2017 Investor Roundtable:General Discussion

[neroden]

OK - back to LEDs. Ever get headaches from too much LEDs? My daily LED monitors leave me with a whole lot of eye-strain at the end of the day.

Nope. Though I have noticed the LEDs do *not* like unstable-frequency electric current. I wonder if a Powerwall could be used to help stabilize the frequency in my house.

My girlfriend is photosensitive and gets headaches from incandescents, and worse headaches from CFLs -- due to the UV content, which LEDs don't have.

 

[racer26]

Oh yeah. In the long run as the incentive programs eventually disappear this will all even itself out. I figure as the US tax credit disappears for Tesla, values of used Teslas will rise each quarter (or more likely, will stay flat rather than depreciating until they have effectively risen).

Right now you can pretty accurately model the used car value for a Tesla by (a) assuming that options other than battery size, AWD, and dual chargers are worthless, (b) subtracting the national incentives from the original price of the car (without options), and (c) using an absolutely standard average car depreciation curve. It's actually a bit higher than this but not too much. I expect this to continue to be true for a while (until electric cars are very common), which means the changes in incentives cause drops and jumps in the used car value.

Yup - unlike the US Fed tax incentive American Model 3 buyers are so concerned about, the Ontario incentive is not tied to any particular number of vehicles manufactured or incentives paid, and has no expiry date legislated into it. Its here for the forseeable future, though I suspect if they start needing to fork out thousands and thousands of incentives, they'll rethink it. As of right now, there are only around 10,000 registered EVs in Ontario.

True if you enjoy the fast-flicker of LED lights. CFL and LED lighting works well to lower the energy costs in the summer (lowers AC needs too) if you use a lot of lighting - such as leaving lights on house-wide and outdoor lighting too. Some of us have gone to trying to go both LED and off - turning off lighting entirely when not in a room. For me, I have seen $450-500 electric bills in the 2007-2008 timeframe to now about $50/month average with Solar PV, using LEDs and not being ignorant of power usage.

One more thing that matters - when kids move out of the house (off to college or whatever) - that lowers the home costs. However, does not improve a family's carbon footprint. Efficiently living in a single house with a large family is far better overall than once they "explode out into remote communities". I'm just waiting to see a guy I know has done to the environment by having 10 "middle-class" kids who will all end up going through college, living in families, building new homes, etc. Getting into the weeds here but actually LEDs are fine - as long as we are also paying attention to other things like family size and all that. A young couple who is very green who then goes and has some of their own kids (instead of adopting) blow their whole reason for going green. Green is living small. And adopt when possible over having your own kids. Sorry to go from LEDs to family planning - but that is a huge difference in world effect of one's green-ness versus some low-energy light bulbs.

OK - back to LEDs. Ever get headaches from too much LEDs? My daily LED monitors leave me with a whole lot of eye-strain at the end of the day. Why does it matter? Because if I hadn't had kids - I would be retired now and not need to stare at LED screens all day.

Not all LED lighting is created equal. Quality LED lights do not flicker perceptibly and throw light in a similar pattern to old incandescents. They're way better than the crappy CFL's they tried to force onto us a few years ago.

 

[racer26]

Huh?
Tesla Building Next Gen Maps through its Autopilot Drivers

Our statements are not at odds.

Waymo's maps are extremely high-res (think: where exactly in 3D space is that traffic light), and Waymo's cars depend on it to be able to function.
Tesla's are not nearly so detailed from what I can tell, and Tesla EAP does not depend on the map to function - not yet clear if FSD will, though.

 

[neroden]

The 29k+ screens is consistent with Elon mentioning in the recent e-mail to employees about Tesla hiring third shift personnel. As I speculated couple of days back, since BIW production capacity greatly exceeds production capacity of the final assembly line (about 3700 cars/week vs 2500 cars/week) adding third shift to final assembly would theoretically allow to approximately match nominal throughput of the BIW lines and general assembly line.

OK, now how many weeks per year are they operating now? 40?

NUMMI was operating 50 but I'm pretty sure Tesla shuts down for tooling more often than that. At 50, they could put out about 185K cars per year. At 40, only 148K...

I think we're actually hitting limits on demand for Model S at current prices (new markets should expand the numbers but unfavorable exchange rate trends should shrink them), but certainly not yet for Model X. If we assume 60K model S and 60K model X, then with this level of rampup Tesla will actually have excess capacity. While there are many ways to stimulate demand, the most obvious is cutting the price a little. Can Tesla cut costs enough to cut the price while still making the same profit per car? Probably. Shifting battery production to the Gigafactory alone should cut the costs meaningfully.

I read somewhere, but it was totally unsourced so it could be wrong, that cutting the price of an auto model by $5000 typically doubled sales. So that would be too much. We only want to raise sales by 20-55%. So maybe a $1000 price cut...

 

[EnzoXYZ]

With my experience importing a handful of 10-20k shipments into China, they let you pay tariffs/duties on internal estimated cost of the goods rather than the final retail price. So maybe this is along those lines.

Couldn't Tesla, Inc. sell the cars at wholesale cost to Tesla China (a wholly owned subsidiary of Tesla Inc, registered in China) who then sells them on to the customer?

I'm not exactly up on all of my international tariff rules. I could be wrong.

Just texted my friend that owns a fright forward company on this question.
My text to him: If you were to export a new car into china from USA can you declare the cost of the car vs. actual sale price of the car to minimize China's import tariffs? Have you dealt with similar high priced goods declarations?
His response: We export new car before like tesla. Export value is for USA custom, they can use lower value for importing to china. US custom & China custom is not connected so they don't share data at all. As long as the declared value is reasonable it will be accepted. The only risk is in the event of damage or total loss. You risk the difference in the value.

So I guess it customary to lower the declared value.

 

[FredTMC]

OK, now how many weeks per year are they operating now? 40?

NUMMI was operating 50 but I'm pretty sure Tesla shuts down for tooling more often than that. At 50, they could put out about 185K cars per year. At 40, only 148K...

I think we're actually hitting limits on demand for Model S at current prices (new markets should expand the numbers but unfavorable exchange rate trends should shrink them), but certainly not yet for Model X. If we assume 60K model S and 60K model X, then with this level of rampup Tesla will actually have excess capacity. While there are many ways to stimulate demand, the most obvious is cutting the price a little. Can Tesla cut costs enough to cut the price while still making the same profit per car? Probably. Shifting battery production to the Gigafactory alone should cut the costs meaningfully.

I read somewhere, but it was totally unsourced so it could be wrong, that cutting the price of an auto model by $5000 typically doubled sales. So that would be too much. We only want to raise sales by 20-55%. So maybe a $1000 price cut...

I don't think that formula of "cutting price by $5k doubles sales" applies in the full size sedan luxury segment. Just a guess.

 

[DurandalAI]

Nope. Though I have noticed the LEDs do *not* like unstable-frequency electric current. I wonder if a Powerwall could be used to help stabilize the frequency in my house.

My girlfriend is photosensitive and gets headaches from incandescents, and worse headaches from CFLs -- due to the UV content, which LEDs don't have.

I have a few different types of LED lighting in my home. I enjoy the "filament" type bulbs from a visual aesthetic of the bulb itself, however they typically only have resistors and two diodes as the circuitry, with no capacitors. This leads to a really heavy 60hz flicker. I suppose one option I could do is modify the light fixture to supply DC current by building my own rectifier circuit with a decently sized capacitor, which would also help with the power fluctuations. I could easily do that with the fixture above my dining room table, but for all the other fixtures I'm not sure it would be worth it.

My T8 LED (fluorescent replacement) tubes seem to have small switching power supplies in them to drop the voltage down, which include small capacitors. They have no flicker, however they do tend to suffer when there are minor brownouts in the home. (We have a number of those.) I've started slowly acquiring AVR (Automated Voltage Regulator) units for the more pricey electronics in the home, such as my studio equipment, and likely will branch that out to other devices, but that of course will not be extended to lighting.

But to answer your question, a Powerwall would not help stabilize the power for your home for anything but severe brownouts and outages, unless you utilize a dual conversion setup. In this, you would rectify the AC into DC to the appropriate voltage for the power pack, then you invert that into the AC for your home. This can be expensive, but it's possible. Many data centers use this kind of setup, as it provides perfect power all the time. Google

 

[techmaven]

Just texted my friend that owns a fright forward company on this question.
My text to him: If you were to export a new car into china from USA can you declare the cost of the car vs. actual sale price of the car to minimize China's import tariffs? Have you dealt with similar high priced goods declarations?
His response: We export new car before like tesla. Export value is for USA custom, they can use lower value for importing to china. US custom & China custom is not connected so they don't share data at all. As long as the declared value is reasonable it will be accepted. The only risk is in the event of damage or total loss. You risk the difference in the value.

So I guess it customary to lower the declared value.

So... we're guessing it's COGS, not ASP?

 

[vgrinshpun]

OK, now how many weeks per year are they operating now? 40?

NUMMI was operating 50 but I'm pretty sure Tesla shuts down for tooling more often than that. At 50, they could put out about 185K cars per year. At 40, only 148K...

I think we're actually hitting limits on demand for Model S at current prices (new markets should expand the numbers but unfavorable exchange rate trends should shrink them), but certainly not yet for Model X. If we assume 60K model S and 60K model X, then with this level of rampup Tesla will actually have excess capacity. While there are many ways to stimulate demand, the most obvious is cutting the price a little. Can Tesla cut costs enough to cut the price while still making the same profit per car? Probably. Shifting battery production to the Gigafactory alone should cut the costs meaningfully.

I read somewhere, but it was totally unsourced so it could be wrong, that cutting the price of an auto model by $5000 typically doubled sales. So that would be too much. We only want to raise sales by 20-55%. So maybe a $1000 price cut...

Historically Fremont operates 12 weeks per quarter, 48 weeks per year. I believe that MS sales could be sustained at current level, but MX sales will steadily grow, perhaps to as much as 2x of MS. It will take time, but that where I believe this is heading. So I think that MS/MX 2017 sales will grow about 50% over 2016, and another 50% in 2018.

 

[Turing]

3Q16 ASP was ~$91k USD.

Where did you find this and is there a breakdown that separates ASP of S from X?

 

[neroden]

For reference, from reading through the 10-Ks, Tesla has not decided when they will adopt the new accounting standards. SolarCity hasn't decided for sure either, but tentatively said Jan 2018 for revenue and Jan 2019 for leases.

It's actually kind of a mess if you don't adopt revenue and lease changes at the same time, and IFRS is encouraging adopting them at the same time. Given that both changes improve the books of Tesla, but are very accounting-heavy, I'm going to personally guess they'll try to adopt them both for Jan 2018 if they can.

 

[RobStark]

I don't think that formula of "cutting price by $5k doubles sales" applies in the full size sedan luxury segment. Just a guess.

Rule of thumb in the auto industry is not that $5k price reduction doubles sales but doubles addressable market.

 

[neroden]

Rule of thumb in the auto industry is not that $5k price reduction doubles sales but doubles addressable market.

Ah, thank you. For Tesla, that's currently going to be damn near equivalent to doubling sales due to the lack of competition.

 

[Stretch2727]

Just texted my friend that owns a fright forward company on this question.
My text to him: If you were to export a new car into china from USA can you declare the cost of the car vs. actual sale price of the car to minimize China's import tariffs? Have you dealt with similar high priced goods declarations?
His response: We export new car before like tesla. Export value is for USA custom, they can use lower value for importing to china. US custom & China custom is not connected so they don't share data at all. As long as the declared value is reasonable it will be accepted. The only risk is in the event of damage or total loss. You risk the difference in the value.

So I guess it customary to lower the declared value.

I deal with this situation everyday and the custom declarations are the agreed transfer price (COGS plus allowances for IP, R & D, corporate overhead) between the 2 affiliate companies. In this case US Tesla company and the China Sales arm. It is definitely not full ASP as you generally have to allow for some margin and profit at the selling company. It would be the amount the China Sales arm is expected to pay for each car which would generally be below the end customer selling price. If you don't allow some margin you will get in trouble with the local taxing authority for shifting profits.

 

[racer26]

Where did you find this and is there a breakdown that separates ASP of S from X?

3Q16 Financials. I had a more specific source when it was part of my prediction for 4Q16 results a couple weeks ago.

No breakdown. Quite simply - Automotive Revenue / units delivered = ASP.

 

[esk8mw]

For reference, from reading through the 10-Ks, Tesla has not decided when they will adopt the new accounting standards. SolarCity hasn't decided for sure either, but tentatively said Jan 2018 for revenue and Jan 2019 for leases.

It's actually kind of a mess if you don't adopt revenue and lease changes at the same time, and IFRS is encouraging adopting them at the same time. Given that both changes improve the books of Tesla, but are very accounting-heavy, I'm going to personally guess they'll try to adopt them both for Jan 2018 if they can.

I wonder if this is the type of thing someone could get a response from Tesla IR on? I know the better analysts will be asking Tesla if they will be adopting so they can update models, and IR usually tries not to give more info to analysts than the public...

Does anyone here have a history of getting responses? I would try myself but I don't grasp the topics well enough to form an intelligent question.

 

[Turing]

3Q16 Financials. I had a more specific source when it was part of my prediction for 4Q16 results a couple weeks ago.

No breakdown. Quite simply - Automotive Revenue / units delivered = ASP.

That's what i figured, so we can only make rough estimates at what the ASP is for S or X separately and how they might be trending QoQ.

 

[jhm]

Rule of thumb in the auto industry is not that $5k price reduction doubles sales but doubles addressable market.

Can't wait to see how going from $145k for a MS P100D to $75k M3 P80D will address a bigger market. Our kidneys may never be the same.

 

[MitchJi]

Is this accurate or FUD?:
Tesla Drivers Are Paying Big Bucks to Test Flawed Self-Driving Software

On a recent Saturday evening in South Carolina, Ward Mundy had his hands firmly on the steering wheel of his Tesla Model S when his wife let out a shout. The car was zig-zagging wildly across the road.

“It’s about the most dangerous ride you can imagine when you turn on Autopilot,” says Mundy, a lawyer and car enthusiast. He managed to get the Model S back under control, but it wasn’t the only time he’s had to rely on quick reflexes to avoid an accident. “You can be sailing along at 50 mph and the radar spots [an approaching] bridge and immediately slams on the brakes.”

Sometimes, the car doesn’t react when it should. “The other extreme is that you approach a stoplight with a car already stopped, and the Tesla doesn’t apply the brakes at all,” says Mundy.

Autopilot is the catch-all term for Tesla’s self-driving technologies. Its suite of hardware and software aims to match a car’s speed to traffic conditions, keep a car within or automatically change lanes (called Autosteer), park a car in a nearby spot, or allow it to be summoned from a garage. Elon Musk has said that by the end of this year, an Autopilot-enabled Tesla will drive from Los Angeles to New York without a human needing to touch the wheel at all.

To Mundy, that day seems a long way off. “It’s really a pretty scary experience,” he says. “If you’d ridden in the car with my wife, you would know how many times she’s screamed to turn it off.”

The Mundys’ recent Autopilot experiences are echoed by other Tesla owners in online forums and in YouTube videos of veering cars and near misses. Tesla has been building new Autopilot hardware and software into every car that’s rolled off the production line since November, rapidly rolling out self-driving capabilities before they’re fully tested. Even though many Tesla drivers seem willing to play along, the company’s strategy has some of them worried.

The stakes are high for Tesla as it gambles on this aggressive approach to testing. Owners who want to activate their cars’ Autopilot feature have to pay thousands of dollars extra. If drivers opt not to, the company loses out in its efforts to recoup its costs. But with a growing record of unexpected swerves, fish-tails and other miscalculations, Tesla is risking not only a hit to its largely sterling reputation, but also the lives and safety of some of its biggest fans.

 

[techmaven]

From Electrek:
SK Innovation is doubling its battery production for electric vehicles to 4GWh/year to support demand from Mercedes and others

original source article at the Korea Herald:
SK Innovation set to expand EV battery capacity

SK Innovations announcing that they are doubling their production capacity to 3.9 GWh by the end of 2018. They already supply Kia and their batteries have not aged well on the Kia Soul EV. They have a Mercedes supply contract. This supports about ~45,000 long range BEVs with 90 kWh packs similar to the Jaguar E-Pace (220 mi EPA range, 90 kWh pack) which is basically what Mercedes showed with the Concept EQ. Which means Mercedes is looking at building maybe 20,000 EQ long range BEVs for launch in 2019?