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

[izemize]

What can Tesla do to get out of the SP death spiral it is stuck in ?

Running out of ideas for now but will add more to this. Suggestions, thoughts?

1., Sell 360-400k cars in 2019
2., Have a positive cashflow
3., Demonstrate strong demand in 2020

 

[davecolene0606]

It’s actually a back-handed complement. The author is really talking about ALL EVs, but he calls them "Teslas". The two have become synonymous.

Kleenx

Fire Away!

 

[ReflexFunds]

I'm delighted Tesla implemented version of "my idea" (please note quote marks), but a bit concerned they did it.

For anyone not following, It's now possible to buy SR(-) limited to 150km (86miles) for $CAD 44.999K and SR+ for $CAD 53.7K.
Introduction of SR- was necessary to qualify SR+, but due to the rules(max $CAD55K), no other higher model (LR, 3P) qualifies for $5K federal incentive.

This will tilt demand towards SR+ from LR and P models, and I doubt Tesla would have done it, except that they're trying to stimulate demand; so things might be dicy for them to make this move.
Let the disagreements fly!

When they are fighting so hard on the game of pennies, a few hours working on a new software limit for the battery pack to get a CAD$5k price reduction seems like a good deal. If they are really worried this will reduce LR mix, they can increase SR+ price to CAD$55k and reduce LR price slightly.

 

[GeckOo]

What can Tesla do to get out of the SP death spiral it is stuck in ? I have some thoughts.

Running out of ideas for now but will add more to this. Suggestions, thoughts?

Good points.now the question is how u make sure EM listen to you?

 

[UncaNed]

We are still questioning what is going on after 5 years of this insanity?
We are still trying to see the reason in this insanity?
We still want to believe TSLA is an investment?

Wake up and smell the coffee... TSLA is an ideological battlefield and SP is not showing who is winning.
What is certain is there will be tears and lives lost.

During those 5 years we often saw SP above 350's.... less than a year ago... Is there any reason to think we wouldn't see that again? Even if shorts DID drive SP to zero?

Perhaps Cock bros., GM, fossil fuelers, and others finally seeing their imminent demise via Teslification, will finance short selling down to SP zero, only giving up after they've long gone bankrupt, making this a very long term investment indeed. I'm still all in, happy to still be investing in Elon's mission.

Now...where exactly IS that bottom...?

 

[RobStark]

6) Advertise free of cost: Tesla should post video clips (30 -120 second) of testimonials from real customers, autopilot features, and employee interviews. YouTube and twitter could be good starting places.

Save for employee interviews, Tesla already does this on their Youtube channel.

Mostly preaching to the choir, plus trolls come in to hit the dislike button.

 

[RobStark]

Telekinesis. My powers are quite limited though. I haven't gotten Jim Chanos head to explode yet no matter how hard I've tried.

 

[UncaNed]

If SP doesnt bounce soon, getting in a dangerous position....

Hugs.

Also... DON'T BUY ON MARGIN. EVER. Unless maybe you've got something very liquid sitting right there in your desk drawer like a Gummint' savings bond to back it up with. An occasional review of a PBS video on the great depression would likely save some of us from ruin.

 

[UncaNed]

It must really hurt for Elon to negotiate with the SEC at this time. Hopefully his lawyers are doing the talking, but we hear ya Elon.

I wish he'd dug in and fought the bastadges instead.

 

[Singer3000]

This move could not be more huge, but it's not so sexy and will take a while to mature. Hence not much interest.

Killing off the entire 1on1 sales process is the key to residential solar moving back into exponential growth. Tesla essentially just did that.

It'll be VERY interesting to see Sunrun's response to this move. I expect none, but we'll see.

I don't know this industry well so excuse my back of a ciggie packet calcs. But...

Excluding the "customer agreements and incentives line", Sunrun's gross margins for "Solar energy systems and product sales" have been fairly consistent for a few years in the mid-teens percent. Their revenue from this item hit a record $350m for the year of 2018, from a market leading share of 9.5%. So that's $61m gross profit in 2018 (COGS of $294m). [NB: I've excluded the customer agreements and incentives line because it is said that Tesla now want a direct sales model, which I presume excludes this as an income stream].

For context, the sales and marketing bill in Opex of Sunrun was $294m, which is presumably the line that you think Tesla can be much more competitive on?

Lets say the US residential market has a good run, Credit Suisse predict a potential increase from c. 10 GW per year in 2018 to 40 GW per year by 2025. And that Tesla can regain Solarcity's lost market share (back to 33% from 9%), while maintaining a similar gross margin to Sunrun. That's still less than $1bn gross profit per year.

Many of us here are talking fairly seriously about a pathway to Tesla being a $100bn revenue company from their conventional auto business alone (i.e. excluding Tesla Network), with gross margins between 20-30%. Is this latest solar news not just a lot of noise from an investor perspective? Although it is definitely good for "the mission" and potentially also for Tesla auto demand, by bringing down total cost of ownership.

 

[SPadival]

Caught up - worst posts ever - you should all be ashamed of yourselves. You people make me look good which is an entirely new concept for me. Unsettling and not as welcome as you might think.

Too many Maxwell posts - they are not really that important from a SP perspective. Jeff Dahn hasn't been sitting on his hands for the past few years. Million miles battery will be just one outcome from his innovation factory. Should be in battery thread.

Too many posts shooting down shorts - this is an investors roundtable - not a crook's hangout.

Too many posts on April deliveries - what can we hope to learn without at least collating them in a separate thread and summarising the overall view?

Too many random posts.

This thread is relatively active and solely focussed on solving the problems on this thread:
Managing the volume of the Investor's Roundtable thread
Please help.

Easy problem to solve ... just ignore a lot of Care Bears... entire pages disappear

Also note that Twitter short trolls keep track of this thread and frequently paste screenshots of posts - @EVNow saw your post this way... so it’s good that we are keeping them busy

 

[UncaNed]

At this point it's about sufficient cash on hand to fund future projects. No way can they fund Semi, Roadster, Y, Pickup even if the lines were free. $2b in cash is insufficient for these projects. I bet he has funding lined up for a cap raise

If he did would that be a bad thing?

 

[Artful Dodger]

It took you two posts? I named that tune in under one.

Ha! I called him a duck on Monday morning after his 2nd post! Lame!

Exhibits all the usual signs:

• new account
  
• High frequency 1st time commenter
• binge starts immediately after another commenter is BANNED
• frequently conflates unrelated points
  
• ignores any point he loses
• takes frequent cheap shots

In short, quacks like a duck. Likely is a duck. MODS really need to start checking some IP addresses.

Cheers!

 

[Fact Checking]

This is incorrect. HW2 cars with NVidia chips were produced from Oct 2016 on.

True, thanks for the correction.

Regardless of the reason (which could have been money), the actual time of divorce was set by the incident in which a Tesla on Autopilot ran in to a crossing Semi in which the CEO of MobilEye publicly denounced Tesla's approach. So yes, that part was certainly unplanned but maybe not inevitable (had the incident not happened, MobilEye might not have reacted publicly giving negotations more time to come to an agreement).

Since Keller+Bannon was hired half a year before to create a replacement ASIC for MobilEye's ASIC, exactly how would an "agreement" have looked like? For Tesla to disband the chip team? Or for MobilEye to tolerate Tesla building a replacement? Both outcomes look rather naive.

I suspect MobilEye used the incident to trash Tesla: at this point they probably saw them as competitors already and were waiting for an excuse to part and discredit a rival.

The MobilEye divorce might also have been uglier than reported: MobilEye might have refused to ship AP1 after the contract expired - putting Tesla under big strain to save Autopilot.

 

[ReflexFunds]

I didn't say collecting the "right" data was the easy part. I said collecting data is. Don't change my words then argue with yourself over what I didn't say.

Collecting data generally is not easy either if you haven't rolled out a $40bn fleet of 500k vehicles with supercomputers and recruited 500k training drivers. But collecting the right data is a fairly critical part of collecting data, and the process for this is often the key competitive advantage of AI companies. Collecting useless data is definitely easier relatively speaking.
Anyway, I wasn't meaning to argue or twist your words, just pointing out that the smartest guys at Tesla AI like Karpathy will spend 90% of their day thinking about data collection, while the human annotation of the filtered data is currently largely outsourced (but i presume thoroughly checked back at HQ).

I am interested in exactly how Tesla's data labelling process works. Particularly as they transition towards using imitation learning for parts of driving policy. Here learning from bad drivers and accidents is potentially just as useful as learning from great drivers - The AI can learn a better understanding of the right behaviour if it also knows examples of the wrong behaviour. This will likely require grading of driving examples rather than simple annotation. I would like a training data set where various driving examples are graded by driving experts (-10 to +1 score etc, more negative penalties for bad behaviour often works best in reinforcement learning). Software guys will be leading Tesla’s AI teams, but I would be surprised If Tesla doesn’t already have driving policy experts on the team, or at least as consultants. Driving instructors are also much cheaper than machine learning experts for data labelling! Interested on your thoughts too @neroden for how they should best go about this?

 

[schonelucht]

Since Keller+Bannon was hired half a year before to create a replacement ASIC for MobilEye's ASIC, exactly how would an "agreement" have looked like? For Tesla to disband the chip team? Or for MobilEye to tolerate Tesla building a replacement? Both outcomes look rather naive.

More along the lines of having both chips on the board (as was initially planned)

 

[ReflexFunds]

More along the lines of having both chips on the board (as was initially planned)

What would the point of having both chips be? Tesla's chip was designed as a better more tailored made substitute for the Mobileye chip. It was also designed to vertically integrate a key part of the solution rather than relying on a potential future competitor.
Mobileye's EyeQ5 chip isn't even supposed to enter series production until March 2021 (so using this chip would have delayed Tesla's computer launch 2 years), and even then Q5 specs don't compete with Tesla's chip for Tesla's software.
I think no way Tesla ever intended to continue to use Mobileye or Nvidia after they finished their own chip.

 

[mongo]

What would the point of having both chips be? Tesla's chip was designed as a better more tailored made substitute for the Mobileye chip. It was also designed to vertically integrate a key part of the solution rather than relying on a potential future competitor.
Mobileye's EyeQ5 chip isn't even supposed to enter series production until March 2021 (so using this chip would have delayed Tesla's computer launch 2 years), and even then Q5 specs don't compete with Tesla's chip for Tesla's software.
I think no way Tesla ever intended to continue to use Mobileye or Nvidia after they finished their own chip.

Tesla could use the Mobileye chip output as a labeling engine for their new NN. It would also avoid the feature cliff that occured going to AP2. Once the NN could do everything the Mobileye chip could, they would stop installing it.

 

[Leo9]

FORM S-3
https://ir.tesla.com/static-files/bbe6381f-a241-4cb7-94c5-3df4a56d1f55

JUST IN: #Tesla files for potential mixed shelf offering, size undisclosed
Reuters Business on Twitter

Capital raise coming up.

 

[ReflexFunds]

Tesla could use the Mobileye chip output as a labeling engine for their new NN. It would also avoid the feature cliff that occured going to AP2. Once the NN could do everything the Mobileye chip could, they would stop installing it.

Yeah I can see Tesla may have initially planned to combine the Mobileye chip and Nvidia system for HW2. It would make a smoother transition while also getting the flexibility from Nvidia to build their own vision NNs.
I thought the above comment was regarding Tesla planning to use Mobileye in their own HW3 - at which stage Mobileye's chip specs and vision outputs were always going to be redundant.