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

[mongo]

https://twitter.com/x/status/1777586441649828205

2023 also had low retail weeks
YoY accumulated deltas (now to past in thousands): -9, -4, -7, -4, +2, +6, +7, +7, +10

 

[cliff harris]

I really don't think there is a go-slow on semi due to cash. Tesla has a staggering amount of cash in the bank right now. The impression I get is that the semi release is slow because they are collecting as much real-world long term data as they can on the existing prototype built models.
Also its very hard, without being part of the project management team, to look at a construction project and say 'there is no urgency'. Maybe they are delayed temporarily by something completely out of their control involving permitting or infrastructure.
I do think once the production facility is complete people will be surprised how many semi's Tesla builds. Its a perfect market for them, based on numbers and economics and not consumer sentiment and FUD.

 

[jerry33]

That's the key. In an RT there won't be any "regen unavailable" situations . . . because it won't let itself get into those.

Does this mean RT will never drive in cold weather?

 

[jerry33]

I really don't think there is a go-slow on semi due to cash. Tesla has a staggering amount of cash in the bank right now. The impression I get is that the semi release is slow because they are collecting as much real-world long term data as they can on the existing prototype built models.
Also its very hard, without being part of the project management team, to look at a construction project and say 'there is no urgency'. Maybe they are delayed temporarily by something completely out of their control involving permitting or infrastructure.
I do think once the production facility is complete people will be surprised how many semi's Tesla builds. Its a perfect market for them, based on numbers and economics and not consumer sentiment and FUD.

The Semi has to be ready when it's shipped. Fleets won't tolerate a truck that doesn't do the job, so breakdowns, poor tire wear, etc. have to be fixed before wide release. It can't be an iterative release the way it can be for consumers. This means a lot of testing and refining. Also there is not really any charging infrastructure for Semi at this point, so that needs to be remedied and enough short distance fleets need to commit to purchasing a charger and installing it so the Semis made can be sold immediately. Long distance fleets will wait until there is a charging network.

 

[nativewolf]

One option: subsidize rides to mass transit and subsidize pooled rides while charging a congestion fee to trips toward the center? Cities have different levers at their disposal and the smart ones will use robotaxi to reduce overall car traffic and make cities more livable. Hopefully at least some of them decide to be smart about this.

It will be a period of change for folks, pooled rides is the only way RT will reduce congestion. Basically the benefits people cite have unfortunate side effects that are very negative. It will be just like the uber drivers constantly circling areas and causing traffic jams. We already know what will happen, how they plan around that I don't know. Pooled rides seems to be the most logical.

 

[elasalle]

Elon managing cash flow. Nvidia chips will hurt this year.

Or focus is now in the new CATL battery line?

 

[CitizenKane]

It will be a period of change for folks, pooled rides is the only way RT will reduce congestion. Basically the benefits people cite have unfortunate side effects that are very negative. It will be just like the uber drivers constantly circling areas and causing traffic jams. We already know what will happen, how they plan around that I don't know. Pooled rides seems to be the most logical.

If RTs are "fleet aware" i.e. if they can get instant information from and about every other RT, then couldn't that significantly reduce jams?

 

[Musskiah]

Tesla (NASDAQ: TSLA) doubled its sales of electric vehicles (EVs) in China in March compared to February, with exports from its Shanghai plant down slightly.

Huge month = soft EV sales

https://twitter.com/x/status/1777586441649828205

No biggie, just one of the biggest weeks of all time. Must be soft demand?

 

[jschwefel]

Tesla EV sales in China double to 62,398 in Mar

In the first quarter, Tesla sold 132,420 vehicles in China, contributing 34.23 percent of its global deliveries.

cnevpost.com

Tesla (NASDAQ: TSLA) doubled its sales of electric vehicles (EVs) in China in March compared to February, with exports from its Shanghai plant down slightly.

The US EV maker sold 89,064 China-made vehicles in March, including 26,666 exported, according to data released today by the China Passenger Car Association (CPCA).

That means Tesla sold 62,398 vehicles in China in March, up 107.02 percent from 30,141 in February, though down 18.61 percent from 76,663 a year earlier.

View attachment 1036915

Note: Chinese New Year was in February.

 

[GhostSkater]

If RTs are "fleet aware" i.e. if they can get instant information from and about every other RT, then couldn't that significantly reduce jams?

Tesla disrupting Waze now, one more to add to the list, specially if they provide traffic data for others, way more accurately

 

[FEERSUMENDJIN]

Great thoughts!

I bet RT first gig will be moving personnel around Giga Texa

You forgot energy.

The energy is just to feed the AI

 

[traxila]

Huge month = soft EV sales

No biggie, just one of the biggest weeks of all time. Must be soft demand?

For this week we were looking at:

Export wave for beginning of quarter.

End of discounting available by Tesla until end of March brought sales forward.

Four day Holiday in China for the week.

Dismiss until next week.

Lol.

 

[elasalle]

Sub-optimal

1 reason

https://twitter.com/x/status/1777599145676615872

 

[FEERSUMENDJIN]

View attachment 1036766

here is one person's experience. This can't be dismissed as "well, this is just one person, it works great for me". To be Robotaxi ready, everything needs to be green for everyone.

The fastest Tesla has gone between announcement and production is about a year, so the vehicle itself won't be ready for at least a year, then they have to make sure the software works well on that particular vehicle (the software just doesn't work on everything eg Cybertruck). Plus they have to make it more robust. All those things add up. They'll get there, but it isn't imminent.

They have already announced Gen3 and have been working on PRODUCTION for over a year. this time it is not the same.

 

[unk45]

This is great data! I didn't know this existed.

Especially if we look at just inventory, not inventory/sales ratio, it is clear that there is substantially more production than sales lately in the US. However, I don't think "channel stuffing" is a good description, because inventory is bouncing back from almost zero, and it's still about 3-7x less than it used to be. That being said, the total amount of inventory buildup left would be at most a few weeks worth of sales.

Interestingly, I see that inventory was already plummeting in the years before 2020. I wonder why.

New vehicle sales are still about 15% less than pre-pandemic. Meanwhile, used vehicle sales post-2020 remain at an all-time high, and show massively greater seasonality for some reason.

Inventory/Sales
View attachment 1036871

Inventory
View attachment 1036873

Sales
Note that I zoomed in on 2005-present for this one
View attachment 1036874

Used Car Sales
View attachment 1036875

The St Louis Fed does some good graphics on numerous subjects. For those who want to evaluate their source data and also have much more detail go here:

BEA Interactive Data Application

The BEA data shows sourcing and detailed components for each element. For example: US auto inventory can be assessed at OEM level, at dealer level or both. Further the BEA also has aggregated final delivery data from State level reporting.

There are myriad resellers of such data, which is a waste of money when it’s all free from the BEA and FRED, among others.
Many of the variations reported are actually reporting anomalies, so it’s crucial to look carefully.

 

[FEERSUMENDJIN]

I don't see how a Robotaxi can be a compact car. I barely fit in a Model 3, and it's difficult for the old and out of shape to get into and out of. There is no way a robotaxi for all can be based on a compact model.

higher off the ground, easier to get into than a model 3 for sure. It should be like a sci fi looking more aerodynamic Honda Jazz that all the old folks have

 

[replicant]

The Semi has to be ready when it's shipped. Fleets won't tolerate a truck that doesn't do the job, so breakdowns, poor tire wear, etc. have to be fixed before wide release. It can't be an iterative release the way it can be for consumers. This means a lot of testing and refining.

We heard Tesla operated about ~100 units (incl. with their Pepsi partner). How many more could they need to properly test the semi in real-word conditions?

If they can speed up construction for the semi plant, why wouldn't they not accelerate today so they can produce it in volume before 2025? They may have reasons, but if volume production isn't reached by 2026, this is not up to Tesla's speed I got used to back in the early days of the company.

 

[Discoducky]

Not an AI expert by any means, but this behavior is strange, hard for me to explain, and makes me question the overall method. When I use the turn signals on FSD, it seems to take that command under advisement, starts to execute the request, but frequently changes it's mind after starting the lane change and ultimately ignores my command. This makes me question how much control Tesla has over the "black box" that is the NN....

Is your concern just when there's no nav for FSD to follow?

IMO they are prioritizing solving for the main use case first, which is no human input. The next priority is to capture enough details to learn from any disengagements. This is best done with a nav route and why the current builds are showing issues when the nav route changes. This hand-off isn't the top priority, but is most likely 2nd.

 

[hobbes]

There are two competing visions of a sustainable future.

The first is built on the idea that the only way to survive as a species is to limit growth, cut consumption, and live austerely.

The second is built on the idea that technological innovation can enable both sustainability and growth.

Tesla stands for the second one.

I don´t see why you can´t combine both approaches.

Requires unpolarized thinking though which seems to be wildly unpopular these days :/

 

[hobbes]

Maybe the Robotaxi is the van not the M2. You heard it here first.

Would definitely avoid osborning existing Models.. And could also be used for last mile parcel deliveries. Like the idea.