Near-future quarterly financial projections

[mongo]

I’m layman and I don’t know, how accurate these stats
Tesla Shines In Global Plug-In Electric Car Sales Stats For July
are, but according to those, Tesla sold 5976 s+x combined in july. So with that pace combined s+x sales q3 would be 17 928. Your model predicts combined s+x sales 27 000. I don’t have any competence to evaluate your number, but I just point out, that the difference is large.

Plant was shut down for the 4th of July weekend. So July numbers are not a good monthly representation of production (I know you are talking sales). 12 weeks at 2k rate is closer to the 27k number than 18k.

If they draw down the 3,900 in transit number, that will help too. The 5,976 number implies only about 2k of new July production was sold. Seems low.

 

[Esme Es Mejor]

I’m layman and I don’t know, how accurate these stats
Tesla Shines In Global Plug-In Electric Car Sales Stats For July
are, but according to those, Tesla sold 5976 s+x combined in july. So with that pace combined s+x sales q3 would be 17 928. Your model predicts combined s+x sales 27 000. I don’t have any competence to evaluate your number, but I just point out, that the difference is large.

First months of quarters always have lower sales, so assuming that pace for August & September will lead to a massive underestimate. Instead, compare the first month of Q3 (July) with the first months of Q1 & Q2 (Jan & Apr).

I don’t have global numbers, just US numbers, so I’m using those. If someone has access to global sales numbers for Jan & Apr 2018, please let me know.

In the US:

January 1500 S/X
April 2275 S/X
July 2525 S/X

All numbers from InsideEVs. July’s numbers suggest higher S/X sales in Q3 than in Q1 or Q2. How much higher I don’t know, but 26-27k seems reasonable.

 

[Fact Checking]

Is this based in the belief that Tesla reduced S/X deliveries in Q2 for tax credit reasons and pushed them into Q3? Q2 2017 was 22000, while Q2 2018 was 22264, so I don't think this is justified. Or is it simply based on the expected 100K/year, and the difficulty of putting out 31000 cars in Q4? Because that would seem hard....

Tesla is 18650 limited for the Model S/X, and I think they made it pretty clear a number of times that they could make more than 100k/year in Fremont - and they probably will once they move to 2170 batteries.

Meanwhile 27k for Q3 doesn't sound implausible to me. +2k of SX would mean an additional +$212m of revenue and it would also improve cash flow and income by about +$53m.

So if Tesla can do that then Q3 would be the right moment to demonstrate it.

This would also explain the slightly lower M3 production guidance: the paint shop is shared between the M3 and the MSX, max is around 6k-7k/week, which adds a ceiling of somewhere around 80k-ish units.

So yes, I think this is what is happening in Q3.

 

[Fact Checking]

other than 9% of the workforce, but that was tin the mid of the quarter, so effect would be only about 50%

I believe the 9% layoffs were announced at around June 9th, and IIRC some were effective only in July.

I.e. the layoffs aren't in Q2 at all, they should affect Q3.

That is partially countered in Q3 by new hirings: line workers for the M3 lines, and for sales (deliveries).

 

[Fact Checking]

in some cases they don't even have to get the direction right to get paid, as long as the stock underperforms the index they still get paid.

How would that work? Here's how the calculation goes I think, please point out mistakes:

Fund manager break-even performance is (roughly) measured against index returns: the long term S&P 500 return is almost 10% (before taxes).

If the manager goes long:

• initial position ties up 100% of cash
• stock rises 10% annual, long term average

When going short:

• short position collateral ties up about as much cash as when going long, so the cash opportunity cost is similar,
• a stock borrowing fee is paid, which is around 2% for TSLA,
• this is partially cancelled out by cash interest rate income the broker pays on the cash collateral (say 1%), which totals to a net 1% interest cost
• the short position pays money if the stock's price is falling. The short position's returns are break-even with a long position if the stock's price falls by 11%.
• a low volatility stock's price will fall by 10% roughly if the company underperforms S&P 500 companies by 10%. (Assuming equilibrium growth and price) For high-growth stocks even a minor miss will be leveraged by the drop in growth expectations.

This looks mostly symmetric, but note how time is working against short positions: every quarter the company performs average it will follow the market with a ~2.5% rise in the price.

The more the timing of the short thesis is off, the harder it gets to get purely break-even with a long position: 1 year delay means a 10% increase, so the subsequent drop has to be 22%. (Interest costs increase with time.) After 2 years it's 33%.

The long position has no interest costs, plus it will rise 10% with an average company.

So the "reliable" way to make money by shorting is to:

• Every new short share creates a new long share out of thin air, which has two effects:
  • it decreases the share price by selling pressure,
  • plus it dilutes the float, which increases volatility, as more longs means a larger pool of investors can panic-sell on a negative event,
• This means it makes sense for shorts to "herd": more shorts against the same company depresses the price and increases the eventual drop: win-win. 30% of TSLA is shorted, these effects are significant.
• Note how short sellers like to talk about their positions after establishing them, they are pimping for the herding effect. Longs don't "need" the herding nearly as much: just average execution gives 10% return.
• On significant herding such as TSLA, the drop in share price has partial self-fulfilling effects:
  • worse equity financing,
  • more difficult retention of key employees, who are often stock options compensated,
  • lower effective wages for employees with equity compensation, which increases opex,
  • lower credit rating in extreme cases
• Shorts are picking low market cap, high growth, high volatility stocks,
• Companies with significant debt structure (such as leasing programs) are targeted, because they are more exposed to credit downgrades or to uncertainty in general,
  
• If the short thesis does not materialize fast enough, in today's "facts are strictly optional" media world negative news can be magnified or simply made up.
• Note the additional collateral damage persistent levels of FUD can cause: 2-3 years of non-stop news about imminent Tesla bankruptcy has material negative effects on customers, suppliers, partners and sources of financing. I believe it's a small miracle that Tesla is able to sell so well despite the level of FUD against it.

This is what happened to Solar City, and this is how Tesla inherited this aggressive, notorious pack of NY hedge funds (Chanos & co) shorting it - and that attracted even more shorts.

Complaining about shorts is entirely justified IMO, because the damage they are inflicting is significant. Elon also knows how corporate life without shorts is: he has overseen SpaceX growing from zero to $25b+ valuation.

 

[Fact Checking]

but there were other things too that should have reduced opex, barnacle scrub, expense scrutiny, etc.

I understood those in large part as capex reduction measures: a lot of contractors tend to be in construction and in the installation and bringing up of new equipment, and in troubleshooting: work that Tesla might not have in-house expertise for and has no need to have permanent positions for.

Tesla has reduced capex guidance for 2018 significantly.

Also, even if it impacts opex, it was done very late in Q2.

 

[Fact Checking]

Problem with the coming recession is that it will originate in the tech sector which tesla is in.

This would actually be helpful to Tesla: while Tesla operates in the tech sector in terms of corporate culture, R&D and valuation, 90%+ of its sales are in the automotive, housing and energy sectors.

Check out vehicle sales during the 2000 dot-bomb recession:

Light Weight Vehicle Sales: Autos and Light Trucks

U.S. car sales didn't drop nearly as much as during 2008, and recovered much faster. I don't think Europe saw a measurable drop.

EV sales could also have additional levels of recession resistance: the segment is so supply limited to such extreme levels that it might be able to weather a moderate non-financial and non-housing recession without much of a drop in sales.

A housing boom (and U.S.-wide bust) is still many years away.

The main Tesla-negative in a high-tech bust that I can see is that Tesla sales have a higher weight in California and on the west coast, and California would be on the frontline of drop in demand.

 

[Causalien]

This would actually be helpful to Tesla: while Tesla operates in the tech sector in terms of corporate culture, R&D and valuation, 90%+ of its sales are in the automotive, housing and energy sectors.

Check out vehicle sales during the 2000 dot-bomb recession:

Light Weight Vehicle Sales: Autos and Light Trucks

U.S. car sales didn't drop nearly as much as during 2008, and recovered much faster. I don't think Europe saw a measurable drop.

EV sales could also have additional levels of recession resistance: the segment is so supply limited to such extreme levels that it might be able to weather a moderate non-financial and non-housing recession without much of a drop in sales.

A housing boom (and U.S.-wide bust) is still many years away.

The main Tesla-negative in a high-tech bust that I can see is that Tesla sales have a higher weight in California and on the west coast, and California would be on the frontline of drop in demand.

I believe both luvb2b and me are pessimistic about the prospect of the recession. It's better not having one or not being in a cash crunch when it hits. It'd be great to be recession resistent, but for bankroll management, I will prepare for the worst case scenario.

 

[Reciprocity]

InsideEVs estimates have a tendency to show low deliveries in the first month of a quarter and high deliveries in the third month of a quarter. (Tesla only releases quarterly data, so InsideEVs numbers are guesswork. In this case they're running off Jose Pontes's sources, which are still guesswork but tend to be pretty accurate.)

So far it looks like US deliveries for S & X are slightly ahead of deliveries last year, so barring something really weird going on internationally, we should expect Q3 to look like Q3 last year. Q3 S&X global deliveries last year was 25930, so 27000 does seem a little high. Is this based in the belief that Tesla reduced S/X deliveries in Q2 for tax credit reasons and pushed them into Q3? Q2 2017 was 22000, while Q2 2018 was 22264, so I don't think this is justified. Or is it simply based on the expected 100K/year, and the difficulty of putting out 31000 cars in Q4? Because that would seem hard....

I'd probably conservatively estimate 26000 S & X delivered (same as last year). Anyway, we'll see in a month.

Tesla is making 2500/w S/X this qtr. Could see deliveries over 30k with the backlog from Q2-200K.

 

[neroden]

So Fred's latest leaked production numbers at Electrek are indicating about 2100 Model S/X per week. This comes out to about 27000 per quarter, so I guess luvb2b's model is probably correct here.

 

[ValueAnalyst]

So Fred's latest leaked production numbers at Electrek are indicating about 2100 Model S/X per week. This comes out to about 27000 per quarter, so I guess luvb2b's model is probably correct here.

assuming it doesn't accelerate further.

 

[bdy0627]

Tesla is making 2500/w S/X this qtr. Could see deliveries over 30k with the backlog from Q2-200K.

I'm seeing about 25,600 S/X deliveries for Q3. According to this latest Electrek article, after 2 months of Q3 production,
"Tesla produced about 53,000 vehicles including over 34,700 Model 3 vehicles in Q3 as of Friday (Aug 31)."
In the last week of production,
"the automaker built about 6,400 vehicles during the last week (last 7 days) of August (from 24th to 31st midday) including about 4,300 Model 3 vehicles."
Tesla misses Model 3 production goal of 6,000 units per week, but on track for overall Q3 goal

53,000 - 34,700 = 18,300 S/X so far. That's a weekly production rate of 2,033 if we assume 9 weeks of production thus far. If we subtract 1/2 week of production due to the July 4th holiday week, that yields a weekly production rate of 2,150. With the Labor Day holiday, we have about 3 1/2 weeks of production left to go in the quarter. Let's assume 2,100 production per week going forward, which was the production rate for the last week in August. That amounts to 7,300 for the rest of September. Adding this to 18,300 gives a total production of 25,600 for the quarter.

What about the in transit vehicles from Q2? There were 3,892 S/X in transit from Q2, so we can add those to the deliveries but we also have to subtract the in transit vehicles at the end of Q3. What will that number be? The in transit number from Q3 2017 was 4,820. There were 3,500 in transit at the end of Q2 2017. We have to just guess at a reasonable in transit number that will ultimately prove wrong, but let's assume 3,892 S/X in transit at the end of this quarter. That's quite a bit lower than Q3 2017, but it seems plausible and makes the math simple. That would offset the in transit from Q2. This means total S/X deliveries should be around 25,600. I think that's a very reasonable estimate, with the actual number probably falling somewhere between 25,000 - 26,000. LUV is assuming 27,000, so that is really close but may prove to be slightly optimistic. If it ends up being 26,000 rather than 27, that reduces revenue by about $100M.

 

[geneclean55]

I'm seeing about 25,600 S/X deliveries for Q3. According to this latest Electrek article, after 2 months of Q3 production,
"Tesla produced about 53,000 vehicles including over 34,700 Model 3 vehicles in Q3 as of Friday (Aug 31)."
In the last week of production,
"the automaker built about 6,400 vehicles during the last week (last 7 days) of August (from 24th to 31st midday) including about 4,300 Model 3 vehicles."
Tesla misses Model 3 production goal of 6,000 units per week, but on track for overall Q3 goal

53,000 - 34,700 = 18,300 S/X so far. That's a weekly production rate of 2,033 if we assume 9 weeks of production thus far. If we subtract 1/2 week of production due to the July 4th holiday week, that yields a weekly production rate of 2,150. With the Labor Day holiday, we have about 3 1/2 weeks of production left to go in the quarter. Let's assume 2,100 production per week going forward, which was the production rate for the last week in August. That amounts to 7,300 for the rest of September. Adding this to 18,300 gives a total production of 25,600 for the quarter.

What about the in transit vehicles from Q2? There were 3,892 S/X in transit from Q2, so we can add those to the deliveries but we also have to subtract the in transit vehicles at the end of Q3. What will that number be? The in transit number from Q3 2017 was 4,820. There were 3,500 in transit at the end of Q2 2017. We have to just guess at a reasonable in transit number that will ultimately prove wrong, but let's assume 3,892 S/X in transit at the end of this quarter. That's quite a bit lower than Q3 2017, but it seems plausible and makes the math simple. That would offset the in transit from Q2. This means total S/X deliveries should be around 25,600. I think that's a very reasonable estimate, with the actual number probably falling somewhere between 25,000 - 26,000. LUV is assuming 27,000, so that is really close but may prove to be slightly optimistic. If it ends up being 26,000 rather than 27, that reduces revenue by about $100M.

It looks like Fred's report was inclusive up to midday Friday. So maybe need to add 1 more shift of S/X production to those numbers? Also, I read somewhere that production wont necessarily shutdown due to the long holiday weekend. There are workers who will take those holiday hours. It might be at a reduced rate though.

 

[bdy0627]

It looks like Fred's report was inclusive up to midday Friday. So maybe need to add 1 more shift of S/X production to those numbers? Also, I read somewhere that production wont necessarily shutdown due to the long holiday weekend. There are workers who will take those holiday hours. It might be at a reduced rate though.

If we add back the 1/2 week of production for September removed for the holiday, there could be an additional 1,000 produced. That would bring up the total to 26,600. The other thing to consider that could bring the number down is if Tesla is able to ramp model 3 to 6,000/week. That might have a "cost" to S/X production. There have been suggestions that the paint shop does not currently allow 8,000 vehicles per week. They might have to limit S/X production to accomplish higher 3 production.

 

[ValueAnalyst]

Tesla is making 2500/w S/X this qtr. Could see deliveries over 30k with the backlog from Q2-200K.

Could US take "over 30k" units in one quarter, given that Norway has been unusually low in July & August? Not sure about other markets.

 

[generalenthu]

Could US take "over 30k" units in one quarter, given that Norway has been unusually low in July & August? Not sure about other markets.

China should be weaker as well after the tariffs. Hope this mess gets reversed soon. There's a bit of pulling forward of demand due to the US tax credit winding down after q4.

 

[Reciprocity]

So Fred's latest leaked production numbers at Electrek are indicating about 2100 Model S/X per week. This comes out to about 27000 per quarter, so I guess luvb2b's model is probably correct here.

29.5 + 17.5 = 47k

17.5/7 = 2500/w up to last with 1 week down out of 8 total and 5 full weeks remaining.

5 x 2500 = 12,500 + 17,500 = 30,000.

This is based on the pre-friday update but all there have been some crazy number of VINs register over the last 3 qtrs, at times sustaining 3k/w and at others, under 1500 but in total more then 2500/w. Something is amiss.

53000 - 34700 = 18300 S/X

18300 - 17500 = 800 in one week.. something stinks with these leaks. 2500 a week for 7 weeks then 800. Maybe the leaker just not being precise with S/X.

Time will tell.

Original article..

Tesla is on track for a production record this quarter but slightly behind ambitious Model 3 goal

New article

Tesla misses Model 3 production goal of 6,000 units per week, but on track for overall Q3 goal

 

[Fact Checking]

18300 - 17500 = 800 in one week.. something stinks with these leaks. 2500 a week for 7 weeks then 800. Maybe the leaker just not being precise with S/X.

I believe the apparent discrepancy comes from two probable facts:

• there's only 5 days of production added to the second article, not 7,
• the first number was probably closer to 17,000, not 17,500

These together give S/X incremental production of 1,300 for 5 days, which extrapolates to about 1,820 for 2 days - much closer to the 2,100 average in Q3.

The error bars are also significant:

• if due to rounding to 100 the real number for S/X is 1,400, that extrapotes to 1,960 per week
• if the Friday snapshot missed the night shift, that would mean the real interval was 4.666 days - which increases the weekly rate to 2,100

So I don't think there's necessarily any discrepancy, we are just dealing with large error bars.

 

[Reciprocity]

I believe the apparent discrepancy comes from two probable facts:

• there's only 5 days of production added to the second article, not 7,
• the first number was probably closer to 17,000, not 17,500

These together give S/X incremental production of 1,300 for 5 days, which extrapolates to about 1,820 for 2 days - much closer to the 2,100 average in Q3.

The error bars are also significant:

• if due to rounding to 100 the real number for S/X is 1,400, that extrapotes to 1,960 per week
• if the Friday snapshot missed the night shift, that would mean the real interval was 4.666 days - which increases the weekly rate to 2,100

So I don't think there's necessarily any discrepancy, we are just dealing with large error bars.

Even if the first number Is 17100 / 7 weeks of actual production, that's 2442. 5 weeks remaining at 2440 = 29300 total for S/X. Large error bars and all, something is going on with S/X that is being over looked. Also, the in-transit numbers are 11k+ for model 3 alone, there were almost 4k S/X also in transit. It's highly likely that 30k+ deliveries happen for model S/X over the next two qtrs. I know that there is a thought that Tesla is production limited to 100k S/X because of cell production, but there is no reason for that to be written in stone, for example Panasonic could have agreed to step it up a bit for the end of the tax credits push. Lastly Tesla is behind the 105k Pace from last year (just over 44k delivered Q1/Q2) by about 60k total, so they need to deliver 30k each quarter at a minimum. Why people think Tesla had no plan for the 7500 phase out is beyond me, as they have had a plan for every tax credit/incentive around the world since day one. All I'm saying is that something is going on with S/X for the second half of the year and many are missing it by focusing 100% in model 3 numbers.

5000 extra model S/X each qtr is $1B+ in revs and ~$300M in gross margin. Pretty important when you want to be profitable. I also assume higher ASP and higher margin due to model 3 eroding the demand for cheaper, lower margin model S 75Ds. Also this might be when X finally surpasses S in total deliveries for the last two quarters. This is also bolstered by more X than S VINs registered this year and I'm general a huge glut in s/X VINs registered in general, averaging about 2500/w.

126918 - 89999 = 36919 / 29 weeks = 1273/w
282640 - 246329 = 36311 / 29 weeks = 1252/w
2525/w over the last 29 weeks. I don't have data going back farther then Feb 15.

Extrapolate that to even just 48 weeks it's over 120k VINs. Why so many over such a long period of time if there are no intentions of making the cars?

Feb 15

Model S:
Highest: 246329
1 Day Range (max-min): 47
7 Day Range (max-min): 1427

Model X:
Highest: 89999
1 Day Range (max-min): 26
7 Day Range (max-min): 1027

Yesterday

Model S:
Highest: 282640
1 Day Range (max-min): 452
7 Day Range (max-min): 2479

Model X:
Highest: 126918
1 Day Range (max-min): 284
7 Day Range (max-min): 2237

On a side note.. recent large spike in model S/X VINS. Nearly 7k in the last week or so. But for a month before that, it was under 2k/w total for S/X for some time.

 

[neroden]

I am beginning to suspect that Panasonic Japan has found a way to make a few more 18650s and remove the production constraint on the S & X. The hints seem to be piling up.