Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

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

This site may earn commission on affiliate links.
Tesla convinces judge to dismiss Model 3 production lawsuit

Sounds like it might not be over: "The plaintiffs can still amend their complaint"

[edit: doh! gotta check them dates better]

Here's the RECAP archive of the case:


The ruling also dismisses the main lawsuit with prejudice:

"The Court thus GRANTS the Motion to Dismiss. Further, because the Court concludes that any further amendment would be futile, the SAC is DISMISSED WITH PREJUDICE AND WITHOUT LEAVE TO AMEND."​

I.e. the loss looks pretty permanent to me. Not sure where Reuters got the "plaintiffs can amend" part - they cannot. They can only appeal.
 
Last edited:
Last edited:
They are independent of each other. A body and/or interior refresh is usually based on the number of years since the last one. Typically four or five. The X is not due (only three), the S is closer. Reducing the number of options, and the price, is one way to have fewer complaints (I purchased a week before a refresh).
Yeah, and:
  • HW3 costs less than HW2/2.5
  • Telsa buys the chip from themselves (margin goes up)
  • nVidia cut out of supply chain (more vert. integration)
So even if HW3 wasn't 1,000x faster executing the NN, it'd still be good for business.

Cheers
 
Here's the RECAP archive of the case:


The ruling also dismisses the main lawsuit with prejudice:

"The Court thus GRANTS the Motion to Dismiss. Further, because the Court concludes that any further amendment would be futile, the SAC is DISMISSED WITH PREJUDICE AND WITHOUT LEAVE TO AMEND."​

I.e. the loss looks pretty permanent to me. Not sure where Reuters got the "plaintiffs can amend" part - they cannot. They can only appeal.

I like this part:

Plantiffs have already been given two bites at the apple; the Court concludes that any further amendment would be futile.

And the reason that are allowable to deny the right to amend:

The final issue is whether or not to grant Plaintiffs leave to amend their Complaint. Such leave is to be given “liberally,” Sonoma Cty. Ass’n of Retired Employees v. Sonoma Cty., 708 F.3d 1109, 1117 (9th Cir. 2013), but it may be withheld where “there is strong evidence of ‘undue delay, bad faith or dilatory motive on the part of the movant, repeated failure to cure deficiencies by amendments previously allowed, undue prejudice to the opposing party by virtue of allowance of the amendment, [or] futility of amendment, etc.’”
 
Yeah, and:
  • HW3 costs less than HW2/2.5
  • Telsa buys the chip from themselves (margin goes up)
  • nVidia cut out of supply chain (more vert. integration)
So even if HW3 wasn't 1,000x faster executing the NN, it'd still be good for business.

Cheers

It's all that, and it's also much, much better:
  • NVidia was probably selling the gp106 chip for around $100, but marginal cost of producing one more Tesla AI chip ASIC is in the few dollars range.
  • Nvidia's latest generation chips would have been significantly more expensive, reducing the margins of the Model 3 or increasing its price above $35k.
  • Tesla can add two or four chips to a single board with very little additional cost other than power use - should they require more computing power in the future. For example the Tesla Semi, a much larger vehicle, could possibly run 20 cameras on 4 AI chips at full FPS.
  • Tesla can also increase the frequency of the chip - it's reportedly around 400 MHz now, which is pretty conservative.
  • Tesla could, in principle, license tens of millions of AI chips to other carmakers who don't have the processing power but would like to run their own NNs. This is probably why Intel bought MobilEye for 15 billion dollars last year.
  • Tesla could in principle license AutoPilot with their own AI chip - i.e. license both the software and the hardware. Other carmakers would find it very difficult to replace that hardware with their own R&D results and Tesla could charge very, very healthy margins.
I.e. having their own chip turns Tesla into a chip-maker, giving the infinite flexibility to their platform, and enabling high margin business opportunities that chip makers generally enjoy.
 
delivery timelines are location dependent. for my zip code everything but SR (6-8 weeks) is within 2 weeks.

Look, does anyone take those vague, hand-updated delivery times on the website that seriously? Because we know from past experience that you shouldn't. Sure, people in Fremont are probably getting cars shoved at them this week to minimize inventory over the break between quarters, but if you're not benefitting from Tesla's delivery scheduling system, you're being hurt by it, and there are still people complaining about erratic and unpredictable delivery times over in the Model 3 forum (although *less* so).

The website estimates are very, very, very rough and very sloppy.

Given the state of the European order backlog -- I think less than half of the backlog was handled in Q1 --- I fully expect US deliveries to start drying up again on April 1st as more ships get loaded for Europe. There will probably be a shift back to the US before the next tax credit expiration at the end of June, of course.

What I'm saying is: that "2 weeks" isn't real. For someone in Fremont who orders the "right" configuration, it's probably 2 days. For someone in Texas who orders the "wrong" configuration, it's probably June.
 
I believe that some traders believe in technical trading :D
Of course, that only yields some predictive power as we don't know how much of $TSLA is being traded on a technical basis...
All the bots are technical trader, by definition. Although some bots also take into account other factors, like keywords grep'ped from a newsfeed. High frequency trading depends on bots. HF trading might be 90% of volume on the NASDAQ recently.