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

[acoste]

JFYI, the Tesla Semi has similar weight to comparable Class 8 trucks, which is ~17,000 lbs plus 200-300 gallon of fuel which is another 1,400-2,100 lbs of weight: a modern tractor trailer can easily weigh ~20,000 lbs.

Diesel engines, the steel chassis to distribute all that force plus fuel weighs a lot.

That the Tesla Semi is going to weigh too much has been a TSLAQ lie from day 1.

Do you have some numbers to support this? I'm just guessing here on the electric parts and some other numbers:

diesel engine 3000 lbs
transmission 700 lbs
300gall fuel 2000 lbs
- electric motor + inverter -1500lbs

that is 4200 lbs for the battery

current battery is at ~243Wh/kg

4200 lbs = 1900kg
243*1900/1000 = 462 kWh battery capacity doesn't seem enough for a semi trailer.

Diesel truck:
average 18-wheeler gets 5.9 mpg,
300 gallon x 5.9 = 1770 miles

Electric truck
Model S 100D would get 40-42mpg on diesel at those speeds (~7 times less consumption than truck). Range might reach 350 miles on a 100kWh battery.
That is 350 miles * 462kWh / 100kWh / 7 = 231 miles

 

[uselesslogin]

Do you have some numbers to support this? I'm just guessing here on the electric parts and some other numbers:

diesel engine 3000 lbs
transmission 700 lbs
300gall fuel 2000 lbs
- electric motor + inverter -1500lbs

that is 4200 lbs for the battery

current battery is at ~243Wh/kg

4200 lbs = 1900kg
243*1900/1000 = 462 kWh battery capacity doesn't seem enough for a semi trailer.

Diesel truck:
average 18-wheeler gets 5.9 mpg,
300 gallon x 5.9 = 1770 miles

Electric truck
Model S 100D would get 40-42mpg on diesel at those speeds (~7 times less consumption than truck). Range might reach 350 miles on a 100kWh battery.
That is 350 miles * 462kWh / 100kWh / 7 = 231 miles

The low CD will help and then an extra 3,000 pounds of batteries is probably ok too.

 

[printf42]

Seems Tesla increased the Model 3 prices by $400 in the US. Not sure what to read from it. Could be a margin play or could be tinkering with demand.

It’s just preparation for the next price increase, and that one would be $90, followed by another one at $9.

 

[anthonyj]

Where is advanced summon? FSD w/ driver by end of year? Really? Maybe Neroden is right. Tesla’s achievements are overshadowed by their disappointments which is sad. Since Tesla is a public company, Musk needs to be a bit more like Steve Jobs, just surprise us when it’s ready

 

[jerry33]

The low CD will help and then an extra 3,000 pounds of batteries is probably ok too.

Unfortunately, not for weight bound trucks, volume bound trucks will be fine, as will trucks that never max out. My understanding, which might be wrong, is that the majority of trucks are weight bound. 3000 lbs extra vehicle weight means that much less cargo, so less income.

 

[bdy0627]

I think those are aggressive long-term numbers.

If the base Model 3 costs $38k to make currently, then current average cost across all models is going to be much higher than that.

Shanghai should obviously shift the price of the cheaper SR way down, so that should be factored in. If global demand is about 400k, Shanghai could be producing nearly half of the model 3s, particularly if model Y is built at Fremont, which would stay busy even with a lower demand for high trim model 3s. The variants coming out of Fremont would stay above $50k ASP with the lower trim versions coming out of Shanghai with much better gross margins or lower priced if necessary to drive demand. We really should be modeling production from Fremont separately from production from Shanghai. At this point, it's not clear that there is enough demand for the long range trim for Fremont to be above 5,000/week once Shanghai is producing the SR. If so, we might have to shift the Fremont numbers to about 250,000 model 3 annually at an ASP of $50k+. It seems obvious based upon the plan for production well above that that the Y would then need to fill the production gap at Fremont.

 

[capster]

Ohhhh ! Care to share, or even hint ? !

Ha ha! Alas, I don't have a specific situation in mind. I merely meant that insight might make one of its glorious, if rare, appearances to this heedful eye.

 

[kbM3]

I don't think this is actually correct.

SP went down to $300 when the Q4/Q1 guidance was lowered, went down to $260 after bad Q1 delivery report and is now at $225 after bad Q1 ER.

Should it have gone down so much because of a bad quarter - probably not. But a bad quarter emboldens shorts and they sell more. Counterintuitively, shorting is highest now at $227, than it was at $350.

Better results in rest of the quarters this year + FSD feature complete will bring up SP. When NOA for city streets is released, I expect significant value assigned to FSD.

My guess would be that FSD will be similar to NOA. They’ll ship it this year, but it will require lots of interventions and not work too well. Then it will improve rapidly in 2020. NOA is getting very, very good on interstates, but kind of flying under the radar, because it wasn’t that great when first released.

 

[kbM3]

Unfortunately, not for weight bound trucks, volume bound trucks will be fine, as will trucks that never max out. My understanding, which might be wrong, is that the majority of trucks are weight bound. 3000 lbs extra vehicle weight means that much less cargo, so less income.

Jerome Guillen has stated they’re targeting the same payload capacity as a diesel.

Also @Artful Dodger (IIRC) did an estimation based on acceleration figures that came to the same conclusion.

Somehow Tesla’s pulling off a miracle. I’d guess they saved a lot of weight in the traditional parts of the tractor.

 

[EVNow]

And the downside is that they're not very flexible because they have no "understanding" of what they're looking at. We tend to understand complex scenes through conversion of broad shapes to a mental model of 3d geometry. Neural nets tend to understand them through textures and the contours of edges. Eg. if you take a 3d model of a cat and texture it with elephant skin, a human will still say it's a cat, but a neural net will tend to say it's an elephant. You could train it to learn the difference between an elephant-textured cat and an elephant, but that's not going to be its initial interpretation of the novel scene.

The short of it is that you have to train neural nets as though they're idiots just looking for any excuse to mess up. They're not going to understand something that's not like things they've seen before. They can't "reason it out". They're pattern detectors.

Replying to an old post - but I thought this aspect of NN is worthwhile.

Neural Networks

Out of shape? Why deep learning works differently than we thought​

After training a deep neural network on thousands and thousands of these images with arbitrary textures, we found that it actually acquired a shape bias instead of a preference for textures! A cat with elephant skin is now perceived as a cat by this new shape-based network. Moreover, there were a number of emergent benefits. The network suddenly got better than its normally-trained counterpart at both recognizing standard images and at locating objects in images; highlighting how useful human-like, shape-based representations can be. Our most surprising finding, however, was that it learned how to cope with noisy images (in the real world, this could be objects behind a layer of rain or snow) — without ever seeing any of these noise patterns before! Simply by focusing on object shapes instead of easily distorted textures, this shape-based network is the first deep neural network to approach general, human-level noise robustness.

Wow, that is so interesting about the NN with a shape bias! Great post, heltok, kudos.

Karpathy tweeted about the texture bias thing and jimmy_d had some interesting thoughts on it.

Karpathy and jimmy_d both argue that neural networks are not inherently texture biased; the ImageNet dataset just happens to be texture biased, and neural networks trained on this dataset internalize the bias.

​

 

[EVNow]

that's for google maps. but there are tons of driving related captchas now. it's not a coincidence.

Just saw this reply from Karpathy about captchas.

 

[EVNow]

Now I think delivery Wave was actually a scapegoat for bad Q1. On the contrary, the real reason was not able to achieve the wave in Q1 end. The root cause, of course, was bad planning.

Ofcourse hindsight is 20/20 and all problems can be blamed on bad planning (I've seen teams do it in multiple companies over decades).

As we can see, so many inventory m3 were unable to be delivered in Europe and China. In a sense, if tsla's claimed goal is to end wave, they already largely achieved it in Q1 because the end of Q1 result(total delivery) is actually very similar to what it would be if there is no wave.

They did some unwinding of the wave in Q1 unwittingly, but this needs to be a commitment and they need to continue to work on it. For example in June they still need to continue sending cars to EU/China. If they don't do that, the wave will be back.

 

[elasalle]

Now I think delivery Wave was actually a scapegoat for bad Q1. On the contrary, the real reason was not able to achieve the wave in Q1 end. The root cause, of course, was bad planning.

As we can see, so many inventory m3 were unable to be delivered in Europe and China. In a sense, if tsla's claimed goal is to end wave, they already largely achieved it in Q1 because the end of Q1 result(total delivery) is actually very similar to what it would be if there is no wave.

Tesla tried to make it's US customers be happy in Q4 18, and hence it did not send vehicles overseas. I think buffer of vehicles in transit has been established, and now the wave approach is not necessary. ~ cheers!!

 

[elasalle]

Trump: We'll know in three to four weeks if China trade talks are successful


+ 2 Funnies - If I wasn't loosing my monies, I'd find this hilarious too

 

[sparcs]

Someone just posted on reddit that there was a $400 price increase on Model 3. Tesla increases US pricing by $400 : teslamotors

Wonder if it is due to the tariffs and the increase in costs of items like the FSD hardware.

 

[sundaymorning]

Someone just posted on reddit that there was a $400 price increase on Model 3. Tesla increases US pricing by $400 : teslamotors

Wonder if it is due to the tariffs and the increase in costs of items like the FSD hardware.

As an investor I welcome this news with open arms.

 

[Todd Burch]

Talk about commenting without bothering to know the context.

I was responding to someone who said CR rated it as the best as if CR did some comparative testing per their usual reports.

I don’t have anything against JDP, but they are fairly biased surveys. Many of them are “sponsored” by dealers as an incentive like with free tank of gas or services.

I knew the context, I read your comments, and I don’t know why you assume I didn’t.

You said the CR result was worthless because it was a consumer survey. I have done CR surveys before. They are unsponsored and unbiased.

 

[lascavarian]

Tesla’s achievements are overshadowed by their disappointments which is sad. Since Tesla is a public company, Musk needs to be a bit more like Steve Jobs, just surprise us when it’s ready

I can understand that. However I also love it when someone asks for "Dog Mode" and a couple months later it appears with minimal fanfare. That is how things happen.

I know EM has a healthy dose of human frailities but I don't perceive him as being dishonest. I think he puts forth a plan and an ask. It is understandable that people responding to the "ask" would like a timeline. But innovation is hard to put on a timeline. Innovation is tough to make easy.

EM is asked a real question and he gives his thinking at the time based on the known unknowns. He is often late but mostly delivers and I find that refreshing and tolerable because I think it is real.

 

[EVNow]

It’s just preparation for the next price increase, and that one would be $90, followed by another one at $9.

May be this is the March of 9s.

ps : To make this clear, March of 9s in May.

 

[Pras]

China current demand size:

In China, model 3 sells for almost $15-20k more than US. I would assume that most of this additional cost is due to some tariff and shipping. Any other theories?

Another point. Could it be that in anticipation of GF3, the China model 3 sales are being held back? I think this might be a drag specially around the time the GF3 is near complete.