2017 Investor Roundtable:General Discussion

[tander]

Very unlikely. The economies of scale in railroads (and barges/ships, for that matter) are monumental and can't be matched by trucks. All the low-volume and all the high-urgency stuff is already being done by trucks.

It kind of sounds like we might end up with underground trains or the 21st century form of them. Maybe a hypoloop for freight? Doesn't sound as cool as hyperloop.

 

[neroden]

50% participation rate for a $15,000+ per year opportunity without lifting a finger is too high?

Of course it's way too high.
* Some large percentage of all Teslas, say roughly half, are sold in places where Tesla Network won't even work (like mine!).
* A large number of existing owners, say roughly half, have already stated that they will never lend their car out to strangers.
Multiply this, you'd get more like a 25% participation rate.

This isn't actually the big problem with your estimate.

Much more importantly, you've grossly overestimated the total *demand* for taxi trips, which appears to be the primary source of your bogus numbers. Your original assumptions imply over 600 million trips per day -- basically literally everyone in the US every single day. Which is just not happening for a network restricted to a short list of metropolises, and probably not for a network covering the whole US. Have you ever looked at patterns of travel demand? Obviously not.

And you can't use Tesla Network for any significant number of peak hour commutes. In order to do so, you'd have to have someone who owns a car and does NOT have a peak hour commute -- think about how uncommon those people are. (On top of that, someone who doesn't have a peak hour commute and owns a car is likely to be more protective of it and less willing to let strangers drive it.) So exclude peak hour commutes from your calculations.

This is a case where it's quite possible to hit the demand limits. Assume that Tesla takes 100% of the addressable market, but be realistic about the addressable market.

It is boring trying to explain Transportation Economics 101 stuff to you. Please go read a book on the statistics of transportation demand.

If anyone else is curious about realistic potential, I'd be happy to discuss it with them.

 

[neroden]

If the costs get low enough, many will choose not to own a car at all and use the network.

Which is fine but *someone* has to buy the cars. There's a numbers problem with relying on random owners to supply the cars: it has limits to how far it can scale up (and in particular, it fails in the peak commute hour, because the vast majority are using their own car during that hour).

If Tesla is simply supplying its own cars, then you have to include the cost of the cars in the profit calculation.

And that would be a reasonable way to estimate the profits from "Tesla Network" -- to figure that manufacturered cars could be sold, leased, or alternatively "rented" through Tesla Network, and calculate that as a third form of revenue -- one which is possibly higher-revenue than sale or lease. But loaded with the cost of manufacturing the car.

The other alternative is to assume that Tesla Network will just be cars owned by other people -- then you have to really try to figure out what the demand for *off-peak* taxi service looks like (it's not that high), and what percentage of the payments have to go to the car owner (it's most of them), to figure out what's left for Tesla (it should still be a respectable profit, just not the absurd stuff VA has been claiming).

 

[neroden]

The primary reason why Gigafactory 1 will have taken six years when completed, lack of capital, will ease as a constraint going forward.

I do not see evidence that capital is the reason it took this long. In fact, I am pretty sure capital is NOT the reason it took this long; Musk has never claimed that capital was the reason for delay here.

I believe it took this long because it had to be designed from the ground up. Design takes time and you can't speed it up -- it's mythical-man-month territory.

Now, this means subsequent Gigafactories should go much faster, since a lot of the blueprints are basically ready now.... provided they don't go all "Model X" and start redesigning everything.

I do suspect there will be some redesign. What with building an entirely new factory design for Model Y and all -- I think it's clear Musk loves his innovation. I don't think he's disciplined enough to do Intel's "exact copy" factory protocol when he sees ways to improve it...

 

[FredTMC]

Elon seems to be in good mood.
The least we can say is that there is no bad surprises concerning the M3 until now

I've given up on trying to desern if Elon's "good mood" tweets mean than the impending tesla quarterly numbers will be good or not.

Only tweet from EM that seems to correlate with an increase in SP is "stormy weather"

 

[Reciprocity]

Which is fine but *someone* has to buy the cars. There's a numbers problem with relying on random owners to supply the cars: it has limits to how far it can scale up (and in particular, it fails in the peak commute hour, because the vast majority are using their own car during that hour).

Yes and no. There will be less people buying cars because there use will be more efficient. Instead of being parked 23 hours per day. Even if that number only goes to sitting 15 hours a day, it can the support 3x or more peoples needs. It's not just going to be individual owners. It's going to be fleet owners, Uber drivers might have several cars and never drive them themselves for example. You keep focusing on how people used cars today and assume it won't change just because the cars drive themselves. It's going to change a lot. I think younger people and elderly who have been limited in where and when they can travel. People will go more places if it's less expensive and more comfortable to do so.

 

[Reciprocity]

I do suspect there will be some redesign. What with building an entirely new factory design for Model Y and all -- I think it's clear Musk loves his innovation. I don't think he's disciplined enough to do Intel's "exact copy" factory protocol when he sees ways to improve it...

The way Elon had talked about the Y, you would almost think you have to design the factory before the car to know how to design the car to be built faster by an order of magnitude. At the very least they might be designed in concert in a simulation before finalizing the design of either.

 

[Starno]

I've given up on trying to desern if Elon's "good mood" tweets mean than the impending tesla quarterly numbers will be good or not.

Only tweet from EM that seems to correlate with an increase in SP is "stormy weather"

Not about quaterly numbers.

But rather the state of Tesla's operations seem to be good. Meaning : if a supplier would be late by now, or some machinery issues or etc. that would considerably slow the M3 production. I don't think he'd tweet those " ironic " and " comic " statements.

 

[ValueAnalyst]

I do not see evidence that capital is the reason it took this long. In fact, I am pretty sure capital is NOT the reason it took this long; Musk has never claimed that capital was the reason for delay here.

I believe it took this long because it had to be designed from the ground up. Design takes time and you can't speed it up -- it's mythical-man-month territory.

Now, this means subsequent Gigafactories should go much faster, since a lot of the blueprints are basically ready now.... provided they don't go all "Model X" and start redesigning everything.

I do suspect there will be some redesign. What with building an entirely new factory design for Model Y and all -- I think it's clear Musk loves his innovation. I don't think he's disciplined enough to do Intel's "exact copy" factory protocol when he sees ways to improve it...

When do you expect Gigafactories 3,4,5, and 6 to start coming online? And finished?

 

[FredTMC]

Not about quaterly numbers.

But rather the state of Tesla's operations seem to be good. Meaning : if a supplier would be late by now, or some machinery issues or etc. that would considerably slow the M3 production. I don't think he'd tweet those " ironic " and " comic " statements.

Oh. I missed those tweets. I'll go review

 

[ValueAnalyst]

Of course it's way too high.
* Some large percentage of all Teslas, say roughly half, are sold in places where Tesla Network won't even work (like mine!).
* A large number of existing owners, say roughly half, have already stated that they will never lend their car out to strangers.
Multiply this, you'd get more like a 25% participation rate.

This isn't actually the big problem with your estimate.

Much more importantly, you've grossly overestimated the total *demand* for taxi trips, which appears to be the primary source of your bogus numbers. Your original assumptions imply over 600 million trips per day -- basically literally everyone in the US every single day. Which is just not happening for a network restricted to a short list of metropolises, and probably not for a network covering the whole US. Have you ever looked at patterns of travel demand? Obviously not.

And you can't use Tesla Network for any significant number of peak hour commutes. In order to do so, you'd have to have someone who owns a car and does NOT have a peak hour commute -- think about how uncommon those people are. (On top of that, someone who doesn't have a peak hour commute and owns a car is likely to be more protective of it and less willing to let strangers drive it.) So exclude peak hour commutes from your calculations.

This is a case where it's quite possible to hit the demand limits. Assume that Tesla takes 100% of the addressable market, but be realistic about the addressable market.

It is boring trying to explain Transportation Economics 101 stuff to you. Please go read a book on the statistics of transportation demand.

If anyone else is curious about realistic potential, I'd be happy to discuss it with them.

You're putting a lot of words in my mouth that I never said, and it's very frustrating.

The demand for rides will rise multiple times as some people choose ridesharing over car ownership.

In my original numbers, daily 600m rides were assumed in 4Q25. I hope we agree that level 4/5 autonomy will be achieved by then in most metropoles, so your implication that I assumed all Americans would use Tesla network for everything by 2018 is wrong on so many levels.

I am convinced that arguing with you is a waste of time because of your inability to critique my actual assumptions, instead of the ones you think I made, as well as your childish insults.

 

[22522]

At the very least they might be designed in concert in a simulation before finalizing the design of either.

Lisa Margonelli wrote something about how IKEA designed their product in Business 2.0... that put new factory process capability first. The design teams competed measured on addressing a specific market need using that new capability.

I think your statement is where it has to be, if looking for orders of magnitude.
Would like a shot at trying tilt up as it has worked wonders in construction for ... centuries.

 

[Curt Renz]

 

[neroden]

Yes and no. There will be less people buying cars because there use will be more efficient. Instead of being parked 23 hours per day. Even if that number only goes to sitting 15 hours a day, it can the support 3x or more peoples needs.

For the last time, I am going to explain this very slowly for the benefit of those who don't understand Transportation Economics 101.

The reason cars are sitting around most of the time is that PEOPLE DO NOT WANT TO GO OUT AT 4 AM.
The reason cars will not be used more efficiently is that EVERYONE IS DRIVING FROM 8 AM TO 9 AM. To get to their JOBS. The only way to get the cars used more efficiently is to actually get the commuters out of their cars (onto bicycles or something).

This isn't going to change due to magical wishful thinking. Jobs aren't going to change their hours (this has been tried, once on a national scale by order in the USSR, and several times by various companies in the US, and it has never made any progress; people don't like it). People are still going to be asleep at 4 AM.

Now, I am really sick of explaining these basics. If you can't figure out why this makes most of these "more efficient car use" fantasies into FANTASIES, I am done with you.

I am not saying there won't be money made from a "Tesla Network". They have a decent chance of replacing Uber, Lyft, carshare, and taxis. It'll never get bigger than that; it can't.

 

[neroden]

In my original numbers, daily 600m rides were assumed in 4Q25.

You can't figure out why this is completely ridiculous?

The reason I criticized the assumptions I *thought* you had made originally... is that I didn't conceive that you could make such appallingly terrible assumptions as "Tesla market larger than entire addressable world off-peak travel market" and "90% profit margin". I overestimated you. I apologize for assuming that you were wiser than you actually are.

It's an error I've made before. I should learn to assume the worst of people.

 

[AlMc]

For the last time, I am going to explain this very slowly for the benefit of those who don't understand Transportation Economics 101.

The reason cars are sitting around most of the time is that PEOPLE DO NOT WANT TO GO OUT AT 4 AM.
The reason cars will not be used more efficiently is that EVERYONE IS DRIVING FROM 8 AM TO 9 AM. To get to their JOBS. The only way to get the cars used more efficiently is to actually get the commuters out of their cars (onto bicycles or something).

This isn't going to change due to magical wishful thinking. Jobs aren't going to change their hours (this has been tried, once on a national scale by order in the USSR, and several times by various companies in the US, and it has never made any progress; people don't like it). People are still going to be asleep at 4 AM.

Now, I am really sick of explaining these basics. If you can't figure out why this makes most of these "more efficient car use" fantasies into FANTASIES, I am done with you.

I am not saying there won't be money made from a "Tesla Network". They have a decent chance of replacing Uber, Lyft, carshare, and taxis. It'll never get bigger than that; it can't.

While I realize some people are put off by your delivery of information I totally agree with the underlying principle.

 

[JRP3]

Yes and no. There will be less people buying cars because there use will be more efficient.

It really needs to be thought of in terms of miles driven, not just number of vehicles. Fewer vehicles will mean more miles per vehicle per day/week/month/year, which means the fleet will wear out that much faster and be replaced that much faster. (Ignoring of course longevity improvements possible with EV's.)

 

[FredTMC]

Oh. I missed those tweets. I'll go review

I went back and reviewed Elon's Tweets over the last day or so.

Sorry, there's nothing in those tweets about Model3 production at all.

And, as I said earlier, I'm done with thinking that 'happy tweets' mean that everything is great with Tesla execution and TSLA results.

Elon has many "irons in the fire" (Tesla, Spacex, Boring Co, Nueralink) and on a personal level he's got kids, girlfriend, B-days etc... many reasons for happy tweets.

 

[elasalle]

For all we know "the floor" could be the factory floor. But if the floor has a couple of release M3, the floor would look beautiful.

 

[Waiting4M3]

I went back and reviewed Elon's Tweets over the last day or so.

Sorry, there's nothing in those tweets about Model3 production at all.

And, as I said earlier, I'm done with thinking that 'happy tweets' mean that everything is great with Tesla execution and TSLA results.

Elon has many "irons in the fire" (Tesla, Spacex, Boring Co, Nueralink) and on a personal level he's got kids, girlfriend, B-days etc... many reasons for happy tweets.

The only thing I got from last night's tweet is that Elon wasn't working in China (no twitter) on his birthday.