Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

[Nocturnal]

Do you have a source for the claim that semi's in platoon are more efficient than rail? Is that a cost/mile metric, an energy / mile metric, or something else? I went looking for a site that provided comparable metrics between the two and couldn't find anything. What I could find is that the train industry is moving a ton of cargo 430-450 miles per gallon of diesel pretty consistently (that's an average calculated from total industry miles and total industry cargo (in tons) moved) and includes all of the switching and shuffling trains and cars that doesn't directly contribute to ton-miles.

If a commercial truck is moving 30 tons (60,000 pounds) at 6 mpg, that sounds like each ton is being moved 5 miles per gallon of diesel (30 tons / 6 mpg = 5 tons / gallon-mile). I readily admit I couldn't find a source for this and I could be doing the math quite badly - if I'm approximately right, then trucks are about 2 orders of magnitude short of trains on energy efficiency. Truck platooning will be a big help but it's not going to increase the second truck's fuel efficiency to 1200 mpg from 6 mpg (so overall pair of truck efficiency is ~600 mpg).

I ask because I'm dubious - the rolling resistance of steel on steel is awfully hard to beat, and train cars tucked in so closely between each other makes for some seriously good wind resistance with each incremental car.


On a cost/mile metric, in the US, I wouldn't be surprised to find that a platoon of semis can compete with trains. In the US though, society subsidizes truck expenses (road maintenance) pretty heavily, while companies that operate trains have to fully fund their track maintenance themselves.


Some obvious limitations of train - trains do a poor job of stopping at every address to make deliveries (they don't ). And apparently trains can still burn relatively high sulfur fuel, while semis have made the switch to low sulfur fuel (leading to lower pollution from trucks than trains). The difference in fuels is changing or has changed.

I couldn't find an exact comparison but did some back of the napkin math a while back. I wound up estimating that Rail Freight is 5.1 cents per ton/mile, current trucking is 15.6 cents per ton mile, and an FSD EV truck might be in the 8.6 cents per ton mile.

So still not there (I did not factor platooning, just FSD electric) price wise but a 24-48 hour transit time vs 2 weeks.

 

[Mike Smith]

Remember last year when Q3 earnings came out there was a similar initial move higher, but the short covering continued into mid January, where short interest bottomed just below 25M. TSLA grinded higher the entire time. In fact it didn't stop until in became apparent Q1 was going to disappoint. I expect the same to happen now, with at least another 5M net short shares being covered over the next few months. The reason this happens is that most shorts can actually afford the loss so they don't need to cover immediately, but they didn't want their position to have increased so they're still looking to cover over time.

 

[ggies07]

JPMorgan Chase & Co. Reaffirms “Sell” Rating for Tesla (NASDAQ:TSLA)

Gotta keep Tesla capped in a trading range during the historic bull run

Wait, JPMorgan has a "sell" rating, but.....

Oh, nice didn't come across this guy before.

Looks like he laid out his thesis in 2016, before
The Osborne Effect On The Auto Industry | CleanTechnica
and

Hosts a guy who talks about the fast transition to EVs?....................................................................................................................................................................................................................................................................................................................................................................................................

 

[FreqFlyer]

I’m sure they’ll do whatever makes the most sense from a cost and efficiency standpoint. In the long-term all the ideas we’re spitballing would have to be evaluated, but obviously we’re a few years out from any of them.

Are we really a few years away? If EM is stating FSD will be feature ready next year but awaiting governmental regulations, fine, use the feature to transport cars from the factory a couple of miles to the port. Use surface streets, stay off the freeway, and do it when the streets are empty. Imagine the costs saved if Tesla can load and unload cars on ships or railcars without labor? Or is it a union issue?

FSD should be able to do relatively easy. The route, either to the port, onto or off a ship or railcar is set.

 

[anthonyj]

Hedge Fund Scraps Tesla Short in New Letter After Surprise Rally
Bloomberg - Are you a robot?

 

[WarpedOne]

A lot of things have to come together for carpooling of any kind to work. First you and the other carpoolers have to live near each other and work near each other. You also have to have a reasonable chance of getting along. Then the hours have to be in sync and never ever vary. Also the driver of the day has to be reliable. This is a lot to ask.

This is too much to ask of people, but is trivial for software ... I need a seat at address X and time Y, a car arrives.

 

[SpaceCash]

Doubt it. People don’t carpool now when they’d be saving even more money. What percentage of Uber rides are Uber pools? Whatever it is, it will go down when the savings are less.

A possible exception might be a robo-taxi with separate compartments, so you can listen to your own music, have privacy...

Yup. I hate carpooling (especially with a rideshare app) and so does everybody else. When I'm en route I don't want to talk to the random driver, the random passengers, none of it. I'm an extrovert but not when I'm catching a ride from point A to point B; at those times I'm catching up on TMC, coordinating plans, just anything other than basic smalltalk with strangers which will not ever be followed up on

 

[UncaNed]

Do you have a source for the claim that semi's in platoon are more efficient than rail? Is that a cost/mile metric, an energy / mile metric, or something else? I went looking for a site that provided comparable metrics between the two and couldn't find anything. What I could find is that the train industry is moving a ton of cargo 430-450 miles per gallon of diesel pretty consistently (that's an average calculated from total industry miles and total industry cargo (in tons) moved) and includes all of the switching and shuffling trains and cars that doesn't directly contribute to ton-miles.

If a commercial truck is moving 30 tons (60,000 pounds) at 6 mpg, that sounds like each ton is being moved 5 miles per gallon of diesel (30 tons / 6 mpg = 5 tons / gallon-mile). I readily admit I couldn't find a source for this and I could be doing the math quite badly - if I'm approximately right, then trucks are about 2 orders of magnitude short of trains on energy efficiency. Truck platooning will be a big help but it's not going to increase the second truck's fuel efficiency to 1200 mpg from 6 mpg (so overall pair of truck efficiency is ~600 mpg).

I ask because I'm dubious - the rolling resistance of steel on steel is awfully hard to beat, and train cars tucked in so closely between each other makes for some seriously good wind resistance with each incremental car.


On a cost/mile metric, in the US, I wouldn't be surprised to find that a platoon of semis can compete with trains. In the US though, society subsidizes truck expenses (road maintenance) pretty heavily, while companies that operate trains have to fully fund their track maintenance themselves.


Some obvious limitations of train - trains do a poor job of stopping at every address to make deliveries (they don't ). And apparently trains can still burn relatively high sulfur fuel, while semis have made the switch to low sulfur fuel (leading to lower pollution from trucks than trains). The difference in fuels is changing or has changed.

I agree with your reasoning, and indeed I misspoke, but according to Elon out 3 trucks convoying beats rail, not necessarily platooning. See 20:09 on

 

[FreqFlyer]

I couldn't find an exact comparison but did some back of the napkin math a while back. I wound up estimating that Rail Freight is 5.1 cents per ton/mile, current trucking is 15.6 cents per ton mile, and an FSD EV truck might be in the 8.6 cents per ton mile.

So still not there (I did not factor platooning, just FSD electric) price wise but a 24-48 hour transit time vs 2 weeks.

There is a reason BNSF is transporting alot of containers, UPS,JB HUNT,Scheidner,etc. between CHI-LAX via the "racetrack".

The railroads are upping their efficiency by using exta locomotives two thirds into the train or on the rear, remote control from the lead locomotive allowing longer trains two to three miles long moving at 70 mph (BNSF across the Arizona,NM, again, the "racetrack"). VERY hard to beat that kind of efficiency.

 

[EVNow]

With Tesla Semi’s running back and forth on autopilot?

When FSD becomes real, Tesla won't be shipping many cars to customers

 

[EVNow]

having just moved from Michigan to the Bay I agree with your sentiment on living out west being better. But moving one family is a task. Moving an entire labor force is not as easy as “it’s better out there”. People have spouses with lives, kids with friends. Not to mention the cost both financial and psychological of the act of moving.
Not sure what your job is but imagine someone told your entire office they were moving 3000 miles away. It would not go well.

If people from mid-west don't want to move, we can always get people from Mexico

Seriously, for a lot of people it might be the difference between having a job or not. They will move esp. the younger ones where the old auto plants are closing.

 

[Nocturnal]

There is a reason BNSF is transporting alot of containers, UPS,JB HUNT,Scheidner,etc. between CHI-LAX via the "racetrack".

The railroads are upping their efficiency by using exta locomotives two thirds into the train or on the rear, remote control from the lead locomotive allowing longer trains two to three miles long moving at 70 mph (BNSF across the Arizona,NM, again, the "racetrack"). VERY hard to beat that kind of efficiency.

True. My assumptions were at current state on average for freight so any improvements in that area would change the math.

 

[FreqFlyer]

I agree with your reasoning, and indeed I misspoke, but according to Elon out 3 trucks convoying beats rail, not necessarily platooning. See 20:09 on

I can forsee a time in the near future where western states with less populations will allow FSD for semis, such as AZ, NM and Texas (I-10 west Texas to San Antonio). It will be weather dependent but the southwest states have better weather most of the year.

 

[Steve m]

Tony gets a lot right, but he exaggerates and is also off in 2 ways:

1) Having fewer cars because of Robo taxis will not reduce traffic. Even though there are fewer cars, the overall mileage will stay the same or even probably increase which will make traffic the same or worse.

2) He applies the battery cost decline curve to the entire cost of the vehicle, when it should only apply to the battery, (and maybe electric drivetrain) alone, versus the gasoline engine/transmission/emission systems. When battery prices were by far the largest aspect, his projections were fairly accurate, however as battery prices decline, his projections will get further and further off. However, since Tesla is reinventing manufacturing, they may make his estimates look better than they actually are.

Having self driving cars that communicate with each other will increase road utilization which implies faster transit times for the same number of cars.

For example. We humans should space our cars about 1 car length apart for every 10 mph we are driving. Cars that communicate could be driven closer because they would all slow down in unison when a car in front needs to break. We humans need a couple of 10ths of seconds to start slowing down. Similarly at stop lights we humans add a delay before we accelerate. If you have 10 cars in a line to get through the intersection the last car starts 9 times the delay after the first car starts.

If you couple in smart roads. Where the road bunches cars to gether such that no two bunches using the same intersection areive at it until it is clear the cars would not stop at all.

 

[dww12]

Having self driving cars that communicate with each other will increase road utilization which implies faster transit times for the same number of cars.

For example. We humans should space our cars about 1 car length apart for every 10 mph we are driving. Cars that communicate could be driven closer because they would all slow down in unison when a car in front needs to break. We humans need a couple of 10ths of seconds to start slowing down. Similarly at stop lights we humans add a delay before we accelerate. If you have 10 cars in a line to get through the intersection the last car starts 9 times the delay after the first car starts.

If you couple in smart roads. Where the road bunches cars to gether such that no two bunches using the same intersection areive at it until it is clear the cars would not stop at all.

How about FSD cars just having less accidents for better road utilization.

 

[jerry33]

This is too much to ask of people, but is trivial for software ... I need a seat at address X and time Y, a car arrives.

Yes, it gets rid of a couple of items, but you still need the others. Carpoolers living near each other and work near each other. They also have to have a reasonable chance of getting along.

 

[Rashomon]

If a commercial truck is moving 30 tons (60,000 pounds) at 6 mpg, that sounds like each ton is being moved 5 miles per gallon of diesel (30 tons / 6 mpg = 5 tons / gallon-mile). I readily admit I couldn't find a source for this and I could be doing the math quite badly - if I'm approximately right, then trucks are about 2 orders of magnitude short of trains on energy efficiency. Truck platooning will be a big help but it's not going to increase the second truck's fuel efficiency to 1200 mpg from 6 mpg (so overall pair of truck efficiency is ~600 mpg).

I believe you’re dividing where you should be multiplying. Try 180 ton-miles/gallon for diesel trucks, or only a factor of 2.5 or so from rail. Now add aero and efficiency improvements from Tesla semi ...

 

[NakedLong]

Tony gets a lot right, but he exaggerates and is also off in 2 ways:

1) Having fewer cars because of Robo taxis will not reduce traffic. Even though there are fewer cars, the overall mileage will stay the same or even probably increase which will make traffic the same or worse.

2) He applies the battery cost decline curve to the entire cost of the vehicle, when it should only apply to the battery, (and maybe electric drivetrain) alone, versus the gasoline engine/transmission/emission systems. When battery prices were by far the largest aspect, his projections were fairly accurate, however as battery prices decline, his projections will get further and further off. However, since Tesla is reinventing manufacturing, they may make his estimates look better than they actually are.

Regarding point one, I wonder how much traffic is caused by parking and looking for parking. Also, I wonder how much a good app could reduce traffic by incentivising ride sharing. Also, if Tesla manages to corner the FSD market, I think we can expect some tunnels with the profits in the longer term...

 

[jerry33]

I believe you’re dividing where you should be multiplying. Try 180 ton-miles/gallon for diesel trucks, or only a factor of 2.5 or so from rail. Now add aero and efficiency improvements from Tesla semi ...

And don't forget to add the cost of the transfer to trucks at the train station. It's the total cost of the transport that the shipper pays for, not just the train part.

 

[adiggs]

I believe you’re dividing where you should be multiplying. Try 180 ton-miles/gallon for diesel trucks, or only a factor of 2.5 or so from rail. Now add aero and efficiency improvements from Tesla semi ...

Thank you - I believe you're right.

Also an indicator that as long as US society is willing and even eager to subsidize commercial trucking, we're going to get an awful lot of it. The convenience of smaller loads and point to point delivery is going to outcompete rail over anything vaguely resembling short distances. I don't think trucks need to be even with trains on efficiency - they just need to be close.


What I'm worried about - throw in hefty subsidization (roads) compared to trains, and we might even succeed in putting our big train networks out of business (oops).