General Discussion: 2018 Investor Roundtable

[ggr]

FWIW, I’m in a long cue to get into the tesla store in century city ca (los angeles). I believe this is the first location to get a model 3 in the showroom! I’ve seen plenty of videos but this is my first chance to actually sit inside one !!! Very excited.

No, the showroom in Stanford was first, last week. First in LA area though.

 

[Menifeer]

...
Based on these 2 factors, I believe that people living in apartments will simply subscribe to car sharing services where they can summon a car upon request (using a voice assistant like Google, Alexa, or Siri). All they need to say is something like "Alexa, I need to go from A to B, get me a car at T time", and whatever service the person subscribes to will dispatch the vehicle to their door. The customer won't have to worry about the charge, or finding a charge spot, because the system will know which cars in the network have enough charge, and the cars will park and charge themselves.

Basically, I think people who disbelieve that EVs can be practical in urban areas, should think about usage models beyond what is available today.

I question the idea that these scenarios will work. If just a few people are involved, maybe so. If large numbers of people start summoning vehicles, there likely won't be enough available to satisfy the demand (to say nothing of the likelihood that a car won't be there at the specified time). A large number of cars would be required at the start and end of work shifts, and most would be unused the rest of the day (since most people stay at work). The traffic situation would be made worse due to the extra trips made by all those empty cars en route to pick up all those passengers.

A case might be made for it working in high-density urban centers where lots of people need to go in all different directions, but not in a typical sprawling suburbia.

I see little threat to the concept of car ownership.

 

[anticitizen13.7]

I question the idea that these scenarios will work. If just a few people are involved, maybe so. If large numbers of people start summoning vehicles, there likely won't be enough available to satisfy the demand (to say nothing of the likelihood that a car won't be there at the specified time). A large number of cars would be required at the start and end of work shifts, and most would be unused the rest of the day (since most people stay at work).

What you're missing here is that many people who live in urban areas don't use cars to get to work in the first place. I know people who walk to their office, people who bike, and people who take public transportation. People who live in an urban core and use a car to get to work are a minority. Mostly they use a car when they need to pick up things that are impractical to carry (like a TV), or to visit places where there is no easy public transit access.

The number of vehicles needed for commuters should be about the same, even in an autonomous world.

The traffic situation would be made worse due to the extra trips made by all those empty cars en route to pick up all those passengers.

Autonomous cars should actually ease congestion, because they don't get distracted and slow down to look at irrelevant things a human might find interesting. They can be networked to drive more efficiently in areas where there are stoplights.

A case might be made for it working in high-density urban centers where lots of people need to go in all different directions, but not in a typical sprawling suburbia.

I see little threat to the concept of car ownership.

In "sprawling suburbia", people typically do not live in apartment buildings with unassigned street parking. Your original argument was that it was the apartment dwellers who would refuse to buy EVs because it would be a pain to find a charge spot and then move once charged. I addressed that concern.

I largely agree that in suburbia, car ownership won't change, because cars in suburbia are more necessary to daily life and not just an occasional use item like they are for many urban dwellers.

 

[Cartegena]

Couple points:
It's not a tax credit, it's an accelerated business depreciation deduction. Instead of the standard 5 year depreciation schedule, you can claim an additional 25k the first year. There is no free lunch in that the business spent the money on the vehicle (an expense) so has less net profit. Also, the percentage of non business use is excluded. And the difference between the adjusted basis and sales prices comes back as income if the vehicle is eventually sold. One other thing is that if you are doing depreciation, you don't get to claim the mileage rate for use (it includes depreciation). Can claim supercharger fees though...

The criteria for 179 is GVWR not curb weight. The X100D has a curb weight of 5,421 and GVWR of 6,658. So a load capacity of 1,237 lbs. Guessing the load capacity of Roadster will be less, but curb may be more due to additional battery capacity and third motor/inverter.

Doesn't the Tax Cuts and Jobs Act of 2017 allow full expensing in the year acquired of capital equipment used in a business?

 

[BornToFly]

About TE, I'm personally closing my eyes until at least 2nd part of 2019 (Q3,orQ4) to see real substantial results to the income statement.

Yeah, this one has been a mystery. Those of us that have been around for several years where expecting TE to have had a major ramp and become a significant portion of the income a year ago (based on Elon's past statements). At this point I have thrown my hands up into the air and just say that I have no clue when it will contribute at least 5-10 Billion in sales.

 

[ValueAnalyst]

2018 Model S/X

How many Model S/X combined units will Tesla deliver in 2018?

 

[gene]

Just have to mention I had to work very hard to keep my Model S at a lifetime, over course of 80,000 miles , at only 304 wh/mile. In my Model 3, I just did 160 miles normal freeway speeds with quite a lot of moderate elevation changes using a mere 211 wh/mile!! That is super efficient.

 

[mongo]

Doesn't the Tax Cuts and Jobs Act of 2017 allow full expensing in the year acquired of capital equipment used in a business?

Good point, I think it does. Too bad I don't need to drive much for work, the first year of owning an X would be great income wise (the following years, not so much).

 

[bdy0627]

Would be nice to have more infos on the progress Tesla is making around self driving tech, and even AI in general.

Yes, it would. Tesla must have a gag order on EM at this point. Where are they on the tech now? Nobody knows. Roughly 1 1/2 years ago, Tesla abruptly changed their tech strategy with the MobilEye fallout. Roughly 1 year ago, they changed their leaders of their autonomous driving team. My WAG is that we will hear an update in Q2, or Q3 at the latest, once M3 production is going well. If the tech slips into 2019, other automakers will almost certainly be gaining serious ground there. AP2.5 has been disappointing so far, to say the least. It is arguably at parity with AP1 at this point, but AP1 goes back well over 2 years now. AP2.5 seems to be roughly 1 year behind EM's estimate that it would be at parity with AP1 within a few months. Where does that put full self driving?

There is the theory that Tesla is marching ahead with their full self-driving software, which they are developing concurrently to AP2.5. Regardless if that's the case, it strikes me as strange that they would struggle so much getting AP2.5 to parity, or almost to parity, with AP1, but be far ahead with FSD software in secret. If they are completely different software approaches, then perhaps that's the case. It just seems somewhat unlikely. Then again, if GM rolls out full self driving next year, even just within an urban center, you could say exactly the same thing about them. They seem to have perhaps leapfrogged a couple of generations via NVIDIA tech. We'll see how it plays out, but count me as becoming skeptical about Tesla's autonomous tech lead at this point. Remember, EM was initially promising a coast to coast self-driving trip by the end of 2017. He then backed off of that a bit, saying it may come a few months later. We haven't heard anything about it lately. Is "a few months later" for EM more like 9+? I sure hope not. I would love to see them get it done in the next 6 months.

 

[bdy0627]

It’s like an S curve lots of exciting announcements ahead to boost it even further: Falcon Heavy launch, FSD demo, Gigafactory announcements, Model Y and pickup reveals, more Boring tunnels, and of course, last but not least, the upcoming stormy weather in Shortville.

50+ million by the end of 2018 is likely.

Nice call, but I'd rather see your prediction for 2018 Model 3 production be on the money.

 

[BioSehnsucht]

Roughly 1 1/2 years ago, Tesla abruptly changed their tech strategy with the MobilEye fallout.

Not so much abruptly changed strategy, but were forced to skip the intermediate step Tesla planned which would have included AP1 hardware (MobilEye) in AP2 vehicles so that they could use AP1 still while developing and doing data gathering for AP2, and then once AP2 was ready transition to not including MobilEye hardware anymore. MobilEye cut them off (depending on who you ask, because they wanted access to Tesla's data or else, or because they felt Tesla was moving too fast and didn't want to be associated with any incidents, or because they were being vindictive about the eventual replacement, or ... ), so Tesla had to go straight into an AP2 hardware revision that didn't have software ready yet.

Tesla's strategy was always to move away from MobilEye, they were just forced to do it suddenly and without being prepared for it.

Eventually they intend to develop a custom AI chip, and move away from Nvidia. I don't expect Nvidia to pull the same stunt because Tesla using their hardware gives some validity to their own (separate, and unrelated other than portions of the hardware) autonomy software/hardware stack they're pushing. Plus, if necessary, the software Tesla is running on Nvidia's hardware can be ported to AMD's hardware (probably about 80-90% of the work can be converted automatically using tools AMD developed a while ago, but there will be some manual rework required to finish the port - and this assumes they're not using a portable intermediate to begin with that would make it relatively painless). I think Nvidia would rather hang onto Tesla as a customer as long as possible than give the prestige away to anyone else sooner than they have to.

 

[dmckinstry]

per share ? A bit too much isn't it ?

Well I'd be happy with 5 million. I'm not greedy. Even $5000.

 

[MP3Mike]

There is the theory that Tesla is marching ahead with their full self-driving software, which they are developing concurrently to AP2.5. Regardless if that's the case, it strikes me as strange that they would struggle so much getting AP2.5 to parity, or almost to parity, with AP1, but be far ahead with FSD software in secret. If they are completely different software approaches, then perhaps that's the case. It just seems somewhat unlikely.

I don't think it seems unlikely at all. There has been evidence shown that AP 2/2.5 essentially are putting out the same data that the EyeQ3 chip did. So Tesla's first attempt may have been to make Tesla Vision look like an EyeQ3 to speed things up. That obviously won't work for FSD, so I think they have parallel programs going. (Probably with some shared pieces.)

 

[avoigt]

Yes, I forgot to mention the recent increase yoy in sales but still very low compared to the size of market and unlikely to continue now that Tesla is tossed from getting tax incentives. Maybe they come back with model 3?

Many of the items you listed with a few exceptions is happening everywhere so I'm still at a loss. $5.50 gas and green minded German consumers should have enough incentive to learn more about EVs not made in Germany. Do non German cars sell well in Germany. Like Fords, Toyotas and the like? Might just be a very German centric market and why not, German cars are pretty good. Maybe model 3 can crack that nut. Just rename it for the German market as the grohmann model 3 and hope they don't notice.

Non German cars are not sold anywhere near German brands. But its not that much the loyalty why people are waiting with an EV decision but a combination of many factors as listed.

Gas prices are just one factor to consider and green minded consumer in the middle class often shy away as a Tesla requires large upfront payment. I know many people who are very excited about Tesla and would love to own one but can't afford it.

The moment EVs are cost competitive in the first year(s) people will change their minds more quickly. Right now Tesla is seen as a luxury car maker for the "rich" who can afford it to drive a non emission green car.

Germany tends to be a slow market if it comes to purchasing decisions but even if the lag behind today they will catch up.

 

[Reciprocity]

Non German cars are not sold anywhere near German brands. But its not that much the loyalty why people are waiting with an EV decision but a combination of many factors as listed.

Gas prices are just one factor to consider and green minded consumer in the middle class often shy away as a Tesla requires large upfront payment. I know many people who are very excited about Tesla and would love to own one but can't afford it.

The moment EVs are cost competitive in the first year(s) people will change their minds more quickly. Right now Tesla is seen as a luxury car maker for the "rich" who can afford it to drive a non emission green car.

Germany tends to be a slow market if it comes to purchasing decisions but even if the lag behind today they will catch up.

That doesn't explain why rich people aren't buying Tesla's. There are a lot is rich people in Germany and there are but S class and 7 series and A8s about 20:1 vs Tesla's where as in America Tesla beats those brands combined. And I would get it if Germany didn't seem to be so progressive when it comes to the environment. Is there anything more environmentally friendly then a Tesla charging during Germany's solar duck curve? Are model S's more expensive in the first year then an S class or 7 series in Germany? The only thing that makes beams loyalty and really just the love of powerful German motor that makes lots of noise and rattles your bones.

 

[bdy0627]

I don't think it seems unlikely at all. There has been evidence shown that AP 2/2.5 essentially are putting out the same data that the EyeQ3 chip did. So Tesla's first attempt may have been to make Tesla Vision look like an EyeQ3 to speed things up. That obviously won't work for FSD, so I think they have parallel programs going. (Probably with some shared pieces.)

I agree that they do appear to have 2 processes going on in parallel. What seems unlikely to me is that they would be far ahead with the FSD software, in secret, while struggling to get AP2.5 to actually work as well as the much older AP1 tech. I certainly hope they are way ahead with the FSD software but the likeliest answer, given what we can see and know, is that they are probably struggling with FSD software. Not a surprise to many of us, but FSD is certainly well behind EM's prediction. AP2.5 is well behind EM's prediction. I'm sure there are quite a few disappointed Model S/X AP2.5 buyers from a year ago who thought their cars would be driving themselves right about now, or at least performing way beyond AP1. The optimistic assessment is that Tesla is secretly doing really well with the FSD software, and will roll it out very soon, just a few months delayed from Musk's prediction. Pessimistically, they are over a year behind on Musk's predictions for AP2.5, and on their way to being that much behind on FSD as well. Most likely, reality may fall somewhere in between. If they are about a year behind on FSD, what does that mean for their moat on that tech? GM in particular seems to be closing the gap swiftly. With NVDIA's tech, Audi and others may be there in the next 18-24 months too.

 

[SteveG3]

On twitter tonight,

Question: "Can we look forward to more voice control before level 5 to maximize safety while driving? Since nearly everything you control on the Model 3 requires input on the screen. Thus, taking your eyes off of the road more than most other cars. Can’t wait for mine! Thanks!"

Elon: "Definitely. You will be able to do pretty much anything via voice command. Software team is focused on core Model 3 functionality right now, but that will be done soon, then we will add a lot more features."

Elon Musk on Twitter



This has been one of the few persistent criticisms of the Model 3 in many reviews.

Not sure if voice command would work for changing autopilot speed/follow distance. There is the possibility of using the right scroll wheel, currently quite underused, to allow adjustment for those settings without using the touchscreen.

 

[azaz]

On twitter tonight,

Question: "Can we look forward to more voice control before level 5 to maximize safety while driving? Since nearly everything you control on the Model 3 requires input on the screen. Thus, taking your eyes off of the road more than most other cars. Can’t wait for mine! Thanks!"

Elon: "Definitely. You will be able to do pretty much anything via voice command. Software team is focused on core Model 3 functionality right now, but that will be done soon, then we will add a lot more features."

Elon Musk on Twitter



This has been one of the few persistent criticisms of the Model 3 in many reviews.

Not sure if voice command would work for changing autopilot speed/follow distance. There is the possibility of using the right scroll wheel, currently quite underused, to allow adjustment for those settings without using the touchscreen.

Because so many interior design decisions were made on a much more optimistic FSD timeline (3mos,6mos). just simple stuff like the screen not being tilted (screen is optimized for video viewing) towards driver, no secondary info source, steering wheel only having two buttons. and now after a couple leadership changes with the AP group and FSD tech being at least say 2 years out and who knows how long regulation wise we aren't gonna be in a taxi relationship for a long time like Elon assumed (see: his twitter explanation for not having an info source in line of sight).

I hope they can quickly iterate and make some design changes software and maybe even hardware wise based on our reality that we won't be in a taxi relationship with out vehicles for at least say 4 years? A proper hud/second info source or much more functional steering wheel would solve a lot of these problems JMO

 

[Waiting4M3]

I get where everyone is coming from and its somewhat depressing because these are all linear ramps for the most part. Where is the S Curve in these estimates? Could it be more like 8k in Q1, 24K in Q2, then 60k in Q3 as the curve goes more vertical and flattens out to 70-80k in Q4. My point is that we could all see a bit more pain this quarter as wait for the vertical part of the S Curve. With each bottleneck that is resolved, you should see a step up almost instantly. You kind of see that Q1 where the pack assembly fixes should lead to 5-8x the production over Q4-2017 based on the estimates here, which would be a significant increase. Maybe 5x is more likely for Q1, but then another 3x for Q2 and Q3, flattening into the end of the year at a rate closer 5k/w.

My main point is that any estimates should try to figure out where the S Curve goes vertical and these estimates are all linear. Maybe the S Curve is a myth anyway, but it seems logical that you would have noticeable increases in production as production bottlenecks are removed.

You say linear, I call it a "S", if you squint

 

[Bet TSLA]

I'm sure that I'll love driving the EV, but my ICE ... yadda, yadda, yadda, ....

Sure. Right. In the real world, just like the rest of us, you'll look for any excuse to drive your EV. You'll make up reasons. You won't drive your ICE unless you absolutely must.

It will be quite sufficient to use the excuse that it's much safer, which it is. Welcome to the club.