Short-Term TSLA Price Movements - 2016

[30seconds]

Can you elaborate a bit on this. Thanks in advance.

You have to folllow GAAP and auditor guidance. There is some flexibility, but needs to be marked against incremental progress towards fulfillment. Just like you can't rexognize the full 100% upfront, neither can you defer the full 100% if you have already started to deliver a portion of the functionality

 

[Fallenone]

You have to folllow GAAP and auditor guidance. There is some flexibility, but needs to be marked against incremental progress towards fulfillment. Just like you can't rexognize the full 100% upfront, neither can you defer the full 100% if you have already started to deliver a portion of the functionality

I'm really curious about the flexibility. Based on their quarterly filings, AP1.0 revenue wasn't deferred much actually, something on the level of $400-500.

 

[MitchJi]

Right, that's where I come out on the thing. Still, a 2k increase (and about 1.5% in margin I'd expect) is a significant boost in revenue/profit on an option that will probably have a 90%+ take rate.

Minus the added cost of the new CPU and sensors, probably under ~$500. Probably less than the AP1 sensors cost in 2015.

They have one of the best systems but are taking a much more conservative and cautious approach.

That's only true if their underlying system is safer and if Tesla's system isn't safer than manual driving as they roll out new features.

 

[Johan]

Minus the added cost of the new CPU and sensors, probably under ~$500. Probably less than the AP1 sensors cost in 2015.


That's only true if their underlying system is safer and if Tesla's system isn't safer than manual driving as they roll out new features.

No, don't get me wrong, Tesla's system is the best. I said Volvo is probably high up though, at least top 5 together with Daimler, Honda, GM, Lexus/Toyota

 

[racer26]

Let me throw out an investor question. Is the Model 3 now under priced? It is likely a nicer car than Tesla envisioned when musk came up with the $35K promise. Hardware 2.0, is successfully transitioned without much drama, will likely ensure Tesla will sell every M3 it can make well into the next decade (I say this assuming no level 4/5 approved by regulators).

I think the M3 is about $10K too low. That is $5B of revenue for every 500,000 cars sold.

Over the next five years $140K is probably too much for a well appointed electric sedan. But $50-60K may be too little.

The way I see it, its not AP2.0 Hardware that breaks the equation, but Tesla Network.

2.0 Hardware is an incremental improvement. Extra, but shrinking, hardware cost, added features to compensate. Presents way better value for money than a non-level 5 capable machine.

Tesla Network though, completely breaks the economics of car ownership. If I can just send my car out to make enough money to pay for it, the value proposition of how much I'm willing to spend on the car itself is completely different. If TN is lucrative enough for the vehicle owner, it makes the price tag of the car completely irrelevant. But then it runs into the fiduciary duty of Tesla to its shareholders, because at that point, the ROIC is so good that Tesla ought to stop selling the cars to the public and just own them itself. I have no idea how to predict what that looks like, and when the tipping point happens.

 

[MitchJi]

I'm really curious about the flexibility. Based on their quarterly filings, AP1.0 revenue wasn't deferred much actually, something on the level of $400-500.

It wouldn't be unreasonable to assume that most of the additional value comes from the new hardware. How much more is the newer model worth today?

 

[Alketi]

Wow, quite a drop in shares available for shorting at Fidelity - now at 955,550.

That's 747,400 shares difference in just 3 hours of trading. Were all these shares really shorted this am? If true, that would have been more that 50% of today's volume.

Where in Fidelity do you see the number of shares available to short?

Are yours still on loan? Many are reporting that all of theirs have been returned.

 

[racer26]

No, don't get me wrong, Tesla's system is the best. I said Volvo is probably high up though, at least top 5 together with Daimler, Honda, GM, Lexus/Toyota

The trouble with the way other automakers are reluctant to make such features public is that abhorrent attitude that Elon hates though.

The others are afraid to put it out to the public for fear of the bad press when something goes wrong.

Elon, and by extension Tesla, don't care. They're pragmatic about it. If I have the data to prove that the autonomous technology is safer on average than humans, it is morally reprehensible to withhold it for fear of the backlash. The backlash is just noise, if I can prove that my technology is safer, even if it still kills people. It should be in the public's hands, racking up millions upon millions of miles, so that it can learn how to deal with all the niggling edge cases over time. People who die in car accidents today are dying needlessly, and their deaths do nothing to better society. People who die in car accidents in an AP equipped car are still dying, but at least their deaths aren't in vain. Their deaths will mean something, and provide a very real tangible benefit to everyone who comes after them. Everyone else's travels will be made safer for their loss. The sooner we can bring about that advent the better.

The point is, there is an acceptable failure rate for the technology in its infancy. It doesn't need to be flawless, it just needs to be better than the flesh bags that have been piloting 2000 lbs of steel down public roads for the last 100+ years. Then, over time, it will get better still, and we'll stop letting the flesh bags drive at all.

 

[EinSV]

For next week's ER, does anyone see any potential wild cards (positive or negative), aside from the obvious (EPS/margin update) that could have a significant effect on the SP?

A few things that might come up either in the report or call:

• PowerPack/PW orders/reservations and/or guidance for Q4 2016 or 2017
• China sales up 300+% (they provided China numbers with the Q3 2015 ER)
• Model 3 update
• GF cell production start date and GF production ramp update
• Q3 sales significantly higher than initially reported
• Projected Fremont production capacity with doubling factory footprint

Anything else?

 

[racer26]

For next week's ER, does anyone see any potential wild cards (positive or negative), aside from the obvious (EPS/margin update) that could have a significant effect on the SP?

A few things that might be includedt:

• PowerPack/PW orders/reservations and/or guidance for Q4 or 2017
• China sales up 300+% (they provided China numbers in Q3 2015)
• Model 3 update
• GF cell production start and GF production ramp update
• Q3 sales significantly higher than initially reported
• Projected Fremont production capacity with doubling factory footprint

Anything else?

~$100M worth of ZEV credits?

 

[Ameliorate]

Let me throw out an investor question. Is the Model 3 now under priced? It is likely a nicer car than Tesla envisioned when musk came up with the $35K promise. Hardware 2.0, is successfully transitioned without much drama, will likely ensure Tesla will sell every M3 it can make well into the next decade (I say this assuming no level 4/5 approved by regulators).

I think the M3 is about $10K too low. That is $5B of revenue for every 500,000 cars sold.

Over the next five years $140K is probably too much for a well appointed electric sedan. But $50-60K may be too little.

Hm, without any features especially including autonomous driving $35,000 is probably fine.
With autonomous driving and tesla fleet, a value of $45,000 does feel low until production reaches...tens of millions per year?
Wasting production on a non optioned car would be unfortunate, although I assume it would eventually be returned to Tesla and upgraded. Similar to myusername's argument of vehicles being outdated within years, I do not believe the current or even 5 years out production is even remotely close to demand (Including competitors stepping up and producing autonomous/electric vehicles)

I am very strapped for cash with nearly everything in TSLA/SCTY but might try to reserve another M3 for tesla fleet. The ROI looks too promising to pass up.

You may very well be right that Tesla is missing out on a good chunk of potential profit, but this Entirely depends on the portion Tesla charges for income generated via Tesla fleet. A $45,000 M3 is worth a lot more to the consumer if they get 80% of Tesla fleet income, versus them getting 40% and Tesla taking 60%. Tesla can easily justify selling at a lower price if they make $150,000 over the lifetime of the car!

Will be interesting to see what they do with autonomous semi's, they too could generate income/mile driven.

-edit, these numbers are just my musings, not based on any concrete evidence

 

[30seconds]

interesting article on new California requirements on autonomous vehicles. basically the DMV wants to see a year of test data first. Something Tesla will not have a problem in collecting, but still may be overkill

Google, automakers object to California rules for self-driving cars

 

[MitchJi]

For next week's ER, does anyone see any potential wild cards (positive or negative), aside from the obvious (EPS/margin update) that could have a significant effect on the SP?

A few things that might come up either in the report or call:

• PowerPack/PW orders/reservations and/or guidance for Q4 2016 or 2017
• China sales up 300+% (they provided China numbers with the Q3 2015 ER)
• Model 3 update
• GF cell production start date and GF production ramp update
• Q3 sales significantly higher than initially reported
• Projected Fremont production capacity with doubling factory footprint

Anything else?

Impact of the new AP Hardware on Q4 margins.

Q4 demand, production and margins.

 

[Fallenone]

It wouldn't be unreasonable to assume that most of the additional value comes from the new hardware. How much more is the newer model worth today?

The thing is, without the software, the hardware can't provide the function customers bought. So I really don't know how they would split the value between these two but in the meantime really want to have a rough idea on this.

 

[racer26]

I'll take a whack at checking your math by estimating a different way.

Aerodynamic & rolling resistance, power & MPG calculator - EcoModder.com

Rolling resistance - Wikipedia, the free encyclopedia

That link shows the ecomodder drag and rolling resistance calculator for an 80000lb fully loaded semi with optimistic Crr and Cd, with a frontal area of a little over 9 m^2 https://www.the-blueprints.com/blue...-4x2-semi-trailer-tractor-streamspace-cab.png

Using that table, somewhere right around 69mph it takes 150kW to overcome the drag.
150kWh = 69mi, therefore that's a fair bit more pessimistic than your example, at 2.17kWh/mi

Because the trailers are a standardized size and shape, I don't think you can appreciably change the A component of the drag equation, and .0045 is already pretty optimistic on the Crr. And sure, 80k lbs is a fully loaded worst case. Tesla does have experience with producing vehicles with low Cd, though, so maybe they can do better than my optimistic estimate of 0.6

Truck drivers are allowed to drive for 11 hours in the US, and 69mph * 11 hours = 759mi.

759mi * 2.17kWh/mi = 1647kWh. You could probably have 8x 200kWh modules.

If I use the Panasonic NCR18650B datasheet http://www.batteryspace.com/prod-specs/NCR18650B.pdf which is the cells Model S used to (still does?) use: 243Wh/kg, and you seem to be WAY optimistic on cycle life.

200kWh modules would be 823kg or 1810 lb. 8 of them would consume 14,485lb or about 18% of the load capacity.

The data sheet suggests 300 cycles to 80% capacity. This article Tesla Model S Battery Life: How Much Range Loss For Electric Car Over Time? claims 94% capacity after 50k mi on 85kWh Model S. S85 had 240mi EPA range, so that would suggest 208 cycles to that point. 80% cycles, times a more conservative 500 cycle life = 400kWh lifetime output per kWh capacity.

400kWh / 2.17kWh/mi = 184.3mi lifetime / kWh capacity. Using a more typical semi average fuel economy of 6mpg, that's about 30.7 gal of fuel displaced.

30.7gal * $2.50/gal = $76.80. - ($40/MWh energy cost * 0.4MWh) = $60.80 -- not enough to even reach optimistic gigafactory cell pricing.

Slowing down to 55mph cuts the energy requirement to around 89kW. Or 1.61kWh/mi
55mi * 11 hours = 605mi
605mi * 1.61kWh/mi = 974kWh
Use 8x 120kWh modules instead.
120kWh = 493kg or 1086lb * 8 = 8691lb or around 11% of the payload capacity.
400kWh / 1.61kWh/mi = 248.4mi lifetime / kWh capacity. Displaces 41.4 gal of fuel
41.4gal * $2.50/gal = $103.51 - $16 energy cost = $87.51 -- closer, but still not quite there.

I don't think you need to pay for the cell cost with the fuel savings though. Semi's are expensive vehicles, and I'm guessing the cost savings of not needing a big diesel engine will buy a fair number of batteries.

I absolutely agree with you. Tesla will use batteries for their Semis, because its the variable they can control.

I also agree they have a clear path to doing that affordably.

I also agree their battery costs are (probably) less than we know, and your $85/kWh figure may not be that far out of line. You said conservative. Using the information that is publicly available, and without assuming anything, a conservative estimate should be more like $135/kWh was all I was suggesting. I'm personally expecting real world GF cost to be around $75-100/kWh.

I agree driverless is part of the plan, and relatively easy to solve, but I don't think regulators will allow it for quite some time, so I believe a Tesla Semi will be able to make its value proposition without being driverless.

If you paid attention to the formulas I used, the hours without stopping was only used to estimate a reasonable pack size. Pack size doesn't change the economics of it in terms of fuel consumption. I wasn't actually suggesting the economics don't work, necessarily. I was just trying to use physics to come up with reasonable estimates of what it would take to do it. No amount of battery technology or intelligent design will change that you need a certain amount of power to move a big vehicle through the air and across the road surface.

Conventional - Sleeper Trucks For Sale
This seems to suggest that the ballpark price range of existing diesel sleeper cab trucks is $130-150k USD. As others have suggested, the trucking world is all about the dollars and cents. You wont sell a more expensive truck unless you can make an economic argument for it.

Several searches comes up with long-haul trucks averaging about 100,000 mi/yr. Lets say 5 year life of the truck (though its probably closer to 10 year). 500,000 mi. 2014 regulations state new sleeper cab trucks have to achieve at least 7.2mpg. So lets base on 8mpg. 500,000 mi / 8mpg = 62,500gal of fuel.

62,500 gal of fuel * $2.50/gal = $156,250 in fuel costs
Cost per mile work sheet for Owner Operators
suggests about $12,500 in repairs, maintenance and tires per year is $62,500
plus the $150k cost of the truck, is $368,750 total cost of ownership for 5 years.

I can't find a new price, but used DD15 Detroit Diesel with low mileage goes for around $15,000 on used markets. Lets estimate $25k new.

I assume you need to go 11 hours straight, because recharging takes time, which can be done while the driver sleeps.

So if the rest of the truck is worth around $125k, Based on yesterday's math, I needed around 1000kWh of batteries to make it run 11 hours.

1000kWh * pack cost of... $100/kWh assuming decent gains from GF is a $100,000 battery.

That would make it about a $225,000 truck.

500,000 mi @ 1.61kWh/mi = about 805MWh. 805MWh * $40/MWh wholesale electricity cost is $32,200.

Tires and maintenance costs will probably be around the same price, so $62,500

That totals to $319,700 total cost of ownership for 5 years.

Certainly seems like it could be viable.

For those interested in the Semi analysis I did back in July

 

[MitchJi]

interesting article on new California requirements on autonomous vehicles. basically the DMV wants to see a year of test data first. Something Tesla will not have a problem in collecting, but still may be overkill

Google, automakers object to California rules for self-driving cars

Alphabet Inc's (GOOGL.O) Google unit and automakers objected on Wednesday to California proposals to set new, mandatory rules for testing self-driving cars in the state, which industry officials said could hobble their efforts in the home to much of self-driving vehicle testing and development.

Automakers and Google raised a litany of concerns about California's proposal at a hearing in Sacramento on Wednesday. They expressed opposition to the state proposal to require compliance with guidelines that federal regulators issued last month, but made voluntary.

They questioned why California would require a new autonomous vehicle data recorder and what data they would be required to test, and they objected to a proposal they said would force a 12-month delay between testing a vehicle and deploying it on public roads.

Automakers also questioned whether police should be able to get any self-driving data within 24 hours without seeking a warrant or subpoena.

California regulatory policy is important to automakers and technology companies because of its impact on operations in the state, and because the policies enacted in the most populous U.S. state often influence what other states and other countries do.

 

[vgrinshpun]

Where in Fidelity do you see the number of shares available to short?

Are yours still on loan? Many are reporting that all of theirs have been returned.

I do not loan my shares. To see the shares available to short and interest go to sell --> sell short.

 

[MitchJi]

The thing is, without the software, the hardware can't provide the function customers bought. So I really don't know how they would split the value between these two but in the meantime really want to have a rough idea on this.

I agree that it's important to know how Tesla plans to handle this. I guess we need to wait for the ER-CC.

 

[adiggs]

Let me throw out an investor question. Is the Model 3 now under priced? It is likely a nicer car than Tesla envisioned when musk came up with the $35K promise. Hardware 2.0, is successfully transitioned without much drama, will likely ensure Tesla will sell every M3 it can make well into the next decade (I say this assuming no level 4/5 approved by regulators).

I think the M3 is about $10K too low. That is $5B of revenue for every 500,000 cars sold.

Over the next five years $140K is probably too much for a well appointed electric sedan. But $50-60K may be too little.

I've started wondering that a little bit myself. I'm also thinking that the auto pilot / self-driving options pricing we are seeing on S/X will be identical on Model 3. So you can still get your minimum battery and safety features only Model 3 for $35k, but the self driving / no options/ minimum battery car is up to $43k.

I figure a car you intend to lend out to Tesla Network, is a car you'll want the big battery for, so it spends less time going back for more electrons. I'm thinking that 50-60k number is looking realistic today, and you may be right about it being too little.

Then again, if Tesla operates Tesla Network as a narrow money maker for the owner, and a big money maker for Tesla, then Tesla can recoup more of the autonomous driving value themselves on the back end and keep the price for autonomous mode "low"(er).


We're also going to see major deflation come to ride sharing service prices with volumes of autonomous driving cars. On day 1 and maybe even the first year, prices for an autonomous ride may be the same as a ride with a human driver. As more autonomous cars are available though, I expect we'll see autonomous rides undercutting the human driver price. At some point, human driver rides will be squeezed out of the market, and with competition, I expect we'll see autonomous cars squeezing down to the lowest operating and ownership cost vehicles. I have no idea what that'll look like when it evolves. (Except I'm eager to see what it looks like)

 

[Muaddid]

Let me throw out an investor question. Is the Model 3 now under priced? It is likely a nicer car than Tesla envisioned when musk came up with the $35K promise. Hardware 2.0, is successfully transitioned without much drama, will likely ensure Tesla will sell every M3 it can make well into the next decade (I say this assuming no level 4/5 approved by regulators).

Possibly, but then again TM's official mission statement is not to make money but to accelerate the advent of electric vehicles... selling Model 3s for less than they are "worth" is a good way to ensure more people buy it no ? Plus it hurts the competition (ICE) even more as they have to compete on functionality AND cost.