TSLA Market Action: 2018 Investor Roundtable

[Fact Checking]

Yah, I'm not sure. My (less that perfect) memory of previous conference calls makes me think they should have already paid their suppliers for the existing line. GF3 should be showing up, it seems.

Would the equipment for Shanghai, Semi, or Y show up already (deposits)?

So my problem with GF3 is timing:

• they issued the capex and profitability guidance for the rest of 2018 in May already,
• but GF3 was not finalized until August IIRC,
• plus I find it unlikely they'd have chosen in August to burden Q3 with GF3 deposits - why not wait just 1.5 more months until Oct 1, to not load it into the critical Q3 time frame?
• Model Y and Semi appears unlikely, as Elon promised a more careful, capex-light and self-funded approach. They won't repeat the Model 3 mistake of overlapping product and tooling design with factory construction and new robotic line orders.

... and there's no other options I can see for the $650m+$650m capex cash spending: store and Supercharger expansion is slow, body shops are a fresh decision.

Contractually mandatory payments for existing Model 3 lines seems the only explanation to me, but maybe I'm missing something?

 

[lklundin]

September ‘18 Top 30 best selling passenger cars

Model 3 was number 4 best selling car in the US for September!!!!!!

It is also really interesting that for September the Model 3 was the number 13 best selling car in the US, when including all vehicle types, including SUVs:

September 2018 – The Best-Selling Vehicles In America – Every Vehicle Ranked | GCBC

PS. Personally, I don't care what is a 'Toyota Highlander' or a 'Ford Explorer' (family) or a 'Jeep Grand Cherokee', but I bet the makers of these vehicles start to wonder...

 

[lklundin]

According to carsonight (who seems reliable) responding to an article by Fred Lambert:

carsonight 7 days ago
Fred, you are a trusted source and I enjoy reading your articles very much, but on GF1 production capacity I question your information.

First and foremost, and my pet peeve, is that planned GF1 cell production is NOT 105 GWh at the cell, 150 GWh at the pack, and this only sows confusion. The numbers to discuss are at the cell.

Second, adding 3 new lines will NOT increase production to 35 GWh, not even close. Production will more likely be increased to 24 GWh per year with 13 lines.

The information I get from the "factory floor", including the problems with labor shortages, check out with what I read concerning Tesla's actual production numbers. You can plug these numbers in for any number of lines of machines.

There are 14 machines in a line, and 10 lines currently at GF1. Each machine has a theoretical top speed of 25,000 cells per day, but reality is more like 20,000 cells per day. Plug those numbers into a 7 day run rate and devide by 4416 cells in a Model 3, and you come up with ~4400 Model 3s per week which is at the top of Tesla's guidance for the quarter. The first 2 weeks of the quarter their production were not that high and Q3 production is likely to be in the 52,000 range for the LR Model 3.

With 13 lines of machines, and using the same numbers, the math is ~ 5700 Model 3s per week by the end of the quarter. Because the SR Model 3 uses fewer cells, 20 lines would likely produce 10,000 per week of a mix of LR and SR Model 3s, and I am told there is room for 20 lines.

When GF1 is complete it will have 77 lines of machines, which at current production rates and using 18 Wh per cell is good for 140 GWh per year, AT THE CELL LEVEL. That's reality in my neck of the woods.

Well, If Tesla starts selling the SR variant, then that is a strong indication that this estimate is accurate...

 

[Big Earl]

Your units are probably wrong.

In principle, the production could be 20 GWh every hour (i.e. 20 GW), but that is too much.

Maybe an annual production of 20GWh ? (that is a production speed of 2.3 MW)

(We all make typos, but I think on TMC we should try to get the units rights).

Your units are a bit wonky, as well. Battery production should be measured in units of energy - GWh/hour or GWh/year. At no point is it correct to refer to battery production in GW, since that is a unit of power.

 

[mongo]

So my problem with GF3 is timing:

• they issued the capex and profitability guidance for the rest of 2018 in May already,
• but GF3 was not finalized until August IIRC,
• plus I find it unlikely they'd have chosen in August to burden Q3 with GF3 deposits - why not wait just 1.5 more months until Oct 1, to not load it into the critical Q3 time frame?
• Model Y and Semi appears unlikely, as Elon promised a capex-light approach. They won't repeat the Model 3 mistake of overlapping product and tooling design with factory construction and new robotic line orders.

... and there's no other options I see: store and Supercharger expansion is slow, body shops are a fresh decision.

Model 3 lines seems the only explanation to me, but maybe I'm missing something?

My thought is that the Q3 CapEx guidance will end up being way high. The trend this year has been lowed estimates due to the not expanding the Model 3 line/ production capacity. Combine that with your thinking that GF3 cap ex will start next quarter and Elon's push for positivity, and Q3 with be even more favorable...

I do need to push back on the non-overlapping bullet point: if they don't overlap, then they extend the timing until all the pieces are usable. They maybe have less robots, but they still need them to be installed at the same time.

 

[beachbum77]

In the past I've seen SA remove quite a few positive Tesla articles due to factual errors, so I sent in a complaint about Bill Maurer's latest article as a test to see if they'll remove a negative Tesla article with an important factual error.
Tesla: Q4 Now The Hard Part - Tesla, Inc. (NASDAQ:TSLA) | Seeking Alpha

I wrote:
Regarding profitability in Q3, the article states:
"It will be a tremendous failure if Tesla falls short, especially on the profitability side, given how many times management (especially CEO Elon Musk) guaranteed it would happen."
In fact neither Tesla nor Musk has ever guaranteed profitability in Q3 even once, much less 'many times'. It has only ever been stated as guidance with varying levels of confidence. Never guaranteed.

Let's see if the article gets removed. I predict not.

SA gives the author the chance to rebut a reader's challenge to a senior editor. If necessary, the editor will add an edit comment while the author makes necessary corrections. SA does not pull down an article unless the author fails to respond to the reader's challenge. The author usually makes that decision if the flaws are structural in nature and cannot be fixed quickly.

 

[Right_Said_Fred]

Your units are probably wrong.

In principle, the production could be 20 GWh every hour (i.e. 20 GW), but that is too much.

Maybe an annual production of 20GWh ? (that is a production speed of 2.3 MW)

(We all make typos, but I think on TMC we should try to get the units rights).

Should have known better. Fixed it.

According to carsonight (who seems reliable) responding to an article by Fred Lambert:

carsonight 7 days ago
Fred, you are a trusted source and I enjoy reading your articles very much, but on GF1 production capacity I question your information.

First and foremost, and my pet peeve, is that planned GF1 cell production is NOT 105 GWh at the cell, 150 GWh at the pack, and this only sows confusion. The numbers to discuss are at the cell.

Second, adding 3 new lines will NOT increase production to 35 GWh, not even close. Production will more likely be increased to 24 GWh per year with 13 lines.

The information I get from the "factory floor", including the problems with labor shortages, check out with what I read concerning Tesla's actual production numbers. You can plug these numbers in for any number of lines of machines.

There are 14 machines in a line, and 10 lines currently at GF1. Each machine has a theoretical top speed of 25,000 cells per day, but reality is more like 20,000 cells per day. Plug those numbers into a 7 day run rate and devide by 4416 cells in a Model 3, and you come up with ~4400 Model 3s per week which is at the top of Tesla's guidance for the quarter. The first 2 weeks of the quarter their production were not that high and Q3 production is likely to be in the 52,000 range for the LR Model 3.

With 13 lines of machines, and using the same numbers, the math is ~ 5700 Model 3s per week by the end of the quarter. Because the SR Model 3 uses fewer cells, 20 lines would likely produce 10,000 per week of a mix of LR and SR Model 3s, and I am told there is room for 20 lines.

When GF1 is complete it will have 77 lines of machines, which at current production rates and using 18 Wh per cell is good for 140 GWh per year, AT THE CELL LEVEL. That's reality in my neck of the woods.

Yep, that’s where I read it: Electrek. Note to self: don’t believe everything Fred writes.

 

[Lessmog]

...
Yep, that’s where I read it: Electrek. Note to self: don’t believe everything Fred writes.

Umm.

 

[Rarity]

According to carsonight (who seems reliable) responding to an article by Fred Lambert:

carsonight 7 days ago
Fred, you are a trusted source and I enjoy reading your articles very much, but on GF1 production capacity I question your information.

First and foremost, and my pet peeve, is that planned GF1 cell production is NOT 105 GWh at the cell, 150 GWh at the pack, and this only sows confusion. The numbers to discuss are at the cell.

Second, adding 3 new lines will NOT increase production to 35 GWh, not even close. Production will more likely be increased to 24 GWh per year with 13 lines.

The information I get from the "factory floor", including the problems with labor shortages, check out with what I read concerning Tesla's actual production numbers. You can plug these numbers in for any number of lines of machines.

There are 14 machines in a line, and 10 lines currently at GF1. Each machine has a theoretical top speed of 25,000 cells per day, but reality is more like 20,000 cells per day. Plug those numbers into a 7 day run rate and devide by 4416 cells in a Model 3, and you come up with ~4400 Model 3s per week which is at the top of Tesla's guidance for the quarter. The first 2 weeks of the quarter their production were not that high and Q3 production is likely to be in the 52,000 range for the LR Model 3.

With 13 lines of machines, and using the same numbers, the math is ~ 5700 Model 3s per week by the end of the quarter. Because the SR Model 3 uses fewer cells, 20 lines would likely produce 10,000 per week of a mix of LR and SR Model 3s, and I am told there is room for 20 lines.

When GF1 is complete it will have 77 lines of machines, which at current production rates and using 18 Wh per cell is good for 140 GWh per year, AT THE CELL LEVEL. That's reality in my neck of the woods.

Thank you for linking to that. It clears up a mystery for me with regard to the math. And here is a more recent comment from carsonight that is on point.

carsonight NickJ 13 hours ago

Thank you for your kind comments. I stand by my comment that I was going out on a limb, but it was not pure guesswork.

Panasonic is producing enoough cells for 4200 Model 3s per week. Over a 13 week quarter that is ~55,000 Model 3s per quarter. Panasonic is adding three new lines between now and the end of the year, and I have no ideaa when those will be up and runnng.

Someone who works those machines and sounds like he knows what he is talking about claims that these new lines are something special, and will be able to produce more cells per machine then previous ones. If true these three new lines could produce as many as 2000 additional Model 3s per week.

The question then, is how soon those new lines can be up and running. From what I hear it will be a ramp up between now and the end of the year so maybe 65k or 70k, but I'm clearly marking that as a guess.

 

[Brando]

Cell and pack assembly will be at the same location as vehicle assembly (per previous calls/letters).
Speaking of, will the Y launch in China first due to plant availability? (no replies in this thread...)

NOT with Trump 25% import duties on China. Stupid trade wars from Trump may crash the world again - 2007 replay.
I wish someone would go bribe Trump to change this trade war nonsense.
OF COURSE our Congress just lets him do it. President is just an executive and must get money from Congress, enforce laws/policies created by Congress and Congress approves his appointments. But few seem to understand our constitution and the empire continues to spiral down.

<sarcasm> When can we invade Iran? That will fix everything! </sarcasm>

 

[Tslynk67]

Either I'm misreading or misunderstanding, or they actually reduced the effect of the previous mandate? It says that the previous mandate was reduction of 30% by 2021, now it is 20% by 2025?

Cumulative...

 

[brian45011]

If they are profitable, the scales wouldn't care that much about growth, at least at first.

Worked after 1Q13, not so much after 3Q16.

 

[lklundin]

Your units are a bit wonky, as well. Battery production should be measured in units of energy - GWh/hour or GWh/year. At no point is it correct to refer to battery production in GW, since that is a unit of power.

I know it is unconventional, but it is technically correct:

An annual production of 20 GWh of batteries is an average (production speed) of about 2.3 MW (counting 8760 hours/year).

This can be visualized as a steady flow of trucks moving twenty three 100 kWh packs through a factory gate every hour, i.e. a 100 kWh pack going out the gate every 156 seconds. (Or since GF1 makes 80 kWh LR packs, about 28.5 LR packs every hour, i.e. one every 126 seconds).

I know this is not super useful, except maybe as an exercise to understand what the physical units mean.

 

[Dan Detweiler]

Once you buy a car without the high pressure sales, hours of negotiation and sleazy tricks you'll never want to go back to the dealership model either.

While I agree with you on principal and speaking in generalities, in light of the delivery tactics experienced by many (Including me) last week it may not be the best time to be discussing the delivery experience. Just sayin'.

Dan

 

[Fact Checking]

With 13 lines of machines, and using the same numbers, the math is ~ 5700 Model 3s per week by the end of the quarter. Because the SR Model 3 uses fewer cells, 20 lines would likely produce 10,000 per week of a mix of LR and SR Model 3s, and I am told there is room for 20 lines.

Great information, sounds plausible!

Btw., the SR estimate looks pessimistic:

• LR is 80 kWh, SR is 55 kWh, so 5,700/week 100% LR rate with 13 lines corresponds to a 100% SR rate of 8,300/week
• At a 50% LR + 50% SR mix the throughput would be 6,780/week
• To reach 10,000/week they'd need 19 lines.
• But there are lower capex expansion methods as well: it appears their low, 80% capacity utilization of the lines is mostly staffing related: not enough night shift workers. This could be improved by increasing the night shift bonus.
• Also, more lines allow smoother, more finegrained production: if there's a shortfall of workers then stopping 1 line out of 10 reduces output by 10%. At 13 lines the reduction is only 7.7%. The more lines, the better the overall utilization and the cheaper the run costs due to various economies of scale of maintenance, etc.
• If capacity utilization improves from 80% to 90% then LR pack output can increase to 5,850/week, and SR+LR to 6,930/week.
• If they add 3 more lines in 2019 then max output increases to 7,000/week for LR and 8,500/week for the 50%/50% mix.
• At a 75%/25% mix the rate is 5,750/week with 10 lines, 7,500/week with 13 and 9,200/week with 16 lines.

So everything depends on the SR/LR mix, and with 13 lines plus minimal Fremont capex they'll already be able to claim close to 10k/week capacity - even if a higher LR mix reduces the actual effective output.

 

[Artful Dodger]

Well the SEC uses that word: SEC.gov | Elon Musk Settles SEC Fraud Charges; Tesla Charged With and Resolves Securities Law Charge

Google is your friend; the SEC is not:

Does the SEC have criminal authority?

The SEC generally conducts investigations in private. ... The SEC can bring a civil action in a U.S. District Court, or an administrative proceeding which is heard by an independent administrative law judge (ALJ). The SEC does not have criminal authority, but may refer matters to state and federal prosecutors.
​

Now here's the answer from the SEC in 2005: (from page 1, here)

Enforcement Overview - SEC.gov

Linda Chatman Thomsen
Deputy Director
Division of Enforcement
U.S. Securities and Exchange Commission

I. AN OVERVIEW OF ENFORCEMENT
A. The Division of Enforcement

The Division of Enforcement ("Division") administers the Securities and Exchange Commission's Enforcement Program. The Division is responsible for detecting and investigating a wide range of potential violations of the federal securities laws and regulations. The securities laws prohibit fraudulent conduct both criminally and civilly, but the Commission is responsible only for civil enforcement and administrative actions.​

So here's a possibility you might not have considered; The SEC is taliking fast and loose, shooting from the hip, and had no prospect of winning a civil case. So they talk trash on their twitter feed, and embellish their actions to appease the groups that pressured them into taking this unwarranted action.

Anyway, this is behind TSLA now, and so I'm done commenting on the issue. Cheers!

 

[Krugerrand]

Oh, I'm not shocked... I'm challenging the assertion because that's exactly what it is.

Apparently it’s just gonna be another pie quarter.

 

[Off Shore]

PS. Personally, I don't care what is a 'Toyota Highlander' or a 'Ford Explorer' (family) or a 'Jeep Grand Cherokee', but I bet the makers of these vehicles start to wonder...

The bread and butter of the respective manufacturers, except for Ford which has already thrown in the towel. I'll bet they're wondering too.

 

[beachbum77]

While I agree with you on principal and speaking in generalities, in light of the delivery tactics experienced by many (Including me) last week it may not be the best time to be discussing the delivery experience. Just sayin'.

Dan

Hey Dan, Did you get your Model 3?

 

[Brando]

side note: arithmetic is a subset of mathematics which is a subset of Philosophy.
Thanks to all for the arithmetic on making/deliveries. Accurate details are often helpful.

PS - Things change fast in today's world. Especially at Tesla. GF-1 still being planned/built and many of the goals may well be increased over time. Timeline may only be close estimates, right? What a show to watch.