2017 Investor Roundtable:General Discussion

[22522]

SO theoretically, after Tesla's credit expires, someone in the above list can buy the M3 at $35,000, re-badge it and sell it for $40,000, make $5,000 per car. Their customers can get a M3 for $40,000-7,500=$32,500, $2,500 cheaper than buying it from Tesla. It could almost work except for the supercharging network and SW updates. Tesla can control the value of the car with access control on those features.

Edit: what if whoever is doing this is willing to buy it from Tesla at $37,500 per car in exchange to have Tesls support the cars? So Tesla gets $2,500 extra, the 3rd party makes $2,500, and the consumer saves $2,500 also. Win/win/win, at the expense of the US government/taxpayers...

If Mazda makes something that will charge on the VW Network with Toyota self drive. Then designed something a little larger than the 3.... 5 series like, or a CUV.

They would have 9 months of a $7.5K price advantage. This could create a drought in US showrooms.

Tesla needs to alter the tax credit structure so this does not happen.

 

[mongo]

Each company and their subsidiaries has 200k Federal Credits total plus the phase out period.

Tesla can't buy Mazda and sell Mazda EVs with Federal Credit after Tesla's credits expire.

Mazda can right now if Mazda and Tesla were so willing. Mazda can buy Model 3s and rebadge them as Mazda 6 EV, Mazda 7 or whatever.

Subaru can do it. Volvo, Jaguar etc.

Or a startup called Smith Motors. Or Bollinger Motors.

The fuzzy one is Nissan and Mitsubishi. Nissan owns a 34% controlling interest in Mitsubishi but it is not a wholly owned subsidiary. Not sure if Mitsubishi's credits count toward Nissan totals but Mitsubishi did not sell many i-MiEV and hell will freeze over before Outlander PHEV arrives in California. Apparently.

Based on my understanding of CFR 49 sec 567 and related, I don't think that is the case. In the rebranding situation, Tesla would still be the final manufacturer. Tesla also cannot sell a complete kit to Mazda to assemble for which the credits would restart (can't find exact passage on the kit at the moment).
I believe that Tesla could sell power trains/ batteries to whomever, and they could build cars with their own frames for the ev credit.

Could even go so far as a Tesla rolling chassis similar to Freightliner or Navistar.

 

[Dagger]

Very interesting read:
Hawthorne city council special meeting on the Boring Company

Also this from Electrek:
Elon Musk plans to build his own hyperloop technology for his Boring Company’s tunnel projects

 

[brian45011]

Speaking of Fedex and UPS, I know from experience that both of these guys are watching the Semi as closely as they are allowed to, and running the numbers diligently to see if there is a business case for a fleet order.

Maybe Tesla could pick the bones of the Smith Electric carcass a la NUMMI? All Electric Trucks. Probably not going to happen. Ever. Why not?

 

[Hogfighter]

Uh oh, here comes the competition:
The Karma Revero Is a Very Bad Car

'I actually thought the parking brake was still engaged when I first pushed the gas'

Ouch. I genuinely feel for these guys.

 

[brian45011]

Very interesting read:
Hawthorne city council special meeting on the Boring Company

Also this from Electrek:
Elon Musk plans to build his own hyperloop technology for his Boring Company’s tunnel projects

That's hilarious. Not one one of those politicians asked about financial recourse or personal guarantees from the promoters, just in case the fin hits the shan. Those politicians walk among us and are convinced they are better to control your future than you.

 

[RobStark]

Based on my understanding of CFR 49 sec 567 and related, I don't think that is the case. In the rebranding situation, Tesla would still be the final manufacturer. Tesla also cannot sell a complete kit to Mazda to assemble for which the credits would restart (can't find exact passage on the kit at the moment).
I believe that Tesla could sell power trains/ batteries to whomever, and they could build cars with their own frames for the ev credit.

Could even go so far as a Tesla rolling chassis similar to Freightliner or Navistar.

For the purpose of the PEV tax credit the governing definition of "manufacture" is not from NHTSA but from the EPA under the Clean Air Act.

I am not a lawyer and I don't go reading administrative law as a hobby.

My understanding of the laws comes from years of reading EV and Auto magazines,forums,blogs etc.

I could be wrong but I doubt it.

 

[mmd]

Looking at the ER numbers again and I think something significant is happening. The SG&A portion for the auto business as a % of auto rev dropped significantly. This is a good thing, a very good thing. Historical SG&A/Rev was too high for the company to be ever profitable. There are a chunk of it for Model 3, I know I know. But they need to keep this % under control, which is something happening.

What numbers are you looking at and where? Can you show your numbers?
Is there a breakdown of SG&A for just the auto business?
In the SH letter, SG&A for q2'17 is $537M compared to Q1's $603M. But that's mainly because Tesla is shrinking the SCTY business and fired most of the sales people.
The SG&A in Q3'16, last quarterly report without SCTY, was $337M for a higher number of cars sold (24500).

 

[Fallenone]

What numbers are you looking at and where? Can you show your numbers?
Is there a breakdown of SG&A for just the auto business?
In the SH letter, SG&A for q2'17 is $537M compared to Q1's $603M. But that's mainly because Tesla is shrinking the SCTY business and fired most of the sales people.
The SG&A in Q3'16, last quarterly report without SCTY, was $337M for a higher number of cars sold (24500).

Q4 2016 ER said $85M increase OpEx caused by SCTY and that's before fully integration. SCTY when standing alone was running $200M+ each Q on SG&A. So I think it is contributing $100M+ on the current TSLA SG&A. So Q2 17 auto related SG&A would be in the neighborhood of $400-450M. Compare this with Q2 16, which was solely auto, $321M. So SG&A increased about a third, while auto rev increased 100%. So, good progress.

 

[mmd]

Q4 2016 ER said $85M increase OpEx caused by SCTY and that's before fully integration. SCTY when standing alone was running $200M+ each Q on SG&A. So I think it is contributing $100M+ on the current TSLA SG&A. So Q2 17 auto related SG&A would be in the neighborhood of $400-450M. Compare this with Q2 16, which was solely auto, $321M. So SG&A increased about a third, while auto rev increased 100%. So, good progress.

OK, I see how you came to this conclusion. Thanks for the details.
You cherry picked your comparable to make your point. Compared to Q2'16 (the lowest selling quarter last year), the SG&A/Auto sales looks good. No doubt.
But if you pick other quarters with similar sales levels as now, like Q3'16, it is worse now. If you think Q3'16 was not solely auto, it means auto only SG&A was even less in Q3'16, making current ratio look even worse.

If we compute the SG&A per car sold for Q2, it is $400M(your number)/22000 = $18k per car!
This has to improve a lot for Model 3 if Tesla hopes to make any profit on it ever.

 

[Fallenone]

OK, I see how you came to this conclusion. Thanks for the details.
You cherry picked your comparable to make your point. Compared to Q2'16 (the lowest selling quarter last year), the SG&A/Auto sales looks good. No doubt.
But if you pick other quarters with similar sales levels as now, like Q3'16, it is worse now. If you think Q3'16 was not solely auto, it means auto only SG&A was even less in Q3'16, making current ratio look even worse.

If we compute the SG&A per car sold for Q2, it is $400M(your number)/22000 = $18k per car!
This has to improve a lot for Model 3 if Tesla hopes to make any profit on it ever.

Picking Q3 16 is also cherry picking in some sense as it had the lowest SG&A to rev ratio in years. Reason for Q3 16 being an outlier is also quite obvious IMO. It was the quarter when Tesla was converting 3 reservation holders for S60 and having fire sales of inventory vehicles. These factors temporarily decreased the cost of customer acquisition (they knew who were interested in their products and have their contact info, sending emails barely cost money).

 

[neroden]

Copy/paste in every quarterly report from Tesla...

Oh, so basically it's the small number of people who paid before they were required to.

This isn't significant. That money is not held for significant amounts of time and few people will be stupid enough to pay hundreds of thousands of dollars earlier than they need to.

You are not required, but you can. Especially in countries where it is not customary to handle (paper) cheques, people will do a money transfer. And generally a bit in advance since it may take an unknown number of days to clear that large amount between different banks (f.e. any transaction over 15k IIRC, European banks have a duty to make sure that the money is not laundered etc...)

Ah, the many reasons why checks are better than this newfangled electronic finance nonsense. The electronic garbage just slows things down.

This quarter we also had testimonies on the forum from people paying the car in full so that Tesla could put it on transport and then count it as delivered as soon as it leaves the lot. We had these testimonies on the forum because as happens with Tesla logistics, things got f*d' up and even though they paid the full amount, the car actually didn't leave the service center. Resulting in angry posters and the usual cycle of answers of 1) disbelief 2) 'my delivery was amazing' 3) blaming the reporter

Yeah, this nonsense happened a couple of quarters ago too. Musk needs to stamp this out. It seems to keep recurring; I think some of the lower-level people want to "make their numbers". :-(

 

[neroden]

Ought to be able to write a AWK script that does it for you...

My understanding of AWK architecture (stream processing with no memory management, and pipes through standard input and standard output) is a key source of disagreement about what is possible.

I'd probably use perl instead. But the fact is updating old code to modern code is not really a fully automatable process. Lot of refactoring involved, which involves essentially artistic decisions.

 

[neroden]

Once upon a time, I worked for a sub-prime financing company that who's parent company was owned by one of the big 3 automakers. This finance company had a great track record of doing the following:
Finance a used Cadillac Escalade to a customer who has previously filed for bankruptcy, and charge them 19-20% interest on their $50k+ purchase. (Not kidding here.) The customer usually manages to make 6-9 payments before it is repossessed due to non-payment.
Finance company sells the vehicle at auction, writes off the immediate loss of the financed principal, interest, and associated fees for tax purposes.
Finance company then takes customer to bankruptcy court. As they cannot file for bankruptcy again due to being a recent bankruptcy, they garnish wages on the difference of the total owed minus the auction sale for the remainder of the bankruptcy period, typically at 1/5 or so of the original car note. That vehicle that was sold at auction might go through this process a few more times during its life. Rinse and repeat.
These loans, combined with the sub-prime mortgages that the parent company had a large hand in, would be securitized into packages to be sold to potential investors, while charging a service fee for each account that they maintained on the investor's behalf within that package. There were some "quality" guarantees within the packages, so occasionally some bad accounts would be rolled out of the package and some "good" ones would be rolled in to replace them.

This was 2006. We all know what happened in 2008...

The inherent problem with these security packages are that they remove the immediate hazard for those issuing the loans. Those buying the securities should not take the originators for their word that the customers contained within the packages are actually what they say they are. A lot of it is smoke and mirrors, with people lying on income and assets, and the originating underwriters not verifying because they know they will wash their hands of it in 2-3 months when it gets rolled into a package.

Loose credit is hazardous for sure, but when you allow the original underwriters to roll the loans into packages to sell to investors, that's where things go completely awry, and in my opinion, shouldn't be legal beyond a certain amount. Yes, we will see another 2008 drop in the future if creditors continue to play fast/loose with sub-prime credit.

Thank you for a concise description of the underlying problems with repackaged / resold loans. Having studied the causes of the 2008 crash intensively, *exactly this sort of stuff* is the basis of the *entire* 2008 crash. Mostly with mortgages rather than car loans, but it was all stuff like this.

The most egregious offender was Countrywide, which resold its loans so fast that they didn't care about creditworthiness at all and would just fabricate the paperwork (all fraudulent) before selling the mortgages to the next sucker (usually another bank). They collected their money through "fees" and didn't care if the borrower defaulted *immediately*. They cared so little about the investors in the mortgages that they actually shredded the notes and mortgages rather than registering them with the local county clerk's office. Legally, of course, this meant that the borrowers owed nothing, but the borrowers didn't generally *know* this -- and the bank which *thought* they had bought a mortgage, generally proceeded to foreclose completely illegally. Upon discovering that they had no note and no mortgage, the foreclosing banks hired corrupt law firms to fabricate forgeries of the notes.

Really gross stuff. Mozillo actually made a profit off his criminal operation at Countrywide even *after* the criminal fines. They got away with it mostly.

 

[neroden]

SO theoretically, after Tesla's credit expires, someone in the above list can buy the M3 at $35,000, re-badge it and sell it for $40,000, make $5,000 per car. Their customers can get a M3 for $40,000-7,500=$32,500, $2,500 cheaper than buying it from Tesla. It could almost work except for the supercharging network and SW updates. Tesla can control the value of the car with access control on those features.

Edit: what if whoever is doing this is willing to buy it from Tesla at $37,500 per car in exchange to have Tesls support the cars? So Tesla gets $2,500 extra, the 3rd party makes $2,500, and the consumer saves $2,500 also. Win/win/win, at the expense of the US government/taxpayers...

This is such a poorly designed credit system. It should really phase out for all manufacturers at the same time.

 

[neroden]

Based on my understanding of CFR 49 sec 567 and related, I don't think that is the case. In the rebranding situation, Tesla would still be the final manufacturer. Tesla also cannot sell a complete kit to Mazda to assemble for which the credits would restart (can't find exact passage on the kit at the moment).
I believe that Tesla could sell power trains/ batteries to whomever, and they could build cars with their own frames for the ev credit.

Could even go so far as a Tesla rolling chassis similar to Freightliner or Navistar.

What if Tesla sold bodies and powertrains and allowed the other company to put in its own interior? Still seems like this is awful loophole-prone.

 

[WarpedOne]

What if Tesla sold bodies and powertrains

Those same bodies and powertrains tesla cannot keep up assembling fast enough already?
Yes, good plan.

 

[ValueAnalyst]

What is everyone's cost estimate of building a 50 GWh Gigafactory?

I think many of us (incl. myself) used to think $5B, in-line with original announcement, and shared by Tesla and Panasonic (halvsies?).

Tesla, however, has been trending under that estimate, and I doubt they will really have to spend $2B in 2H17.

Further, it is unlikely to cost 2x-3x to double or triple Gigafactory 1 production by 2020.

Finally, advancements from Tesla Automation (formerly Grohmann) and the experience of having built the first gigafactory will likely lead to a lower cost per kWh of output for building subsequent Gigafactories.

So, what is your cost estimate of building a 50 GWh Gigafactory?

Mine is $3B, shared equally between Tesla and its partners, to bring the first 50 GWh online (which took four years from 2014 to 2018).

Another $3B, shared equally between Tesla and its partners, to increase output to 100-150 GWh (2018 to 2020).

With Tesla Automation advancements ("our best engineers are working on automating production"), as well as further imporevements in energy density, I would expect significant improvements on cost per GWh for subsequent Gigafactories.

So I very roughly estimate that $5B, shared equally between Tesla and its partners, will be the approximate cost of building subsequent Gigafactories assuming each will produce 100-150 GWh at the pack level.

This would mean the cost of the all-important 100 Gigafactories can be $500B, which Apple and Tesla alone will likely be able to do by 2035.

 

[mongo]

For the purpose of the PEV tax credit the governing definition of "manufacture" is not from NHTSA but from the EPA under the Clean Air Act.

I am not a lawyer and I don't go reading administrative law as a hobby.

My understanding of the laws comes from years of reading EV and Auto magazines,forums,blogs etc.

I could be wrong but I doubt it.

I think that the EPA/ IRS go by the manufacturer listed on the VIN/ manufacturer's sticker. Case in point: the early Ford Fiesta's manufacturer sticker was "MFD BY KIA MOTORS CO FOR FORD MOTORS CO". So Ford owned the certification, and Kia just built them. Tesla would need to be a contract manufacturer for some else's design (with full FMVSS cert and crash testing by that other company). Which goes back to selling powertrains.

Law reading is not a hobby, nor amI a lawyer. I was previously exposed to the nitty gritty of "whose vehicle is it" when working on a NGDV proposal. I try to confirm my rememberances with the relevant supporting documents so other can verify/ refute my conclusions.

 

[mongo]

What if Tesla sold bodies and powertrains and allowed the other company to put in its own interior? Still seems like this is awful loophole-prone.

My interpretation is that powertrains would be okay, but a full body that just needs trim pieces would not.

For strange loopholes, look at the import of the Ford Transit. To avoid the commercial vehicle "chicken tax", the plantern overseaswould install seats to make it a passenger vehicle, thus cutting import tax by 90%. Then the seats would be removed before delivery to dealers. Government caught on to that one.