General Discussion: 2018 Investor Roundtable

[Waiting4M3]

Sounds like a huge investment in a low margin business.

Define huge investment vs low margin. I don't know what it costs to build a garage, but would be shocked if it's much more than a few million $ if you locate off city center and get cheaper real estate. Autonomous driving is the key here since the cars would be able to drive into the denser area when needed, but park in cheaper areas to save cost on building the garage. Also it allows you to use much fewer super chargers than parking spot, say 10:1 ratio. Each car can autonomously drive to the charging stall to charge for 0.5-1hr, then make room for the next. So say you have a 500 car parking garage, you only nee 50 super chargers, lets call it $10M for the garage+SC. If you charge $100/month parking+charging, that's $600K/yr, it would take 16 yrs to pay the $10M initial investment back. In large cities like San Francisco, people pay way more than $100/month for parking. I'm not saying it's a slam dunk money making idea, but I see some potential.

 

[immunogold]

or even one step beyond, use the automated garage to house fleet of Tesla Network cars that Tesla owns and operates. Private owners can opt in

I like your idea. For robust adoption of the Tesla Network, owners will want to feel confident that their vehicle will be charged and clean whenever it is returned to them. A Tesla Hub facility may be easiest way to accomplish this service for individually-owned vehicles and for Tesla-owned fleet vehicles.

In addition, as a subscription service the Tesla Hub could provide overnight storage and charging for the urban owners who only have access to on-street parking. Urban parking is expensive day or night, suggesting that this business model may have merit. You should pitch it; you could be VP of Tesla’s Real Estate and Netwok Hub Division... It would surprise me if they had not given this at least some thought already.

 

[Waiting4M3]

I like your idea. For robust adoption of the Tesla Network, owners will want to feel confident that their vehicle will be charged and clean whenever it is returned to them. A Tesla Hub facility may be easiest way to accomplish this service for individually-owned vehicles and for Tesla-owned fleet vehicles.

In addition, as a subscription service the Tesla Hub could provide overnight storage and charging for the urban owners who only have access to on-street parking. Urban parking is expensive day or night, suggesting that this business model may have merit. You should pitch it; you could be VP of Tesla’s Real Estate and Netwok Hub Division... It would surprise me if they had not given this at least some thought already.

Ha, right! VP at Tesla. The main reason I'm here on the TSLA investment forum is so that someone else at Tesla can work hard to make ME money. Also I can yell at Tesla any time they miss an Elon target. Why would I want to be on the other side of this?

 

[dandurston]

G Poseidon about ready to depart and MarineTraffic now reporting destination for these 1300+ Tesla cars is Shanghai.

Does it say when it'll arrive? Will these cars be delivered in Q1?

 

[GoTslaGo]

Based on the link it just left and is outside of SF harbor near Farallon Islands. Here’s a screenshot (11 pm PST):

Looks like should get to Shanghai March 10, 2018

 

[hobbes]

G Poseidon about ready to depart and MarineTraffic now reporting destination for these 1300+ Tesla cars is Shanghai.

kenriko made a great drone video of the ship and the parking lot just before it left port:

Looks like a few cars didn´t fit .

 

[j0hn]

and a image i found on reddit. supposidly from the 10th of februari. someone care to count

 

[Runarbt]

View attachment 282519

and a image i found on reddit. supposidly from the 10th of februari. someone care to count

Quick count - 3-4000 cars.

 

[hobbes]

Quick count - 3-4000 cars.

No way. You counted the non-packaged employee cars, too, right?!

If this
General Discussion: 2018 Investor Roundtable
are 1310 cars, in the new pic it is still below 2000.

 

[wnorris]

Loup Ventures (Gene Munster of Apple fame) has a good blog post on the future of transportation. Tesla is well positioned here.

Key Questions on the Evolving Future of Transportation | Loup Ventures

 

[RubberToe]

Self driving cars to hit California roads in April. Only caveat is that a remote "operator" needs to be available for edge cases the car can't handle.

Self-driving cars with 'remote' drivers could test on Calif roads in April: DMV
Self-driving cars with 'remote' drivers could be tested on California roads starting in April

Seems like a good interim solution. One person sitting in a control room to handle 500 roaming cars. Googles cars are down to 1 required intervention every 5,000 miles. If any given car drives say 200 miles per day, 25 cars need help once a day. 500 cars need help 20 times per day.

You replace 500 drivers with one person in a control room who handles an edge case once an hour. All thats needed is reliable data link to the car, and the network is becoming ubiquitous.

This scales well, and eliminates the need to go from 99.999% to 100% reliable. Seems like a game changer. Will be interested to see which companies apply.

RT

 

[Runarbt]

No way. You counted the non-packaged employee cars, too, right?!

If this
General Discussion: 2018 Investor Roundtable
are 1310 cars, in the new pic it is still below 2000.

yes. If only the white wrapped. my guess is just shy of 2000. the largest blocks are 5×30 = 150 cars.

 

[mongo]

Googles cars are down to 1 required intervention every 5,000 miles. If any given car drives say 200 miles per day, 25 cars need help once a day. 500 cars need help 20 times per day.

I'm wary of statistics using small numbers. If the intervention is location based (construction zone) vs semi-random, then you need to remote drive all Google cars past the problem point.

At a high level, Starlink would let Tesla implement this.

@Waiting4M3
Regarding the parking lot idea: It may be cheaper (from an owner POV) to just have the autonomous cars cruise on straight low speed streets when not used. (And even cheaper to ditch the SF apartment and live in the car)

 

[Reciprocity]

Self driving cars to hit California roads in April. Only caveat is that a remote "operator" needs to be available for edge cases the car can't handle.

Self-driving cars with 'remote' drivers could test on Calif roads in April: DMV
Self-driving cars with 'remote' drivers could be tested on California roads starting in April

Seems like a good interim solution. One person sitting in a control room to handle 500 roaming cars. Googles cars are down to 1 required intervention every 5,000 miles. If any given car drives say 200 miles per day, 25 cars need help once a day. 500 cars need help 20 times per day.

You replace 500 drivers with one person in a control room who handles an edge case once an hour. All thats needed is reliable data link to the car, and the network is becoming ubiquitous.

This scales well, and eliminates the need to go from 99.999% to 100% reliable. Seems like a game changer. Will be interested to see which companies apply.

RT

It depends on the intervention required. My guess is that have very good records on each one including video. I mean they would need that data to lower the rate of human interventions. The question I would have, was the car making a guess that it knew was borderline or was it fully confident and would have killed someone. There is a huge range there where 9 in 10 of those interventions could pop-up on a person's screen to help with a decision but that other 1 could be problematic. Accidents will happen. They happen with people at some rate. But injuries or a single fatality would be catastrophic for autonomous cars in general.

One place this works great is if the car approaches anything that it does not recognize. A simple interface could pop up on an operators screen to help make the decisions the machine can't. You don't need to remotely drive the car, it can drive but it doesn't always know exactly what to do. Humans do this all the time. You ever see someone who doesn't know if they should exit the freeway and they stop in the median and take some time to decide? If the self driving cars sensors are impaired the operator can pull the car over but in most cases the car might get confused by something like construction and would just need help, bot remote driving but simply help making the right decisions. This data would also be gathered and used to train.

There could be geofenced areas that require active monitoring including known construction areas. Dense pedestrian traffic and schools. You probably need more the 1 operator per 500 cars, but if you start with 1 for ever 10 it's a great start towards 1 per 100 and eventually 1 per 500 at some point. I believe you will always need this monitoring because cars will have issues even on the Tesla network and the owner may not be able to react quickly so Tesla will need to have a command center that can react quickly, diagnose issues and get the car out of traffic.

 

[Cattledog]

Define huge investment vs low margin. I don't know what it costs to build a garage, but would be shocked if it's much more than a few million $ if you locate off city center and get cheaper real estate. Autonomous driving is the key here since the cars would be able to drive into the denser area when needed, but park in cheaper areas to save cost on building the garage. Also it allows you to use much fewer super chargers than parking spot, say 10:1 ratio. Each car can autonomously drive to the charging stall to charge for 0.5-1hr, then make room for the next. So say you have a 500 car parking garage, you only nee 50 super chargers, lets call it $10M for the garage+SC. If you charge $100/month parking+charging, that's $600K/yr, it would take 16 yrs to pay the $10M initial investment back. In large cities like San Francisco, people pay way more than $100/month for parking. I'm not saying it's a slam dunk money making idea, but I see some potential.

You'll pay above $30,000/space for basic structured parking + design/engineering fees + an upcharge for electrical infrastructure and any inhabitable space you want (control room, wash room, etc.). So call it $40,000/space for most of America outside the high $ construction zones and exclusive of land costs.

 

[Thumper]

Sooner or later there will be more simultaneous interventions needed than there are monitors, no matter what ratio of monitors to cars you start with.

 

[hobbes]

Self driving cars to hit California roads in April. Only caveat is that a remote "operator" needs to be available for edge cases the car can't handle.

Self-driving cars with 'remote' drivers could test on Calif roads in April: DMV
Self-driving cars with 'remote' drivers could be tested on California roads starting in April

Seems like a good interim solution. One person sitting in a control room to handle 500 roaming cars. Googles cars are down to 1 required intervention every 5,000 miles. If any given car drives say 200 miles per day, 25 cars need help once a day. 500 cars need help 20 times per day.

You replace 500 drivers with one person in a control room who handles an edge case once an hour. All thats needed is reliable data link to the car, and the network is becoming ubiquitous.

This scales well, and eliminates the need to go from 99.999% to 100% reliable. Seems like a game changer. Will be interested to see which companies apply.

RT

Problem I see is that the critical edge cases need very quick action - reaction time at a few seconds the most. I can hardly imagine any monitoring person reacting that quick to a completely new situation that just pops up without prior notice. Job would be somewhat similar to an air traffic controller I guess, but in air traffic you have much more space and fewer vehicles, and pilots still take the absolute time critical actions themselves.

 

[Reciprocity]

Problem I see is that the critical edge cases need very quick action - reaction time at a few seconds the most. I can hardly imagine any monitoring person reacting that quick to a completely new situation that just pops up without prior notice. Job would be somewhat similar to an air traffic controller I guess, but in air traffic you have much more space and fewer vehicles, and pilots still take the absolute time critical actions themselves.

You are not going to have video feeds from thousands of cars. It's just not going to happen. The cars will have to know when they encounter an issue that confuses them. It could as mundane as a garbage can blows over into the road. If you need interventions to save pedestrians on cellphones that wonder into the road, it's a non starter. My thinking is that the car would stop and not know what to do next with the garbage can in the road and need help to make a decision to cross into the other lame to get around. It could certainly do this on it's own in future but my guess is what early on, that will be a no-no that requires intervention. A true level 5 car must be able to navigate those issues but not this. Remember level 4 is the same as level 5 except that you must have someone available to take over with a 20 sec warning. Not instantly. This would be a level 4/5 hybrid with the person taking over being remote. You would also put in a kill switch for the passenger which would instantly bring on a remote human controller to help. Remember the car can drive, just might get confused once every 1000 miles.

 

[dc_h]

yes. If only the white wrapped. my guess is just shy of 2000. the largest blocks are 5×30 = 150 cars.

I think the color cars in foreground are imports.

 

[Bobfitz1]

This is an FYI for what it is worth. I have no way to confirm it. Yesterday while visiting in NYC I stopped into the new Tesla Gallery on the West side of Manhattan to see their Model 3. Gorgeous as many have reported. Display car had 19 inch wheels.
A nice young lady associate gave me the run down on differences with MS and how center screen works. It's resolution and speed while doing navigation was outstanding.

While talking with the associate I mentioned I planned to wait to order until dual motor M3 was available, which I thought would be fairly soon. I said that based on my recollection of various posts reporting VINs being issued for dual motor cars and a few seeing preview of configurator screens which will be activated once dual motor is available. The associate told me the standard M3 would become available in a few months while dual motor not until later this year. I pushed back with the VIN report and she then replied they had a sales associate meeting only a few days earlier, where they were told by management the standard (short range) M3 would precede dual motor cars. While we've seen plenty of 'anecdotal' info from sales associates that was way wrong, she did not appear to be confused on this point. I don't know if anyone on forum has reported any new info or evidence for AWD M3 being offered before the base version.