding SBC and ZEV credit – non-GAAP   $704,225   $504,188

So the "excluding non-cash items" wording is probably simply an expression that the Model 3 could already generate significant cash flows in Q2 already. The "positive in Q2" is probably around 5%.

There are several other factors, beyond improvement in M3 GM%, that contribute to the Q over Q increase in non-GAAP Automotive Gross Profits:

Non-Zev Regulatory Credit Sales (GHG/CAFE) increased by $23.7 MM.
The net change in pre-existing warranties improved by $48.6 MM.
Deferred Revenue recognition grew by $16.4 MM.
Automotive Leasing Gross Profit increased by ~$34 MM

 

[Fact Checking]

There are several other factors, beyond improvement in M3 GM%, that contribute to the Q over Q increase in non-GAAP Automotive Gross Profits:

Non-Zev Regulatory Credit Sales (GHG/CAFE) increased by $23.7 MM.
The net change in pre-existing warranties improved by $48.6 MM.
Deferred Revenue recognition grew by $16.4 MM.
Automotive Leasing Gross Profit increased by ~$34 MM

Indeed - and that's a major oversight in my quick estimate:

• in Q2 22.3k S+X and 18.4k 3's were delivered, so assuming similar GHG/CAFE per unit credits earned, the split would be 55% for the S+X, i.e. 55%*$23.7m = $13m of the gross profit is not 3 related,
• it's unclear (to me) how warranty reserves split, but assuming a rough estimate of 50%/50% (S+X has higher ASP but probably better expected reliability at this early stage of the 3's production) that would be another $24.3m of gross profit not 3 related,
• deferred revenue recognition is probably mostly related to leasing, I didn't see any AutoPilot revenue recognition language in their Q2 report, so this would be $16.4m non-3 related gross margin,
• leasing would be exclusively S+X, so another $34m non-3 income.

That's 13+24.3+16.4+34=87.7m off the gross profit, i.e. this should make the Model 3 gross margin improvement about $100m - with error bars of several percentage points.

That's a cash gross margin of ~$5,400 per unit in Q2, which with its ASP of ~56k would be a gross Model 3 cash margin of 9-10% in Q2.

Do you agree with this estimate, or is it still too inaccurate?

 

[brian45011]

Indeed - and that's a major oversight in my quick estimate:

• in Q2 22.3k S+X and 18.4k 3's were delivered, so assuming similar GHG/CAFE per unit credits earned, the split would be 55% for the S+X, i.e. 55%*$23.7m = $13m of the gross profit is not 3 related,
• it's unclear (to me) how warranty reserves split, but assuming a rough estimate of 50%/50% (S+X has higher ASP but probably better expected reliability at this early stage of the 3's production) that would be another $24.3m of gross profit not 3 related,
• deferred revenue recognition is probably mostly related to leasing, I didn't see any AutoPilot revenue recognition language in their Q2 report, so this would be $16.4m non-3 related gross margin,
• leasing would be exclusively S+X, so another $34m non-3 income.

That's 13+24.3+16.4+34=87.7m off the gross profit, i.e. this should make the Model 3 gross margin improvement about $100m - with error bars of several percentage points.

That's a cash gross margin of ~$5,400 per unit in Q2, which with its ASP of ~56k would be a gross Model 3 cash margin of 9-10% in Q2.

Do you agree with this estimate, or is it still too inaccurate?

Probably in the ballpark

Tesla's terminology in seemingly inconsistent statements about Q2 M3 GM% and guidance for Q3 is too cryptic to have much confidence in the accuracy of anyone trying to unpack it precisely.

$37.1 MM of the warranty reserves difference relates to the change in revenue recognition accounting standard for RVG vehicles; the other ~$11 MM may relate to Roadsters/S/Xs, whose warranties are expiring.

The increase in Depreciation and Amortization between quarters was $69 MM. Since S & X production capacity is relatively flat, most of the increase probably relates to new plant and equipment placed in service for M3 production capacity.

It's unclear how Tesla allocates regulatory credit sales among models. IMO, all regulatory credit sales should be reported in Other Income to reduce confusion about trends in GM% for products sold to its retail customers.

 

[Fact Checking]

The increase in Depreciation and Amortization between quarters was $69 MM. Since S & X production capacity is relatively flat, most of the increase probably relates to new plant and equipment placed in service for M3 production capacity.

So in Q1 and Q2 they had capex cash outflow of $655m and $610m. If we assume that most of that is Model 3 related, then they could be installments of equipment put into service previously.

The way I understand it is that as they make the payments they capitalize the increase in "owned" assets and put that on a depreciation/amortization schedule - most of it on a straight line basis.

If the average amortization period is say 3 years (12 quarters), then the extra capital of +~$655m in Q1 would add about ~$55m to straight line amortization cost starting in Q1 and lasting 12 quarters. That's in the ballpark of the $69m increase.

Total equipment on the balance sheet started going up in late 2016, muddied by the SCTY merger somewhat. If they have 2-3 years of amortization schedules then this drag on GAAP income would start decreasing sometime in 2019. (Of course only if there's no new capex spending - which won't happen.)

BTW., Model X expansion depreciation/amortization costs might start running out just about now, if they are on a 2-3 years schedule. (Which they might not.)

Has Tesla disclosed amortization schedules in one of the reports perhaps?

 

[Singer3000]

Just re-reading Macquarie, who have an H2 base case of $50m ZEV Credit income but a ceiling of $600m. What is the view here on whether there might be a ZEV upside surprise? To what extent would this sort of upside surprise info already be sitting in the lap of Mr Market, given necessarily there are multiple external parties involved (at least one buyer and presumably the state administrator)?

 

[SBenson]

Tesla launches new Model 3 with ‘mid-range’ battery for $45,000, changes pricing structure

So this bolsters the short thesis that order book ran out at end of Q3. I expected some form of demand levers to be pulled. But this seems a bit drastic as it lowers the entry price by $4K.

Also, FSD is gone... That is quite something. Well, in perspective, its a miracle it ever was a purchase option at all. It had Theranos written all over it. At some point I genuinely started feeling it is fraud because my 9yr old son truly believed that Tesla's could drive themselves anywhere as he saw the clip right on the website - with no warning, no footnote ever saying its a demo video.

 

[generalenthu]

Tesla launches new Model 3 with ‘mid-range’ battery for $45,000, changes pricing structure

So this bolsters the short thesis that order book ran out at end of Q3. I expected some form of demand levers to be pulled. But this seems a bit drastic as it lowers the entry price by $4K.

Also, FSD is gone... That is quite something. Well, in perspective, its a miracle it ever was a purchase option at all. It had Theranos written all over it. At some point I genuinely started feeling it is fraud because my 9yr old son truly believed that Tesla's could drive themselves anywhere as he saw the clip right on the website - with no warning, no footnote ever saying its a demo video.

1. Panasonic clearly said they were a bottleneck.
2. There are indications (from Fred@electrek) that Tesla is comfortably able to go past 5k per week, but obviously no cells.
3. Why not send it to Mexico or Australia if they need to sell at higher ASP? Coz, no cells.

 

[neroden]

I'm familiar with the concepts although not in any detailed manner. But how big is the fixed cost? If you look at the cash flow statements from Q2 of 2017 to Q2 of 2018 the Depreciation & Amortization line went up only 96M$, so that's before model 3 and after. Let's say the equipment is depreciated over 5 years so that's 20 quarters and that 100M$ would add up to 2B$ which sounds ballpark reasonable. If that's approximately correct then delivering 20k is 5k$ per unit, 50k is 2k$ per, and 100k is 1k$ per. In other words, after Q3 there isn't a significant dilution remaining.

What do you figure is the size of this fixed number and what source do you get it from?

I'm sorry, but I don't have a current estimate. I tend to back-compute it from the financial statements using whatever additional information I have. Thankfully financial statements do separate out most of the fixed costs; figuring out when they're putting fixed costs in the variable section or vice versa is tricky.

I would go along with your guess regarding depreciation; it seems sound enough in the general scheme of things.

Also since you have accounting skills. Is there any chance the S/X margins improve simply because the model 3 reuses quite a bit of the same capital equipment (Body shop, paint, Factory itself, etc)? I don't think this has been included in any model I've seen.

In a word, yes.

I mean fundamentally when I think bottom-up about it, I have a hard time understanding why the 35k$ would have such a hard time being profitable when a Camry starts at 24k$. The battery might cost 7.5k$ at MOST, but then the relative complexities\cost\sheer material of everything else *seems* about the same.

 

[neroden]

Here's a pretty good analysis of the structure of Tesla's labor costs:

Peeking Behind Tesla’s Labor Curtain | CleanTechnica​

It has an extensive list of sources and clearly laid out methodology. Here's the estimated composition of labor, by type:

That's with a headcount of 37.5K.

So if those numbers are accurate, and if most new hires were related to 'making cars', then that's an opex increase of about ~$200m per quarter. If these new hires started in early Q3 and the 45k headcount was for end of Q3, and if the headcount increase was linear, this is an opex upside of ~$100m for Q3 and ~$200m for Q4.

This is... fascinating? My immediate conclusion is that there are too many employees at stores and too many solar sales, and not nearly enough service mechanics. And mail room? Seriously? That should be "Information Technology", surely? This data seems to be mostly educated guesswork and should be taken with a large grain of salt.

 

[neroden]

Musk suggested at shareholder meeting in May their battery costs were connected to market prices of raw materials:

"We think at the cell level probably we can do better than $100/kWh maybe later this year … depending upon [stable] commodity prices"

Cobalt spot price has dumped this year (to my surprise): from 40$\lb in June to 28$\lb today: Cobalt Prices and Cobalt Price Charts - InvestmentMine

Nickel has dropped as well: From 15.5k in June to 12.5k today: Nickel PRICE Today | Nickel Spot Price Chart | Live Price of Nickel per Ounce | Markets Insider

Lithium has dropped as well: From 130 to 117. Lithium | 2016-2018 | Data | Chart | Calendar | Forecast | News

Not sure what graphite did.

Those 4 are definitely the most significant inputs. That's pretty monumental importance. Back when I was less lazy I figured out the raw materials were ~50$-60$\kWh. I lost that work, but if you take 10% off that and get 5$\kWh, you are looking at 400$ a car or up to 30M$ for the quarter. That's worth even more than an ill-considered tweet.

Tesla does buy most of its minerals on long-term contracts, not on the spot market. But this will help their cost structure going forward. Likely will not be a dramatic, instant effect though.

 

[neroden]

Might want to add another $13 million to one time SG&A expense for 4Q18

Oregon claws back $13 million from Tesla over inflated solar credits, report says

Not sure how the claw back washes through the VIEs.

Based on one of the *main functions* of the VIEs, which is to pass tax credits through to entities which can actually use them, I expect nearly all of this will end up being allocated to non-controlling interests, unless SolarCity had some kind of deal to compensate them if the government clawed back the money.

 

[Fact Checking]

Tesla does buy most of its minerals on long-term contracts, not on the spot market. But this will help their cost structure going forward. Likely will not be a dramatic, instant effect though.

Yeah. Also note that Tesla secured two long-term lithium-carbonate supply deals within the past couple of months (with enough lithium supply for millions of 80kWh battery packs for the next 3-6 years), one of which contract is delivering this year already - which deals probably already took advantage of the (predicted) drop in spot lithium prices.

Which was a good move I think, because while there's a lot of new lithium capacity coming online in the next two years, I think it's not nearly enough to support all the EV demand that is present, and then we haven't even included the GWh of sales that Tesla Energy is going to generate ...